View or print this publication - Northern Research Station - USDA ...

"

ABOUTTHE COVER 

On the cover are artistic renditions of five forest insects and/or their damage. None are drawn to scale. The forest tent 

caterpillar, Malacosoma disstria (Lepidoptera: Lasiocampidae), (blue with white marks on its back) is a very common 

outbreak defoliator of several species of deciduous trees in eastern North America. The red-headed pine sawfly, Neodiprion 

sertifer (Hymenoptera: Diprionidae), is an outbreak species in young forests of various eastern American pines. The pine 

sphinx, Hyloicus pinastri (Lepidoptera: Sphingidae), is a common (green and white-striped) defoliator of pines and spruces in 

Europe. The balsam twig aphid, Mindarus sp. (Homoptera: Aphididae), feeds in the spring on balsam fir buds and shoots. 

The dead, wilted jack pine shoot, is typical damage caused by the white pine weevil, Pissodes strobi (Coleoptera: 

Curculiondiae), a transcontinental species attacking not just pines but also several western spruces in North America. Gyorgi 

Csoka provided a photograph of the pine sphinx for artistic inspiration. Kristine A. Kirkeby is the artist. 

The use of trade, firm, or corporation names in this publication is for the information and convenience of the reader. 

Such use does not constitute an official endorsement or approval by the U.S. Department of Agriculture of any 

product or service to the exclusion of others that may be suitable. Statements of the contributors from outside the 

U.S. Department of Agriculture may not necessarily reflect the policy of the Department. 

North Central Forest Experiment Station 

Forest Service--U.S. Department of Agriculture 

1992 Folwell Avenue 

St. Paul, Minnesota 55108 

Manuscript approved for publication 

1996

DYNAMICS OF FOREST HERBIVORY: QUEST FOR 

PATTERN AND PRINCIPLE 

Editors: 

William J. Mattson 

North Central Forest Experiment Station 

Pesticide Research Center 

Michigan State University 

East Lansing, MI 48824 USA 

Pekka Niemelii 

Faculty of Forestry 

University of Joensuu 

FIN-80101 Joensuu, Finland 

Matti Rousi 

Punkaharju Research Station 

Finnish Forest Research Institute 

FIN-58450 Punkaharju, Finland

PREFACE 

This proceedings is the result of an international symposium that was held February 2-6, 1994 in Maui, Hawaii. It was 

organized under the guidelines of the International Union of Forestry Research Organizations (IUFRO) by the North 

Central Forest Experiment Station of the U.S. Forest Service, and the Finnish Forest Research Institute. Two IUFRO 

Working Parties ($7.01-02, Mechanisms and genetics of plant resistance against insects, and $7.01-03, Mechanisms and 

genetics of plant resistance against mammals) convened for this international conference on the dynamics of herbivory 

and the elucidation of plant defenses against herbivores. 

The proceedings is published and distributed by the USDA Forest Service, North Central Forest Experiment Station in 

recognition and unstinting endorsement of the IUFRO goals of promoting strong world partnerships, particularly in 

forest science, for the sustained management, conservation, and the betterment of the World's forest and woodland 

ecosystems. 

ACKNOWLEDGMENTS 

The editors gratefully acknowledge the generous support of the North Central Forest Experiment Station and the Finnish 

Forest Research Institute for underwriting the costs of organizing, assembling, editing, printing, and distributing this 

publication. The editors also wish to thank all of the particpants whose penetrating and spirited contributions helped 

make this scientific exchange highly worthwhile.

TABLE OF CONTENTS 

Seeking The Rules of Plant-Herbivore Interactions 

l. Why are tree responses to herbivory so variable? 

E. Haukioja and T. Honkanen .............................................................................................................................................. 1 

2. Damage-induced nutritional changes in pine foliage: an overview. 

R Lyytikainen ..................................................................................................................................................................... I1 

3. Stand and landscape diversity as a mechanism of forest resistance to insects. 

T. Schowalter ...................................................................................................................................................................... 21 

4. On neighbor effects in plant-herbivore interactions. 

J. Tuomi, M. Augner, and R Nilsson .................................................................................................................................. 28 

5. Defense theories and birch resistance. 

M. Rousi, J. Tahvanainen, W. Mattson, and H. Sikanen .................................................................................................... 39 

6. A redox-based mechanism by which environmental stresses elicit changes in plant defensive chemistry. 

D.M. Norris ........................................................................................................................................................................ 46 

7. Fungal endophytes: Contrasting effects in trees and grasses. 

S.H. Faeth ........................................................................................................................................................................... 58 

8. Propagating the effects of herbivore attack in an object-oriented model of a tree. 

J. Perttunen, H. Saarenmaa, R. Siev_inen, H. Salminen, A. Pouttu, and J. V_ikev_i............................................................ 60 

Variable Plants and Herbivores: Idiosyncrasies of Individual Cases 

9. Defoliation of Fagus crenata affects the population dynamics of the beech caterpillar, Quadricalcarifera punctatella. 

N. Kamata, Y. Igarashi, and S. Ohara ................................................................................................................................ 68 

10. The impact of flowering on the suitability of balsam fir for the spruce budworm varies with larval feeding behavior. 

E. Bauce, and N. Carisey .................................................................................................................................................... 86 

11. Staminate flowering and tree phenology affect the performance of the spruce budworm. 

W.J. Mattson, B.A. Birr, and R.K. Lawrence ..................................................................................................................... 97 

12. Folivore feeding on male conifer flowers: defence avoidance or bet-hedging? 

T.S. Jensen ........................................................................................................................................................................ 104 

13. The blackmargined aphid as a keystone species: a predator attractor redressing natural enemy 

imbalances in pecan systems. 

M. K. Harris and T. Li ...................................................................................................................................................... 112 

14. Ponderosa pine response to nitrogen fertilization and defoliation by the pandora moth, Coloradia pandora Blake 

(Lepidoptera: Saturniidae). 

B.E. Wickman, R.R. Mason, and H.G. Paul ..................................................................................................................... 118 

15. Phytochemical protection and natural enemies: what regulates a root and stem boring swift moth? 

A.V. Whipple, and D.R. Strong ........................................................................................................................................ 127

Genetic Variation in Efficacy of Plant Resistance Against Herbivores 

16. Resistance of hybrid and parental willows to herbivores: hypotheses and variable herbivore responses over three years. 

R.S. Fritz, B.M. Roche, and C.M. Nichols-Orians ............................................................................................................ 132 

17. FI hybrid spruces inherit the phytophagous insects of their parents. 

W.J. Mattson, R.K. Lawrence, and B.A. Birr .................................................................................................................... 142 

18. Recent advances in research on white pine weevil attacking spruces in British Columbia. 

G.K. Kiss, A.D. Yanchuk, R.I. Alfaro, J.E. Carlson, and J.E Manville ........................................................................... 150 

19. A sugi clone highly palatable to hares: screening for bioactive chemicals. 

H. Hirakawa, T. Nakashima, Y. Hayashi, S. Ohara, and T. Kuwahta .............................................................................. 159 

20. Genetic and phenotypic variation in induced reaction of scots pine, Pinus sylvestris L., to Leptographium wingfieMii: 

reaction zone length and fungal growth. 

F. Lieutier, B. L_ngstr6m, H. Solheim, C. Hellqvist, and A. Yart .................................................................................... 166 

2l. Can phloem phenols be used as markers of scots pine resistance to bark beetles. 

F. Lieutier, F. Brignolas, V. Picron, A. Yart, and C. Bastiern ............................................................................................ 178 

22. Differential susceptibility of white fir provenances to the fir engraver and its fungal symbiont in northern California. 

G.T. Ferrell, and W.J. Otrosina ......................................................................................................................................... 187 

Environmental Variation in Efficacy of Plant Resistance Against Herbivores 

23. Mild drought enhances the resistance of Norway spruce to a bark beetle-transmitted blue-stain fungus. 

E. Christiansen, and A.M. Glosli ...................................................................................................................................... 192 

24. Approaches to studying environmental effects on the resistance of Pinus taeda L. to Dendroctonusfrontalis 

Zimmermann. 

P. Lorio ............................................................................................................................................................................. 200 

25. Effects of root inhabiting insect-fungal complexes on aspects of tree resistance to bark beetles. 

K.F. Raffa, and K. Klepzig ............................................................................................................................................... 211 

26. Douglas-fir and western larch defensive reactions to Leptographium abietinum and Ophiostoma pseudotsugae. 

D.W. Ross, and H. Solheim ............................................................................................................................................. 224 

27. Interruption of bark beetle aggregation by a vigor-dependent Pinus host compound. 

K.R. Hobson ..................................................................................................................................................................... 228 

28. Will global warming alter paper birch susceptibility to bronze birch borer attack? 

R.A. Haack....................................................................................................................................................................... 234 

29. Variations in spruce needle chemistry and implications for the little spruce sawfly, Pristiphora abietina. 

C. Schafellner, R. Berger, J. Mattanovich, and E. Ftihrer .............................................................................................. 248 

30. Acidic deposition, drought, and insect herbivory in an arid environment: Enceliafarinosa and Trirhaba geminata in 

southern California. 

T.D. Paine, R.A. Redak, and J.T. Trumble ........................................................................................................................ 257 

31. The resistance of scotch pine to defoliators. 

V.I. Grimalsky ................................................................................................................................................................... 263

32. Companion planting of the nitrogen-fixing Gliricidia sepium with the tropical timber species, Milicia excelsa, and its 

impact on the gall forming insect, Phytolyma lata. 

M.R. Wagner, J.R. Cobbinah, and D.A. Ofori .................................................................................................................. 264 

33. Functional heterogeneity of forest landscapes: How host defenses influence epidemiotogy of the southern pine beetle. 

R.N. Coulson, J.W. Fitzgerald, B.A. McFadden, RE. Pulley, C.N. Lovelady, and J.R.Gardino ..................................... 272

CONTRIBUTORS and PARTICIPANTS 

ALFARO, R.I., Canadian Forest Service, Pacific Research Centre, Victoria, B.C., Canada 

AUGNER, MAGNUS, Depamnent of Ecology, Theoretical Ecology, Lund University, T Ecology Bldg., S-22362 Lund, 

Sweden 

BASTIEN, C., Station d'Am61ioration des Arbres Forestiers, Institut National de la Recherche Agronomique, Ardon, 45160 

Olivet, France 

BAUCE, l_,.,D6partement des sciences foresti&es, Facult6 de Foresterie et de G6omatique, Universit6 Laval, Ste-Foy 

(Qu6bec) G1K 7P4, Canada 

BERGER, R., Institute of Forest Entomology, Forest Pathology, and Forest ProtectionUniversit_it ftir Bodenkultur, 

Hasenauerstrasse 38, A 1190 Vienna, Austria 

BIRR, BRUCE A., USDA Forest Service, North Central Forest Experiment Station, B-7 Pesticide Research Center, Michigan 

State University, East Lansing, M148824, USA 

BOWERS, WADE W., Canadian Forest Service, Natural Resources Canada, P.O. Box 6028, St. John's, N.F. A 1C 5X8, 

Canada 

BRIGNOLAS, E, Station de Zoologie Foresti_re, Institut National de la Recherche Agronomique, Ardon, 45 160 Olivet, 

France 

CARISEY, W., D6partement des sciences foresti_res, Facult6 de Foresterie et de G6omatique, Universit6 Laval, Ste-Foy 

(Qu6bec) G1K 7P4, Canada 

CARLSON, J.E., University of British Columbia, Biotechnology Laboratory, Vancouver, B.C. 

CATES, REX G., Department of Botany, 425WIDB, Director, Chemical Ecology Laboratory, Brigham Young University, 

Provo, UT 84602 

CHRISTIANSEN, ERIK, Norwegian Forest Research Institute, N- 1432 As, Norway 

COBBINAH, JOSEPH R., Forestry Research Institute of Ghana, UST P.O. 63, Kumasi, Ghana 

COULSON, ROBERT N., Knowledge Engineering Laboratory, Departments of Entomology and Geography, Texas A&M 

University, College Station, TX 77843 

FAETH, STANLEY H., Department of Zoology, Arizona State University, Tempe, Arizona 85287-1501, USA 

FERRELL, GEORGE T., USDA Forest Service, Pacific Southwest Research Station, 2600 Washington Avenue, Redding, 

California 96001, USA 

FITZGERALD, JEFFREY W., Knowledge Engineering Laboratory, Departments of Entomology and Geography, Texas 

A&M University, College Station, TX 77843 

FRITZ, ROBERT S., Department of Biology, Box 133, Vassar College, Poughkeepsie, NY 12601, USA 

F_)HRER, E., Institute of Forest Entomology, Forest Pathology, and Forest Protection, Universit_it ftir Bodenkultur, 


GIARDINO, JOHN R., Knowledge Engineering Laboratory, Departments of Entomology and Geography, Texas A&M 


GLOSLI, ANNE MARIE, Norwegian Forest Research Institute, N- 1432 As, Norway

GRIMALSKY, ¥.I., Research Institute of Forestry, Gemol, Belorussia 

HAACK, ROBERT A., USDA Forest Service, North Central Forest Experiment Station, 1407 S. Harrison Road, East 

Lansing, MI 48823, USA 

HARRIS, MARVIN K., Department of Entomology, Texas A&M University, College Station, TX 77843 

HAUKIOJA, ERKKI, Laboratory of Ecological Zoology, Department of Biology, University of Turku, FIN-20500 Turku, 

Finland 

HAYASHI, YOSHIOKI, Forestry and Forest Products Research Institute, PO Box 16, Tsukuba-Norin, Ibaraki 305, Japan 

HELLQVIST, C., Division of Forest Entomology, Swedish University of Agricultural Sciences, S-77073 Garpenberg, 

Sweden 

tlELSON, BLAIR, Forest Pest Management Institute, Canadian Forest Service, Natural Resources Canada, Sault Ste. Marie, 

Ontario, P6A 4J, Canada. 

HERMS, DANIEL A., The DOW Gardens, 1018 W. Main Street, Midland, MI 48640 

HIRAKAWA, HIROFUMI, Forestry and Forest Products Research Institute, PO Box 16, Tsukuba-Norin, Ibaraki 305, Japan 

HOBSON, KENNETH R., Department of Entomology, 345 Russell Labs, University of Wisconsin, Madison, WI 53706, 

USA 

HONKANEN, TUI,IA, Laboratory of Ecological Zoology, Department of Biology, University of Turku, FIN-20500 Turku, 

Finland 

IGARASHI, YUTAKA, Laboratory of Forest Entomology, Tohoku Research Center, Forestry and Forest Products Research 

Institute, Nabeyashiki 72, Shimokuriyagawa, Morioka, Iwate 020-01, Japan 

ISAEV, ALEX, International Forestry Institute, Udalcov Str. 24.87, Moscow, Russia 

JENSEN, THOMAS SECHER, Institute of Biological Sciences, Department of Zoology, Aarhus University, DK-8000 

Aarhus C, Denmark 

KAMATA, NAOTO, Laboratory of Forest Entomology, Tohoku Research Center, Forestry and Forest Products Research 

Institute, Nabeyashiki 72, Shimokuriyagawa, Morioka, Iwate 020-01, Japan 

KISS, GYULA K., B.C. Forest Service, Kalamalka Forestry Centre, 3401 Reservoir Road, Vernon, B.C. V1B 2C7 

KLEPZIG, KIER D., Department of Plant & Soil Science, Southern University and A&M College, Baton Rouge, LA 

70813, USA 

KUMAPLEY, PHILOMENA A., Institute of Renewable Natural Resources, University of Science & Technology, Kumasi- 

Ghana, W. Africa 

KUWAItATA, TSUTOMU, Forestry and Forest Products Research Institute, PO Box 16, Tsukuba-Norin, Ibaraki 305, Japan 

LANGSTRt_M, B., Department of Plant & Forest Protection, Swedish Univ. of Agric. Science, Box 7044, S-75007 Uppsala, 

Sweden 

LARSSON, STIG, Department of Plant & Forest Protection, Swedish Univ. of Agric. Science, Box 7044, S-75007 Uppsala, 

Sweden 

LAWRENCE, ROBERT K., USDA Forest Service, North Central Forest Experiment Station, B-7 Pesticide Research 

Center, Michigan State University, East Lansing, MI 48824, USA

LI, T., Department of Entomology, Texas A&M University, College Station, TX 77843 

LIEUTIER, E, Station de Zoologic Foresti?_re, Institut National de la Recherche Agronomique, Ardon, 45160 Olivet, France 

LORIO, PETER L., Jr., USDA Forest Service, Southern Forest Experiment Station, 2500 Shreveport Highway, Pineville, 

LA 71360, USA 

LOVELADY, CLARK N., Knowledge Engineering Laboratory, Departments of Entomology and Geography, Texas A&M 


LIEUTIER, F., Station de Zoo[ogie Foresti6re, Institut National de la Recherche Agronomique, Ardon, 45160 Olivet, France 

LYYTIK,g, INEN, P)/,IVI, Finnish Forest Research Institute, Department of Forest Ecology, EO. Box 18, FIN-01301 Vantaa, 

Finland 

MANVILLE, J.E, Canadian Forest Service, Pacific Research Centre, Victoria, B.C. 

MASON, R.R., Forestry and Range Sciences Laboratory, 1401 Gekeler Lane, La Grande, OR 97850 

MATTANOVICH, J., Institute of Forest Entomology, Forest Pathology, and Forest ProtectionUniversit_it fi.irBodenkultur, 


MATTSON, WILLIAM J., USDA Forest Service, North Central Forest Experiment Station, B-7 Pesticide Research Center, 

Michigan State University, East Lansing, MI 48824, USA 

McFADDEN, BRYAN A., Knowledge Engineering Laboratory, Departments of Entomology and Geography, Texas A&M 


NAKASHIMA, TADAKAZU, Forestry and Forest Products Research Institute, PO Box 16,Tsukuba-Norin, Ibaraki 305, 

Japan 

NIEMEL,_, PEKKA, Faculty of Forestry, University of Joensuu, FIN-80101, Joensuu, Finland 

NILSSON, PATRIC, Department of Ecology, Theoretical Ecology, Lund University, T Ecology Bldg., S-22362 Lund, 

Sweden 

NORRIS, DALE M., Department of Entomology, 642 Russell Laboratories, University of Wisconsin, Madison, WI 53706, 

USA 

OFORI, DANIEL A., Forestry Research Institute of Ghana, UST EO. 63, Kumasi, Ghana 

OHARA, SEIJI, Laboratory of Chemical Conversion, Wood Chemistry Division, Forestry and Forest Products Research 

Institute, Matshunosato 1, Kukizaki, Inashiki, Ibaraki 305, Japan 

ORIANS, COLIN M., Department of Biology, Williams College, Williamstown, MA 01267, USA 

OTROSINA, W.J., USDA Forest Service, Southeastern Forest Experiment Station, 320 Green Street, Athens, Georgia 

30602, USA 

PAINE, 'I.D., Department of Entomology, University of California, Riverside, CA 92521 

PAUL, H.G., Forestry and Range Sciences Laboratory, 1401 Gekeler Lane, La Grande, OR 97850 

PERTTUNEN, JARI, Finnish Forest Research Institute, Unioninkatu 40A, SF-00170, Helsinki, Finland 

PICRON, V., Station de Zoologie Foresti_re, Institut National de la Recherche Agronomique, Ardon, 45160 Olivet, France

POUTTU, ANTTI, Finnish Forest Research Institute, Unioninkatu 40A, SF-00170, Helsinki, Finland 

PULLEY, PAUL E.. Knowledge Engineering Laboratory, Departments of Entomology and Geography, Texas A&M UI_iversity, 

College Station, TX 77843 

QUIRING, DANIEL T., Pop. Ecol. Group, Department of Forest Resources, Univeristy of New Brunswick. Fredericton, 

N.B. E3B6C2 

RAFFA, KENNETH f'., Department of' Entomology, 642 Russell Laboratories, University of Wisconsin, Madison, Wt 

53706. USA 

REDAK, R.A., Depamnent of Entomology, University of California, Riverside, CA 92521 

ROCHE, BERNADETTE M., Department of Biology, Vassar College, Poughkeepsie, NY 12601, USA 

ROSS, DARRELL W., Department of Forest Science, Oregon State University, Corvallis, Oregon 97331-7501. USA 

ROUSI, MATTI, The Finnish Forest Research Institute, Punkaharju Research Station, SF-58450 Punkaharju, Finland 

SAARENMAA, ttANNU, Finnish Forest Research Institute, Unioninkatu 40A, SF-00170, Helsinki, Finland 

SALMINEN, HANNU, Finnish Forest Research Institute, Unioninkatu 40A, SF-00170, Helsinki, Finland 

SCHAFELI,NER, CHRISTA, Institute of Forest Entomology, Forest Pathology, and Forest Protection, Universit;:it f/Jr 

Bodenkultur, Hasenauerstrasse 38, A 1190 Vienna, Austria 

SCHOWAI2FER, TIM D,, Department of Entomology, Oregon State University, Corvallis, Oregon 97331-2907, USA 

SIEVANEN, R|STO, Finnish Forest Research Institute, Unioninkatu 40A, SF-(X) IT0, Helsinki, Finland 

SIKANEN, ttANNI, The Finnish Forest Research Institute, Punkaharju Research Station, SF-58450 Punkaharju, Finland 

SOLHEIM, HAI,VOR, Norwegian Forest Research Institute, Section of Forest Ecology, Division of Forest Pathology, N- 

1423 As-NLH, Norway 

STRONG, DONALD R., Bodega Marine Laboratory, University of California, Box 247, Bodega Bay, CA 94923 

TAttVANAINEN, JORMA,University of Joensuu, Department of Biology, P.O. Box 111, SF-80101 Joensuu, Finland 

TRUMBLE, J,T., Department of Entomology, University of California, Riverside, CA 92521 

TUOMI, JUHA, Department of Ecology, Theoretical Ecology, Lund University, T Ecology Bldg., S-223 62 Lund, Sweden 

VAKEVA, JOUNI, Finnish Forest Research Institute, Unioninkatu 40A, SF-00170, I-telsinki, Finland 

WAGNER, MICIIAEL R., School of Forestry, Northern Arizona University, RO. Box 15018, Flagstaff, AZ 86011, USA 

WHIPPLE, AMY V., Bodega Marine Laboratory, University of California, Box 247, Bodega Bay, CA 94923 

WICKMAN, B.E., USDA Forest Service, Silviculture Laboratory, 1027 NW Trenton Avenue, Bend, OR 97701 

YANCHUK, A.D., B.C. Forest Service, Research Branch, 31 Bastion Square, Victoria, B.C. V8W 3E7E 

YART, A., Station de Zootogie Forestibre, Institut National de la Recherche Agronomique, Ardon, 45160 Olivet, France 

YOSHIKAWA, KEN, Fac. of Agriculture, Okayama Univ., Tsushima-naka 1- l- 1, Okayama 700, Japan

WHY ARE TREE RESPONSES TO HERBIVORY SO VARIABLE? 

ERKKI HAUKIOJA and TUIJA HONKANEN 

Laboratory of Ecological Zoology, t)epartment of Biology, University of Turku 

FIN-20500 Turku, Finland 

INTRODUCTION 

Following defoliation the quality of tree leaves can become 'worse, better, or remain unchanged for herbivores. To some 

extent, such variation results from tile different types of foliage which are studied. For example, after severe defoliation, 

trees may reflush in the same season, but these new leaves differ from normal, mature foliage, and leaves which grew at the 

time of partial defoliation may become different fi'om leaves in an undamaged tree. Similarly, defoliated and undefoliated 

trees may produce qualitatively different foliage in the next growth season. The fk)liage produced in the year after defoliation 

occurred has been found to be especially different from that of undefoliated trees both in chemical traits, and herbivore 

bioassays, as is the case for birch (Neuvonen and Haukioja 1991), larch (Benz 1974), and oak (Rossiter et al. 1988). All of 

_:hese trees are deciduous and there seems to be a general trend that defoliation alters the quality of deciduous foliage more 

noticeably than that of evergreen f_liage. 

We have studiec/defoliation-induced responses in mountain birch, Betula pubescens ssp tortuosa, and Scots pine, Pinus 

s)'Ivestris, i.e., in a deciduous species responding strongly to herbivory (Haukioja et al. 1985), and in an evergreen species 

tending toward the other end of the continuum (Niemelfi et al. 1984, 1991, Watt et al. 1991). In this paper, we briefly review 

the treatments which are beneficial (induced amelioration, IA hereafter) or harmful (induced resistance, IR hereafter) to 

herbivores. Furthermore, we discuss the validity of several hypotheses which attempt to explain how and why trees respond 

to herbivory as they do: the carbon-nutrient (C/N) balance hypothesis (Bryant et al. 1983, 1991; Tuomi et al. 1991), the 

growth--differentiation (G/D) balance hypothesis (Herrns and Mattson 1992), and the sink-source (S/S) hypothesis (Haukioja 

1991b, Honkanen et al. 1994, Honkanen and ttaukioja 1994). 

TIlE TARGET TISSUE AND TIMING OF THE DAMAGE 

Whether IA or IR is elicited after damage does not depend upon the species or individual concerned. The same tree 

may become either better or worse depending on the type of tissue which was damaged (Haukioja et al. 1990), and also the 

timing of damage (Neuvonen et al. 1988). 

IR can be separated into rapid (RIR) and delayed (DIR) induced resistance (Haukioja 1982). In the former, poor diet 

quality affects the generation of herbivores which caused the damage, while the latter influences the next generation(s) of 

herbivores. Note that this division is relevant as regards the consequences of the damage on herbivore populations dynamics: 

RIR is a stabilizing and DIR a destabilizing factor because it introduces time lags. Furtherrnore, the division into RIR and 

DIR is meaningful only in the context of a herbivore generation time; a certain response by the plant may be experienced as 

DIR by a short-lived herbivore species but as RIR by a long-lived one. From the plant's point of view these rapid and 

delayed responses do not necessarily differ, A corresponding distinction between rapid and delayed responses also applies to 

induced amelioration. 

The mountain birch system provides an example of how herbivore-induced damage results in a multitude of changes 

in foliage quality. Poor quality leaves are produced after laminae of growing or newly expanded leaves are damaged by real 

or simulated herbivory. "['he RIR in birches tends to be mild when measured by its effects on herbivores (Haukioja and 

Neuvonen 1987), while DIR may severely curtail the potential increase of the herbivore population (Haukioja et al. 1985). 

Attempts to trigger RIR or DIR by damaging mature, late summer foliage have led to mild or no responses at all (Neuvonen 

et al. 1988, Tuomi et al. 1988b, Hanhim_iki and Senn 1992). On the other hand, the effects of early season damage may 

Mattson, W.J., Niemel_,1, R, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 

For. Serv. Gen. Tech. Rep. NC-183, N.C. For. Exp. Sta., St. Paul, MN 55108.

persist in mature leaves and affect subsequent herbivores, even from different guilds (Neuvonen et al. 1988, HanhimS.ki 

1989). The delayed effects (DIR) can remain effective on folivores consuming mature leaves produced the following 

summer (Hanhimfiki 1989). Induced resistance in mountain birch foliage is triggered most effectively by damage at the time 

when there is peak feeding by the most destructive insect pest of mountain birch, the geometrid, Epirrita autumnata (Tenow 

1972, Haukioja et al. 1988). Since defoliation tends to effect all members of the birch chewing guild similarly (Hanhim_iki 

1989), the outcomes of defoliation in birch foliage seem to represent a general response, and are not specifically targeted 

against Epirrita autumnata or other early season defoliators. 

Induced amelioration in mountain birch is triggered either by removal of the apical tips of dormant twigs (Haukioja et 

al. 1990) or by damage to expanding buds or tiny leaves (Senn and Haukioja 1994). In practice, the former type of damage 

may result from winter browsing by mammals and the latter from insect damage very early in the season. Winter browsing 

makes the forthcoming foliage more preferable for a suite of different types of herbivores (Danell and Huss-Danell 1985), 

although the effects on insect performance tend to be mild (Neuvonen and Danell 1987, Haukioja et al. 1990). Larvae of 

Epirrita autumnata start feeding at, or even before, budbreak, and are known to destroy apical buds from branches (Haukioja 

et al. 1990). Early leaf damage might affect foliage quality in the same year while bud removal influences growth in the 

following season. In practice this means that depending on the type of target tissue and timing of the damage, larvae of 

Epirrita autumnata are potentially capable of inducing either IA, RIR or DIR in birch foliage (Haukioja 1991a). 

The case with the pine is different. Most studies conducted on pines have not reported strong qualitative changes in 

needle chemistry after defoliation and, accordingly, the effects on insect performance also have been mild or insignificant 

(e.g., NiemelS. et al. 1991, Watt et al. 1991, Reich et al. 1993). However, some studies have revealed defoliation-induced 

chemical changes in pines. Wagner and Evans (1985) reported that defoliation of Pinus ponderosa seedlings increased the 

production of both phenols and proteins in mature and immature foliage. Honkanen and Haukioja (1994) demonstrated 

increased nitrogen needle concentration after severe defoliation of Scots pine. These results indicate that the outcomes of 

defoliation in pines can vary substantially, and so are the consequences on herbivores. 

Pruning or removal of pine buds or twigs causes more lush growth of needles and shoots (e.g., Nuorteva and Kurkela 

1993, Honkanen et al. 1994), just like in mountain birch. Pine browsers, particularly moose, are known to cause IA 

(L6yttyniemi 1985), but it is not known whether insect herbivores of pine are able to induce such a response. 

In summary, both in the case of mountain birch and Scots pine it is possible to induce variable effects - better or 

worse foliage quality, or no change at all - by selectively damaging different types of tissues and by varying the timing of the 

treatment. 

THE RESOURCE STATUS OF BIRCH AND PINE FOR INDUCED RESPONSES: 

ADEQUACY OF THE C/N HYPOTHESIS 

The simplest explanation for the induced resistance is that chewing or browsing of plant biomass removes nutrients 

and such losses lower foliage quality for later herbivores (Tuomi et al. 1984). If the loss of nutrients is critical for induced 

resistance, the tree's status can be remedied by fertilization which usually increases the nutrient concentration in foliage. The 

crucial question here is whether fertilization decreases the resistance of the tree and cancels any negative effects of defoliation. 

In the case of the mountain birch, Haukioja and Neuvonen (1985) reported that nitrogen fertilization did not alleviate 

the detrimental effects of defoliation (DIR) on Epirrita larvae. This contradicts the predictions of the carbon-nutrient balance 

(C/N) hypothesis, which ascribes the resistance of deciduous trees to carbon-based secondary compounds which generally 

decrease when trees obtain extra nutrients. However, Bryant et al. (1993) found that fertilized paper birches, Betula 

papyrifera, lost their ability for DIR. They criticized the mountain birch experiment by Haukioja and Neuvonen (1985) for 

simultaneous application of the fertilizing and defoliation treatments; the trees might not have had time to benefit from 

increased nutrient availability. However, another experiment with the mountain birch-Epirrita system (Ruohom_iki et al. in 

press), accounted for the criticism by Bryant et al. (1993) and yet reported the same result of Haukioja and Neuvonen (1985). 

Furthermore, the strength of the DIR did not decline with shading, as predicted by the C/N hypothesis (Ruohom_iki et al. in 

press). It is possible that mountain birch and paper birch behave differently, or the insects used in the two experiments 

showed differential tolerance to the altered carbon-based compounds, or that changes in carbon and nitrogen were not crucial 

for the herbivore. In any case, it is obvious that for birches the C/N hypothesis did not provide the general explanation for 

induced responses. 

2

The C/N hypothesis does not adequately predict fertilization induced changes in pine foliage, either. Honkanen et al. 

(1994) and Hor_kanen and Hauki_ja (1994) demonstrated that branches in fertilized Scots pi_es produced more and bigger 

needles irrespective of whether the branch or the whole tree was del'\)liated or not. This is hard to reconcile with tile predic- 

tion that carbon deficiency after defoliation restricts growth of foliage in evergreens (e.g., Bryant et al. I983, Tuomi et al. 

19882). Furthermore, fertilization does not generally decrease, and may even increase, allocation to carbon-based secondary 

compounds in Scots pine (Bj/.Srkman et al. 199t, Edenius 1993). 

in s_._mmary, the C/N hypothesis does not adequately explain the effects of changes in the availability of mineral 

nutrients or shade for DIR in birch, or the former in pine. Sit-_ce the growth-differentiation balance (G/D) hypothesis is based 

on the C/N hypothesis with respect to resource availability, we doubt the generality of both these to explain herbivore- 

induced responses of trees on the basis of resource availability (Iason and Hest,er 1993). 

THE FUNCTIONAL ORGANIZATION OF THE TREE 

Next we briefly discuss the functional orga_aization of trees to demonstrate that some very basic physiological 

mechanisms behind tree structure and function actually contribute to a simple, ahnost mechanistic explanation for induced 

responses of: birch. The same basic explanation may also contribute to understanding the differences between responses of 

the deciduous mountain birch and the evergreen Scots pine.. The tkfllowing reasoning forms the basis of tile source-sink (S/S) 

hypothesis. First, trees are modular organisms. Usually this statement is interpreted as a description of the morphological 

desigrl of trees: they are built of repetitive, multicellular units comprising a stem segment, leaf and meristem (White 19791). 

This emphasis does not help much in explaining tree functions although it vaguely makes understandable that tree parts may 

respond semioindependently, contrary to parts of unitary organisms like humans or fruit flies (Vuorisalo and Tuomi 1986). 

The term integrated physiological unit (IPU) was coined for this purpose (Watson and Casper 1984, Watson 1986). That 

concept emphasizes more the independence of IPUs than the integration among IPU's. However, it is the latter that makes a 

tree function like one individual. Haukioja (1991b) proposed that a simple set of rules can generally determine the integrity 

of tree functions and help not only to understand the relative independence of IPU's but also the temporal and spatial variabil- 

ity in their independence. Furthermore, it explicitly acknowledges other parts of the tree besides the canopy, i.e., the stem 

and roots. 

Integration within an individual tree simply requires that IPUs at the "best" locations gain control over other, less 

optimally located IPUs. Within the canopy, this happens if the apical rneristems, which have the best access to light, remain 

or become metabolically active and simultaneously suppress nearby, less tavorably located, meristems (Haukioja 1991b, 

Sachs e,ral. 1993). For this outcome, resources should be preferentially transported to the winning meristems. This takes 

place by their elevated activity relative to that of nearby meristems; high meristematic activity is coupled with the production 

of pertinent growth hormones which, in turn, preferentially direct the flow of resources to these strong meristems. We 

contend that this incidentally creates the integration among above-ground parts of the plant. Integration within a tree also 

varies seasonally (Sprugel et al. 199t). We assume this results from the existence of other meristems than apical ones in the 

canopy. Trees have lateral meristems in branches, trunks and roots. When the sink strength of these exceeds that of apical 

meristems, resources are preferentially conveyed to the trunk and to the below-ground parts. When this happens, the whole 

tree is well integrated, and the morphologically defined IPU's have temporally lost their functional semi-independence. 

The responses of IPU's ultimately are based on genetic rules which are facultative and subject to position-dependent 

influences. Thus, irrespective of the similar genetic basis, the behavior of an IPU depends on its position within the plant 

(which takes into account the possible dominance of other, especially nearby, IPUs) and on the external environment. The 

latter includes availability of resources. Therefore, the potential and the actual behavior of an IPU may be constrained by 

either internal (genetic and position-dependent) or external (resources, competition) factors, or by both. The production of a 

tree-like architecture requires that the genetically determined facultative and position-dependent rules include instructions for 

achieving, maintaining and relinquishing dominance among competing IPUs. However, we assume that, within the limits of 

the genetically based conditional and position-dependent rules, competition among IPUs within an individual plant is 

physiologically genuine yet these rules still transtk_rm the population of IPUs into an integrated "individual plant". 

Variance in the strength of apical dominance (in addition to traits like dormancy/activity of n_teristems, internode 

length and branching angle, Suthertand and Stillman 1990) explains a lot of differences among different tree types. Since the 

relative strength of apical and other meristems has a hereditary basis (Kozlowski 1971), characteristics like the shape of the 

tree (resulting from the intermodular interactions) are susceptible to artificial and natural selection. 

3

THE SINK-SOURCE HYPOTHESIS: TYPE AND TIMING OF DAMAGE ARE CRUCIAL 

A crucial point underlying variable responses after damage to different types of tissues is that different meristems are 

provisioned seasonally at different times (Fig. 1). We claim that such a predictable temporal variation in allocation of plant 

resources to different sinks, as well as shifts of individual meristems from sinks to sources within the course of the season, 

provides an important explanation for the variability in damage-induced responses of birch and pine foliage (Table 1). 

MOUNTAIN BIRCH 

Early spring sink leaves < _-- resource pool 

Early summer young source leaves -> local sinks 

Late summer mature source leaves > resource pool 

SCOTS PINE 

Early spring source leaves ---- _ resource pool 

Early/late summer apical source leaves _ local sinks 

basal source leaves-- > resource pool 

Figure l.--The assumed seasonal variation in resource flows in the mountain birch and Scots pine. 

Flushing leaves are strong sinks, and we assume that their success in drawing resources depends on their sink 

strength. Within few days young leaves shift from sinks to sources (Valanne and Valanne 1984) and start to preferentially 

provision nearby buds which will contain the leaf primordia for the next season. Simultaneously the leaves themselves 

mature, and when the growth of buds in short shoots, and of meristem-supporting leaves has been completed, we assume 

lateral meristems and meristems in the root system to be the dominant sinks tbr resources. 

Damage to Physiological Sinks in Birch 

Consequences of damage to sink organs depend on the strength and the position of the sinks. Sinks can be eliminated 

or weakened by clipping the tips of dormant shoots, by removing the apical dormant buds only (Haukioja et al. 1990), or, 

later in spring, by damaging flushing apical leaves which still are in the strong sink phase (Senn and Haukioja 1994). This 

releases extra resources for normally suppressed basal meristems (Lehtil_ et al. ms.). The key is the apical dominance and 

not the amount of nutrients lost in the damage because when small but equal amounts of basal, instead of apical, meristems 

were removed, no changes were found in the extant foliage (Haukioja et al. 1990). 

Damage to Physiological Sources in Birch 

The consequences of damage to birch source leaves depend on the meristems which the leaves were feeding at the 

time of the damage (and which they would have provisioned had the damage not taken place). Damage to young source 

leaves reduces the sink strength of meristems in the leaf and stem primordia which the relevant leaves were feeding at the 

moment of the damage. For birch long-shoot leaves, those meristems are located in their axillary buds and for shorz-shoot 

leaves (which form the vast majority of all leaves in mature mountain birches) they are in short shoot buds from which the 

next season leaves flush. Both suffer significantly from even very limited local damage; the shoot from the axillary bud a of 

damaged long shoot leaf grows short the next year (Haukioja et al. 1990, Ruohom_iki et al. ms). Similarly, leaves as well as 

4

Table l.---Typical changes in %liage growth vigor, nitrogen concentration and the amount of secondary compounds after 

different types of damage. + = increase, 0 = no change and - = decrease in the studied character. 

Damaged tissue Growth Nitrogen Secondary Reference 

vigor compounds 

BIRCH 

Buds/flushing leaves + + 1,2 

Young source leaves - - + 2,3,4,5 

Mature source leaves 0 0 0 5,6 

SCOTS PINE 

Buds/flushing leaves + + 9 7,8 

Source needles 

feeding current 

year foliage 0,+ ;' 0 7,9,10,11 

Source needles 

feeding general 

stores 0 0 0 9,10 

l) Danell et al. (unpblished data); 2) Niemelfi er al. 11979;3) Haukioja er al. 1985; 4) Haukioja er al. 1990; 5) Tuomi er al. 

1988; 6) Hartley and Lawton 1991; 7) Honkanen et al. 1994; 8)Lgmgstr6m e/al. 1990; 9) Ericsson er al. 1980; 10) Honkanen 

and Haukioja ( 1994); 11) Niemel/i et al. 1991 

_Depends on the scale of defbliation (Honkanen and Haukioja 1994) 

catkins in short shoots, whose leaves were damaged the previous year, remain smaller than in undamaged shoots (Ruohom_iki 

egal. ms.). Since these outcomes follow from damage to even a single leaf, reduction of the total resource pool of the tree 

cannot be the relevant explanation in such a case. Instead, the meristems fed by damaged leaves presumably were unable to 

efficiently compete tot resources in the common pool (Ruohom_iki er at. ms). 

Damage to late season foliage does not affect the sink strength of the following year's buds because they had been 

completed earlier. Instead, this may :reduce stored resources. 

The above scenario offers the simplest explanation for DIR: early summer defoliation during the previous growth 

season led to "weak" buds which, when still in the sink phase, were unable to get their normal load of nutrients and photosynthates. 

The poor quality of such foliage for herbivores may follow both from reduced amounts of nutrients and from increased 

secondary compounds. The same basic explanation might apply to local RIR, too: when foliage is damaged in a 

branch, the developing foliage loses part of its sink strength, and the whole branch might therefore become a poorer competitor 

of resources compared to other branches. 

Damage to Scots Pine Sinks and Sources 

In Scots pine, damage to apical buds has similar effects as in birch (Table 1), and we assume that these effects also 

can be explained in the same way (Honkanen et al. 1994). However, the effects of defoliation are more complex than in the 

case of the mountain birch. This can be explained by the simultaneous existence of several needle generations which are in 

different positions on a shoot, and which partially feed different meristems (Fig. 1). Defoliation of apical needles in the oneyear-old 

needle-class that supply the developing current year foliage early in the season, decreases the growth of new shoots. 

However, defoliation of older needles at the same time does not affect the growth of new shoots, presumably because older 

needles preferentially supply lateral meristems of the tree (Ericsson 1978, Honkanen and Haukioja MS). In contrast to birch, 5

late season defoliation decreases the growth potential of next year's pine needles more than early season defoliation 

(Honkanen et al. 1994). This presumably relates to the long developmental time of pine buds (Agren and Axelsson 1985) 

and that defoliation removes the local storage in pine needles. We assume that late season defoliation of birch mainly hinders 

provisioning of the general resource pool (Fig. 1). 

In conclusion, both the type of the damaged tissue andtiming of the damage contribute to the post-defoliation 

responses of mountain birch and Scots pine. The important points are whether the damaged tissue was a physiological sink 

or a source, and which sinks the damaged foliage was provisioning at the time of the damage. Thus IA, DIR, RIR, as well as 

local changes in growth activity, can all be understood as outcomes of the same basic explanation. 

DISCUSSION 

The system of sink-source gradients is part of the mechanism regulating the allocation of resources within a plant. 

Therefore, it offers a basis for explanations about spatial and temporal patterns of resources, and therefore also for chemical 

compounds which are potentially important for herbivores and about changes in these compounds after herbivory. This is not 

to say that we regard resource availability unimportant; the availability of carbon and mineral nutrients obviously represents 

modifying and constraining factors for tree-herbivore interactions. Among other things, resources modify sink strengths of 

meristems and plant apical dominance (Cline 1991). 

The G/D and, especially, the C/N hypotheses, emphasize plant quality and different ways that shortages of resources 

may cause chemical changes which are assumed to be important for herbivores. The S/S hypothesis, instead, emphasizes the 

ability of meristems to manufacture new biomass via spatially and temporally variable allocation of resources to specific 

meristems. Limitations in resource allocation may alter foliage quality either via the primary or the secondary leaf chemistry, 

or, most probably, both. The SIS hypothesis does not make a priori predictions of the importance of these chemical changes 

for herbivores. 

We assume that the basic explanation for DIR, and perhaps for localized RIR, is the weakened state of meristems. 

This leads to a lower competitive ability in the plant sink-source system. Altered resistance is an automatic outcome of these 

changes. There are four aspects of induced tree responses which support this. First, defoliation/clipping-induced changes in 

plant growth activity are basically local phenomena. This also refers to the largely unspecific changes in primary and 

secondary chemistry which produce effects classified as RIR, DIR and IA in trees like birches. This clearly contrasts with 

some damage-induced systemic responses which effectively spread within and among plants (Walker-Simmons and Ryan 

1977, Farmer and Ryan 1990, Takabayashi et al. 1991). They concern well known and obviously strictly defensive traits like 

production and transfer of the proteinase inhibitor inducing factor (PIIF) e.g. in many herbaceous plants (McFarland and 

Ryaa 1974), or the resin duct system of conifers (e.g., Blanche et al. 1992). Second, the responses of birch foliage quality to 

defoliations or clippings is of a general nature and affects all species of chewing insect herbivores tested so far in a similar 

way (Hanhimfiki 1989). Third, the existence of IA demonstrates that reasons other than defense (or recovery) must be sought 

when interpreting plant responses to herbivory. Such seemingly nonadaptive responses are real and presumably reflect 

constraints in plant design. Fourth, birch and pine are surprisingly insensitive to the amount of lost biomass: limited as well 

as intensive defoliations cause fairly similar degrees of responses (Haukioja and Neuvonen 1987, Honkanen and Haukioja 

1994). This is consistent with sink strength regulated photosynthesis which is common in plants (e.g., Wardlaw 1990). All 

these aspects indicate that defoliation-induced changes in carbon-based potentially defensive chemistry, as well as in the 

primary chemistry, have explanations other than defense, or the relative availability of different types of resources. Note that 

acceptance of this explanation does not deny the importance of plant chemistry on herbivores. However, the question of 

whether induced tree resistance, measured as a decrease in insect performance after previous foliar damage, is a true defense 

or not, is outside the scope of this paper. 

The S/S hypothesis has some obvious practical implications. For instance, branch or ramet specific responses of trees 

to defoliations (Haukioja et al. 1983, Tuomi et al. 1988b, L_ngstr6m et al. 1990, Hanhim_iki and Senn 1992, Honkanen et al. 

1994, Honkanen and Haukioja 1994) become easily comprehensible. This fact has made it possible to use individual ramets 

of the same multi-stemmed tree as units on which different treatments have been applied. The question of whether such 

branch-specific defoliations are as effective in inducing induced resistance as more extensive tree-wide defoliations has an 

ahnost paradoxical answer: at least in pine, branches show stronger, not weaker, responses after branch-wide than after treewide 

defoliations (Honkanen and Haukioja 1994). This result agrees with the S/S hypothesis and, to our understanding, 

6

cannot be derived from any other hypothesis: within the canopy, competition of resources among meristems puts a single 

branch with defoliation-weakened meristems into a poorer relative position if all other branches are intact than if they also are 

i weakened. Second, the S/S hypothesis provides a simple explanation for the multiannual "memory" of DIR (Haukioja el al. 

i 988): it might simply result from the time :needed to restore the vigor of meristems. Furthermore, the implications of the S/ 

S hypothesis also make it worthwhile to study whether plants have evolved mechanisms to prevent herbivores causing IA. 

For instance, it has been reported that developing current year foliage of Scots pine produces 13-keto-8(14)-podocarpen-18oic 

acid which deters pine sawflies and decreases use of the young current year foliage (Niemel_i et al. 1982). A strong 

deterrence of herbivores from such current year foliage may reduce the danger of induced amelioration. 

Perhaps the most important empirical message of the S/S hypothesis is a call for extreme caution in damage simulations. 

For instance, when triggering the DIR in the foliage of the mountain birch, even small variations in timing (perhaps a 

day or two) of defoliations may shift the treatment from sink leaves to source leaves. That may alter the response from IA to 

RIR or DIR, i.e., even the direction of the response. We doubt whether any study published so far has reported both the 

timing of defoliations in relation to tree phenology, and the exact defoliation practice, in such detail that it would allow the 

treatments to be precisely repeated. Actually, we are lacking a proper classification of damages to plants. Therefore, it is not 

astonishing that reports of plant responses to defoliations are notoriously variable (Lawton 1991). Although insects and 

plants may differ with reference to species, populations and even individuals, we regard the main problem the superficial 

knowledge of the system under study: we have to understand and to be able to record what we are doing with our own trees 

when we simulate herbivore damage. 

SUMMARY 

The debate about damage-induced changes in food plant quality has concentrated on their active vs incidental 

defensive roles. We show for birch and pine that the feedback consequences for herbivores vary from favorable to harmful 

depending on the treated plant part and on timing of the damage. Our results cannot be directly derived from the nutrient 

status of the tree. Instead, damage-induced alterations in meristematic sink strength provides a general explanation. In birch, 

damage to dominant sinks allows other parts to obtain a greater than normal share of resources which leads to their vigorous 

growth and better tbrage quality for herbivores. However, damage to young source leaves that are supplying local meristems 

for next year, weakens such meristems and reduces their success in competing for resources. This leads to less vigorous 

growth and poorer plant quality for herbivores. Damage to mature leaves presumably disturbs provisioning of the general 

resource pool of the tree but such effects, especially if they are monitored locally, may be hard to reveal. An emphasis on 

damage-induced effects on plant sinks and sources makes induced resistance, induced amelioration, as well as localization of 

responses more understandable. 

ACKNOWLEDGEMENTS 

We thank all the persons who have helped in the practical work on which this paper is based, as well as those who 

have helped in formulating the ideas and who have criticized earlier versions of the present paper (Sinikka Hanhim_iki, Seppo 

Neuvonen, Pekka Niemelfi, Matti Rousi, Kai Ruohom_.ki, Janne Suomela, John Vranjic, and Timo Vuorisalo). John also 

checked our English. Financially this study has been made possible by grants from the Academy of Finland, Maj and Tot 

Nessling Foundation, and the Lapland Forest Damage Project. 

LITERATURE CITED 

*GREN, G.I. and AXELSSON, B. 1985. Correlations between needle and shoot growth characteristics in Scots pine (Pinus 

sylvestris L.) stands. Forestry 58: 199-205. 

BENZ, G. 1974. Negative Riickkoppelung durch Raum- und Nahrungskonkurrenz sowie zyklische VerSnderung dr 

Nahrungsgrundlage als Regelprinzip in der Populationsdynamik des Grauen L_irchenwicklers, Zeiraphera diniana 

(Guende) (Lep., Tortricidae). Z. ang. Ent. 76: 196-228. 

BJC)RKMAN, C., LARSSON, S., and GREE R. 1991. Effects of nitrogen fertilization on pine needle chemistry and sawfly 

performance. Oecologia 86: 202-209. 7

BLANCHE, C.A., LORIO, EL., Jr., SOMMERS, R.A., HODGES, J.D., and NEBEKER, T.E. 1992. Seasonal cambial 

growth and development of loblolly pine: xylem formation, inner bark chemistry, resin ducts and resin flow. Forest. 

Ecol. Manage; 49:151-165. 

BRYANT, J.E, CHAPIN, F.S., III, and KLEIN, D.R. 1983. Carbon/nutrient balance of boreal plants in relation to vertebrate 

herbivory. Oikos 40: 357-368. 

BRYANT, J.E, KUROPAT, EJ., REICHARDT, EB., and CLAUSEN, T.E 1991. Controls over the allocation of resources by 

woody plants to chemical antiherbivore defense, p. 83-102. In Palo, R.T. and Robbins, C.T., yds. Plant Defenses 

Against Mammalian Herbivores. CRC Press, Boca Raton, Florida. 

BRYANT, J.E, REICHARDT, EB., CLAUSEN, T.E, and WERNER, R.A. 1993. Effects of mineral nutrition on delayed 

inducible resistance in Alaska paper birch. Ecology 74: 2072-2084. 

CLINE, M.G. 1991, Apical dominance. Bot. Rev. 57:318-358. 

DANELL, K. and HUSS-DANELL, K. 1985. Feeding by insects and hares on birches earlier affected by moose browsing. 

Oikos 44: 75_81. 

EDENIUS, L. 1993. Browsing by moose on Scots pine in relation to plant resource availability. Ecology 74:2261-2269. 

ERICSSON, A. 1978i Seasonal changes in translocation of 14Cfrom different age-classes of needles of 20-year-old Scots 

pine trees (Pinus sylvestris). Physiol. Plant. 43: 351-358. 

FARMER, E.E. and RYAN, C.A. 1990. Interplant communication: Airborne methyl jasmonate induces synthesis of proteinuse 

inhibitors in plant leaves. Proc. Natl. Acad. Sci USA 87: 7713-7716. 

HANHIM,g, KI, S. I989i Induced resistance in mountain birch: defence against leaf-chewing insect guild and herbivore 

competition: Oecologia 81: 242-248. 

HANHIMJkKI, S. and SENN, J. 1992. Sources of variation in rapidly inducible responses to leaf damage in the mountain 

birch-insect herbivore system. Oecologia 91:318-331. 

HARTLEY, S.E. and LAWTON, J.H. 1991. Biochemical aspects and significance of the rapidly induced accumulation of 

phenolics in birch foliage, p. 105-132. In Tallamy, D.W. and Raupp, M.J., yds. Phytochemical Induction By Herbivores. 

John Wiley, New York. 

HAUKIOJA, E. 198Z Inducible defences of white birch to a geometrid defoliator, Epirrita autumnata, p. 199-203. In 

Visser, J.H, and Minks, A.K., yds. Insect-Plant Relationships. PUDOC, Wageningen. 

HAUKIOJA, E. I991a! Cyclic fluctuations in density: interactions between a defoliator and its host tree. Acta Oecol. 12: 

7-88 

HAUKIOJA, E. 1991b_ The influence of grazing on the evolution, morphology and physiology of plants as modular 

organisms. Philos. Trans. R. Soc. Loud. B. 333: 241-247. 

HAUKIOJA, E. and NEUVONEN , S. 1985. Induced long-term resistance in birch foliage against defoliators: defensive or 

incidental? E_61ogy 66:1303-1308. 

HAUKIOJA, E. and NEUVONEN, S. 1987. Insect population dynamics and induction of plant resistance: the testing of 

hypotheses, p _11'432. In Barbosa, E and Schultz, J.C., yds. Insect Outbreaks. Academic Press, San Diego. 

HAUKIOJA, E., KAPIAINEN, K., NIEMELA, P., and TUOMI, J. 1983. Plant availability hypothesis and other explana- 

8

HAUKIOJA, E., SUOMELA, J., and NEUVONEN, S. t985. Long-term inducible resistance in birch foliage: triggering 

cues and efficacy on a defoliator. Oecologia 65: 363-369. 

HAUKIOJA, E., NEUVONEN, S., HANHIMAKI, S., and NIEMEL._, R 1988. The autumnal moth in Fennoscandia, p. 

163-178. In Berryman, A.A., ed. Dynamics of Forest Insect Populations. Patterns, Causes, Implications. Plenum 

Press, New York. 

HAUKIOJA, E., RUOHOMAKI, K., SENN, J., SUOMELA, J., and WALLS, M. 1990. Consequences of herbivory in the 

mountain birch (Betula pubescens ssp tortuosa): importance of the functional organization of the tree. Oecologia 82: 

238- 247. 

HERMS, D.A. and MATTSON, W.J. 1992. The dilemma of plants: to grow or defend. Quart. Rev. Biol. 67: 283-335. 

HONKANEN, T. and HAUKIOJA, E. 1994. Why does a branch suffer more after branch-wide than after tree-wide defoliation? 

Oikos 71 : 441-450. 

HONKANEN, T., HAUKIOJA, E., and SUOMELA, J. 1994. Effects of simulated defoliation and debudding on needle and 

shoot growth in Scots pine (Pinus syIvestris): implications of plant source/sink relationships for plant-herbivore 

studies. Funct. Ecol. 8: 631-639. 

IASON, G. and HESTER, A.J. 1993. The response of heather (Calluna vulgaris) to shade and nutrients - predictions of the 

carbon-nutrient balance hypothesis. J. Ecol. 81: 75-80. 

KOZLOWSKI, T.T. 1971. Growth and Development of Trees. Academic Press, New York. 

LA.NGSTROM, B., TENOW, O., ERICSSON, A., HELLQVIST, C., and LARSSON, S. 1990. Effects of shoot pruning on 

stem growth, needle biomass, and dynamics of carbohydrates and nitrogen in Scots pine as related to season and tree 

age. Can. J. For. Res. 20: 514-523. 

LAWTON, J. 1991. Ecology as she is done, and could be done. Oikos 61: 289-290. 

LOYTTYNIEMI, K. 1985. On repeated browsing of Scots pine saplings by moose (Alces alces). Silva Fenn. 19: 387-391. 

MCFARLAND, D. and RYAN, C.A. 1974. Proteinase inhibitor-inducing factor in plant leaves: a phylogenetic survey. Plant 

Physiol. 54: 706-708. 

NEUVONEN, S. and DANELL, K. 1987. Does browsing modify the quality of birch foliage for Epirrita autumnata larvae? 

Oikos 49: 156-160. 

NEUVONEN, S. and HAUKIOJA, E. 1991. The effects of inducible resistance in host foliage on birch-feeding herbivores, 

p. 277-291. In Tallamy, D.W. and Raupp, M.J., eds. Phytochemical Induction By Herbivores. John Wiley and Sons, 

New York. 

NEUVONEN, S., HANHIMA.KI, S., SUOMELA, J., and HAUKIOJA, E. 1988. Early season damage to birch foliage 

affects the performance of a late season herbivore. J. Appl. Entomol. 105: 182-189. 

NIEMEL,_., R, ARO, E.-M., and HAUKIOJA, E. 1979. Birch leaves as a resource for herbivores. Damage-induced increase 

in leaf phenols with trypsin-inhibiting effects. Rep. Kevo Subarct. Res. Stat. 15: 37-40. 

NIEMEL*, R, MANNILA, R., and M,_.NTSA.LA, R 1982. Deterrent in Scots pine, Pinus sylvestris, influencing feeding 

behavior of the larvae of Neodiprion sertifer (Hymenoptera, Diprionidae). Ann. Ent. Fenn. 48: 57-59. 

NIEMELJ_, R, TUOMI, J., MANNILA, R., and OJALA, R 1984. The effect of previous damage on the quality of Scots 

pine foliage as food for Diprionid sawflies. Z. Ang. Ent. 98: 33-43.

NIEMELA, R, TUOMI, J,, and LOJANDER, T. 1991. Defoliation of the Scots pine and performance of diprionid sawflies. 

J. Anim. Ecol! 60! 683-692. 

NUORTEVA, H. and KURKELA, T. 1993. Effects of crown reduction on needle nutrient status of scleroderris-cankerdiseased 

and greefi,pruned Scots pine. Can. J. For. Res. 23:1169- 1178. 

REICH, RB., WALTERS; M.B., KRAUSE, S.C., VANDERKLEIN, D.W., RAFFA, K.F., and TABONE, T. 1993. Growth, 

nutrition and gas exchange of einus resinosa followingartificial defoliation. Trees 7: 67-77. 

ROSSITER, M.C., SCHULTZ, J.C., and BALDWIN, I.T. 1988. Relationships among defoliation, red oak phenolics, and 

gypsy moth growth and reproduction. Ecology 69: 267-277. 

RUOHOMAKI, K.i eHAPIN, F. S., HAUKIOJA, E., NEUVONEN, S. and SUOMELA, J. Delayed inducible resistance in 

mountain birch in resp°nse to fertilization and shade. Ecology. In press. 

SACHS, T., NOVOPLANSKY, A., and COHEN, D. 1993. Plants as competing populations of redundant organs. Plant Cell 

Environ. 16:765-770. 

SENN, J. and 1994. Reactions of the mountain birch to bud removal: effects of severity and timing, and 

implications for herbivores. Funct. Ecol. 8: 494-501. 

SPRUGEL, D.G.,HINCKLFX, %M., 

Ecol. Syst. 221309;334. 

,,d SCHAAP, W. 1991 The theory and practice of branch autonomy, annu. Rev. 

SUTHERLAND, W. and STILLMAN, R.A. 1990. Clonal growth: insights from models, p. 95-111. In van Groenendael, J. 

and de Krooni H_i eds. Clonal Growth in Plants: Regulation and Function. SPB Academic Publishing, The Hague. 

10 

vian mountaifi eh_h and northern Finland 1862-1968. Zool. Bidr. Uppsala 2: 1-107.

DAMAGE-INDUCED NUTRITIONAL CHANGES IN PINE FOLIAGE: 

AN OVERVIEW 

PAIVI LYYTIK_INEN 

Finnish Forest Research Institute, Department of Forest Ecology 

RO. Box 18, FIN-01301 Vantaa, Finland 

INTRODUCTION 

Plant resistance can be divided into constitutive and induced defenses. Constitutive defense can deter, repel, intoxi- 

cate or interfere with insect herbivores° This can change in response to external or internal influences (e.g., weather, site 

factors, plant age), bt_t is t.lsua]ly stable (Schtdt 1988). Induced responses are a form of phenotypic plasticity and can shift 

plant resistance as a result of herbivore attack (Haukioja and Neuvonen 1987, Haukioja 1990). Induced changes in foliage 

quality can take place in a density-dependent fashion and modify the stability of insect populations. Responses which are 

caused by one insect generation and act during subsequent generations have been called delayed induced resistance (DIR). 

Those responses which _)ccur in the same insect generation have been called rapid induced resistance (RIR) (Haukioia and 

Neuvonen 1987_ Hauki_4ja 1990). Emphasis has usually been placed on the defensive role of chemical compounds, leading to 

underestimation of the ro_e of some *_positive" compounds (e.g,, mineral nutrients, water, sugars). It is often questioned 

whether chemical changes (low quality to herbivores) represent an active defense or are only incidental by-products or reflect 

the: abiotic environment (Tuomi et al. 1984, Haukioja and Neuvonen 1985). 

Plant species or individuals may respond differently to insect-caused or artificial def_._liation. The induced responses 

of deciduous trees have been studied more than those of evergreen trees. The most studied tree genera are Bemla, Quercus, 

A[nus and SMix (Schu]t and Baldwin 1982, Jeker 1983, Raupp and Denno 1984, Haukioja and Neuvonen 1987, Neuvonen et 

al. 1987). Defoliation by Zeiraphera diniana altered the contents of nitrogen and fiber in European larch, Lc_rix decidua, for 

a four-year period (Baltensweiler et al. 1977). Responses to insect defoliation have been detected in the foliage of Pinus 

ponderosa (Wagner and Evans 1985), P contorta (Leather et al. 1987) and P sylvestris (Lyytik_iinen 1993). Induced reac- 

tions have also been reported in P. radiata after damage by spider mites (Karban 1990). 

A very important question in insect ecology is whether certain tree species can exhibit induced responses and whether 

such responses have an impact on defoliating insect pests. The main aim of this study is to find those circumstances when 

induced responses have been expressed and affected insect performance or perhaps population dynamics. The discussion is 

focused on induced reactions in pine foliage and experiments with defoliating insects. 

QUALITATIVE CHANGES IN PINES 

Rather few studies have been carried out on real or artificial defoliation in conifers, especially pines. The very first 

investigations dealt with modified polyphenol metabolism in P. sylvestris after attack by Neodiprion sertifer (Thielges 1968), 

but the conclusions were erroneous due to indirect comparison between the occurrence of the compound and sawfly attack. 

Some of the reported studies have involved insect rearings without chemical analyses, only leading to hypotheses about the 

causal factors. One difficulty is the wide variability in experimental design, which makes comparisons difficult. One basic 

problem seems to be the analysis of needle year-classes of different age. The nutritive status, i.e., the value of foliage as 

photosynthetic machinery and the ability to react after damage, are not the same in different year-classes (Bernard-Dagan 

1988, Haukioja 1990). The timing of treatments during the growing season can also affect the expression of the reactions 

(Hartley and Lawton 1991). 

The nutrients most studied are nitrogen, minerals, sugars, and water content (Table 1). However, needle resin acids 

that are obviously harmful for diprionid sawflies (Larsson et al. 1986) have been analyzed only once after defoliation. Rapid 

Mattson, W.J., Niemel_i, E, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest fbr pattern and principle. USDA 

For. Serv. Gen. Tech. Rep. NC-183, N.C. For. Exp. Sta., St. Paul, MN 55108. 

11

induced reactions have been studied in P ponderosa and P. sylves_ris. In P. po_derosa the nitrogen and tannin concentrations 

decreased, but those of proteins, phenols, procyanidins and certain nutrients increased in P. sylvestris. Needle water content 

tended to rise shortly after damage. After one or more years (Table 1) the nitrogen, minerals, and monoterpene concentrations 

in P. contorta increased, but only that of P decreased. A similar trend for the nutrient and water contents in P. sylvestris 

has also been reported. The most interesting aspect is the increase in diterpene acids and sugars, but there are only two 

reports of such findings (Buratti el al. 1988, Lyytik_iinen 1994). 

Artificial and natural defoliation induce either rapid or delayed chemical changes in pine foliage. The response to 

insect-caused defoliation is usually stronger than that to artificial damage (Neuvonen et al. 1987, Harttey and Lawton 1991), 

but opposing reactions have also been reported (Baldwin 1988). It would appear that pines are more insensitive to defoliation 

than deciduous trees. The damage level and its timing are probably more important factors for the expression of the reactions 

than the damage mode. Deciduous trees can respond to low foliage damage (e.g., tearing) (Haukioja and Neuvonen 1987), 

but conifers need at least moderate damage (Wagner and Evans 1985, Leather et al. 1987, Lyytikfiinen 1992a). In some cases 

the qualitative changes in late summer were different than those iraearly summer (e.g., Lyytik_iinen 1994). Ericson et al. 

(t980a,b) observed that Scots pines defoliated late in the season suffered from lowered starch concentrations, and the trees 

with the highest degree of damage lost their current year's shoots and new buds. The studies referred to in Table 1 have 

mostly been concerned with responses to sawfly dalnage, induced reactions have unfortunately been investigated with very 

few other pine insect species, even though there are many other harmful pests on pines in Europe, e.g., Panolisflammea 

(Barbour 1987). 

NEEDLE QUALITY AND THE PERFORMANCE OF DEFOLIATING INSECTS 

Rapid Responses 

Qualitative changes do occur in pines, but are these changes targeted at herbivores and are the species used sensitive 

enough to respond to induced reactions? Rapid reactions and insect response have been detected in two pine species: Pinus 

ponderosa and P sylvestris (Table 2). Artificial damage adversely affected the growth of Neodiprionfulviceps, which should 

indicate negative reactions in P. ponderosa. The response of Neodiprion sertifer was mainly positive on P. sylvestris shortly 

after insect-caused defoliation, but after artificial damage there were no differences between control and experimental trees. 

Another sawfly species responded either negatively or not at all. Furthermore, the results with Scots pine were not parallel. 

It would appear that Diprion pint was the least sensitive to qualitative changes in the needles. On the other hand, the low 

defoliation level for the whole canopy may have an effect on the indifferent performance of different sawfly species (Niemel_i 

et at. 1984). In general, slight defoliation caused no changes in insect performance. In some cases slight defoliation may 

even have adverse effects on insect success compared to heavy defoliation (LyytikE.inen 1992a). There were some positive 

responses in early summer and soon after damage, but mainly negative responses in late summer. The lack of impact on 

sawfly survival indicates no effect on population dynamics. On the basis of these reports (Table 2), it would appear that 

Scots pine possesses no effective, rapid resistance reactions against diprionid sawflies. 

Delayed Responses 

More studies have been carried out on delayed reactions in pines than on rapid reactions. In P. contorta the investiga- 

tions deal only with experiments on seedlings (Table 3). The negative effect of treatment was obvious after defoliation by 

Panolisflammea, reflecting the persistence of induced reactions harmful to insects. One possible explanation for this may be 

that current-year needles were used in experiments and they obviously contain more secondary compounds than older needle 

year-classes (Buratti et al. 1988, 1990). There was a trend for positive or neutral responses after previous early summer 

defoliation. However, the responses were mostly negative following late summer damage in both pine species (Table 3). The 

results for P flammea only are dissimilar due to the different experimental design. In general the results for P sylvestris were 

negative or neutral. There were also experiments with N. sertifer and Gilpinia pallida in which there were no responses. As 

in studies on rapid responses, delayed responses did not usually affect larval survival, thus indicatingnegligible effects on 

population dynamics. 

The importance of defoliation level was also obvious in the reports of delayed induced reactions in pines. The 

direction of the response was not similar at different levels of defoliation (Lyytik_iinen 1993, 1994), or else the response in 

sawfly performance only occurred after moderate damage (Lyytik_iinen 1992a). In some cases, the sawfly response was 

13

nonlinear (Lyytik_iinen 1992a), indicating some kind of optimality perhaps in nutrient status due to stress (Wagner and Frant 

1990). The results in these reports (Table 3) testify that lodgepole pine obviously responds negatively to defoliating insects, 

but in Scots pine the responses are weaker and are only expressed under certain conditions. 

DISCUSSION 

Foliage of different ages may respond differently to defoliation. Several reports indicate that qualitative changes 

induced after spring defoliation in deciduous trees are more obvious those occurring after late summer defoliation (e.g., 

Haukioja and Niemelfi 1979). Young, expanding foliage acts as a sink for metabolites, and plant resources are converted to 

foliage development and expansion (Coleman 1986). Major changes in the biochemistry and physiology can occur as a result 

of foliage damage. On the other hand, damage to older foliage may decrease the amount of resources available to young 

foliage (Coleman 1986). However, no studies have been carried out on the gradual changes in needle quality after detbliation 

taking place during successive growing season(s). Needle quality is usually based on a single sampling. In Pinus pinaster, 

sugar, lipid, starch, and terpene concentrations have been found to vary according to season (Bernard-Dagan 1988). This 

phenomenon may offer one explanation for the findings concerning different responses following early or late summer 

defoliation. 

Plant resource allocation and partitioning change in relation to age. Seedlings allocate most of their resources to 

growth, and mature plants allocate to reproductive organs (Kolowski 1971). Induced reactions have been found in three pine 

species: P. contorta, P ponderosa and P radiata (Leather et al. 1987, Wagner 1988, Karban 1990). The growth of P. radiata 

saplings is the fastest (Kolowski 1971). P. ponderosa and P. contorta are regarded as slow-growing species, but the growth 

of P. sylvestris iseven slower than that of R contorta (Lyytik_iinen 1993). The lack of induced reactions is common in 

inherentlyslow-growing conifers, relying on constitutive resistance (Rhoades and Cates 1976). The induced reactions 

detected inthese pinespecies were linked with age and fast growth because seedlings and saplings primarily showed the 

responses. In lodgepole pine, trees over 20 years old showed no induced resistance against herbivores (Leather et al. 1987, 

Watt et al. 1991), and only at ages > 12 years, did saplings suffer from outbreaks of Panolis flammea (Barbour 1987). In 

Scots pine the reactions were more pronounced in seedlings than in saplings (Lyytik_,inen 1992b, 1993, 1994). The defensive 

strategy of Scots pine is obviously also age-specific, i.e., the seedlings resemble fast-growing species in showing induced 

responses. 

Plants usually show flexibility in their physiology within the bounds set by environmental limitations (e.g., drought, 

shading, nutrient deficiency). Changes in resource allocation may be directed towards equalization of all the factors limiting 

plant growth (e.g., Hunt and Nicholls 1986). Environmental stresses may cause shifts in the partitioning of resources to 

et al. 1987). Bryant et al. (1983) suggested that the phenotypic responses of secondary chemistry 

are carbon-nutrient balance. Carbon-based, secondary metabolites are supposed to accumulate in 

is limited by mineral nutrients (Tuomi et al. 1988, Tuomi 1992). The defense bears no cost, because 

supported by the resources acquired in excess of primary metabolic requirements (Tuomi et al. 

lbundant nutrient status of the growing site in many studies could have reduced the probability of 

c production of secondary metabolites (e.g., Niemel_i et al. 1991, Lyytik_iinen 1993, 1994). 

for only weak rapid or delayed reactions in pines may be due to physiological differences between 

trees. In deciduous trees the stems and roots provide the major carbon reserves (Bryant et al. 1983). 

In and current photosynthates in evergreen trees are stored in old foliage (Tuomi et al. 1984, 1988), 

e.g., by sawfly larvae. These physiological differences may contribute to the different pattern of 

Ion (Niemel_i and Tuomi 1993) because of a shortage of mineral nutrients or carbon (Niemel_i et al. 

1984)_ 

used in the experiments covered here did not respond in a similar manner to damage. Defoliation 

of and saplings (Table 3) affected the success of G. pallida, but not that of IV.sertifer (Lyytikfiinen 

the effect of damage affected the performance of R flammea on P contorta (Leather et al. 1987) and 

N. (Wagner 1986). One explanation for these findings could be the different sensitivity of the 

(Hanski 1987, Haukioja 1990, Niemel_i and Tuomi 1993). Outbreak species such as N. sertifer may 

be :es in food quality than less common species, e.g., G. pallida on the same host species. Another 

ex in larval feeding periods. The larvae of N. sertifer consume only mature needles in early

summer, but the other sawfly species can also eat the current foliage during rniddle or late summer. Needle year-classes may 

also react differently after needle damage (Niemel_ and Tuomi 1993), or interactions among needle year-classes may occur 

(Buratti et al. 11988). Thirdly, as earlier rnentioned, there could also be phenological changes in the same year-classes during 

the summer (Bernard-Dagan 1988). The quality of mature needles is different m early season species compared to late 

season species. 

The existence of induced responses in P. contorta and R potzderosa may in part ofter one explanation for the population 

dynamics of the species feeding on them (e.g., Barbour 1988). In Scots pine the weak induced reactions may be the 

reason for the tack of cydicity in sawfly outbreaks (Niemela and Tuomi 1993). The lack of damage-induced reactions in the 

needles of mature and pole-size Scots pine enables outbreaks of N. sertifer to occur for several years at the same site. The 

explanations for the different phases in population dynamics must be due to factors (e.g., parasitoids, diseases) other than 

qualitative changes in the foliage. 

SUMMARY 

In pine foliage, induced responses against defoliating insects have been detected in P#zus ponderosa, P. contorta and 

P. syh,estris. Contents of nitrogen, minerals, and water tended to increase after damage. The damage level and timing are 

important factors for the expression of these reactions. P. ponderosa showed evidence of rapid induced reactions, but not in 

E sylvestris. Delayed induced reactions were obvious in P. cotztorta, but in P. sylvestris the responses were weaker and 

expressed only under certain conditions. Foliage of different ages may respond differently, because young foliage acts early 

on as a sink for metabolites. Furthermore, foliage quality can change according to season. The induced reactions detected in 

certain pine species are linked to their seedlings and fast growth. The defensive strategy of pines is probably highly agespecific: 

seedlings resemble fast-growing species. The explanation fbr the typically weak reactions in pines may be due to 

physiological differences between them and deciduous trees. The weak induced resistance in Scots pine may explain lack of 

cyclicity in pine sawfly outbreaks. 


I wish to thank Pekka Niemel_i _br his critical and helpful comments on the manuscript and John Derome for kindly 

revising the English text. The work was financially supported by The Academy of Finland (project no. 209103 1). 


BALDWIN, I.T. 1988. The alkaloidal response of wild tobacco to real and simulated herbivory. Oecologia 77: 378-381. 

BALTENSWEILER, W., BEN, G., BOVEY, R and DELUCCHI, V. 1977. Dynamics of larch bud moth populations. Annu. 

Rev. Entomol. 22: 79-100. 

BARBOUR, D.A. 1987. Pine beauty moth population dynamics. General considerations and life-table work. In Leather, 

S.R., Stoakley, J.T., and Evans, H.E, eds. Population biology and control of the pine beauty moth. For. Comm. Bull. 

67: 7-13. 

BARBOUR, D.A. 1988. The pine looper in Britain and Europe, p. 291-308. In Berryman, A.A., ed. Dynamics of Forest 

Insect Populations: Patterns, Causes, and Management Strategies. Plenum Press, New York. 

BERNARD-DAGAN, C. 1988. Seasonal variations in energy sources and biosynthesis of terpenes in maritime pine, p. 93- 

115. In Mattson, W.J., Levieux, J., and Bernard-Dagan, C., eds. Mechanisms of Woody Plant Defenses Against 

Insects: Search for Pattern. Springer-Verlag, New York. 

BRYANT, J.R, CHAPIN, F.S.III, and KLEIN, DR. 1983. Carbon/nutrient balance of boreal plants in relation to vertebrate 


17

BURATTI, L., ALLAIS, J.E, and BARBIER, M. 1988. The role of resin acids in the relationships between Scots pine and 

the sawfly, Diprion pini (Hymenoptera: Diprionidae). I Resin acids in the needles, p. 171-187. In Mattson, W.J., 

Levieux, J., and Bernard-Dagan, C., eds. Mechanisms of Woody Plant Defenses Against Insects: Search for Pattern. 

Springer-Verlag, New York. 

BURATTI, L., ALLAIS, J.E, GERI, C., and BARBIER, M. 1990. Abietane and pimarine diterpene acid evolution in Scots 

pine, Pinus sylvestris, needles in relation to feeding of the pine sawfly, Diprion pini L. Ann. Sci. For. 47:161-171. 

CHAPIN, ES.III, BLOOM, A.J., FIELD, C.B,. and WARING, R.H. 1987. Plant responses to multiple environmental 

factors. Bio Science 37: 49-57. 

COLEMAN, J.S. 1986. Leaf development and leaf stress: increased susceptibility associated with sink-source transition. 

Tree Physiol. 2: 289-299. 

ERICSSON, A, HELLKVIST, K., HILLERDAHL-HAGSTROMER, K., LARSSON, S., MATTSON-DJOS, E., and 

TENOW, O. 1980a. Consumption and pine growth - hypothesis on effects on growth processes by needle-eating 

insects. In Persson, T, ed. Structure and Function of Northern Coniferous Forests - an ecosystem study. Ecol. Bull. 

(Stockholm) 32: 537-545. 

ERI , A., LARSSON, S., and TENOW, O. 1980b. Effects of early and late season defoliation on growth and carbohydrate 

dynamics in Scots pine. J. Appl. Ecol. 17: 747-769. 

GERI , L., and ALLAIS, J.E 1988. The role of resin acids in the relationship between Scots pine and sawfly, 

Oiprionpini (Hymenoptera: Diprionidae). II. Correlations with the biology of Diprion pini, p. 189-201. In Mattson, 

and Bernard-Dagan, C., eds. Mechanisms of Woody Plant Defenses against Insects: Search for 


109: 436_447. 

E, and LEVIEUX, J. 1990. Incidence de la consommation du feullage de Pins sylvestres 

filles sur le developpement et la fecodite de Diprion pini L. (Hym., Diprionidae). J. Appl. Ent. 

HANS,, I. I987. Pine sawfly population dynamics: patterns, processes, problems. Oikos 44: 165-174. 

HA_EY, S.E. and LAWTON, J.H. 1991. Biochemical aspects and significance of the rapidly induced accumulation of 

foliage, p. 105-132. In Tallamy, D.W. and Raupp, M.J., eds. Phytochemical Induction by Herbi- 

',wYork. 

HAU Ai E. 1990. Induction of defenses in trees. Annu. Rev. Entomol. 36: 25-42. 

HA1 S. 1985. Induced long-term resistance of birch foliage against defoliators: defensive or 

,gy 66:1303-1308. 

18 

S. 1987. Insect population dynamics and induction of plant resistance: the testing of 

1-432. In Barbosa, E and Schult, J.C., eds. Insect Outbreaks. Academic, San Diego. 

A.D. 1986. Stress and coarse control of root-shoot partitioning in herbaceous plants. Oikos 47: 

einer Defoliation im Vorjahr und des Blattalters auf die Larvenentwicklung von Melasoma 

)tera, Chrysomelidae). Mitt. Schwei. Ent. Ges. 56: 237-244. 

outbreaks on only young trees: testing hypotheses about aging and induced resistance. 

Growth and development of trees. Academic Press, New York.

LARSSON, S., BJ(_RKMAN, C., and GREF, R. 1986. Responses of Neodiprion sertifer (Hym., Diprionidae) larvae to 

variation in needle resin acid concentration in Scots pine. Oecologia 70: 77-84. 

LEATHER, S.R., WATT,A.D., and FORREST, G.I. 1987. Insect-induced changes in young lodgepole pine (Pinus contorta): 

the effect of previous defoliation on oviposition, growth and survival of pine beauty moth, Panolisflammea. Ecol. 

Entomol. 12: 275-281. 

LYYTIKAINEN, R 1992a. The influence of damage level in Pinus sylvestlqs foliage on the performance of diprionid 

sawflies. Scand. J. For. Res. 7: 249-257. 

LYYTIKAINEN, R 1992b. Comparison of the effects of artificial and natural defoliation on the growth of diprionid sawflies 

on Scots pine foliage. J. Appl. Ent. 114: 57-66. 

LYYTIK.AINEN, R 1993. The performance of diprionid sawflies on defoliated pine seedlings. Acta Oecologica 14: 305- 

315. 

LYYTIKAINEN, P. 1994. Effects of natural and artificial defoliations on sawfly performance and foliar chemistry of Scots 

pine saplings. Ann. ool. Fennici (In Press). 

NEUVONEN, S., HAUKIOJA, E. and MOLARIUS, A. 1987. Delayed inducible resistance against a leaf-chewing insect in 

four deciduous tree species. Oecologia 74: 363-369. 

NIEMELJ(, P. and TUOMI, J. 1993. Sawflies and inducible resistance of woody plants, p. 391-407. In Wagner, M.R. and 

Raffa, K.F. Sawfly Life History Adaptations to Woody Plants. Academic Press, New York. 

NIEMELJk, E, TUOMI, J., and LOJANDER, T. 1991. Defoliation of the Scots pine and performance of diprionid sawflies. 

J. Anirn. Ecol. 90: 683-692. 

NIEMEL,i_, R, TUOMI, J., MANNILA, R., and OJALA, R 1994. The effect of previous damage on the quality of Scots 

pine foliage as food for diprionid sawflies. Angew. Entomol. 98: 33-43. 

RAUPP, M.J. and DENNO, R.E 1984. The suitability of damaged willow leaves as food for the leaf beetle, Plagiodera 

versicolora. Ecol. Entomol. 9: 443-448. 

RHOADES, D.F. and CATES, R.G. 1976. Towards a general theory of plant antiherbivore chemistry. Rec. Adv. Phytochem. 

10: 168-213. 

SCHULTZ, J.C. 1988. Plant responses induced by herbivores. Trends. Ecol. Evol. 3: 45-49. 

SCHULTZ, J.C. and BALDWIN, I.T. 1982. Oak leaf quality declines in response to defoliation by gypsy moth larvae. 

Science 217: 149-151. 

THIELGES, B.A. 1968. Altered polyphenol metabolism in the foliage of Pinus sylvestris associated with European sawfly 

attack. Can. J. Bot. 46: 724-725. 

TUOMI, J. 1992. Toward integration of plant defence theories. TREE 7: 365-367. 

TUOMI, J., NIEMELA, R, CHAPIN, F.S., BRYANT, J.R, and SIREN, S. 1988. Defensive responses of trees in relation to 

their carbon/nutrient balance, p. 57-72. In Mattson, W.J., Levieux, J., and Bernard-Dagan, C., eds. Mechanisms of 

Woody Plant Defenses against Insects: Search for Pattern. Springer, New York. 

TUOMI, J., NIEMEL._, E, HAUKIOJA, E., SIREN, S., and NEUVONEN, S. 1984. Nutrient stress. An explanation for 

plant antiherbivore responses to defoliation. Oecologia 61: 208-210. 

19

WAGNER, ',nce of moisture stress and induced resistance in ponderosa pine, Pinus ponderosa Dougl. ex 

y, Neodiprion autumnaIis Smith. For. Ecol. Manage. 15: 43-53. 

83:452-457. 

defenses in ponderosa pine against defoliating insects, p. 141-155. In Mattson, W.J., 

gan, C., eds. Mechanisms of Woody Plant Defenses against Insects: Search for Pattern. 

RD. 1985. Defoliation increases nutritional quality and allelochemicals of pine seedlings. 

P7. 

D.R 1990. Influence of induced water stress in ponderosa pine on pine sawflies. Oecologia 

WATT, and FORREST, G.I. 1991. The effect of previous defoliation of pole stage lodgepole pine on 

plant the growth and survival of pine beauty moth (Panolisflammea) larvae. Oecologia 86:31-35. 

20 

.... _!i5_ ¸i/_

STAND AND LANDSCAPE DIVERSITY AS A MECHANISM OF FOREST 

RESISTANCE TO INSECTS 

T. D. SCHOWALTER 

Department of Entomology, Oregon State University, Corvallis, Oregon 97331-2907, USA 

INTRODUCTION 

Plant vulnerability to herbivorous insects depends on both suitability (determined by nutritional and defensive 

factors) and exposure (determined by plant location relative to herbivore population sources). The importance of plant 

biochemical defenses against herbivores has received much attention (Rosenthal and Janzen 1979, Cates and Alexander 1982, 

Harborne 1982, Coley et al. 1985). Herbivore populations typically are aggregated on particular trees that often differ from 

their neighbors in chemical composition or other factors that indicate greater susceptibility to herbivores (Alstad and 

Edmonds 1983, Lorio 1993). Plant compounds are important as feeding deterrents or toxins for herbivores unable to adapt 

appropriate avoidance or detoxification mechanisms. However, herbivores have adapted various strategies for avoiding or 

detoxifying the chemical defenses of their hosts (Bernays and Woodhead 1982, McCullough and Wagner 1993). 

Variation in suitability among plants may be important both for reducing selection for herbivore adaptations and for 

minimizing host exposure to adapted herbivores. Herbivores seeking new hosts must both distinguish and be able to reach 

suitable hosts. Therefore, plant suitability, apparency, and distance from herbivore populations function interactively to 

determine colonization by herbivores (Courtney 1986, Schowalter and Stein 11987). 

My objective in this paper is to evaluate plant diversity as a means of reducing host exposure to herbivores. I will 

consider diversity at the level of genetic variability within a particular species, and at the level of stand and landscape 

variability in community composition. These three levels constitute a nested hierarchy of diversity that can limit population 

growth of herbivores within stands and across landscapes. 

Genetic Diversity Within Species 

Populations of plants typically vary in genotype and susceptibility to herbivores. Herbivores usually prefer, and show 

highest survival and reproduction on, plants or plant species whose defenses can be tolerated or detoxified, given that these 

plants can be discovered (Courtney 1986). Most herbivores are specialists on relatively few (usually related) plant species, 

because avoidance behaviors and detoxification mechanisms require an appropriate genetic template and are energetically 

expensive. Insects that feed on several plant species with distinct resistance mechanisms are subject to different selective 

forces on each host and may develop sibling species as host-specific demes diverge (Alstad and Edmonds 1983; Via 1990, 

1991; Bright 1993). Conversely, stands of plants with similar resistance mechanisms permit rapid adaptation by herbivores. 

Genetic diversity within a plant species confers resistance to herbivores by presenting a matrix of susceptible and 

non-susceptible hosts, but genetic effects are difficult to evaluate in natural forests where genotypic effects can be confounded 

by plant proximity to herbivore population sources or other factors varying geographically. Conifer seed orchards 

and nurseries provide an opportunity to examine the importance of plant genotype independent of plant location. Herbivore 

impact on different genotypes replicated throughout a seed orchard or nursery can be compared to assess the extent to which 

injury is related to genotype. 

Schowalter and Haverty (1989) examined seed losses to two insect species, Contarinia oregonensis and Megastigmus 

spermotrophus, in a Pseudotsuga menziesii clonal seed orchard and in a progeny plantation in western Oregon. The various 

clones or families showed different degrees of resistance to the two insects (Table 1). Resistance to one insect species was 

Mattson, W.J., NiemelS., R, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


21

Table 1.--Percentage seed lost to two cone and seed insect species among offspring of selected parental crosses in a 

Pseudotsuga menziesii progeny plantation in western Oregon. Means above the diagonal are seed losses to a midge, 

Contarinia oregonensis; values below the diagonal are losses to a chalcid, Megastigmus spermotrophus. 

x Parent 1 2 3 4 5 6 7 8 9 10 11 12 Midge 

Parent Mean _ 

1 u 51 43 55 56 54 68 39 43 67 53 

2 8 -- 58 55 84 75 62 85 76 51 59 67 

3 10 -- 54 61 79 65 60 59 58 62 

4 13 17 6 u 59 63 57 52 75 55 52 62 57 

5 6 7 11 -- 69 62 55 63 68 40 62 

6 10 10 24 6 -- 59 56 60 58 61 67 66 

7 6 6 8 7 16 6 -- 59 

8 4 18 13 7 10 m 65 

9 5 14 13 8 14 _ 66 

10 2 11 6 11 4 20 -- 59 

11 9 12 15 16 10 22 -- 51 

12 6 5 2 15 19 _ 63 

Chalcid 

Mean _ 7 9 12 14 8 14 8 10 11 9 14 9 

195% CI = 3.3-4.1 for the midge and 1.2-1.4 for the chalcid 

22 i 

ii

not related to resistance to the second species. Of ten clones in the seed orchard that deviated significantly (based on 95% 

CI) from mean seed loss to either species, eight that were resistant to one species showed no resistance to the other. Four of 

these clones were highly susceptible to the second species. Only two clones were resistant or susceptible to both insects. 

Similar results were found in the progeny plantation (Schowalter and Haverty 1989). Parental crosses that were 

resistant to one insect generally were susceptible to the other. In this case, resistance appeared to be heritable as a dominant 

trait, based on the generally low seed losses for progeny of resistant and susceptible parents (Table 1). 

Schowalter and Stein (1987) compared the extent of Lygus hesperus feeding on different seed sources (representing 

different genetic backgrounds) in a conifer seedling nursery. This insect is a mobile species that feeds primarily on agricultural 

crops :in a hit-and-run manner. Results indicated significant separate and interactive effects of conifer seed source 

(genotype) and of proximity to Lygus population sources in adjacent agricultural crops. The effect of plant proximity to 

herbivore populations, even at this small scale for a relatively mobile herbivore, is surprising. 

Data from these studies demonstrate that, within monocultures of plants, a diversity of genotypes can limit resource 

availability and suitability for particular herbivore species. This diversity represents a species-level defense that limits initial 

herbivore population growth. Genetic diversity within a monoculture is not sufficient to prevent herbivore outbreaks over 

long time periods, especially when conditions that stress plants and/or inhibit production of plant defenses permit herbivore 

population growth (Waring and Pitman 1983). When environmental conditions increase susceptibility of a given plant 

species, exposure to herbivores can be reduced by surrounding non-host plants and stands. 

Plant Species Diversity Within Stands 

Our view of plant species interactions traditionally has focused on competitive interactions. This view has supported 

the tree farm (monoculture) approach to forestry. Recent studies, however, are indicating that plant species interactions are 

more complex. Plant species often share mycorrhizae, contribute collectively to soil fertility through differential nutrient 

uptake and concentration in litter and rhizosphere, and increase the chemical complexity of the forest aerosol (Visser :1986, 

Hunter and Arssen 1988). These mutualistic aspects of plant species interactions may reduce the likelihood of plant stress 

and apparency to herbivores, at least for some combinations of plant species. 

If genetic diversity within monocultures can affect herbivores, then a diversity of plant species within a community 

matrix should limit herbivory to a greater extent. Studies with several insect species in different vegetation types have 

demonstrated that diverse vegetation limits overall herbivory. For example, Root (1973), Kareiva (1983) and Turchin (1988) 

reported that intermixed crops were subjected to lower levels of herbivory by insects than were monocultures of the same 

crops. Examples from forests are rare, largely because manipulating tree diversity for experimental purposes is difficult, and 

natural variation in diversity is confounded by geographic variation in soils, aspect, and other factors that also affect herbivory 

(Schowalter and Filip 1993). 

Gara and Coster (1968), Johnson and Coster (1978), and Schowalter et al. (198 lb) reported that southern pine beetle, 

Dendroctonusfrontalis, populations appeared to be sensitive to Pinus spp. density. Populations generally grew rapidly in 

dense pine stands and declined in sparse pine stands. However, Johnson and Coster (1978) and Schowalter et al. (1981b) 

reported that large populations (> 100,000 beetles) developing under favorable conditions also were capable of sufficient 

aggregation to colonize sparse hosts. Thinning experiments in western North America indicated that tree spacing also is 

critical to mountain pine beetle, Dendroctonus ponderosae, populations (Sartwell and Stevens 1975, Mitchell et al. 1983). 

The effects of tree species diversity were unclear; either hardwoods compete with pines and aggravate stress-related beetle 

activity (Hicks 1980) or hardwoods interfere with discovery of hosts (Belanger and Malac 1980). 

Schowalter and Turchin (1993) conducted a relatively unique experiment to test effects of Pinus spp. density and 

stand diversity on D. frontalis populations. They manipulated pine and hardwood basal areas in a :2x 2 factorial experiment 

and introduced equivalent D. frontalis populations into replicated plots to prevent confounding effects of plot proximity to 

beetle population sources. Infestations subsequently developed only in the dense pure pine stands and achieved infestation 

sizes (>10 dead trees) sufficient to warrant suppression in this treatment (Fig. 1). Infestations did not develop in dense pine 

stands where hardwoods were present or in low density stands, indicating that tree spacing and stand diversity both function 

to reduce insect population growth. 

23

uaa ,,.4 --" _" - LP/LH p 

ua_t _ 4 -''_''LP/HH 

_ a 

I-- 

ua tu 3 _ ---'c> HP/LH _ r 

.!_ :. :--_::_--- 

...:.. -.:..-. "-----'..:.- ----:-"------ 

I 2 3 4 

TIME(Months) 

Figure 1.mCumulative pine mortality to Dendroctonusfrontalis by pine and hardwood basal area treatments in Mississippi 

and Louisiana during 1989 and 1990. Vertical lines represent 1 SEM. Low pine (LP) = 11-14 m2/ha ba; high pine 

(HP) = 23-29 ba; low hardwood (LH) = 0-4 ba; high hardwood (HH) = 9-14 ba. Two infested trees were introduced 

into each experimental stand at the beginning of study. N-8. 

Schowalter (unpubl. data) compared arthropod abundances among replicated and intermixed plots representing young 

Douglas-fir, Pseudotsuga menziesii, plantations (10-15 years old), mature natural P. menziesii monocultures (100-150 years 

old), intact old-growth P. menziesii/Tsuga heterophylla (450 years, old) and old-growth P. menziesii shelterwood (450 years 

old) treatments in western Oregon. Western spruce budworm, Choristoneura occidentalis, was significantly more abundant 

and caused twofold more defoliation in the mature monoculture than in other treatments. This difference also could reflect 

the greater predator diversity and abundance in old-growth stands (Perry 1988, Schowalter 1989, Torgersen et al. 1990), or 

unmeasured differences in mutualistic endophyte diversity or foliage quality among age classes (McCutcheon and Carroll 

1993). 

Stand Diversity Across Landscapes 

Diversity at the landscape level augments diversity at the stand level. As the diversity of stands representing different 

species composition or age classes increases, the distance between stands containing suitable resources increases (Perry I988, 

Schowalter 1989). Although herbivores are capable of dispersing over considerable distances, several factors reduce the 

likelihood of distant hosts being discovered or colonized. First, insect ability to perceive hosts over long distances is limited 

for most species. Plant cues or other factors conveying host location become dissipated in the forest aerosol, making distant 

hosts and hosts in diverse stands less apparent (Visser 1986). Second, survival decreases with distance as a result of longer 

exposure to mortality agents and exhaustion of energy reserves, reducing insect ability to reach distant hosts. Therefore, 

diverse landscapes should prevent localized outbreaks from spreading to distant hosts. 

Conversely, herbivore populations are promoted in landscapes that provide greater homogeneity of resources and few 

barriers to population spread. Major outbreaks typically occur in relatively homogeneous landscapes. 

Declining forest health in eastern Oregon and Washington is largely the result of change in landscape diversity. 

Forests in this region originally were a diverse matrix in which stands of shade-tolerant mixed-conifer Pseudotsuga/Abies/ 

Pinus forest in moist sites at higher elevations, north aspects, and riparian areas were embedded within an arid landscape 

dominated by sparse fire-tolerant Pinus/Larix forest and savannah. Fir defoliators such as C. occidentalis occurred as local 

populations forced to search for hosts aggregated within a largely inhospitable landscape. As a result of fire suppression and 

selective logging of Pinus and Larix over the past century, the fir forest spread across this landscape. The landscape is now 

relatively homogeneous, dominated by forests of dense P. menziesii and Abies spp. Drought stress of these mesic species in 

addition to resource concentration has permitted C. occidentalis to reach epidemic population levels over most of this large 

area (Hadfield 1988). 

24 

i _

In contrast to the situation in eastern Oregon and Washington, C. occidentalis historically has occurred at innocuous 

population levels in western Oregon and Washington, although localized outbreaks have occurred. Recent (and past) activity 

of C. occidentalis in western Oregon is concentrated around a major pass through the Cascade Range. This area has been 

affected by years of drought but also was accessible to epidemic C. occidentalis populations spilling over the pass from 

eastern Oregon (pets. obs.). Nevertheless, Schowalter (1989 and unpubl, data) found that, near this area, C. occidentalis was 

rare or absent in diverse old-growth forests but was abundant and causing measurable defoliation in mid-successional P. 

menziesii monocultures. Perry and Pitman (1983) compared suitability ofP. menziesii foliage from eastern and western 

Oregon for C. occidentalis. They found that foliage from western Oregon was more susceptible to budworm feeding and 

suggested that the greater diversity of trees, predators, and parasites in western Oregon has limited budworm populations and 

minimized selection for resistance to this insect. If diversity has been a major factor preventing budworm defoliation in 

western Oregon and Washington, then widespread commercial production of mid-successional P. menziesii forests may result 

in increasing C. occidentalis activity in this region. 

Outbreaks of D. frontalis in the southeastern U.S. also result from change in landscape diversity. Most of the land 

area in this region originally was vegetated by sparse woodlands and savannahs dominated by Pinus palustris, a species 

tolerant of the frequent fires and drought of this region and relatively resistant to bark beetles (Schowalter et al. 1981 a). 

Mesic riparian and bottomland forests included a diverse assemblage of intolerant species, including Pinus taeda and 

hardwoods. Bark beetles in this landscape would have been restricted primarily to scattered injured or diseased trees. Land 

conversion followed by eventuat abandonment and reforestation led to establishment of dense stands of rapidly growing and 

commercially valuable P. taeda over most of this region. This species is susceptible to D. frontalis, resulting in devastating 

outbreaks across the region (Schowalter et al. 1981a, Schowalter and Turchin 1993). 

These observations indicate that diversity can be effective in limiting herbivore epidemiology. However, diversity 

: may not provide protection when herbivores reach large population size in a sufficient proportion of the surrounding landscape. 

SUMMARY 

Plant vulnerability to herbivores is a function of both biochemical suitability and exposure to herbivore populations. 

Plant exposure to herbivores can be minimized by plant diversity. Plant diversity at genotypic, species, and landscape levels 

presents herbivores with a mosaic of suitable and non-suitable hosts and non-hosts. This increases the distance herbivores 

must travel to reach new hosts and increases the difficulty of discovering suitable hosts hidden by non-hosts in diverse stands 

and landscapes. By contrast, limited genetic diversity within monocultures of commercially valuable trees provides a 

concentrated resource for adapted herbivores. This resource inevitably will become more vulnerable to herbivores, especially 

during adverse conditions that stress plants and/or limit defensive capability. Localized outbreaks occur where isolated 

stands become vulnerable. These outbreaks can be restricted to isolated vulnerable stands in a diverse landscape that 

provides barriers to herbivore population spread. Regionwide outbreaks occur where a sufficient proportion of the landscape 

is occupied by suitable hosts. Large populations can expand into otherwise resistant stands and affect resistant or sparsely 

distributed hosts. 


ALSTAD, D.N. and EDMONDS, G.F., Jr. 1983. Adaptation, host specificity, and gene flow in the black pineleaf scale, p. 

413-426. In Denno, R.F. and McClure, M.S., eds. Variable Plants and Herbivores in Natural and Managed Ecosystems. 

Academic Press, NY. 

AMMAN, G.D., MCGREGOR, M.D., SCHMITZ, R.F., and OAKES, R.D. 1988. Susceptibility of lodgepole pine to 

infestations by mountain pine beetles following partial cutting of stands. Can. J. For. Res. 18: 688-695. 

BELANGER, R.P. and MALAC, B.E 1980. Silviculture can reduce losses from the southern pine beetle. Washington, DC: 

U.S. Department of Agriculture, Agric. Handb. 576. 

BERNAYS, E.A. and WOODHEAD, S. 1982. Plant phenols utilized as nutrients by a phytophagous insect. Science 216: 

201-202. 

25 

i

BRIGHT, D.E. 1993. Systematics of bark beetles, p. 23-36. In Schowalter, T.D. and Filip, G.M., eds. Beetle-Pathogen 

Interactions in Conifer Forests. Academic Press, London. 

CATES, R.G. and ALEXANDER, H. 1982. Host resistance and susceptibility, p. 212-263. In Mitton, J.B. and Sturgeon, 

K.B., eds. Bark Beetles in North American Conifers: a System for the Study of Evolutionary Biology. University of 

Texas Press, Austin. 

COLEY, ED., BRYANT, J.E, and CHAPIN, ES., III. 1985. Resource availability and plant antiherbivore defense. Science 

230: 895-899. 

COURTNEY, S.E 1986. The ecology of pierid butterflies: dynamics and interactions. Adv. Ecol. Res. 15:51-131. 

GARA, R.I. and COSTER, J.E. 1968. Studies on the attack behavior of the southern pine beetle. III. Sequence of tree 

infestation within stands. Contrib. Boyce Thompson Inst. 24: 77-86. 

HADFIELD, J.S. 1988. Integrated pest management of a western spruce budworm outbreak in the Pacific Northwest. 

Northw. Environ. J. 4: 301-312. 

HANSEN, E.M., GOHEEN, D.J., HESSBURG, EF., WITCOSKY, J.J., and SCHOWALTER, T.D. 1988. Biology and 

management of black-stain root disease in Douglas-fir, p. 63-80. In Harrington, T.C. and Cobb, F.W., Jr., eds. 

Leptographium Root Diseases on Conifers. American Phytophathological Society Press, St. Paul, MN. 

HARBORNE, J.B. 1982. Introduction to Ecological Biochemistry. Academic Press, London. 

HICKS, R.R., JR. 1980. Climate, site, and stand factors, p. 55-68. In Thatcher, R.C., Searcy, J.L., Coster, J.E., and Hertel, 

G.D., eds. The Southern Pine Beetle. Tech. Bull. 1631. Washington, DC: U.S. Department of Agriculture, Forest 

Service. 

HUNTER, A.E and ARSSEN, L.W. 1988. Plants helping plants. BioScience 38: 34-40. 

JOHNSON, EC. and COSTER, J.E. 1978. Probability of attack by southern pine beetle in relation to distance from an 

attractive host tree. For. Sci. 24: 574-580. 

KAREIVA, E 1983. Influence of vegetation texture on herbivore populations: resource concentration and herbivore movemerit, 

p. 259-289. In Denno, R.E and McClure, M.S., eds. Variable Plants and Herbivores in Natural and Managed 

Systems. Academic Press, NY. 

LORIO, EL., Jr. 1993. Environmental stress and whole-tree physiology, p. 81-101. In Schowalter, T.D. and Filip, G.M., 

eds. Beetle-Pathogen Interactions in Conifer Forests. Academic Press, London. 

MCCULLOUGH, D.G. and WAGNER, M.R. 1993. Defusing host defenses: ovipositional adaptations of sawflies to plant 

resins, p. 157-172. In Wagner, M.R. and Raffa, K.E, eds. Sawfly Life History Adaptations to Woody Plants. Academic 

Press, San Diego, CA. 

MCCUTCHEON, T.L. and CARROLL, G.C. 1993. Genotypic diversity in populations of a fungal endophyte from Douglasfir. 

Mycologia 85: 180-186. 

MITCHELL, R.G., WARING, R.H., and PITMAN, G.B. 1983. Thinning lodgepole pine increases tree vigor and resistanc_ 

to mountain pine beetle. For. Sci. 29: 204-211. 

PERRY, D.A. t988. Landscape pattern and forest pests. Northwest. Environ. J. 4: 213-228. 

PERRY, D.A. and PITMAN, G.B. 1983. Genetic and environmental influences in host resistance to herbivory: Douglas-fir 

and the western spruce budworm. Z. ang. Entomol. 96: 217-228. 

26

ROOT, R.B. 1973. Organization of a plant-arthropod association in simple and diverse habitats: the fauna of collards 

(Brassica oleracea). Ecol. Monogr. 43" 995-124. 

ROSENTHAL, G.A. and JANZEN, D., eds. 1979. Herbivores: Their Interaction with Secondary Plant Metabolites. Academic 

Press, NY. 

SARTWELL, C. and STEVENS, R.E. 1975. Mountain pine beetle in ponderosa pine: prospects for silvicultural control in 

second-growth stands. J. For. 73: 136-140. 

SCHOWALTER, T.D. 1989. Canopy arthropod community structure and herbivory in old-growth and regenerating forests in 

western Oregon. Can. J. For. Res. 19: 318-322. 

SCHOWALTER, T.D. and FILIR G.M., eds. 1993. Beetle-Pathogen Interactions in Conifer Forests. Academic Press, 

London. 

SCHOWALTER, T.D. and HAVERTY, M.I. 1989. Influence of host genotype on Douglas-fir seed losses to ContarMia 

oregonensis (Diptera: Cecidomyiidae) and Megasti'gmus spermotrophus (Hymenoptera: Torymidae) in western 

Oregon. Environ. Entomol. 18: 94-97. 

SCHOWALTER, T.D, and STEIN, J.D. 1987. Influence of Douglas-fir seedling provenance and proximity to insect population 

sources on susceptibility to Lygus hesperus (Heteroptera: Miridae) in a forest nursery in western Oregon. 

Environ. Entomol. 16: 984-986. 

SCHOWALTER, T.D. and TURCHIN, R 1993. Southern pine beetle infestation development: interaction between pine and 

hardwood basal areas. For. Sci. 39: 201-210. 

SCHOWALTER, T.D., COULSON, R.N., and CROSSLEY, D.A., Jr. 198la. Role of southern pine beetle and fire in 

maintenance of structure and function of the southeastern coniferous forest. Environ. Entomol. 10: 821-825. 

SCHOWALTER, T.D., POPE, D.N., COULSON, R.N., and FARGO, W.S. 198 lb. Patterns of southern pine beetle 

(Dendroctonusfrontalis Zimm.) infestation enlargement. For. Sci. 27: 837-849. 

TORGERSEN, T.R., MASON, R.R., and CAMPBELL, R.W. 1990. Predation by birds and ants on two forest insect pests in 

the Pacific Northwest. In Morrison, M.L., Ralph, C.J., Verner, J., and Jehl, J.R., Jr., eds. Avian Foraging: Theory, 

Methodology, and Applications. Cooper Ornithological Society, Studies in Avian Biology 13: 14-19. 

TURCHIN, R 1988. The effect of host-plant density on the numbers of Mexican bean beetles, Epilachna varivestis. Am. 

Midl. Nat. 119: 15-20. 

VIA, S. 1990. Ecological genetics and host adaptation in herbivorous insects: the experimental study of evolution in natural 

and agricultural systems. Annu. Rev. Entomol. 35: 421-446. 

VIA, S. 1991. The genetic structure of host plant adaptation in a spatial framework: demographic variability among reciprocally 

transplanted pea aphid clones. Evolution 45: 827-852. 

VISSER, J.H. 1986. Host odor perception in phytophagous insects. Annu. Rev. Entomol. 31 : 12I- 144. 

WARING, R.H. and PITMAN, G.B. 1983. Physiological stress in lodgepole pine as a precursor for mountain pine beetle 

attack. Z. ang. Entomol. 96: 265-270. 

WITCOSKY, J.J., SCHOWALTER, T.D., and HANSEN, E.M. 1986a. The influence of precommercial thinning on the 

colonization of Douglas-fir by three species of root-colonizing insects. Can. J. For. Res. 16: 745-749. 

WITCOSKY, J.J., SCHOWALTER, T.D., and HANSEN, E.M. 1986b. Hylastes nigrinus (Coleoptera: Scolytidae), Pissodes 

fasciatus, and Steremnius carinatus (Coleoptera: Curculionidae) as vectors of black-stain root disease of Douglas-fir. 

Environ. Entomol. 15: 1090-1095. 27

' T 

ON NEIGHBOR EI_FEC S IN PLANT-HERBIVORE INTERACTIONS w e 

g 

JUHA TUOMI, MAGNUS AUGNER, and PATRIC NILSSON 

Department of Ecology, Theoretical Ecology, Lund University, T Ecology Bldg., S-223 62 Lund, Sweden 

INTRODUCTION 

Coevotution involves genetic changes that occur in both plant and herbivore populations over evolutionary time. 

Genetic changes in turn depend on fitnesses associated with given plant and herbivore genotypes. This leads us to the 

ecologicaI time scale because fitness is determined by functional interactions between individual plant units and herbivores 

over their life-spans. Such interactions can be formalized as a dependency of fitness on plant and herbivore phenotypes in a 

giver_context of interaction, 

The basic features of co-evolutionary processes depend primarily on the direct effects that an individual's phenotype 

has of_its fitness. However, as pointed out by Atsatt and O' Dowd (1976) and Rhoades (1979) among others, the actual 

interactions are often much more complicated due to various indirect effects when fitness depends on what the neighbors are 

doing. We label these indirect relations as neighbor effects (Eshel 1972), and argue that they may have a fundamental 

..... 

importance in plant-herbivore interactions (Mattson et al. 1991). 

We discuss two aspects of indirect interactions between plants. First, in some cases palatable plants gain a benefit by 

growing close to unpalatable plants. This possibility was first discussed by Tahvanainen and Root (1972) and Atsatt and 

O'Dowd (1976), and may be called "associational refuge" effect (Pfister and Hay 1988). Several experiments have demonstrated 

a refuge effect (McNaughton 1978, Hay 1986), but not always (McNaughton 1978, Danell et al. 1991, ttj;,ilten et al. 

1993). 

Second, it is frequently assumed that herbivores can choose their host plants depending on their nutritive value and 

toxicity. This presumes explicitly that herbivores can recognize the host plant and "read" cues that correlate with palatability. 

If the herbivore can generalize the cues (Launchbaugh and Provenza 1993), unpalatable plants that repel herbivores will 

naturally gain a benefit themselves. Likewise, the herbivore should avoid other plants with these very same cues. That is 

'_ (3" " 

why we expect that plant defenses may imply syner_:st_c benefits" in the sense that a plant implies a benefit for other 

individuals sharing the same phenotype (Nee 1989,Guilford and Cuthill 1991, Rosenberg 199 l, and Tuomi and Augner 

993), ..... 

We propose that these neighbor effects may well shape some aspects of plant defenses. For this purpose, we present a 

simplified model system where plants are :interactingin pairs. Then we analyze the model system in a game theoretical 

" 

context by assuming three phenotypes: non-defensive type (N), and two defensive types - one which kills the invading 

herbivores (D), and another that provides herbivores a signal over which to generalize (S). 

MODEL SYSTEM 

We assume a plant population where individual plants are interacting in pairs, i.e., in trait groups (Wilson 1975) of 

two plants. Each pair has a risk of herbivory, say m: that is a function of grazing intensity over the whole population and of 

the phenotypic composition of the given trait group. We call the plants the player and the neighbor (or the opponent). An 

invading herbivore selects the player with probability s and the neighbor with probability 1-s. We will, for simplicity, assume 

that the consequent risk of herbivory is equal for both the player and the neighbor so that mps = mp(1-s) = m in all trait 

groups, in other words, invading herbivores can select neither between trait-groups nor between the player and the neighbor. 

Mattso__,,W.J., Niemela, P., and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 

For. Secy. Gen. Tech. Rep. NC-183, N.C. For. Exp. Sta., St. Paul, MN 55108. 

28

We tlhus neglect two potential sources of neighbor effects because experienced herbivores may well be able to evaluate the 

expected value of a patch or a plant from a distance, e.g., by appearance or scent. Hjfilten et al. (1993) have discussed in a 

greater length how selection between and within patches can affect associational refuges. 

In short, our herbivores are naive, but we assume that they can learn to evaluate the defensive status of the plants if 

appropriate cues are present. The consumption process is divided into two parts (bites). From our earlier assumptions, it 

follows that the risk of the first bite is m. Each bite by a herbivore implies a fitness cost for the plant. Plant defenses can in 

various ways reduce the total cost of herbivory: reducing the fitness cost caused by a herbivore per bite, increasing mortality 

(d) of herbivores during a bite, and inducing feeding aversion that can be either unconditional (the herbivore does not take the 

second bite of any food) or conditional (the herbivore does not take the second bite if a specific cue is present). In our case of 

two bites, unconditional feeding aversion is equivalent to increased mortality as in both cases further herbivory is stopped. 

We thus consider only conditional feeding aversion and the parameter a indicates the probability that feeding aversion is 

induced during a bite. We also assume that the cost of herbivory is h if the herbivore survives and no feeding aversion is 

induced, while the cost of herbivory will be h' if the herbivore happens to die or if a feeding aversion is induced during a bite 

(O < h' < h). We assume specifically that non-defensive (N) plants cause no mortality (d - 0, a - 0), while there are two kinds 

of defensive plants; those (D) that will kill the herbivore during the bite (d = 1, a - 0), and those (S) that cause a conditional 

feeding aversion before any toxic effects appear on herbivore survival (a - 1, d = 0). 

Both kinds of defenses (D and S) reduce the average costs that herbivory implies to the player. In addition to these 

direct fitness effects, various neighbor effects can arise if herbivores are allowed to move from the player to the neighbor and 

vice versa. We assume that a surviving herbivore moves after the first bite with a probability e. For simplicity, feeding 

aversion is not allowed to affect this probability, and the parameters h, h', d, and a assume the same values during the second 

bite as earlier during the first bite. Taken together, these assumptions imply that the average cost (H) of herbivory over two 

bites will be 

H=[(1-di)(1-ai)h + (1-di)ah' + d'ih](1 +nii+nij) 

ni_= (1-d,)(1-a)(1-e) 

nij = { (1-di)(1-ai)ei (i,j=S) } 

(1-dj)ej (i v j aS) 

where n_ implies the probability of a second bite for the player by a herbivore initially invading the player (denoted by i), and 

n_j by a herbivore moving from the neighbor (denoted by j) to the player. The latter term will be (1-d)(1-a)ej if feeding 

aversion induced by the neighbor protects the player (i.e., both S), and (1-d)e if the neighbor does not induce a feeding 

aversion (i.e., the neighbor N or D) or if the player does not possess the cue required for the maintenance of feeding aversion 

(i.e., the neighbor S and the player either N or D). In our specific case (ei = e. l = e), H = h[1 + (l-e) + n_j]for a player adopting 

N (d = 0, a = 0), and H = h'(1 + nij)for a player adopting D (d = 1, a = 0) or S (d = 0, a = 1). When H is multiplied by m we 

will get, following Augner et al. (1991), a measure of the herbivory load, Hm, for the payoff matrix. 

This highly simplified model system elucidates clearly that the fitness of the player may well depend on the neighbor. 

In our case, this is so if herbivores can move between the plants (0 < e

¢D 

Neighbor 

N D S 

N -2m -(2-e)m -2m 

>_ D 1 (l+e)m-C 1 m-C 1 

¢_ -_ - _ - _ (l+e)m-C 

, ,, 

S -_1 (l+e)m-C(l+k) -21 m-C(l+k) -21 m-C(l+k) 

Figure 1.--Payoff matrix for a game between two plants showing the payoffs to the player adopting either nondefensive 

strategy (N), lethal defense (D), or defense inducing conditional feeding aversion (S). (m) denotes the risk of 

herbivory, C as well as k indicate the costs of defenses, and e is the probability that a herbivore moves from one plant 

to another. 

the model so that D and S types have similar payoffs when playing against N; the only difference is the cost of defense. Our 

primary interest lies in the qualitative differences between the two defensive types D (lethal defense) and S (conditional 

feeding aversion). We define parameter space where these phenotypes are evolutionarily stable strategies (ESS) in the three 

subgames (Figs. 2-4). Finally, we study the dynamics of the game when all three strategies are taken account (Figs. 5 and 6). 

In all cases, we have random associations and the population frequencies of N, D, and S are q, p, and r respectively. 

A Disadvantage of Lethal Defense 

The first subgame of N and D contrasts a nondefensive strategy (N) against a defensive strategy (D) that kills all 

invading herbivores and that implies a cost, C. When the trait groups are randomly derived from the population frequencies q 

= l - p for N and p for D, the fitnesses will be 

W = Wo - (2-pe) m 

W_ =W 0- C- 1/2 [1 + (1-p)e]m 

where W o is a basic level of fitness. When e = 0, selection will exclusively favor either N or D depending on the balance 

between the cost of defense (C) and the risk of herbivory (m). When e > 0, some herbivores will move from N to D. This 

wifl reduce the herbivory load of N and increase that of D in mixed pairs. As a consequence, we will have three possibilities 

for different combinations of C/m and e (Fig. 2a): 

(1) N will be an ESS for C/m> (3 - e)/2. 

(2) D will be ESS tbr C/m < (3 - 2e)/2. 

(3) Stable polymorphism (0 < _ < 1) is obtained when neither N nor D are ESSs. 

3O

c__ c__ 

1T1 Ill 

a) a) 

0 0 

0 e 1 o e 1 

W N 

Z Z 

i 

I 

b) b) 

! I I 

0 _ p 1 0 9 r 1 

Figure 2.--ESS-conditions in N-D subgame contrasting Figure 3.--ESS-conditions in N-S subgame contrastnondefensive 

strategy (N) and lethal defense (D): ing nondefensive strategy (N) and defense (S) 

[a] The parameter areas where N and D are pure that induces conditional feeding aversion: [a] 

ESSs are shaded, while in the blank area neither The parameter areas where N and S are ESSs 

of them is an ESS; and [b] Fitnesses of N and D are shaded; and [b] Fitnesses of N and S as a 

as a function of population frequency (1_) of D function of the population frequency (r) of S 

when 0 < _ < 1. I_ = stable equilibrium; C, m, and when 0 < _ < 1. _ = unstable equilibrium; C, 

e as in Fig. 1. m, and e as in Fig. 1. 

31

Z 

C ,,,,, ,,,, 

m 

0 

a) 

0 e I 

I 

I 

I 

ws 

b) 

I 

0 _" r 1 

Figure 4.--ESS-conditions in D-S subgame contrasting lethal defense (D) and defense (S) that induces conditional 

feeding aversion: [a] The parameter areas where D and S are ESSs are shaded (k > 0); and [b] Fitnesses of D and 

S as a function of the population frequency (r) of S when 0 < } < 1. T 2= unstable equilibrium; C, m, and e as in 

Fig. 1. 

32

c_ >l.s S S S 

m 

N < DN DN < D 

c DN DN D 

k =0 k = 0.4- k = 1 

Figure 5.--Basic game dynamics when neighbor effects are excluded (e = 0). N, D, and S correspond to populations where 

q = 1, p = 1 and r = 1 respectively. 

The third situation corresponds to "evolutionarily stable" associational refuges as both plant types coexists (Sabelis and de 

Jong 1988, Tuomi and Augner 1993). The coexistence is possible because both types have a selective advantage when rare, 

but a disadvantage when common (Fig. 2b). In our case, stable polymorphism requires that D has a higher fitness value for p 

close to 0. However, when p increases toward 1, WN increases more steeply than W D. The slope of W Nas a function p is em, 

while the slope of W Dis (1/2)era. The reason for this difference is that h' = 1/2. Consequently, the associational protection 

has a greater value for N that suffers h = 1, while D suffers only a half of that value. If both N and D would win equally from 

the neighbors' defenses (i.e., h' = h), associational protection would not maintain stable polymorphism in the present case. 

The major lesson of this subgame is that the associational effects counteract the evolution of lethal defenses. We 

have proposed that this implies a dilemma for plant defenses: the defenses should protect the plant without benefiting 

nondefensive neighbors too much (Tuomi et al. submitted). If the defensive plants kill all herbivores, the nondefensive type 

can easily invade. Consequently, one could expect that "less lethal" defenses could do a better job in the long run. However, 

when adopting sublethal defenses the plant itself will suffer more damage by herbivores. There are at least two solutions for 

the dilemma. First, if a plant is allowed to adopt a lethality-level (d) between 0 and 1, associational refuge effects select for 

sublethal defenses with d < 1 (Tuomi et al. submitted). Second, another solution of the dilemma could be a defense strategy 

(S) that is potentially "lethal" to the herbivores, but that is associated with chemical or visual traits which herbivores can use 

as cues in order to avoid the potentially "lethal" defenses. Below, we analyze this second possibility. Since we have set d = 0 

for S, we thus assume that feeding aversion is induced before the amount of ingested food reaches any toxic level. 

33

C__=l.G S S S 

m 

N DNv _-. DN D 

c o8 S S S 

m- " 

N DN .. DN D 

P P P 

c 04 S S S 

DN l > -DN D 

k =0 k =0.4 k = 1 

Figure 6.--Basic game dynamics when herbivores are allowed to move between the interacting plants (e = 1). N, D, and 

S correspond to populations where q = 1, p = I, and r = 1, respectively. Q is an unstable equilibrium and R is a 

saddle point, while P can be a saddle point or a stable equilibrium. Trajectories will eventually approach the 

points denoted by the dots. 

Conditional Feeding Aversion 

We formulated the present model so that the player is subject to the exactly same herbivory load when adopting either 

D or S, and when the neighbor is N (Fig. 1). There are two major differences when N-S subgame is compared to N-D 

subgame. First, we assume that S and D may imply different costs for the plant. We have defined that the cost of S is 

C(l+k). To be conservative, we assume that k is non-negative indicating that it is equally or more costly to induce and 

maintain feeding aversion than to kill the herbivore. If both D and S require an equal investment in defensive compounds, 

34

:hen k > 0 suggest that the cue or the signal associated with S is costly. Second, the major difference between S and D is that 

does not allow N to escape from herbivory in mixed pairs. That is why the fitness of N, when playing against S, reduces to 

whereas Ws will be 

W N= W0- 2m 

Ws = W0 - C(I + k)-l/2 [l+(1-r)e]m 

where r is the population frequency of S. The population frequency of N is q = 1 - r. Also in this case we have three 

possibilities (Fig. 3a): 

(1) N will be an ESS when C/m > (3 - e)/2(l+k). 

(2) S will be an ESS when C/m < 3/2(1+k). 

(3) There will be a unstable equilibrium 0 < _ < 1 when both N and S are ESSs. 

In other words, no stable coexistence is possible since S will be an ESS above the equilibrium frequency ?, and N will be an 

ESS below the equilibrium (Fig. 3b). 

In summary, this subgame has two interesting aspects. The first is that nondefensive plants do not benefit from S 

because herbivores are assumed to be able to make a distinction between N and S types. The second is that the defensive 

plants will benefit from each other if feeding aversion is maintained by a cue that the neighbors share. As a consequence, the 

fitness of defensive plants will increase as the frequency of S approaches 1, while the fitness of nondefensive plants remains 

constant (Fig. 3b). This situation corresponds to "synergistic selection" discussed by Guilford and Cuthill (1991) in the 

evolution of aposematism, and by Tuomi and Augner (1993) in relation to plant defenses. 

Costs of Defenses 

When allowing D and S to play against each other, a plant adopting D will obtain the same payoffs as when playing 

against N (Fig. 2). A plant adopting S will obtain the same payoffs when the neighbor is D or S. When r is the population 

frequency of S, we get the following fitnesses 

W D=W0 - C - 1/2 m(1 +er) 

W s =W 0- C(1 + k)- 1/2 m 

where W Ddeclines as a function of er, while Ws is independent of both e and r. There are four qualitatively different 

situations: 

(1) S will be a pure ESS either ifk < 0, or ifk = 0 and e > 0. 

(2) S and D are equally fit if k = 0 and e = 0. 

(3) D will be a pure ESS if k > 0 and C/m > e/2k. 

(4) There is an unstable equilibrium 0 < } < 1 if k > 0 and C/m < e/2k. 

Consequently, the costs of defenses play a prominent role here. If S is expensive relative to D, lethal defense is most 

likely to be selected for. Fig. 4a indicates the last two situations obtained for k > 0. The unstable equilibrium 

? = 2Ck/em 

represents the point where the fitness curves intersect each other (i.e., Ws = WD,Fig. 4b). When Fmoves closer to zero, 

the situation becomes more favorable to S. This will happen when k becomes smaller, or when e and m become larger. 

The reverse changes will accordingly favor D. We interpret this so that lethal defenses may be more economical against 

rare herbivores (m small) and sedentary herbivores (e small), while defenses inducing conditional feeding aversion 

should be more effective against common herbivores that are mobile and move from a plant to another (both m and e 

high). 

35

Game Dynamics 

When the above subgarnes are combined, there arises a number of possible dynamic solutions for the game where a_t 

three strategies are taken account. We have chosen to describe some basic situations for e = 0 (Fig. 5) and e = 1 (Fig. 6) whe_, 

k 2 0. The population frequencies of N, D and S are q, p and r respectively (q + p + r = 1), and the equilibria discussed abo'_e 

are denoted by P = (q, fi, O) (the stable equilibrium of the N-D subgame), Q = (_t,0, _) (the unstable equilibrium of the N-S 

subgame), and R = (0, _,'?) (the unstable equilibrium of the D-S subgame). 

When e = 0, them are no neighbor effects and hence no frequency dependency. Now the game dynamics can be 

directly derived from fitness ranks, as 

WN= W_ - 2m 

Ws =W_)- C(1 + k)- l/2m 

Ws = W0-C(1 +k)- l/2m 

When C is sufficiently low relative to m, selection will favor a defensive strategy, either S or D. For k > O, selection 

will always favor D over S. On the other hand, when C is high relative to m, N will always be an ESS (Fig. 5). 

When e = l, the situation is a little bit more complicated as now fitnesses depend on population composition. Using 

the standard methods (Taylor and Jonker 1978, Zeeman 1981,Bomze 1983, Hofbauer and Sigmund 1988), we can analyze 

the local stability of equilibria (Fig. 61).Also in this case, selection will favor a defensive strategy if C is sufficiently low 

relative to m. However, now S will have an advantage over D when k = 0. For a higher values of k (e.g., k = 0.4 or 1.0, Fig. 

6)_the equilibria Q and R appear with the consequence that, depending on the initial state, the population may evolve toward 

S, or alternatively to D (or P). Finally, ifC is high relative to m, INwill have an advantage over both D and S. 

This analysis confin,as the earlier result of the D-S subgame that high m and e favor S (Fig. 6), and low e favors D 

(Fig. 5)_ However, the suggestion that low m favors D over S, does not imply that D is an ESS. When m is low enough, N 

can invade and either out-compete both defensive types or establish stable polymorphism with D. 

CONCLUSIONS 

Plant de:lenses have both direct effects on plant fitness and indirect effects that arise from interactions between 

neighboring plants. Although direct fitness effects obviously are the primary forces of selection, neighbor effects may also 

shape some basic aspects of plant defenses. This is especially so if fitness differences due to the direct effects of plant 

phe_otypes are marginal. 

We analyzed two types of neighbor effects when herbivores are allowed to move from one plant to another: 

(t) Palatable plants can benefit from the defenses of their unpalatable neighbors. Defenses can, for instance, increase 

mortality among herbivores or induce :unconditional feeding aversion with the consequence that the herbivory load of 

palatable plants is low in mixed trait groups. Associational protection can maintain stable polymorphism if palatable plants 

gain more in fitness than unpalatable plants when growing close to an unpalatable neighbor. We thus expect that associational 

protection counteracts the evolution of plant defenses that do not provide herbivores any signal over which to generalize_ 

(2) Unpalatable plants can selectively benefit other such plants if they share a trait that herbivores can use as a basis 

of their food intake and their choice for host plants. In that case, palatable and unpalatable plants that do not share the signal 

do not gain assc_:iational protection. Defenses associated with a signal are most effective against abundant and mobile 

he:rbivores, and their advantage is greatest when they are common. 

Conseque__tly, we expect that there should be some correspondence between the design of plant defenses and the 

sensory and Iearning mechanisms of herbivores. Undoubtedly, thorns and spines are defensive characters that can functiorl 

36

oth as protective weapons and as potential signals. If this also holds for chemical defenses, neighbor effects might have a 

_ndamental importance in shaping plant defenses. This may well be so as, because according to Provenza et al. (1992), 

mmmalian herbivores learn to select food items through two interrelated systems. The affective system integrates the taste 

f food and its post-ingestive consequences, while the cognitive system integrates the odor and sight of food and its taste. 

tecause they have to sample plants in order to adjust food intake to avoid intoxication (Provenza et al. 1992), the expected 

erbivory load of a plant is likely to depend both on its own defenses and on the defensive status of its neighbors. 

SUMMARY 

In order to be evolutionary stable, plant defenses should not benefit palatable neighbors too much. If they do, 

_alatable plants can invade and eventually either out-compete unpalatable plants or establish polymorphic populations. 

Therefore, we expect that plant defenses that can provide signals for herbivores could be superior to defenses that have no 

,alue as signals. A game theoretical analysis is presented in order to explore the evolution of plant defenses when plant 

]tness depends on the defensive status of the neighbor. 


\TSATT, RR.and O'DOWD, D.J. 1976. Plant defense guilds. Science 193: 24-29. 

_.UGNER, M., FAGERSTROM, T., and TUOMI, J. 1991. Competition, defense and games between plants. Behav. Ecol. 

Sociobiol. 29:231-234. 

3OMZE, I.M. 1983. Lotka-Volterra equation and replicator dynamics: A two-dimensional classification. Biol. Cybernetics 

48:201-211. 

)ANELL, K., EDENIUS, L., and LUNDBERG, R 1991. Herbivory and tree stand composition: Moose patch use in winter. 

Ecology 72: 1350-1357. 

ESHEL, I. 1972. On the neighbor effect and the evolution of altruistic traits. Theor. Pop. Biol. 3: 258-277. 

3UILFORD, T. and CUTHILL, I. 1991. The evolution of aposematism in marine gastropods. Evolution 45:449-451. 

HAY, M.E. 1986. Associational plant defenses and the maintenance of species diversity: Turning competitors into accomplices. 

Am. Nat. 128: 617-641. 

HJALTEN, J., DANELL, K., and LUNDBERG, R I993. Herbivore avoidance by association: Vole and hare utilization of 

woody plants. Oikos 68: 125-131. 

HOFBAUER, J. and SIGMUND, K. 1988. The Theory of Evolution and Dynamical Systems. Cambridge Univ. Press, 

Cambridge. 

LAUNCHBAUGH, K.L. and PROVENZA, F.D. 1993. Can plants practice mimicry to avoid grazing by mammalian 

herbivores? Oikos 66: 501-504. 

MATTSON, W.J., HERMS, D.A., WITTER, J.A., and ALLEN, D.C. 1991. Woody plant grazing systems: North American 

outbreak folivores and their host plants, p. 53-84. In Baranchikov, Y., Mattson, W.J., Hain, ER, and Payne, T.L., eds. 

Forest insect guilds: patterns of interaction with host plants. Gen. Tech. Rep. NE-153. Radnor PA: U.S. Department 

of Agriculture, Forest Service. 400 p. 

McNAUGHTON, S.J. 1978. Serengeti ungulates: Feeding selectivity influences the effectiveness of plant defense guilds. 

Science 199: 806-807. 

NEE, S. 1989. Does Hamilton's rule describe the evolution of reciprocal altruism? J. Theor. Biol. 141: 81-91. " 

37

PFISTER, C.A. and HAY, M.E. 1988. Associational plant refuges: Convergent patterns in marine and terrestrial communities 

result from differing mechanisms. Oecologia 77: 118-129. 

PROVENZA, ED., PFISTER, J.A., and CHENEY, C.D. 1992. Mechanisms of learning in diet selection with reference to 

phytotoxicosis in herbivores. J. Range Manage. 45: 36-45. 

RHOADES, D.E 1979. Evolution of chemical defense against herbivores, pp. 3-54. In Rosenthal, G.A. and Janzen, D.H., 

eds. Herbivores: Their Interactions with Secondary Plant Metabolites. Academic Press, New York. 

ROSENBERG, G. 1991. Aposematism and synergistic selection in marine gastropods. Evolution 45:451-454. 

SABELIS, M.W. and de JONG, M.C.M. 1988. Should all plants recruit bodyguards? Conditions for a polymorphic ESS of 

synomone production in plants. Oikos 53: 247-252. 

TAHVANAINEN, J.O. and ROOT, R.B. 1972. The influence of vegetational diversity of the population ecology of a 

specialized herbivore, Phyllotreta cruciferaea (Coleoptera: Chrysomelidae). Oecologia 10: 321-346. 

TAYLOR, RD. and JONKER, L.B. 1978. Evolutionary stable strategies and game dynamics. Math. Biosci. 40:145-1 56. 

TUOMI, J. and AUGNER, M. 1993. Synergistic selection of unpalatability in plants. Evolution 47: 668-672. 

WILSON, D.S. 1975. A theory of group selection. Proc. Natl. Acad. Sci. USA 72: 143-146. 

ZEEMAN, E.C. 1981. Dynamics of the evolution of animal conflicts. J. Theor. Biol. 89: 249-270. 

38

DEFENSE THEORIES AND BIRCH RESISTANCE 

MATTI ROUSI t, JORMA TAHVANAINEN 2, WILLIAM J. MATTSON 3 

and HANNI SIKANEW 

_The Finnish Forest Research Institute, Punkaharju Research Station, SF-58450 Punkaharju, Finland 

2University of Joensuu, Depamnent of Biology, P.O. Box 111, SF 80101 Joensuu, Finland 

3USDA Forest Service, North Central Experimental Station, 1407 Harrison Rd, East Lansing, M148823 USA 

INTRODUCTION 

Different tree species in northern temperate forests clearly vary in their resistance to browsing mammals (Gill 1993 

a,b). In addition to variation at the species level, there are also clear indications of genetic differences in herbivore resistance 

within species. Such variation in resistance has been explained by the duration and intensity of past exposure to herbivores 

and pathogens (Leppik 1970, Bryant et al. 1989). It is believed that the centers of genetic diversity for plant species should 

be the best places in which to find resistance to their common herbivores and diseases (Vavilov 1920). All of these kinds of 

variation have important economic and ecological implications for plantation forestry and forest industries. For example, 

hybridization could be used to elevate the resistance of otherwise high quality but susceptible tree species or individuals (see 

Rousi 1990). 

It has been postulated that there is a trade-off between growth and plant defense (Rhoades 1979, Herms and Mattson 

1992). That being the case, practical forestry might not be interested in slow growing genotypes, although they might be 

more safe to use. Trees are generally attacked by a myriad of herbivores and if there is a trade-off between resistance and 

growth to each of those, then generally resistant genotypes should be very slow growing. Ecological trade-offs in the form 

of negative correlations in resistance to different types of herbivores are also possible. Conventional wisdom suggests that 

trees which have to compete intensely for nutrients and light, can probably be resistant only to restricted number of herbivores. 

In addition to genetic components, environment may also modify plant resistance. If carbon is the limiting factor due 

to shading, or to large amounts of available nutrients and water, then defensive secondary metabolite level may decline so 

long as the plant still has strong sinks capable of usurping the available photosynthates and nutrients (Mattson 1980, Bryant 

el al. 1983). 

To test the effect of environment and birch genotype on hare resistance, we carried out several cafeteria experiments 

using winter dormant shoots of small birch seedlings and plantlets grown in Punkaharju, East Finland. 

RESULTS AND DISCUSSION 

Birch Species and Genotypes Vary In Their Hare Resistance 

Birch species vary strongly in their hare resistance. According to many studies, Japanese white birches, B. 

platyphylla and B. ermanii, are both very resistant. Some other species, such as B. alleghaniensis (Fig. 1) and B. papyrifera, 

are among the most susceptible species, whereas B. pendula is somewhat intermediate in resistance. Interestingly, experiments 

made in Alaska, Finland, and Japan all give essentially the same results (Chiba and Nagata 1968, Bryant et al. 1989, 

Rousi et al. 1989, 1996a). This most probably indicates the wide adaptability of birch: there are no drastic changes in 

resistance even if birches are grown as exotics. In addition, it shows that the Japanese mountain hare, Lepus ainu, the Finnish 

mountain hare, L. timidius, and the snowshoe hare, L. americanus, make their food selection on the same bases. 

Mattson, W.J., Niemel',i, R, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 

For. Serv. Gem Tech. Rep. NC-183, N.C. For. Exp. Sta., St. Paul, MN 55108. 

39

P 

AL 70 

A 60 

T 

A 50 

B 

I 

I. 40 

I 

T 30 

Y 

20 

I 

N 

D 10 

E 

X 0 

B.alleghaniensis 16 x 21 15 x 21 17 x 21 B.pendula clone 39 

Figure I ._Hare palatabi]ity of B. alleghaniensis, three backcrossesof B. pendula andpla_phylla, local forest origin of B. 

pendul= andone micropropagatedc]oneof B. pendu{a. (::lone39 consistsof micropropagated plantlets of B. pendula. 

Mother trees (15,16 and 17) are hybrids between the two species and full sibs. Father trees (18,19 and 21) are South 

Finnish plus trees. Mean +S.E. Palatability index is the result from cafeteria experiments (two nights, six hares. For 

details see Rousi et al. 1991). Mean +S.E. 

Hybridization has been shown to be a very promising method to increase e.g., vole resistance of Larix species (Chiba 

et al. 1982). In addition, hybrids between susceptible and resistant birch species have been intermediate in their resistance to 

hares (Chiba and Nagata 1969). In order to make preliminary tests of the inheritance of resistance, we tested the resistance of 

a backcross between Japanese and European birch [a hybrid between a Japanese white birch (very resistant) and European 

white birch (intermediate) was crossed with several South Finnish birch genotypes]. 

The Japanese white birch used as a parent tree in our crossings was always of the same genotype. Generally, 

resistance of hybrids was at the same level as was that of local forest origins of European white birch (Figs. 1 and 2). Hybrids 

between Japanese and Fennoscandian white birches have shown superior juvenile growth as reported by Johnsson 

(1966). But the first year growth of our backcrossed seedlings was the same as that of European white birch (see Figs. 3 and 

4). Obviously the genetic base of our crossings was too limited for any far reaching conclusions. 

Birch species and origins derived largely from Pleistocene refugia such as Alaska, Siberia or Japan, are supposed to 

be more resistant than ecologically similar birches derived from nonrefugia (Bryant et al. 1989, Rousi 1988). It is difficult to 

test this hypothesis because it is difficult to verify whether particular birch populations came from authentic "refugia" or not. 

Moreover, no one knows the kind of selection pressure such populations have faced since release from their refugia. The 

most resistant and the most susceptible species in our experiments are from Japan, a country that was minimally glaciated 

during the Pleistocene. Moreover, B. pendula origins from Siberia (not glaciated) and Finland (heavily glaciated) are equally 

resistant to hares (Rousi et al. 1996a). Experiments with micropropagated plantlets have revealed very large variation in hare 

resistance among different B. pendula genotypes growing in South Finland. Clone 39 especially has proved to be highly 

resistant under various experimental conditions (Fig. 5, Rousi et al. 1996b). 

Fast Growing Birches Are Also Resistant 

We tested the resistance of persistence-adapted yellow birch, B. alleghaniensis, against fast growing, short-lived 

white birches in order to find out whether inherent growth strategy is related to hare resistance of the species. The first year 

(greenhouse) growth of yellow birch was less (ANOVA, Bonferroni t-test, p

P 45 

A 

L 40 

A 

T 35 

A 

B 30 

I 

L 25 

I 

T 20 

Y 

15 

i 10 

N 

D 5 

E 

X 0 

16x19 16x18 15x18 17x19 17x18 clone 39 

e 2.--Hare palatability of backcrosses between B. pendula and B. platyphylla. Mother trees (15,16 and 17)are hybrids 

between the two species and full sibs. Father trees (18,19 and 21) are South Finnish plus trees. For more details see 

Figure 1. Mean +S.E. 

P 

A 

L 60 

A [] B.alleghaniensis 

T 

A 

50 

B 

I 

L 

40 

I 30 mB.pendula 

T [] hybrid [] hybrid 

Y 

20 

[] hybrid 

I r = -.794 

N 10 p = .109 

D 

E 0 _-- I I I I I I ; I : : 

X 

62 63 64 65 66 67 68 69 70 71 72 73 

SEEDLINGHEIGHT(cm) 

•e 3.--Relationship between seedling growth and hare palatability. Material is the same as in Figure 1- H is for hybrid 

(backcross). Clone 39 is omitted because plantlets were ca. 2-yr-old; seedlings were 1-yr-old. 

41

p 

A 40 

L 

A m 

T u 

A 35 

B 

I 

k 

i 30 

T 

Y 

II 

t 

!1 

25 

i r = -.580 

N 

D 

p = .305 

E 20 ..................... t -- : : : : , ,_ ; _ " 

X 63 64 65 66 67 68 69 70 71 72 73 

SEEDLING HEIGHT (cm) 

Figure 4.--Re|ationship between seedling growth and hare resistance. Material the same as in Figure 2. Clone 39 is 

omitted because plantlets were ca. 2-yr-old; seedlings were l-yr-old. 

P 

A 60 

L 

A 

T 50 

A 

B 

40 

I 

L 

! 30 

T 

Y 

20 

! 

N 10 

D 

E 

X 0 

K 1889 109 39 

CLONES 

Figure 5.--Hare resistance of three micropropagated genotypes of B. pendula. Clone 109 stands for f. bircalensis (lobed 

leaves). Cafeteria experiments as in Figurel. Means +S.E. 

42

However, hares showed strong preference for it. The general trend in our experiments has been that species and genotypes 

resistant to hares showed better than average growth, although the relationship was not statistically significant (Figs. 3 and 4). 

These results support our earlier findings which indicate either a positive relationship between first year growth of birch 

seedlings and their hare resistance, or no relationship at all (Rousi et al. 1989, 1990, 1991, 1996b). 

European white birch genotypes have not shown any relationship between their growth and resistance to hares, 

Microtus voles, Curculionidae weevils, or Melampsoridium leaf rust (Poteri and Rousi 1996, Rousi et al. 1996b). These 

experiments did not support any ecological trade-off either: resistance to any one pest was not related to resistance to others. 

However, vole and hare resistance in some experiments have shown a positive relationship (Chiba and Nagata 1969), and 

have also tended to do so in our own experiments (Rousi et al. t989, 1996b). Interestingly, clone 39 has shown generally 

high resistance to both biotic and abiotic threats (Fig. 5, Rousi et al. 1996b). The same clone also seems to be exceptionally 

resistant to ozone injury (clone 5-M, in P_ifikk6nen et al. 1993), whereas the generally herbivore susceptible clone 36 is 

similarly very susceptible to ozone damage (2-M, in P_i_ikk6nenet al. 1993). 

Environment May Sometimes Affect Birch Resistance 

Our previous experiments with various birch species (8 birch species) and origins have indicated, as expected, that 

growth environment (shading or fertilization) does not have an effect on low resistance species (slow or fast growing), such 

as B. papyrifera, pubescens, maximowicziana, schmidtii or grossa. In high resistance species, fertilization increased the 

palatability of two species, B. platyphylla and B. ermanii, but did not affect another, B. resinifera. Shading had no clear 

effect on any of the tested species (Rousi et al. 1996a). 

European and Japanese white birches are closely related (Dugle 1966). However, fertilization did not diminish the 

hare resistance of European white birch (Rousi et al. 1991, 1993), but lowered that of Japanese white birch (Rousi et al. 

1996a). On the other hand, shading tends to increase hare palatability of European white birch (Figs. 6 and 7), but its effect 

on Japanese white birch is negligible (Rousi et al. 1996a). The effect of shading on European white birch is, however, 

strongly dependent on genotype and site fertility. There were no statistical differences in resistance among three families that 

we tested in low fertility soils (Fig. 6). But, if seedlings are grown in high fertility soil, resistance was significantly different 

in two out of three families (ANOVA, Bonferroni t-test, p

P FERTILIZATION E 

A 

L ( 

A 0 - no shade 

T 60 50 = 50% shade ( 

A 

B 50 

I 40 

L 

I 30 

T 

y 20 

10 

I 

N 0 

D 

E 

0 50 0 50 0 50 

X V 5845 x V 5848 V 5845 x V 5781 V 5845 x V 5846 

Figure 7.--Hare resistance of three B. pendula F2 families in two shade treatments. Seedlings were grown in nursery peat 

and obtained optimum fi3rtilization (For details see Rousi et al. 1996a). Cafeteria experiments as in Figure 1. Mean + 

S.E. 

SUMMARY 

We conclude that prevailing defense ttleories, (e.g., Carbon-Nutrient Balance, Growth-Differentiation Balance, etc.) 

may not be fully applicable to birch. The reaction of birch species, origins, andgenotypes to variable environments produces 

variable resurts, which depend on plant genotype, environment:and the particular mechanism of resistance to each pest type. 

Consequently, generalized conclusions about birch resistance should be made with great caution. 


BRYANT, J.R, CHAPIN, F.S. [[I and KLEIN, D.R. 1983. Carbon/nutrient balance of boreal plants in relation to vertebrate 

herbivory. Oikos 40:357-368. 

BRYANT, J.R, TAHVANAINEN, J., SULKINOJA, M., JULKUNEN-THTTO, R., REICHARDT R, and GREEN, T. 1989. 

Biogeographic evidence for the evolution of chemical defense against mammalian browsing by boreal birch and 

willow. The American Naturalist 134: 20-34. 

CHIBA, S. and NAGATA, Y. 1968. Resistance to field mice and mountain hares in Betula-species and hybrids. (1) Comparison 

between species. Qji Institute t\)r Forest Tree Improvement. Technical Note 68:20-24 (In Japanese). 

CttIBA, S. and NAGATA, Y. 1969. Resistance to field mice and mountain hares in Betula-species and hybrids. (2) Comparison 

between parental species and hybrids. Oji Institute for Forest Tree Improvement. Technical Note 88: 153- 

155 (In Japanese). 

CHIBA, S., NAGATA, Y., and TOMAKL T 11982. Variation of vole resistance in Larix leptolepis and comparison with the 

hybrid L gmelini and L leptolepis. Oji Institute for Forest Tree Improvement. Technical Note 209:6-9 (In Japanese). 

44

DUGLE, J.A. 1966. A taxonomic study of Western Canadian species in the genus Betula. Canadian Journal of Botany 44: 

929-1007. 

GILL, R.M.A. 1992a. A review of damage by mammals in north temperate forests: 1. Deer. Forestry 65(2): 145-169. 

GILL, R.M.A. 1992b. A review of damage by mammals in north temperate forests: 2. Small mammals. Forestry 65(3): 

281-308. 

HERMS, D.A. and MATTSON, W.J. 1992. The dilemma of plants: to grow or defend. The Quarterly Review of Biology 

67(3): 283-335. 

JOHNSSON, H. 1966. Avkommepr6vning av bj6rk. F6reningen Skokstrfidsf6r_idling, Arsbok 1966: 90-135. 

LEPPIK, E.E. 1970. Gene centers of plants as a source of disease resistance. Annual Review of Phytopathology 8: 323-344. 

MATTSON, W.J. 1980. Herbivory in relation to plant nitrogen content. Annual Review of Ecology and Systematics 11: 

119-161. 

PAAKKONEN, E., PAASiSALO, S., ttOLOPAINEN, T., and K._RENLAMPI, L. Growth and stomatal responses of birch 

(Betula pendula Roth.) clones to ozone in open-air and chamber fumigations. New Phytologist 125: 615-623. 

POTERI, M. and ROUSI, M. 1996. Variation in Melampsoridium resistance among European white birch clones grown in 

different fertilizer treatments. Eurp. J. For. Pathol.: In Press. 

RHOADES, D.F. 1979. Evolution of plant chemical defense against herbivores, p. 3-54. In Rosenthal, G.A. and Janzen 

D.H., eds. Herbivores: their interaction with secondary plant metabolites. Academic Press, Orlando. 

ROUSI, M. 11988. Resistance breeding against voles in birch: possibilities for increasing resistance by provenance transfers. 

EPPO (European Plant Protection Organization) Bulletin 18: 257-263. 

ROUSI, M. 1990. Breeding forest trees for resistance to mammalian herbivores - a study based on European white birch. 

Acta Forestalia Fennica 210. 

ROUSI, M., TAHVANAINEN, J., and UOTILA, I. 1989. Inter- and intraspecific variation in the resistance of winterdormant 

birch (Betula spp.) against browsing by the mountain hare. Holarctic Ecology 12:187-192. 

ROUSI, M., HENTTONEN, H., and KAIKUSALO, A. 1990. Resistance of birch (Betula pendula and B. platyphylla) 

seedlots to vole. Scandinavian Journal of Forest Research 5: 427-436. 

ROUSI, M., TAHVANAINEN, J., and UOTILA, I. 1991. A mechanism of resistance to hare browsing in winter-dormant 

European white birch (Betula pendula). The American Naturalist 137: 64-82. 

ROUSI, M., TAHVANAINEN, J., and UOTILA, I. 1993. Effects of shading and fertilization on resistance of winterdormant 

birch (Betula pendula) to voles and hares. Ecology 74(1): 30-38. 

ROUSI, M., MATTSON, W.J., TAHVANAINEN, J., HENTTONEN, H., KOIKE, T., and UOTILA, I. 1996a. Growth and 

hare resistance of birches: testing defense theories. Oikos: In Press. 

ROUSI, M., TAHVANAINEN, J., HERMS, D.H, HENTTONEN, H., and UOTILA, I. 1996b. Clonal variation in susceptibility 

of white birches to herbivores. For. Sci.: Submitted. 

VAVILOV, N.I. 1922. The law of homologous series in variation. Journal of Genetics 12(1): 47-89. 

45

A REDOX-BASED MECHANISM BY WHICH 

ENVIRONMENTAL STRESSES ELICIT CHANGES IN 

PLANT DEFENSIVE CHEMISTRY 

DALE M. NORRIS 

642 Russell Laboratories, University of Wisconsin, Madison, WI 53706, USA 

INTRODUCTION 

It is now clear that most, if not all, lifeforms employ chemicals in communicating with their biotic and abiotic 

environments. Higher plants and animals, including insects and humans, may even use many of the same compounds 

(Rodriquez and Levin 1976, Norris and Liu 1992, Raina et aL 1992). Quinol-quinone redox couples, e.g., p-hydroquinone/ 

p-benzoquinone (Fig. 1), are prime examples of such common messengers. Hundreds of ions, free radicals (Packer and 

Glazer 1990) and molecules have been shown to be messengers. Alkaloids, flavonoids and terpenoids are families of 

compounds which are especially important as messengers (Norris 1986). 

0 O- OH 

II 

II II 

0 0 O- OH 

"A" "B" 

Figure l.--The classical redox couple in physical chemistry, p-hydroquinone / p-benzoquinone. 

The development of knowledge about the chemical senses has frequently been quite disconnected among many 

scientific disciplines. As a result, limited progress has been made in finding unifying principles. However, for several 

decades, I have developed hypotheses pertaining to such unifying principles and tested them experimentally (Norris 1994). 

The results of our and others' efforts have allowed me to propose that a common code for chemical communication exists 

among lifeforms, and them and their abiotic environments. The proposed name for this code is the Environmental Energy 

Exchange (EEE) Code (Norris 1981, 1986, 1988, 1994; Neupane and Norris 1992; Norris and Liu 1992). 

In this paper, the essential characteristics of this EEE code, including its underlying sulfhydryl (thiol, -SH) / disulfid_ 

(-S-S-) -dependent oxidative-reductive, energy-transducing mechanism, are applied to the interpretation of how environmen- 

tal stresses elicit changes in plant defensive chemistry. 

Stress Alteration Of Phytoehemieal Messages 

The chemistry by which plants "communicate" with their herbivores has been classified as either constitutive 

(i.e., pre-stress) or inducible (i.e., post-stress). Plant pathologists and plant physiologists created this classification to 

Mattson, W.J., Niemel_i, P., and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USD_-X 

For. Serv. Gen. Tech. Rep. NC- 183, N.C. For. Exp. Sta., St. Paul, MN 55108. 

46 

!

accommodate findings related to microbial interactions with plants (Bell 1981). Constitutive defensive phytochemistry 

is that expressed in the plant by its genome in the supposed absence of specific stress from the environment; whereas, 

inducible defensive phytochemistry is that expressed in response to a specific stress caused by some factor (e.g., microbial 

infection) in the plant's environment. Although these interpretations were developed for plant-microbe interactions, 

subsequent studies, especially by entomologists and chemical ecologists, have shown that much of the involved phytochemistry, 

and some of the developed concepts, may prove useful to understanding messenger phytochemicals in 

herbivore-plant communications (Kogan and Paxton 1983). Our studies of such phytochemistry indicate that the major 

differences between pre- and post-stress chemistry are largely quantitative (e.g., Neupane and Norris 1991a, Markovic et 

al. 1993). Thus, a plant usually responds to environmental stresses by making more, or less, of individual compounds in 

a given (genome-determined) array of secondary metabolites. These chemicals have sometimes been termed 'phytoalexins' 

by plant pathologists and physiologists. 

Unifying Mechanisms Of Elieitation 

A wide diversity of biotic and abiotic entities have proven effective as exo-elicitors (external to the plant cell) in the 

inducible alteration of plant defensive chemistry against environmental stresses (Sequeira 1983). However, plant pathologists 

and physiologists have especially focused on fragments of microbial-pathogen or plant-cell walls as exo-elicitors. _3-glucans 

from ptant-cel] walls apparently play major elicitory roles (Ryan 1983, Darvill and Albersheim 1984, Templeton and Lamb 

1988, Dixon and Lamb 1990). Heavy metals, including mercury, and sulthydryl reagents are other proven exo-elicitors 

(Stossel 1984). Although the diversity of exo-elicitors might imply that a common mechanism of exo-elicitation is unlikely, 

the suggested involvements of sulfhydryls in the process led Sequeira (1983) to hypothesize their importance in such inductions. 

Other reports of elicitation agents reacting with sulfhydryls in phytoalexin biosynthesis (Stossel 1984) support the 

underlying importance of sulfhydryls in elicitation processes. More recent studies (Neupane and Norris (1990, 1991a) and 

Liu et al. (1992, 1993)) showed that several classical compounds for detecting (i.e., reacting with) sulfhydryls also alter the 

biosynthesis of isoflavanoid phytoalexins in soybeans, Glycine max (L.) Merrill. These induced phytoalexins also alter the 

defense of these plants to some insect herbivores (Sharma and Norris 1991). 

It now seems clear that the elicitory mechanism in plants involves directly, or indirectly, the oxidative-reductive 

(redox) chemistry of sulfhydryl (-SH) / disulfide (-S-S-) interactions (Fig. 2). First, reagents which react specifically with 

sulfhydryls are effective elicitors of stress-alterable defensive phytochemistry (Stossel 1984; Sequeira 1983; Neupane and 

Norris 1990, 1991a; Liu etal. 1992, 1993, 1994; Haanstad and Norris 1992; Norris and Liu 1992; Norris 1994). Second, 

sulfhydryl proteins in the plasma membrane which surrounds each plant cell have been shown to react with elicitory sulfhydryl 

reagents (e.g., p-chloromercuriphenyl sulfonic acid, PMBS) when applied in vivo to the intact plant (Liu et al. 1992, 

1993, 1994). Such reaction, or interaction, between a sulfhydryl reagent and a sulfhydryl protein (i.e., receptor) at the plasma 

membrane level may trigger significant alterations in the defensive chemistry of elicited plant cells (Liu et al. 1992, 1994; 

Norris 1994). Third, the essential mechanism by which sulfhydryl reagents and other proven elicitory entities, both biotic 

and abiotic, function in altering defensive phytochemistry is apparently oxidative-reductive (Neupane and Norris 1991b, 

1992; Markovic et al. 1993; Norris and Liu 1992). The classical antioxidants, L-ascorbic acid (vitamin C) and o_-tocopherol 

(vitamin E), have proven to be effective elicitors when applied to whole plants (Neupane and Norris 1991b, 1992). Fourth, 

i Neupane and Norris (1992) showed that L-ascorbic acid and herbivory by Trichoplusia ni (Hubner) elicited analogous 

defensive responses in soybeans. Thus, the involvement of a common mechanism of elicitation, i.e., oxidative-reductive, is 

supported. Our current interpretation is that herbivory elicits inducible defensive phytochemistry through triggering oxidation 

/ reduction-dependent reactions which are common with those caused in plant tissues by a wide variety of other environ- 

! mental stresses. 

-$H /-5-5- 

Figure 2.--The sulfur-based redox couple, sulfhydryl (thiol, -SH) ! disulfide (-S-S-), in which the addition of reducing 

agent favors the formation of thiols; whereas, the addition of oxidizing agent favors the reverse, i.e., formation of 

disulfides. This redox couple in proteins is considered the functional interface by which energy is exchanged as 

information between living cells and their environments. 

47

METHODS 

Antioxidant Alteration Of Ash Antixenosis To Malacosoma disstria 

Because sulfhydryl groups may readily participate in redox reactions in living cells (Morton 1965), our studies 

evaluated a natural antioxidant (reducing agent), o_-tocopherol(vitamin E), as an elicitor of defensive phytochemistry. Alphatocopherol 

was tested in a basal trunkband on green ash trees to alter their antixenosis to forest tent caterpillar larvae, 

Malacosoma disstria (Hubner). The trees were 4.6m tall 'Summit' green ash, Fraxinus pennsylvanica var. subintegerrinia 

(Vahl) Fernald, and had a mean trunk diameter of 5.1 cm at 1 m above the soil surface. The basal trunk of each tree received 

o_-tocophero[, a proven elicitor (Neupaneand Norris 1991b), as a band application at the dosages of 25.0 or 50.0 IU / ml in 

mineral oil. An international unit (IU) equals 1 mg of all-rac-cx-tocopheryl acetate (U.S. Pharmacopaea 1980). Sixty ml of 

either concentration of ot-tocopherol in mineral oil were placed on a 120cm 2bandage; the control trees received only 60 ml of 

mineral oil. 

Insect Bioassays 

Leaves for bioassay were removed from three distinct trees per treatment or control at each of several intervals after 

treatment, and two-choice feeding assays were conducted with 1.5-cm-diam disks cut with a No. 8 cork borer from such 

leaves and with third-instar forest tent caterpillars. The insect's feeding option thus was between a comparable leaf disk from 

an elicited versus a solvent-control tree. The quantitation of insect feeding was detailed by Markovic et al. (1993). 

Chemical Analyses 

[.,eaves for chemical analyses were collected from three distinct ash trees for each elicitation dosage and the solvent 

control, and for 8 and 16days after treatment. Three compound leaves from each tree were immediately put individually into 

a glass jar containing 80% methanol, and then stored in darkness at -20°C until chemical analysis. Procedures used to extract 

chemicals from ash leaves; hydrolyze extracted chemicals; and analyze the chemicals by high performance thin layer 

chromatography (HPTLC) and high performance liquid chromatography (HPLC) were as detailed by Markovic et al. (1993). 

RESULTS 

Antioxidant Altered Foliar Antixenosis In Ash 

Forest tent caterpillar herbivory was altered on leaf disks from green ash trees elicited with either of the two dosages 

of c_-tocopherol as compared to disks from solvent-treated ash trees (Table 1). High performance thin layer chromatography 

(HPTLC) revealed distinct differences in both the non-hydrolyzed and hydrolyzed portions of the ethyl acetate extractables 

from elicited versus control ash trees (Markovic et al. 1993). Reduced insect preference for foliage due to tocopherol 

elicitation was accompanied by an increased mean total HPLC-resolved peak area of ethyl acetate extractables from the nonhydrolyzed 

leaf sample as compared to control foliage. Whereas, increased insect preference for foliage due to elicitation 

was accompanied by a decreased total HPLC-resolved peak area of such extractables as compared to control. HPLC of the 

above non-hydrolyzed, ethyl acetate extractables showed mainly quantitative, rather than qualitative, differences between ¢xtocopherol-elicited 

versus solvent-treated controls. The differences were especially evident in five major, and three lesser, 

peaks; these eight peaks thus were chosen for a more detailed comparison between the non-hydrolyzed fraction of the ethyl 

acetate extractables from the two cx-tocopherol treatments and solvent-treated controls at two times (8 and 16 days) after 

elicitation. 

At 8 days after elicitation, the foliage from trees receiving 25 IU / ml contained significantly more of HPLC peaks 

1-6, but not of peaks 7 or 8, than did leaves from solvent-control trees (Fig. 3). Foliage from these treated trees also was 

significantly less preferred than that from the control trees (Table 1). Conversely, at 16 days after elicitation, foliage from 

trees that received either 25 or 50 IU / ml was preferred over that from the solvent-control ones (Table 1). The ethyl acetate 

extractables at 16 days after elicitation from preferred foliage from trees that received 25 or 50 IU / ml had a smaller average 

total HPLC-resolved peak area than did those from the leaves of solvent-control trees (Fig. 4). The mean total HPLC- 

resolved peak area was also significantly different between the foliage collected from the solvent-control trees at 8 and 16 

days after elicitation. Thus, time in thegrowing season also affected the chemical composition of the trees. 

48

Table l.--Mean ±S.E. area eaten by the forest tent caterpillar in leaf disks from tocopherol-elicited versus 

control-elicited ash trees. 

Behavioral Area eaten (cm2) b 

response Days _ Treatment Control 

Nonpreferred _ 8 0.18 _+0.04 "*c 0.41 + 0.04 

Preferred 1_ 16 0.49 + 0.07* 0.31 + 0.06 

Preferred 2d t6 0.46 + 0.05"" 0.16 + 0.04 

Days after elicitation. 

b From Markovic et al. (1993). 

Vitamin E dose was 25.0 IU / ml. 

d Vitamin E dose was 50.0 tU / ml. 

Means followed by * or **are significantly different from their control at p < 0.05 and p < 0.01, respectively. 

35000 

30000 

25000 

R Non-preferred 

a E! Control B 

b a 

W 

¢ 20000 b 

w 15000 

a b 

10000 b a 

a 

a 

a a 

5000 b a a 

0 

1 2 3 4 5 6 7 8 

PEAK NUMBER 

gure 3.--Comparative mean areas of the indicated peaks in the standardized HPLC analysis of the nonhydrolyzed fraction 

of the ethyl acetate extractables from ash foliage. Data are for nonpreferred foliage (elicited with 25 IU / ml) and 

control B foliage (elicited only with the solvent). Leaves were collected at 8 days after elicitation. Peak areas not 

followed by the same letter are significantly different at p < 0.05. 

49

35000 j 

30000 

25000 b b 

< 

w 20000 

IZ: < 

b 

m 15000 b 

O. 

10000 

b 

m Preferred 1 

a IW Control A 

r'l Preferred 2 

a a a b 

5000 c a 

0 . . . 

1 2 3 4 5 6 7 8 

PEAK NUMBER 

Figure 4.--Comparative mean areas of the indicated peaks in the HPLC analysis of the nonhydrolyzed fraction of ethyl 

acetate extractables from ash foliage. Data are for preferred 1 (elicited with 25 IU / ml), preferred 2 (elicited with 50 

IU / ml and control A (elicited only with solvent) foliage. Leaves were removed from these trees at 16 days after 

elicitation. Peak areas not followed by the same letter are significantly different at p < 0.05. 

DISCUSSION 

Antioxidant Altered Foliar Antixenosis 

The findings support the interpretationthat preference / non-preference decisions by M. disstria larvae regarding __ 

tocopherol-elicited versus solvent (control)-elicited green ash foliage are probably attributable to quantitative changes in 

several compounds, not just in one chemical. Elicitation by _-tocopherol alters the quantitative contributions of at least 

several chemicals in the mixture of secondary metabolites. The effects are very dynamic, e.g., changing from increased to 

decreased antixenosis between 8 and 16 days after elicitation with 25 IU / ml. These findings seem compatible with the 

knowledge that survivor plants systematically switch their major allocation of nutrient and energy resources between primary 

growth and secondary growth-differentiation including defensive chemistry. Such dynamic switching seems reasonable 

because survivor plants must grow, but they must also defend themselves from potentially lethal environmental stresses, 

50 

b 

i

Sulfhydryl-Disulfide Mechanisms In Plant And Animal Perceptions 

Most, if not all, common chemical messengers between insects and plants, and between them and their environments 

(Rodriguez and Levin 1976, Neupane and Norris 1992, Norris and Liu 1992, Raina et al. 1992) are both elicited and perceived 

by sulfhydryl / disulfide (-SH / -S-S-) -dependent mechanisms (Neupane and Norris 1992, Liu et al. 1992, Norris 

1994). The energy-transduction mechanism in chemical communications by both animals and plants is sulfur dependent. 

The initial studies in this area showed that 1,4-naphthoquinones serve as repellents and deterrents to Periplaneta americana 

L., the American cockroach, and Scolytus multistriatus (Marsh.), the smaller European elm bark beetle, by reacting with 

sulfhydryls in receptor proteins in dendritic membranes of chemosensitive neurons (Norris et al. 1970a, 1970b, 1971; 

Rozental and Norris 1973, 1975; Singer et al. 1975). 

The redox dependency of chemoreception in insects was also shown in studies (Norris 1969, 1970) where p-hydroquinone, 

the reduced partner in the classical redox couple, p-hydroquinone / p-benzoquinone, excited feeding by S. 

multistriatus; whereas, the oxidized partner, p-benzoquinone, inhibited feeding. In spite of these early findings, enthusiasm 

for redox-based receptor and energy-transduction mechanisms in animal and human chemoreception has, until quite recently, 

been limited. This was especially true for insect chemoreception. Although numerous research reports have confirmed that 

insect chemoreception is sulfhydryl-disulfide dependent (Villet 1974; Frazier and Heitz 1975; Singer et al. 1975; Ma 1977, 

1981; Vande Berg 1981), several workers in this entomological specialty (e.g., Kaissling 1971, 1974, 1987; Vogt and 

Riddifbrd 1986) concluded that non-covalent bonding, but not redox reactions, are involved in the transduction mechanism in 

insect chernoreception. Until recently, most researchers working in receptor and energy-transduction mechanisms in living 

cells also had concluded that redox reactions were not involved. In contrast, findings by Norris (1969, 1971, 1976, 1981, 

1985, 1988, 1994) have consistently indicated that redox chemistry is fundamentally involved in receptor function and 

energy-transduction mechanisms. The validity of our results has been reconfirmed throughout the general field of receptors 

and signal transduction (e.g., Storz et al. 1990, Van Der Vliet and Bast 1992, Stamler et al. 1992, Reichard 1993, Pyle 1993, 

Ravichhandran et al. 1993). The sulfhydryl-disulfide dependent:redox chemistry of receptors and energy transduction has 

become a major area of research throughout biology (Van Der Vliet and Bast 1992). Thus, these aspects of insect chemoreception 

have not only been confirmed, but also are now widely viewed as a part of a much larger chemoreception 'whole' 

within biology. 

The Required Elemental Trait 

The fundamental trait required for the evolution of a system for information generation, transfer and use in maintaining 

order is the transfer of chemical groups and energy. This required trait for chemical messengers singles out phosphorus 

and sulfur from all other atoms in the "Periodic System" (Wald 1969). The two basic atomic characteristics which qualify 

sulfur and phosphorus uniquely as agents of chemical group and energy transfers are (a) the possession of d, in addition to s 

and p, orbitals which increases their capacities to form linkages with a variety of energy potentials and (b) an intrinsic 

instability of such linkages, which facilitates the exchange of chemical groups and energy (i.e., informational units) (Wald 

1969). Such roles for phosphorus are well established experimentally, but they are only just now being recognized for sulfur. 

However, sulfur now seems to be the most highly qualified element regarding abilities to exchange energy as information 

among living forms and their abiotic environments (Wald 1969; Norris et al. 1970a, 1970b, 1971; Norris 1971, 1979, 

1981, 1985, 1986, 1988, 1994; Neupane and Norris 1992; Norris and Liu 1992; Van Der Vliet and Bast 1992). First of all, 

regarding such roles, sulfur, as thiol (-SH), brings to living systems a form of organic sulfur which possesses those atomic 

traits which are otherwise limited to organic oxygen and are essential to organisms. However, thiol lacks most, if not all, of 

those traits (or degrees thereof) possessed by oxygen which make oxygen or its free-radical derivatives, especially harmful, 

even deadly to living organisms, unless the oxygen is handled in special ways, e.g., bound to hemoglobin in vertebrate blood 

(Wald 1969, Harold 1986). Second, sulfur as well as phosphorus forms relatively long (i.e., loose) bonds with other atoms; 

such bonds hold the attached atom less tightly and thus it may be more readily transferred, exchanged, as informational 

energy in living systems (Wald 1969, Harold t986). Third, the trait which ultimately sets sulfur, as a vehicle of chemical 

group and energy transfers as information, apart from phosphorus is the former's ability to form a thiol (-SH). This property 

uniquely enables sulfur to readily accept, donate or variously share the basic energy unit in living systems, hydrogen (Wald 

1969, Szent-Gyorgyi 1973). Thus, through sulfhydryl / disulfide (thiol, -SH / disulfide, -S-S-) redox chemistry, sulfur 

uniquely brings to living cells, and their chemical communications, a "yin and yang", a give and take, mechanism for moving 

hydrogens, the basic energy unit of living systems, as information between plants and animals and their abiotic environments 

(Norris 1986, 1988, 1994; Neupane and Norris 1992; Norris and Liu 1992). 

51

The conversion of abiotic informational energy states, as in pheromone and phytochemical messengers, into biotic 

informational energy states occurs in sulfhydryl / disulfide-proteinaceous receptors in the plasma membrane which encloses 

each living cell (Norris 1981, 1986, 1988). Chemical messengers, by directly or indirectly oxidizing sulfhydryls or reducing 

disulfides in proteins of living cells, allow the folding and unfolding, respectively, of the three-dimensional structure of such 

macromolecules in cells (Figs. 5) (Sela et al. 1957, Norris 1981). Such folding and unfolding of proteins (e.g., as in muscles 

during their contractions versus relaxations, or in nerve membranes during impulse generation versus decay) especially bring 

to living systems, motion; a trait which is so commonly associated with life (Harold 1986). 

tt - 

( 

H,,,- } _ 

Figure 5.uA schematic illustrating how the formation of disulfide (-S-S-) bonds in proteins can stabilize the three-dimensional 

structure of such macromolecules in a more aggregated state (see on the right); whereas, the reduction of those 

-S-S- bonds by reducing agents (e.g., some chemical messengers) can yield sulfhydryls (-SHs) and a more "opened", 

non-aggregated, three-dimensional form to the protein in the receptor and energy-transducer macromolecule (see on 

the/eft). In this illustration, each circle in the chains of circles represents an amino acid in the chain of amino acids 

which make up a protein. Each circle containing a sulfur (S) represents the amino acid, cysteine, which bears one 

sul thydryl (-SH). 

As Norris and coworkers (Norris et al. 1970a, 1970b, 1971; Rozental and Norris 1973; Singer et al. 1975) showed 

about 20 years ago, the highly reversible redox system based on sulfhydryl / disulfide interactions with chemical messengers 

yields not only qualitative (i.e., excitatory or inhibitory), but also quantitative (i.e., quantal, unitized) exchanges of energy. 

An informational system requires simply an ability to map an output (i.e., efferent) message or behavior on to an input (i.e .... 

afferent) message or stinmlus (O'Connell 1981). The above described highly reversible sulfhydryl / disulfide redox system 

for transferring energy (hydrogens, or electrons and protons) as information is readily quantifiable in millivolts (mV) 

especially by polarography (Rozental and Norris 1973, 1975) or the electroantennogram (EAG) (Norris and Chu 1974; Norris 

1979, 1986, 1988). Thus, this system fully satisfies the requirements of a code for the informational exchange of energy. 

Theorized Environmental Energy Exchange Code 

An informational code thus must be based on (a) an universal medium (i.e., messenger) and (b) a means (method) of 

using that universal messenger to convey both qualitative and quantitative messages. Such a system allows a special use of 

energy for establishing order (Harold 1986). Norris ( 1971, 1976, 1986, 1988, 1994) has presented theories, based on sulfur 

electrochemistry as the universal messenger, for chemical communications among animals, including humans, and plants; 

and between them and their environments. The most recent theoretical version of such a code is here termed the 'Environ- 

mental Energy Exchange Code'. In the proposed EEE code, sulfur electrochemistry is the universal medium (messenger); 

and the thiot (-SH), its derivatives and its redox couple, disulfide (-S-S-), are the chemical means for using sulfur electrochemistry 

to convey both qualitative and quantitative messages. Thus, through chemical-messenger oxidation of a sulfhydryl 

(thiol, -SH) in each of two molecules of the amino acid cysteine in the receptor protein (energy-transducer), a disulfide (-S-S- 

) may form (Fig. 2 ). Such a disulfide in the three-dimensional protein involves a conformational (shape) change; such 

energy transfer thus allows the protein to become more folded (aggregated) in shape (Fig. 2 ). Conversely, through chemicalmessenger 

reduction of a disulfide, the -S-S- bridge between the two cysteines is broken and the protein may then assume a 

more open (i.e., unfolded) three-dimensional shape (Fig. 2 ). This chemical messenger-driven, highly reversible sulfhydryl / 

disulfide electrochemistry in the receptor protein (energy-transducer) thus brings to organisms quantitative dynamic form, 

which is so essential to function. 

52

Iraa prior experimental analysis of such information exchange between the feeding-inhibitory messenger menadione 

(2-methyt-t ,4-naphthoquinone) from the environment, and sulthydryls in the dendritic membrane of chemosensitive neurons 

in the antenna of Periplaneta americana, the electrochemical transduction of informational energy into organismal response 

(i.e., activity) was described mathematically as a regression with r= 0.997 (Norris 1986, 1988). Thus, one unit of experimentally 

determined input (stimulus) energy from the environment can be mapped (correlated) informationally upon (related to) 

one unit of output (organismal-response) energy using the proposed sulfur-based encoding mechanism. Some key parameters 

of this experimentally elucidated sulflaydryl / disulfide-dependent electrochemical mechanism are further discussed in Table 

2. 

Table 2.--Some key fascets of the sulfhydryl / disulfide-dependent electrochemical mechanism involved in insect perception 

of phytochemical (e.g. 1,4-naphthoquinone) messengers. 

1. Each of the three possible pairings of the chemoreception parameters (a, b, e) yields a linear regression with r > 0.95: (a) 

the moles of messenger required in a standardized insect behavioral assay to cause a > 99% change in that behavior; (b) the 

maximum in vitro polarographic U_/2shift in millivolts by the involved receptor and energy-transducer sulflaydryl / disulfide 

protein from the insect's antennal chemosensory neurons when saturated with the above messenger: and (e) the maximum 

percent inhibition of a standardized excitant-stimulated electroantannogram (EAG) by the above messenger (Rozental and 

Norris 1973, 1975; Norris and Chu 1974; Norris 1979, 1986, 1988). 

2. Simultaneous solution of the three linear relationships among the parameters in (1) showed that the correlation (rZ= 0.95) 

between (a) the maximum U_/2shift elicited in the sulfhydryl / disulfide receptor and energy-transducer protein by the 

messenger and (b) the maximum percent inhibition of the standardized EAG by that messenger is so high that only one of 

these two parameters need be considered in a mathematical description of the transduction of the molar-messenger energy 

into insect behavioral change. Log Y - 3.40 - 0.112 Log X, quantifies the energy-transduction relationship (Norris 1986, 

11988). 

3. Based on behavioral analyses (Rozental and Norris 1975; Norris 1986, 1988), three distinct sets of sulfhydryl (thiol)dependent 

receptor sites for 2-methyl-1,4-naphthoquinone (menadione) messenger exist in the receptor and energy-transducing 

protein in Periplaneta americana. Those sulfhydryls in each of these three sets of receptor sites cause a 4-5 millivolt 

shift in the receptor and energy-transducing protein's U_/2value when they react at the mercury electrode involved in this 

polarographic analysis (Rozental and Norris 1973; Norris 1979, 1986, 1988). 

4. Based on the EAG-inhibition assay, messenger-menadione saturation of one set of receptor sites, as described in (3), 

causes about an 8% inhibition (Norris and Chu 1974; Norris 1986, 1988). Thus, the theoretical maximal inhibition of the 

standardized EAG by saturation of the receptors in all three sites with menadione might be predicted as 3 (sites) times 8%, 

which equals 24%. It is interesting and significant that the experimentally determined range in maximal percent of EAG 

inhibition by saturation with menadione was 23-25%. 

5. Data summarized in (3) and (4) above lead to the interpretation that a receptor and energy-transducer Uz/2shift of 4-5 mV 

equals an 8% inhibition in the EAG. This means that each 4-5 mV shift in the Uu2 is accompanied by an 8% inhibition in the 

standardized EAG. The observed linear relationship between the U_a millivolt shift in the receptor and energy-transducer 

protein and the percent EAG inhibition shows that the primary encoding of the message dictating the whole-insect behavior 

occurs in the energy-transducer protein in the chemosensory sensillum (Norris 1979, 1986, 1988). 

6. Our research explains for the first time, both in electrochemical and electrophysiological parameters, why the EAG is so 

meaningful to an understanding of the chemical senses of insects. EAG does not just measure millivolts of electrical energy, 

but also the energy after it has already been coded as information in the receptor and energy-transducer protein adequately to 

elicit a predictable behavior in (by) the insect (Norris 1979, 1981, 1986, 1988). 

7. Saturation of the receptor and energy-transducer protein with p-chloromercuribenzoate (PCMB), a compound which reacts 

. specifically and irreversibly with sulfhydryls, blocks the above characterized messenger-induced Uu2 shift in the protein. 

Thus, the conversion of molar-messenger energy into electrochemically based information adequate to predict whole-insect 

behavior is blocked by the sulfhydryl-specific reagent, PCMB (Rozental and Norris 1973; Norris 1979, 1981, t988). 

8. Sulfur, as in sulfhydryl / disulfide redox systems in proteins, is the critical dynamic elemental interface between responsive 

cells and the stimulating environment, whether biotic or abiotic. 

53

SUMMARY 

Animals and plants "communicate" with each other, and with their environments, via common ion, free radical, and 

molecular messengers which function by electrochemical energy-transduction mechanisms. Such mechanisms depend upon 

the reversible sulflwdryl-disulfide redox couple in receptor and energy-transducer proteins to convert quantitatively messeaget-based 

energy states from the environment into altered membrane potentials and/or second messengers which may serve 

as signals in the elicited cell, and between it and other cells in an organism. Receptor and energy-transducer proteins are 

associated with the plasma membrane which surrounds each living cell. Within biological constraints, the conversion of a 

molar-messenger energy state into a redox-based energy state in the sulfhydryl-disulfide receptor and energy-transducing 

protein in the plasma membrane is linear (i.e., quantal). This means that messenger-borne energy from the environment is 

converted to informational energy (i.e., units) by the perceiving cell. 

Many entomologists and chemical ecologists have used the electroantennogram (i.e., EAG) to detect compounds 

from the environment which alter the behavior of insects. In using this classical technique, the experimentalist is measuring 

change in the energy (e.g., dendritic-membrane potential) state in the primary peripheral chemosensory neurons which are 

specially "housed and exposed to the external environment" within the antenna of the insect. The experimental use of the 

EAG and the correct prediction, thereby, of the resultant behavioral change elicited in the whole, live insect constitutes 

scientific proof that the information necessary for alteration of the behavior of the whole insect can be encoded in the primary 

peripheral chemosensitive neuron. This encoding of energy into information is dependent upon the element 'sulfur', and 

especially its readily reversible sulfhydryl (i.e., thiol, -SH) / disulfide (i.e., -S-S-) redox couple. This encodement of chemical-messenger 

energy into biologically useful information is blocked (or otherwise altered) in the intact cell or whole 

organism by the application of biological concentrations of reagents which react specifically with sulfhydryls and/or disulfides 

in proteins in plasma inembrane. Recent research has proven that this transduction of environmental energy into 

biologically useful inflmnation also occurs in plant cells. Thus, the sulfhydryl / disulfide-dependent Environmental Energy 

Exchange Code is supported by extensive scientific findings from both animal and plant reahns. 

The unique atomic attributes of the element 'sulfur' for fulfilling this vital role in the conversion of environmental 

energy into biologically useful infornaation figrall cells were clearly described by Wald (1969). It is fortunate that scientists 

can now readily test the role of sulfur, and especially the sulthydryl / disulfide redox couple, in the exchange of environmentally 

based energy into information in any living cell. Chemical ecologists seem especially fortunate in this regard through 

their frequent familiarity with the EAG and other electrophysiological techniques for experimentation. We have also shown 

the usefulness of classical electrochemical (e.g., dropping-mercury-electrode polarography) techniques for asking questions 

about the roles of sultur and its derivatives in the proposed Environmental Energy Exchange Code. Further experiments on 

this exciting energy-exchange interface between living cells, organisms, and their vital environments should yield data which 

significantly improve our abilities to quantify environmental influences on the expressions of phenotypes by genomes, and on 

the functionalities and longevities of such phenotypes. 


Research reported here from the author's laboratories was supported partially by the College of Agricultural and Life 

Sciences, University of Wisconsin, Madison, Wisconsin, USA; in part by numerous research grants from the U.S. National 

Science Foundation and the U.S. National Institutes of Health; and in part by U.S. Hatch Projects 3040 and 3419, McIntire- 

Stennis Project 3127, and USDA Competitive Grants 84-CRCR-1-1501 and 88-37153-4043. 


BELL, E.A. 1981. The physiological role(s) of secondary (natural)products, p. 1-19. In Conn, E.E., ed. The Biochemistry 

of Plants: A Comprehensive Treatise. vol 7. Secondary Plant Products. Academic Press, New York. 

DARVILL, A.G. and ALBERSHEIM, R 1984. Phytoalexins and their elicitors - a defense against microbial infection in 

plants. Annu. Rev. Plant Physiol. 35: 243-275. 

DIXON, R.A. and LAMB, C.J. 1990. Molecular communication in interactions between plants and microbial pathogens. 

54 Annu. Rev. Plant Physiol. Plant Molec. Biol. 41" 339-367.

FRAZIER, J.L. and HEITZ, J.R. 1975. Inhibition ofolfaction in the moth Heliothis virescens by the sulfhydryl reagent 

ftuorescein mercuric acetate. Chem. Senses Flavor 11:271-281. 

HAANSTAD, J.O. and NORRIS, D.M. 1992. Altered elm resistance to smaller European elm bark beetle (Coleoptera: 

Scolytidae) and forest tent caterpillar (Lepidoptera: Lasiocampidae). J. Econ. Entomol. 85:172-181. 

HAROLD, F.M. 1986. The vital force: A study of bioenergetics. W.H. Freeman and Co., San Francisco, 577 p. 

KAISSLING, K.E. t971. Insect olfaction, p. 351-431. In Beidler, L.M., ed. Handbook of Sensory Physiology, vol. 5. 

Springer-Verlag, Berlin. 

KAISSMNG, K.E. 1974. Sensory transduction in insect olfactory receptors, p. 243-273. In Jaenicke, L., ed. Biochemistry 

of Sensory Functions. Springer-Verlag, Berlin. 

KAISSLING, K.E. 1987. R.W. Wright Lectures on Insect Olfaction, p. 1-75. In Colbert, K., ed. Simon Fraser University, 

Burnaby, B.C., Canada. 

KOGAN, M. and PAXTON, J. 1983. Natural inducers of plant resistance to insects, p. 153-17 I. In Hedin, RA., ed. Plant 

Resistance to Insects. American Chemical Society, Washington, D.C. 

LIU, S.H., NORRIS, D.M., HARTWIG, E.E., and XU, M. 1992. Inducible phytoalexins in juvenile soybean genotypes 

predict soybean looper resistance in the fully developed plants. Plant Physiol. 100: 1479-1485. 

LIU, S.H., NORRIS, D.M., and LI, J. 1994. Peroxidase activity as correlated with stress-inducible insect resistance in 

Glycine max. Trends in Agricultural Sciences-Entomology, in press. 

LIU, S.H., NORRIS, D.M., and XU, M. 1993. Insect resistance and glyceollin concentration in seedling soybeans support 

resistance ratings of fully developed plants. J. Econ. Entomol. 86: 401-406. 

MA, W.C. 1977. Alterations of chemoreceptor function in armyworm larvae (Spodoptera exempta) by a plant-derived 

sesqui-terpenoid and by sulthydryl reagents. Physiol. Entomol. 2: 199-297. 

MA, W.C. 1981. Receptor membrane function in olfaction and gustation: Implications from modification by reagents and 

drugs, p. 267-287. In Norris, D.M., ed. Perception of Behavioral Chemicals. Elsevier/North-Holland Biomedical 

Press, Amsterdam. 

MARKOVIC, I., HAANSTAD, J.O., and NORRIS, D.M. 1993. Chemical correlates of alpha-tocopherol (vitamin E) altered 

Malacosoma disstria herbivory in Fraxinus pennsylvanica var. subintegerrinia, green ash. J. Chem. Ecol. 19:1205- 

1217. 

MORTON, R.A. 1965. Biochemistry of Quinones. Academic Press, London. 

NEUPANE, F.R and NORRIS, D.M. 1990. Iodoacetic acid alteration of soybean resistance to the cabbage looper (Lepidoptera: 

Noctuidae). Environ. Entomol. 19: 215-221. 

NEUPANE, F.R and NORRIS, D.M. 1991a. Sulthydryl reagent alteration of soybean resistance to the cabbage looper, 

Trichoplusia ni. Ent. Exp. Appl. 60: 239-245. 

NEUPANE, ER and NORRIS, D.M. 1991b. Alpha-tocopherol alteration of soybean anti-herbivory to Trichoplusia ni larvae. 

J. Chem. Ecol. 17: 1941-1951. 

NEUPANE, ER and NORRIS, D.M. 1992. Antioxidant alteration of Glycine max (Fabaceae) defensive chemistry: analogy 

to herbivory elicitation. Chemoecology 3: 25-32. 

NORRIS, D.M. 1969. Energy transduction mechanism in olfaction and gustation. Nature 222: 1263-1264. 

55

NORRIS, DM. 1970. Quinol stimulation and quinone deterrency of gustation by Scolytus multistriatus. Ann. Entomol. 

Soc. Amer.63: 476-478. 

NORRIS, D.M. 1971. A hypothesized unifying mechanism in neural function. Experientia 27:531-532. 

NORRIS, D.M. 1976. How certain insects take the bitter with the sweet. Bull. Entomol. Soc. Amer. 22: 27-30. 

NORRIS, D.M. 1979. Chemoreception proteins, p. 59-77. In Narahashi, T., ed. Neurotoxicology of Insecticides and Pheromones. 

Plenum Publish., New York. 

NORRIS, D.M. 1981. Possible unifying principles in energy transduction in the chemical senses, p. 289-306. In Norris, 

DM., ed. Perception of Behavioral Chemicals. Elsevier/North-Holland Biomedical Press, Amsterdam. 

NORRIS, DM. 1985. Electrochemical parameters of energy transduction between repellent naphthoquinones and lipoprotein 

receptors in insect neurons. Bioelectrochem. Bioenerget. 14: 449-456. 

NORRIS, DM. 1986. Anti-feeding compounds, p. 97-146. In Haug, G. and Hoffmann, H., eds. Chemistry of Plant Protection, 

vol. 1. Springer-Verlag, Heidelberg. 

NORRIS, DM. 1988. Periplaneta americana perception of phytochemical naphthoquinones as allelochemicals. J. Chem. 

Ecol. 14: 1807-1819. 

NORRIS, D.M. 1994. Phytochemicals as messengers altering behavior. In Ananthakrishnan, T.N., ed. Functional Dynamics 

of Phytophagous Insects. Oxford and IBH Publish. Co. Pvt. Ltd., New Delhi, India, in press. 

NORRIS,D.M., BAKER, J.E., BORG, T.K., FERKOVICH, S.M., and ROZENTAL, J.M. 1970a. An energy-transduction 

mechanism in chemoreception by the bark beetle, Scolytus multistriatus. Contrib. Boyce Thomp. Inst. 24: 263-274. 

NORRIS, D.M. and CHU, H.M. 1974. Chemosensory mechanism in Periplaneta americana: Electroantennogram comparisons 

of certain quinone feeding inhibitors. J. Insect Physiol. 20: 1687-1696. 

NORRIS, D.M., FERKOVICH, S.M., BAKER, J.E., ROZENTAL, J.M., and BORG, T.K. 1971. Energy transduction in 

quinone inhibition of insect feeding. J. Insect Physiol. 17: 85-97. 

NORRIS, D.M., FERKOVICH, S.M., ROZENTAL, J.M., BAKER, J.E., and T.K. BORG. 1970b. Energy transduction: 

inhibition of cockroach feeding by naphthoquinones. Science 170: 754-755. 

NORRIS, DM. and LIU, S.H. 1992. A common chemical mechanism for insect-plant communication, p. 186-187. In 

Menken, S.B.J., Visser, J.H., and Harrewijn, P., eds. Proceedings of 8th International Symposium on Insect-Plant 

Relationships. KluwerAcademic Publisher, Dordrecht, The Netherlands. 

O'CONNELL, R.J. 1981. The encoding of behaviorally important odorants by insect chemosensory receptor neurons, p. 

133-163. In Norris, D.M., ed. Perception of Behavioral Chemicals. Elsevier/North-Holland Biomedical Press, 

ATnsterdam. 

PACKER, L. and GLAZER, A.N., eds. 1990. Oxygen radicals in biological systems, part B, Oxygen radicals and antioxidants, 

p. 1-855. In Abelson, J.N. and Simon, M.I., eds.-in- chief. Methods in Enzymology. vol. 186. Academic Press 


PYLE, A.M. 1993. Ribozymes: a distinct class of metalloenzymes. Science 261:709-714. 

RAINA, A.K., KINGAN, T.G., and MATTOO, A.K. 1992. Chemical signals from host plant and sexual behavior in a naoth. 

Science 255: 592-594. 

RAVICHANDRAN, K.G., BODDUPALLI, S.S., HASEMANN, C.A., PETERSON, J.A., and DEISENHOFER, J. 1993. 

56 Crystal structure of hemoprotein domain of P450BM-3, a prototype for microsomal P450's. Science 261" 731-7 36.

REICHARD, R 1993. From RNA to DNA, why so many ribonucleotide reductases? Science 260: 1773-1777. 

RODRIGUEZ, E. and LEVIN, D.A. 1976. Biochemical parallelisms of repellents and attractants in higher plants and 

arthropods, p. 214-270. In Wallace, J.W. and Mansell, R.L., eds. Biochemical Interaction Between Plants and Insects. 

Rec. Adv. Phytochem. I0: 214-270. 

ROZENTAL, J.M. and NORRIS, D.M. 1973. Chemosensory mechanism in American cockroach olfaction and gustation. 

Nature 244: 370-371. 

ROZENTAL, J.M. and NORRIS, D.M. 1975. Genetically variable olfactory receptor sensitivity in Periplaneta americana. 

Life Sciences 17:105-110. 

RYAN, C.A. 1983. Insect-induced chemical signals regulating natural plant protection responses, p. 43-60. In Denno, R.F. 

and McClure, M.S., eds. Variable Plants and Herbivores in Natural and Managed Systems. Academic Press, New 

York. 

SELA, M., WHITE, F.H., and ANFINSEN, C.B. 1957. Reductive cleavage of disulfide bridges in ribonuclease. Science 

125: 691-692. 

SEQUEIRA, L. 1983. Mechanisms of induced resistance in plants. Annu. Rev. Microbiol. 37: 51-79. 

SHARMA, H.C., and NORRIS, D.M. 1991. Chemical basis of resistance in soya bean to cabbage looper, Trichoplusia hi. J. 

Sci. Food Agric. 55: 353-364. 

SLINGER,G., ROZENTAL, J.M., and NORRIS, l).M. 1975. Sulfhydryl groups and the quinone receptor in insect olfaction 

and gustation. Nature 256: 222-223. 

STAMLER, J.S., SINGEL, D.J., and LOSCALZO, J. 1992. Biochemistry of nitric oxide and its redox-active forms. Science 

258: 1898-1902. 

STORZ, G., TARTAGLIA, L.A., and AMES, B.N. 1990. Transcriptional regulator of oxidative stress-inducible genes: direct 

activation by oxidation. Science 248: 189-194. 

STOSSEL, A. 1984. Regulation by sulfhydryl groups of glyceollin accumulation in soybean hypocotyls. Planta 160: 314- 

319. 

SZENT-GYORGYI, A. 1973. The development of bioenergetics, p. 1-4. In Avery, J., ed. Membrane Structure and Mechanisms 

of Biological Energy Transduction. Plenum Press, New York. 

TEMPLETON, M.D. and LAMB, C.J. 1988. Elicitors and defense gene activation. Plant Cell Environ. 11: 395-401. 

VANl)E BERG, J.S. 1981. Ultrastructural and cytochemical parameters of chemical perception, p. 103-131. In Norris, 

l).M., ed. Perception of Behavioral Chemicals. Elsevier/North-Holland Biomedical Press, Amsterdam. 

VAN DER VLIET, A. and BAST, A. 1992. Effect of oxidative stress on receptors and signal transmission. Chem. Biol. 

Interact. 85:95-116. 

VILLET, R.H. 1974. Involvement of amino and sulfhydryl groups in olfactory transduction in silk moths. Nature 248: 707- 

708. 

VOGT, R.G. and RIDDIFORD, L.M. 1986. Pheromone reception: a kinetic equilibrium, p. 201-208. In Payne, T.L., Birch, 

M.C., and Kennedy, C.E.J., eds. Mechanisms in Insect Olfaction. Clarendon Press, Oxford, U.K. 

WALl), G. 1969. Life in the second and third periods; or why phosphorus and sulfur for high-energy bonds?, p. 156-t67. In 

Kalckar, H.M., ed. Biological Phosphorylations: Development of Concepts. Prentice-Hall, New York. 

57

FUNGAL ENDOPHYTES: CONTRASTING EFFECTS IN 

TREES AND GRASSES 

STANLEY H. FAETH 

Department of Zoology, Arizona State University, Tempe, Arizona 85287-1501, USA 

INTRODUCTION 

Fungal endophytes, fungi that live asymptomatically and intercellularly within tissues of most plants, have recently 

received increasing attention from ecologists (e.g., Strong 1988). The presence of fungal endophytes in plants, especially 

grasses, can increase resistance to herbivores due to the production of alkaloidal mycotoxins, but also may increase resistance 

to drought and flooding stress, increase plant competitive abilities, and, for endophytes transmitted vertically via seeds, deter 

seed predators or increase seed dispersal (Clay 1990, Knoch et al. 1993). Endophytic fungi are typically considered plant 

mutualists because of these potential modes of increased plant fitness (Clay 1990). However, most research on fungal 

endophytes has involved introduced, agricultural forage grasses, such as tall rescue or perennial ryegrass. Little is known of 

the ecological role of endophytes in either native grasses or woody plants relative to plant-herbivore, plant-plant, or plantseed 

predator interactions. Here, I summarize our research to date on endophytes in Quercus emoryi (Emory oak) and their 

effects on a major herbivore, the leafminer Cameraria sp. nov. (Lepidoptera:Gracillaridae). Leafminer larvae spend 11 

months within leaf tissues and are confined to a single leaf chosen by the ovipositing female. Therefore, endophytes should 

alter leafininer performance more than that of mobile, exophytic insects. I contrast these results with endophytes in Festuca 

arizonica (Arizona fescue). I predict stronger effects of endophytes on herbivores and seed predators of Arizona fescue due 

to differences in mode of fungal transmission and specificity, 

i_: 

METHODS :: 

We have monitored seasonal and spatial patterns of fungal endophyte infections and the leafminer in trees of Emory :i 

oak at Oak Flat study area in central Arizona for the past 4 and 10 years (Faeth 1991), respectively. We have isolated at least 

12 species of endophytes from Emory oak, but four species, QE1 (Asteromella sp.), QE2 (Ascomycete:Diaporthales), QE7 

(Plecophomella sp.), and Y1 (filamentous yeast) make up >95% of all infections. All of these endophytes are transmitted 

horizontally via spores, likely carried in rainsplash. In observational studies, we have correlated the presence of living and 

dead larval leafminers with the frequency of infection. In manipulative experiments, we have either increased (spore 

spraying of leaves or spore injection of individual mines) or decreased (enclosing branches with plastic or application of 

fungicides) to test the role of individual fungal endophytes on leafminer developmental and mortality. 

We have begun to investigate the role of fungal endophytes in Arizona rescue populations and their relationship to 

intensity of grazing andsoil nutrients. Arizona rescue harbors two endophytes, Acremonium starrii and a p-endophyte 

(Phialophora-like). Both endophytes are transmitted vertically from maternal to offspring plant via seed, but the p-endophyte 

also sporulates and can be transmitted horizontally fromadult plant to adult plant. We have conducted preliminary 

experiments testing the role of the Acremonium endophyte in reducing seed predation and increasing seed dispersal by seed 

harvesting ants (Pogonomymrex species) by presenting E+ (infected) and E- (uninfected) rescue seeds to ant colonies and 

following proportions collected and discarded into refuse piles. 


Generally, endophyte infections in oaks increase seasonally, with the peak infection level coinciding with summer 

rains in Arizona in July-August. Overall infection levels vary with all spatial scales- between localities, between and within 

Mattson, W.J., Niemel_i, p., and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. LISDA 

For. Serv. Gen. Tech. Rep. NC-183, N.C. For. Exp. Sta., St. Paul, MN 

58

; 

trees, and between and within individual leaves. Differences in infection levels between trees result from dif{Erences in 

infection by individual species, with QE7 comprising differences early in the season and QE1 later in the growing season. 

Leafmining is associated with increased fungal infection, and leaves with dead larvae have higher infections than those with 

live larvae. However, neither mass nor survival of leafminers was affected by spore-injection of the three common endophytes. 

One endophyte (Y 1) increased developmental thne. Other experiments show that oviposition is not associated with 

leaves that are either' more or less likely to become infected with endophytes. Thus, leafmining activity appears to increase 

infection frequency, probably by altering the surface of the leaf, but the endophytes themselves appear to have only weak 

effects on leafminer development and survival. 

Preliminary observational evidence :indicates that Arizona fescue in areas under intensive grazing by cattle and native 

ungulates have higher frequency of Ac'rernonium starrii but not the p-endophyte. These results support our prediction that 

endophytes that are only transmitted vertically are more likely to interact mutualistically with the plant and provide protection 

against herbivores. Furore experiments will test the dependency of the mutualism on available resources to the grass and 

controlled levels of herbivory. Similarly, experiments with seed-h_vesting ants show that Acremoni_ma-infected seeds are 

less likely to be collected. Of the seeds that are harvested, E+ seeds are more likely than E- seeds to be discarded into reff_se 

piles where germination success is higher than in surrounding areas. 

SUMMARY 

Endophytic fungi are diverse and ubiquitous in almost all woody and non-woody plants examined to date. The role 

of endophytes in plant-herbivore, plant-plant, and plant-seed predator interactions in natural systems is still largely unex- 

p]ored. Testable hypotheses and predictions, however, can be made about the direction and strength of the interactions based 

upon mode of transmission and specificity of the fungi and plant. 


I thank Kyle Hammon and Dennis Wilson for preliminary data and Rindy Anderson, Carmen Febus, Dawn 

Hagerman, Catherine Hudson, Terri Kingrey, Myrna Miller, and Franziska Schulthess for technical assistance. This work is 

supported by NSF Grant BSR 91-07296 to SHE 


CLAY, K. 1990.. Fungal endophytes of grasses. Ann. Rev. Ecol. Syst. 21: 275-297. 

FAETH, S.H. 199t. Effect of oak leaf size on abundance, dispersion, and survival of the leafminer, Cameraria sp. nov. 

(Lepidoptera:Gracillariidae). Envir. Ent. 20: 196-204. 

KNOCH, T.R., I_ETH, S.H., and ARNOTT, D.L. 1993. Endophytic fungi alter foraging and dispersal by desert seed harvesting 

ants. Oecologia 95: 470-475. 

STRONG, D.R. 1988. Endophytic mutualism and plant protection from herbivores. Ecol. 69: 1. 

59

PROPAGATING THE EFFECTS OF HERBIVORE ATTACK IN AN 

OBJECT-ORIENTED MODEL OF A TREE 

JARI PERTTUNEN, HANNU SAARENMAA, RISTO SIEV,_NEN, HANNU SALMINEN, 

ANTTI POUTTU, and JOUNI V,_KEV,_ 

Finnish Forest Research Institute, Unioninkatu 40A, SF-00170, Helsinki, Finland 

INTRODUCTION 

Understanding cause and effect in forest health problems requires an understanding of the structure and functioning 

of trees. Before conclusions about a malfunctioning system can be made, we must first understand how a healthy one works, 

In industrial engineering and medicine, diagnostic reasoning about structure have long traditions (e.g., Bratko et al. 1989, 

Torasso and Console 1989), and well defined structural models exist for man-made machines and the human body. Diagnosis 

of disorders in trees and forests is not that advanced. One of the reasons for this is that we have not had good structural 

models of trees at our disposal. 

By structural models, we mean models that make the parts of the system explicit. Mathematical models describing 

photosynthesis, respiration, water pressure, etc. are not enough alone. They have to be bound into some larger structure, like 

shoots, buds, needles, etc. These parts in turn must be explicitly connected with their physical immediate neighbors. Such a 

model of a tree would form a topology of thousands of interconnected parts. L-systems (Prunsinkiewicz and Hanan 1989) 

and fractal trees (Mandelbrot 1977) are probably the original approaches to this problem. However, those models are made 

for visualization, and do not maintain the status of the parts after they have been graphically drawn. Plant and tree growth 

has been studied in detail for research in plant morphology and landscape architecture (e.g., Reffye et al. 1989). However, 

structural models of tree development and survival have emerged only lately because the necessary software technologies 

have only been in place for a few years. Sequeira et al. (1991) have presented an structural model of plant. Ahonen and 

Saarenmaa (1991) and Saarenmaa (1991) have prototyped a topological tree. Nikinmaa (1992) has a presented the models 

for the functioning of a structurally explicit Scots pine tree. 

Object-oriented (OO) programming is rapidly becoming the leading way of constructing software. OO software 

consists of objects that are more or less encapsulated entities. They combine attributes and behavior under the same encapsulation. 

Objects can consist of other objects and they can communicate with each other with messages. They make it possible 

to represent structures such as biological organisms with a great realism. Once an object-oriented tree model has been built, 

it can, hopefully, be reused for many different purposes. These uses may be growth prediction and physiological research. 

Such models should also be useful for making diagnoses and visualizing the effects of disorders. Herbivore-tree interactions 

are often so complicated that they can only be understood through modelling. An OO tree model should be able to propagate 

the effects of a herbivore attack in one part of the tree into others where the symptoms and secondary damage may be 

observed. 

In this paper we describe briefly an OO model of Scots pine, Pinus sylvestris L., that we have built, and demonstrate 

its uses for simulating disorders caused by two herbivores, 

METHODS 

Object-oriented modelling allows building software that mechanistically resembles the real world. This similarity is 

achieved with three basic concepts: encapsulated objects, classification of objects, and message passing between objects 

(Rumbaugh et al. 1991, Saarenmaa et al. 1994). 

Mattson, W.J., Niemel_i, P., and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. LISIDA 


60

Encapsulated objects are data structures that always maintain their identity, have a clear boundary, and show to the 

outside world only selected facets of themselves. Objects have attributes and behavior. They can be concrete physical 

things, abstract ones, or events. Examples of concrete things are trees, insects, vehicles, people, etc. Examples of abstract 

objects are beliefs, statistical information, and time. Examples of events are decisions, insect outbreaks, treatments, etc. The 

attributes contain the internal state of objects in their values. Examples of the attributes of an insect object are length, wing 

color, etc. Examples of behaviors (also called methods) of in an insect object are development, emergence, oviposition, etc. 

Objects fall into two categories: classes and instances. A class is a definition, a blueprint for actual instances. For 

instance, a biological taxon such as the species of "Scots pine" can be a class. Its instances, in turn, are the physical individual 

trees that grow outside this building. Classes often form hierarchies where subclasses inherit the attributes and 

methods of their superclasses. At a lower level, specialized values and new attributes and methods can be introduced in order 

to refine the generic superclasses. 

Objects communicate by sending messages to each other. Different objects can respond to the same message in their 

own ways. For instance, a "time" object may send a message of "lightness change" with a value "dusk" to all the objects in a 

simulation. One insect object may respond to it by activating its flight behavior whereas another, that would only fly in 

sunlight, would deactivate its own. 

Objects may consist of other objects. For instance, a tree object consists of a root system, one or several trunks, and a 

crown, which, in turn, consist of smaller objects. Objects may be associated and connected with other objects by having their 

names for some purpose as the value of an attribute. OO models should be implemented with an OO programming language. 

The present system, called Lignum, has been written in C++ under the Unix operating system, X-Window System, and 3dimensional 

interactive graphics. 

RESULTS 

Description of the Tree Model 

The model describes a single Scots pine tree that consists of small elementary parts. The functions of the model have 

been further elaborated from Nikinmaa (1992), and the model has been described in detail by Perttunen et al. (1994) and 

Salminen et al. (1994). The parts belong currently only into two operational classes: shoots and buds. In addition, we define 

classes for the stem, branch whorls, foliage, bark, phloem, sapwood, and heartwood. Each instance of a stem contains zero or 

more instances of shoot, branch whorl, and it ends with an instance of bud. Each pair of instances of shoot is separated by an 

instance of branch whorl (Fig. 1). Shoots consist of foliage, bark, phloem, sapwood, and heartwood. In this version of the 

model, the objects in a stem are connected to each other using a list structure. 

The tree is originally started from one bud instance. Buds have methods for creating shoot instances, branch whorl 

instances, and new bud instances. The shoots create the parts they consist of. Shoots and branch whorls that do not have 

foliage any longer, are collapsed into the stems. 

As we do not yet have a class for foliage, the implementation of shoot captures the productive metabolism of the tree. 

These include methods for photosynthesis, respiration, and growth. Foliage is implemented as an attribute for the needle 

mass, which must have a value greater than zero for any of the production to take place. Production is handled by a method 

which boasts the following function: where Wf is the foliage mass of the shoot and i the degree of interaction of shoot p 

which is dependent on the shadiness of the location (Sievanen 1992) with the foliage mass x above the height z. Production 

has to be multiplied with factor P0 which describes the photosynthesis under unshaded conditions (Nikinmaa 1992). 

Respiration of buds, foliage, phloem, and sapwood is directly proportional to their mass. In the current implementation, 

the root system of the tree is extremely simple. It is considered to have mass that increases as the result of the production 

in the shoots. Partof the root mass dies annually. It also respires similarly to the foliage and the sapwood. 

After production and respiration the tree can allocate its net production into the new shoots, foliage and roots. The 

equations for these are given by Perttunen et al. (1994). Nutritional status of the parts is only dealt with a single attribute in 

the present model. 

61

Figure I.--Schematic representation of the object tree. (1) 

® Bud 

F--q Shoot 

_

_re3.--The Lignum model can be controlled through windows and sliders and the parameters can be customized. In the 

graph shown, one shoot in the top whorl has turned brown due to damage in sapwood as caused by Tomicus spp. 

Examples of Herbivore Attack on the Model ]¥ee 

Propagation of the effects of herbivore attack in the tree model can only be implemented through material flows from 

object to another. So far, we have only implemented an upward water flow-method for the sapwood part of shoots and 

ds and a downward starch flow-method for the phloem part of shoots and whorls. The status-attribute of these parts can 

values between 0 and 1. The rate of the flow is directly proportional to this status. The status of phloem, sapwood, and 

Lgeis also shown graphically as the color of the part as Figure 4 shows. 

Sapwood.Status _.- Sapwood.Upflow 

,. + +_,,'"_+ 

".X BLUE Foliage.Status 

Shoot.Colour + _ + ..... GREEN f.,,,,,,,.,,._..-,-- 

+ 

, / 

RED - :_ 

/ 

Foliage.Production 

/ + 

Phloem.Status _,. Phloern.Downflow + 

1re 4.--The color of the diseased shoot is shown as the combination of three main status attributes of its parts. 

63

Boring by Pine Shoot Beetles 

Examples of insects feeding on the canopies of trees are the pine shoot beetles, Tomicus spp. This bark beetle is the 

first one to invade dying or felled pines to breed in spring. The progeny emerging from breeding sites in trunks disperse to 

the crowns of pines during mid- and late-summer. In order to become mature, young adults feed in pine shoots burrowing 

intothe pith. The beetles attack mainly current year shoots of upper half of the canopy. The burrows may heal over only in 

the thickest shoots, otherwise they turn brown, break, and fall down. Physiologically, Tomicus attacks in shoots have an 

effect like the sudden cutting of ducts (Fig. 5). 

% 

f:: ' • _ _: ,:,. . Figure 5. After about 10 years of repeated Tomicus piniperda damage the tops of 

• _- _ i,,__. 5 Scots pine have become deformed as on the right. Under a permanent 

__ ': '*'.... " " beetle pressure the spike-like tops die and the pines start looking flat- 

_.w-_ "_'g_ ;:,:_ ,;. ,.(._ '. _:::':: i _: topped. 

, • _,: ": 

The effects of removal of shoots on remaining parts of the tree can be divided into three groups: first, the environmental 

effects, e.g., the changes in light conditions, water availability and release of nutrients from fallen shoots; second, the 

structural reactions like leader shoot changes, formation of adventitious shoots, sapwood-heartwood relation, and the dying of 

roots as the pipe model indicates; and third, the carbohydrate dynamics of the tree, including changes in assimilation 

nutrient allocation, root activity and compensation processes. The last group of effects is the most difficult to include 

model. These effects are discussed thoroughly by Swedish researchers (Fagerstrom et al. 1977, Ericsson et al. 1985, 

Langstrom et al. 1990,Troeng and Langstrom 1991, Langstrom and Hellqvist 1991, 1992). In addition, the reactions 

capacity, 

in a 

seem to 

d 

m 

b 

o 

be different according to age, growing site, and provenance of trees, v 

t 

The distribution of damage classes of crown can be used to make rough estimates of growth losses in stands repeat- 

edly attacked by pine shoot beetles (Kukkola et al. 1994). With a stand of model trees, the spatial pattern of beetle dispersal 

and the distribution of growth losses of pines may be examined, b 

The present model should be able to simulate the pruning pattern of Tomicus attack and the growth loss. The model 

o 

t 

is currently being used this way by letting individual Tomicus object instances to choose a shoot for foraging. The choices 

attacking beetles depend on the shoots already allocate by the previously attacked beetles, and the amount the tree stands 

above or below its neighbors. 

of 

out 

r 

Defoliation by Insect Larvae 

Scots pine is adapted to low nutrient resources and stores its major reserves in foliage. Figure 6 shows the dynamics e 

that follow from a severe defoliation in this situation. A severe defoliation reduces the carbohydrates and nutrients overall in o 

the tree, and also reduces the needle number. This leads to increased nutrient concentration in the remaining needles, as the s 

fine root biomass is not immediately reduced and its nutrient uptake pumps more nutrients into the remaining needles. This a 

increases the quality of the remaining needles for food of the defoliator. Increased nutrition in needles also leads to their d 

increased size and increases production. The carbohydrate concentration in needles increases as well as the concentration of 

the carbon-based protective metabolites (for a review, see Herms and Mattson 1992). 

64 

F 

d

SEVERE DEFOLIATION 

Tree 

Fine Root 

+ 

_ 

/_' 

/ 

........................ _ Carbohydrate 

Reserves 

Biomass - J_ Photosynthesis 14Y'+ 

Nutrient + 

Res erv es 

1 

+ 

Tree 

ee Number 

/ .,._.._ 

Carbohydrate 

Uptake Concentration 

Needle Size 

Capacity 'X_+ Concentration Needle Nutrient 

I + 

Nutrient | Needle C-Based 

Absorption , + Secondary 

Needle Food Quality _ Metabolites 

(No Induction) - 

Figure 6.--Dynamics of nutrients and carbohydrates in the Lignum model, with the effects of"defoliation. 

The present model, as it has explicit descriptions for foliage in each shoot, is capable of duplicating this kind of 

dynamics. However, we are still calibrating the model, and can not yet claim error-free predictions. 

DISCUSSION 

The present model was created for a testbench of ideas in many different disciplines. It should be reusable for 

diagnostic reasoning, studying the effects of herbivore attack, visualization of symptoms, research on growth and development, 

and modelling tree architecture. Although we are still working on the first version of the model, the experiences have 

been promising. Future versions will be built to incorporate simulation on parallel computer hardware, and zooming in and 

out of the tree. The color schemes that we use for illustrating the status of the functions of the tree do not produce realistic 

visualization of the symptoms as of now. We will study the color effects that follow from various injuries, and try to animate 

the suffering and death of trees. 

The model is a part of a larger system for integrated forest health management (Saarenmaa et al. 1994). It has been 

built compatible with object-oriented descriptions of forest insects and fungi. These have attributes that describe what parts 

of trees, represented as object classes, they attack and what kind of disturbances they cause in those parts. Matching OO 

trees with OO insects and other harmful agents will create new kind of ecological simulations and also be useful in diagnostic 

reasoning. 

SUMMARY 

A model of a tree which consists of thousands of small interconnected parts can potentially be useful for studying the 

effects of insect and disease attack on the whole system. We present such a model for Pinus sylvestris, which is based on 

object-oriented programming. The model consists of classes for shoots, buds, the stem, branch whorls, foliage, bark, phloem, 

sapwood, and heartwood. The model simulates primary production, respiration, growth, water flow, and starch flow, and has 

a geometry that can be monitored in 3-D color on computer screen. The effects of attacks by shoot boring bark beetles and 

defoliating insects are studied. 

65


AHONEN, J. and SAARENMAA, H. 1991. Model-based reasoning about natural ecosystems: An algorithm to reduce the 

computational burden associated with simulating multiple biotic agents, 193-200. In Computer Science for Environmental 

Protection. Sixth Symposium. Munich, Germany, December 4-6, 1991. Springer-Verlag, Heidelberg. 

BRATKO, I., MOZETIC, I., and LAVRAC, N. 1989. Kardio. A study in deep and qualitative knowledge for expert systems. 

260 p. The MIT Press, Cambridge, Massachusetts. 

ERICSSON, A., HELLQVIST, B., LANGSTROM, B., LARSSON, S., and TENOW, O_ 1985. Effects on grow*h of simulated 

and induced shoot pruning by Tomicus piniperda as related to carbohydrate and nitrogen dynamics in scots pine. 

LAppl. Ecol. 22: 105-124. 

•FAGERSTROM, T., LARSSON, S., LOHM, U., and TENOW. O. 1977. Growth in scots pine (Pinus sitvestris L.): a 

hypothesis on response to Blastophagus piniperda L. (Col., Scolytidae) attacks. For. Ecol. and Manage. 1:273-281. 

HERMS, D.A. and MATTSON, W.J. 1992. The dilemma of plants: to grow or defend. The Quart. Rev. Biol. 67(3): 283- 

335. 

KUKKOLA, M., ANNILA, E., TIMONEN, M., and VARAMA, M. 1994. The development and yield of the Scots pine 

stand repeatedly attacked by the pine shoot beetles. Manuscript. Finnish Forest Research Institute. 

KURTH, W. 1992. Morphological models of plant growth: Possibilities and ecological relevance. 14 p. ISEM's 8th 

International Conference on the State-of-the-Art in Ecological Modelling, Kiel, Germany, September 28 - October 2, 

1992. 

LANGSTROM, B. and HELLQVIST, C. 1991. Effects of different pruning regimes on growth and sapwood area of Scots 

pine. For. Ecol. and Manage. 44: 239-254. 

LANGSTROM, B. and HELLQVIST, C. 1992. Height growth recovery and crown development in top-damaged Pinus 

sylvestris trees. Scandinavian J. For. Res. 7: 237-247. 

L,_NGSTROM, B., TENOW, O., ERICSSON, A., HELLQVIST, C., and LARSSON, S. 1990. Effects of shoot pruning on 

stem growth, needle biomass, and dynamics of carbohydrates and nitrogen in Scots pine as related to season and tree 

age. Can. J. For. Res. 20: 514-523. 

LINDENMEYER, A, 1968. Mathematical models for cellular interaction in development. Parts I and II. J. Theoret. Biol. 

18: 280-315. 

MANDELBROT, B.B. 1977. The fractal geometry of nature. 468 p. W.H. Freeman and Co., New York. 

NIKINMAA, E. 1992. Analyses of the growth of Scots pine; matching structure with functions. Acta Forestalia Fennica 

235, 68 p. 

PERTTUNEN, J., SIEVANEN, R., SALMINEN, H., VAKEVA, J., NIKINMAA, E., and SAARENMAA, H. (1994). i 

generic growth model for a tree based on simple structural units and object-oriented programming. 30 p. (Subn-_tted). 

PRUNSINKIEWICZ, E and HANAN, J. 1989. Lindenmayer systems, fractals, and plants. Lecture Notes in Biomathernat_ 

its 79, 120 p. 

de REFFYE, P., LECOUSTRE, R., EDELIN, C., and DINOUARD, P. 1989. Modelling plant growth and architecture, p. 

237-246. In Cell to Cell Signalling: from Experiments to Theoretical Models. Academic Press, Oxford. 

66

RUMBAUGH, J., BLAHA, M., PREMERLANI, W., EDDY, E. and LORENSEN, W. 1991. Object-oriented modeling and 

design. 50() p. Prentice-Hall, Engtewood Cliffs, New Jersey. 

SAARENMAA, H. 1991. Object-oriented design of topological tree and forest models, 2 p. In Proceedings of the First 

European Symposium on Terrestrial Ecosystems: Forests and Woodlands. Florence, Italy, May 20-24, 1991. 

SAA_NMAA, H., PERTTUNEN, J., VAKEVA, J., and NIKULA, A. 1994. Object-oriented modeling of the tasks and 

agents in integrated forest health management. AI Applications 8(1): (In print). 

SALMINEN, H., SAARENMAA, H., PERTTUNEN, J., and VAKEVA, J. 1994. Modelling the hierarchical structure of 

trees with an object-oriented scheme, 18 p. In Resource Technology '94. International Symposium in Natural 

Resource Management (In print). 

SEQUEIRA, R.A., SHARPE, RJ.H., STONE, N.D., EL-ZlK, K.M., and MAKELA, M.E. 1991. Object-oriented simulation: 

plant growth and discrete organ to organ interactions. Ecol. Model. 58: 55-89. 

SIEVANEN, R. 1992. Construction and identification of models for tree and stand growth. Helsinki university of technology. 

Automation _1_chnology Laboratory. Series A: Research Reports 9, 155 p. 

TORASSO, R and CONSOLE, L. 1989. Diagnostic problem solving. 243 p. North Oxford Academic, London. 

TROENG, E. and LANGSTROM, B. 1991. Gas exchange in young Scots pine following pruning of current shoots. Ann. 

Sci. Forn. (Paris) 48: 359-366. 

67

DEFOLIATION OF FAGUS CRENATA AFFECTS THE POPULATION 

DYNAMICS OF THE BEECH CATERPILLAR, 

QUADRICALCARIFERA PUNCTATELLA 

NAOTO KAMATA t, YUTAKA IGARASHI I, and SEIJI OHARA 2 

_Laboratory of Forest Entomology, Tohoku Research Center 

Forestry and Forest Products Research Institute 

Nabeyashiki 72, Shimokuriyagawa, Morioka, Iwate 020-01, Japan 

2Laboratory of Chemical Conversion, Wood Chemistry Division 

Forestry and Forest Products Research Institute 

Matshunosato 1, Kukizaki, Inashiki, Ibaraki 305, Japan 

INTRODUCTION 

In spite of remarkable progress in understanding the ecology of plant-herbivore interactions during the past 90 years, 

there is little yet known about the interactions between beeches and their folivores (Edwards and Wratten 1983, Edwards et 

al. 1986). Two beeches, Fagus crenata and E japonica, are dominant in the deciduous, broad-leaved forests of the cooltemperature 

zone in Japan. They are found from the southern part of Hokkaido, the northernmost main island, to the high 

mountains of Kyushu, the southernmost. The beech caterpillar, Quadricalcarifera punctatella (Motschulsky) (Lepidoptera: 

Notodontidae), is a monophagous species which feeds on beech leaves and occasionally completely defoliates the trees. V_e 

studied its population dynamics and tested the maternal effects hypothesis (Rossiter 1992) and the inducible resistance 

hypothesis, assuming that food deterioration occurs in beech following herbivory. A further objective was to determine the 

exact mechanisms by which Q. punctatella body size and density decrease after an outbreak has occurred. 

Life History and Features of Outbreaks 

Q. punctatella is a univoltine species (Igarashi, 1975). The pupa overwinters in the ground, and the adult emerges 

from late May to late July with a peak in mid- to late June (Kamata and Igarashi 1995a). Eggs are laid in masses on the 

underside of beech leaves, the total number per female being about 350. These are laid in several masses, each containing 

20-100 eggs (Kamata and Igarashi 1995b). The larva feeds only on beech leaves and molts 3 or 4 times. The peak appearance 

of the last instar larva is in early August. 

Serious outbreaks have been recorded since 1917, and outbreaks are known from central Honshu, the main island of 

Japan, to southern Hokkaido (Fig. 1) (Kamata and Igarashi 1995d). Most intervals between outbreaks have been 8-12 years 

and the average duration 1-4 years. Outbreaks in many regions have been synchronized even though the period of one or the 

time between two differed by time and place. 

METHODS 

Study Sites 

Larval density was estimated in Hakkohda (Site A), Hachimantai (Site B), and Appi (Site C), located in the northern 

part of Honshu island (Fig. 1). Outbreaks have been recorded at Sites A and B, but no conspicuous defoliation has ever been 

recorded in Site C (Kamata et al. in preparation). 

Mattson, W.J., Niemelii, P., and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. U_IS)_A 

For. Serv. Gen. Tech. Rep, NC-183, N.C. For. Exp. Sta., St. Paul, MN 55108. 

68

oo0 

a b 

ol 

¢' _achiman_i ()® 1,eAppi (C_ 

Figure 1.---Locations of beech forests which have had been conspicuously defoliated by Q. punctatelIa (dark shading), and 

three study sites: Hakkohda (Site A), Hachimantai (Site B), and Appi (Site C). 

Estimation of Larval Density 

The density of Q. punctatella in each plot was estimated by trapping frass pellets of the last instar larvae, because 

defoliation is mostly the result of its feeding (Kamata and Yanbe, 1993). The density can usually be estimated by the 

following equation: Density = Total number of frass pellets per square meter/630, where 630 is mean number of frass 

pellets during last instar stage. Because density is underestimated when larval density becomes high and conspicuous 

defbliation occurs, the density must be calculated by the following equation: log,0 (Density) = 0.988"1og,0 X + 0.898, 

where X is the maximum number of frass pellets per day per square meter (Kamata and Igarashi, 1994a). 

Adult Number and Size 

Adult numbers were measured at Site B. Two light traps were set in an open area in a small valley surrounded by a 

forest of predominantly beech. The distance between the two traps was 10m and both were 20m from the nearest edge of 

the beech forest where the heaviest defoliation had occurred in a 1981 outbreak. Each light trap had two blue fluorescent 

bulbs (20W) and a basin with water and kerosene. The relative numbers were estimated by the total catches during the 

20 days from June 1lth to 30th, because the peak of the adult catches was in mid- to late June (Kamata and Igarashi, 

1994). Body size of the adult greatly influenced fecundity, but the fecundity of a trapped female cannot be estimated 

directly or indirectly from body weight, because both measurements decrease rapidly after emergence due to oviposition 

(Kamata and Igarashi 1995a). Rather, the area of forewing, which showed a good correlation with both fecundity and 

adult weight, was used as a substitute measurement of body size and fecundity (Kamata and Igarashi 1995c). The area 

was therefore measured by digitizer to the nearest mm z and its annual change was compared with population dynamics. 

In Site A, one light trap, which had two blue fluorescent bulbs (20W) and no basin, was set in a bare area ca. 20 m apart 

from the edge of beech forest completely defoliated in 1990. A sheet of white, finely-meshed cotton cloth (1.8m x 1.8m) 

was set behind the trap and adults were caught on it. This survey was conducted several days each year from 1988 to 

1993. Area of the beech caterpillar's forewing was also measured. 69 

\

Defoliation Experiments 

To test the food deterioration hypothesis, the growth of larvae reared on beech saplings which had suffered previous 

artificial defoliation wascompared with controls. It was assumed that the effects of defoliation would be apparent in mid 

August. Treatments were a 2 year regime where all leaves were clipped for 2 years running in 1991 and 1992. Another 

regime was a single year treatment during which all leaves were clipped for only the year (1992) prior to the experiment. No 

clipping was done in thecontrol group. Q. punctatella were reared in screened enclosures on branches of treated and control 

beeches. Larvae were occasionally shifted from one branch to another so that food was always plentiful. Survivorship and 

mature body size were compared by rearing the larvae to maturity on these saplings. The Q. punctatella used in the experiment 

originated from a population on Site A three generations after the previous outbreak. 

Testsfor Food Deterioration and Maternal Effects Hypotheses in the Field 

Three different regional Q. punctatella populations were reared on beech trees in the Tohoku Research Center of the 

Forestry and Forest Products Research Institute in Morioka to evaluate population quality (Fig. 1). Larvae were reared in 

screened enclosures on branches and mature body sizes were compared. 

Eggs collected from these three regions were attached to beech saplings in each region. Larvae were also reared in 

screened enclosures on branches, and survivorship and mature body size compared among the three. From these results the 

influence of I_ostand herbivore quality on population dynamics and associated changes in body size were determined. 

Quantitative and Qualitative Properties of Beech Leaves 

_Ik)learn about changes in leaves which had been defoliated, their number and dry weight were determined. Dry 

weight of l0 leaf disks (2 cm in diameter), one punched from each of 10 randomly selected leaves, was viewed as an index of 

leaf density. Decrease inthe weight and the density of a leaf were compared to determine how leaves change due to defoliation. 

Leaf toughness wasalso measured using a "penetrometer" as described by Feeny (1970). 

We used the protein-precipitation assay as a measurement of the tannins. We measured the precipitation of BSA by 

extracts of beech leaves according to Martin and Martin (1983) and Makkar et al. (1987). Leaf powder (150 mg) was 

extracted three times with 3 ml of 70% acetone, and the total volume of the combined solution of extracts was adjusted to 10 

ml with the solvent. One ml of the above stock solution from leaf powder was added to 2 ml or 3 ml of 0.1% BSA solution 

(containing 2 mg or 3 mg of BSA, respectively) in acetate buffer (pH=5.0). The amount of precipitated BSA was determined 

by ninhydrin assay. 

RESULTS 

Population Dynamics and Body Size 

Larval densities changed with the same pattern in all three areas (Fig. 2). Density was lowest in 1986, and increased 

successively until 1990 (Kamata and Igarashi 1995c). In Site A, slight defoliation was recognized in 1989 and heavy 

defoliation in 1990. The predator, Calosoma maximowiczi (Coleoptera, Carabidae), and an entomopathogenic fungus, 

Cordyceps militatis (Clavicipitales: Clavicipitaceae), greatly increased in abundance during the outbreak periods (Kamata 

and Igarashi 1994c, Sato et al. 1994). The densities in Sites B and C were not as high in 1990, and neither defoliation nor 

natural enemies were conspicuous. However, the density decreased in 1991 in all three areas, then increased again in 1992. 

The resilience of the population density in 1992 was stronger in Sites B and C, where natural enemies were not conspicuous 

in 1990, than in Site A. 

Moth numbers in Site B fluctuated in tandem with larval density (Fig. 3). They were lowest in 1987, then increased. 

Interestingly, moth numbers were almost the same in 1990 and 1991, though the larval density decreased in 1991. Adult 

number began to decrease after 1992. 

70

1000 Conspicuous Defo i at ion 

100 .__. _; Si t e-A 

_.i Hakkohda 

t0 _ .... ,, _ i 

0.1 

o.o_i 

100 , i 

i : , Site-C 

10 _ _" Appi 

i ''_ " i....* 

>" i...*"! i 

,J 

o3 

0.1 

o.ol,iii i i ', i 

I00 i 

0.1 Site-B 

Hachimantai 

O.Ol i 

i 

o.ool ,,.... i ..... i i _ i 

198586 87 88 89 90 91 92 93 

Year 

Figure 2._Annual changes in last instar larval densities of Q. punctatella in the three study plots (mean + 90% confidence 

intervals). Site A (Hakkohda) is where conspicuous defoliation occurred during our study, Site B (Hachimantai) is 

where conspicuous defoliation had been recorded but did not occur during our study, and Site C (Appi) is where no 

conspicuous defoliation of this species has been recorded. 

71

400 -0 

" 350 

r'''" 

= 300 

.¢Z: 

,_ 250 

"='200 

E 

Femal e • : 

• 

'. , • 

"- 150 

o 100 

,5 

z 50 _.--_ 

1600 

1400 

% 1200 

< 1000 

_._ 

% 800 

Male 

! ! 

_..===J 

." "_ 

'-,--o600 _" "", , 

_; 400 -/ 

z 200 

' I _I I II il ! _I 

1985 1986 1987 1988 1989 1990 1991 1992 1993 

Year 

Figure 3.--Number of Q. punctatella adults caught in two light traps in Site B (Hachimantai) 1985-1993. The number of 

adults of each sex in each trap is shown. Nitrogen content of foliage was measured by a CHNS/O Analyzer 

(Perkin-Ehner PE2400 Series II) to the nearest 0.01%. 

Body size of moths gradually increased with larval density, but decreased suddenly in 1991 in Site A where severe 

defoliation had occurred the previous year (Fig. 4). A quantitative food shortage in 1990 was probably the main factor 

causing this size reduction. Body size increased a bit but was still small in 1992 and 1993, but larval density decreased and 

no food shortage occurred after 1991. Thus, factors other than a shortage of food were involved in the size decrease. Because 

size was ahnost the same in the two sites before the outbreak, the great difference between the two after the outbreak 

was not genetically determined. 

72

2.oo 2.oo 

o t 4 t 

¢._ e'- 

1.80 1.80 ! t ! { 

_ _- 

,l-- 

3 3: 

,-,- o° 1.60 Site-A t ,,° o 1, 60 Site-B Female 

I. 40 Female "-- o 1. 40 

a,_ 0.> 

< I. 20 ..................... --'--- ........ < I. 20 ............... 

.--- 1 60 _ 1 60 

"E "E 

"; 140 "; 1 40 | I 

I 

g 

L I L_ 

o 

4-- 

1 30 Male t " 

4"- 0 

Male 

Site-A I ° 1 30 Site-B 

o 120 _ 1 20 

CL_ 

110 ...... _--- ' ' 

89 90 91 92 93 85 B6 87 88 89 90 91 92 93 

Year Year 

Figure 4.--Annual changes in adult size of (2. punctatella caught in light traps in Site A (Hakkohda) and Site B (Hachimantai) 

(mean + 95% confidence intervals). Area of fore wing, which has a strong relationship to both fecundity and adult 

weight, is shown for each sex. 

Defoliation Experiments 

Quantitative Changes in Beech Leaves and Food Limitation During an Outbreak Period 

The mean size and biomass of leaves declined precipitously after defoliation (Fig. 5). The 1-year treatments caused a 

46.1% decline (ca. 33.4 to 18.0 rag) in mean leaf dry weight from 1991-1992, and a 32.3% decline (ca. 41.2 to 30.8 mg) 

from 1992-1993. Following the second year of defoliation (i.e., in 1992), biomass declined further to ca. 12 rag/leaf in 1993, 

a 64.9% drop from the 1991 mean weight (Fig. 5a). Leaf density or dry weight of 10 leaf disks (31.4 cm 2total) from defoli .... 

ated trees followed the same pattern: 1-year treatments caused a 14.0% decline (108 to 92.8 rag/10 disks) from 1991-1992, 

and an 8.0% decline (104 to 93.8 rag/10 disks) from 1992-1993 (Fig. 5b). The 2-year treatment caused a further decline in 

leaf density to ca. 81 rag/10 disks, > 25% down from the initial mass in 1991 (Fig. 5b). Besides a decline in leaf size, and 

leaf density, the number of leaves on each tree also decreased slightly after the 1-year treatment, though two of 12 trees 

exhibited a slight increase in leaves (Fig. 5c). Consequently, total leaf biomass per tree declined by about 50% due to the 1year 

treatment (Fig. 5d). 

In two plots at Site A, slight defoliation was observed in 1989, the amount of leaves fed on by Q. punctatella was 

estimated at ca. 60% of total canopy, and ca. 20% of the beech trees were completely defoliated. The following year, Q. 

punctatella increased in density and defoliated completely ahnost all beech trees. Food limitation was an important mortality 

factor in this generation because many dead larvae, which had not been killed by predators or pathogens and were supposed 

to have died of starvation, were scattered on the ground. Considering the results of the defoliation experiment, the biomass of 

leaves should have decreased in 1990 as a result of the severe defoliation the previous year. A negative feedback occurred: 

as herbivore density gradually increased the defoliation reduced the amount of available food in the following generation, 

thus acting as a limiting factor. 

73

co 100-- d 

> 

_' 10 

01--I 

700 -- 

I I 

60 CD 600-- 

500- 

400c 

_-- o 300 - 

200z 

I00- - -" 

60 

60 

._ 

0--1 

120-- 

1 I 

110 

•--- _ 100 @ Control 

_- E _[_ 1 Yr 

o "--- 90 -- Treatment 

2Yr 

"=: 80 -- Treatment 

(3,} 

_J 

70-I I I I 

50 Control 

o. __ _E 40 

30 

L 

l -""_ 1 Yr 

.= 20-- 

.,.., "---" l _ i l Treatment 

' _ 2 Yr 

"_ 10_ _ Treatment 

0---I I I I 

1991 1992 1993 

Year 

Figure 5.--Quantitative changes in beech leaves in the year following artificial defoliation, a: Dry weight per leaf, b: Index. 

74 

of leaf density measured by dry weight of 10 leaf disks (2 cm in diameter), c: Number of leaves in each tree, d: Dry 

weight of all leaves in each tree.

Qualitative Changes in [,eaves and Larval Growth 

Leaf"toughness, like leaf"density, decreased following clipping of the leaves: it was only 85.7% of the control the 

year following the 1 year treatment and 69.8% of the control after the 2 year treatment (Fig. 6). Nitrogen content was almost 

the same in untreated trees tot 3 years (1991-1993) but decreased after leaves were clipped. It was 2.23-2.26% before the 

treatment, 2.00-2.05% the year following the 1 year treatment (ca. 10% reduction), and 1.88% the year following the 2 year 

treatment (16.6% reduction) (Fig. 7a). The N content following the 1 year treatment was almost the same in 1992 and 1993. 

These results clearly indicate that nitrogen content is reduced by defoliation and is influenced not by the year but by the 

degree of defoliation. Tannin content, however, increased the year following leaf clipping (Fig. 7b). When 2 mg of BSA was 

added, about 0.4 mg of the protein was precipitated in the regimes of 2 year treatment and 1 year treatment, although precipitation 

was not detected in the control regime. Furthermore, when 3 mg of BSA was added, no precipitation was recognized 

in the 1 year treatment or the control regimes, but 0.43 mg of the protein was precipitated following the 2 year treatment. 

These quantitative and qualitative changes in beech following defoliation support that the trees were less vigorous after 

defoliation, and also less suitable for herbivores because of the declining nitrogen and increasing tannins. 

180 

160 I " 

z J 153.5 

[ 

co | 

140 

131.7 

c 

= 120 

0 

F--- 

100 

(:D 

107.2 

L_ ............. 1 

--J T 

80 ' ' ................ 

2 Yr Treatment 1 Yr Treatment C0ntr0] 

Figure 6._Leaf toughness of beechleavesin the year following artificial defoliation. The number is a mean value, bar 

means maximum and minimum value, and upper and lower sides of box are one standard error. 

Survivorship rates on beech saplings were determined using the number of hatched larvae as a base. The x-axis in 

Figure 8 is the number of days following egg attachment on June 30th. Although approximately half of the larvae reared on 

the control saplings survived to maturity, only about 10% of those reared on the 2 year treatment trees survived to maturity. 

Many of the latter died during early stages of larval development. Survivorship on the 1 year treatment was intermediate 

between the 2 year treatment and the control. Both female and male body size declined for larvae reared on clipped saplings, 

as did adult body size (Fig. 9). In the 2 year treatment regime, body size was smaller than of the 1 year treatment regime, and 

was just three quarters of that in the control. Defoliation lowers beech quality which, in turn, lowers larval survivorship and 

body size. 

Population Quality of Herbivore in Three Sites 

Test for Maternal Effects and Food Deterioration in the Field 

The larvae from Site A had smaller body size (p < 0.01 t-test) than those from the other two regions (Fig. 10). This 

tendency was exactly the same for males and females, as well as for mature larvae and for pupae. No difference was found 

between the Site B and C populations. Thus it can be said that even three generations after an outbreak, the quality of the 

Site A population was less than that of the populations from the other two regions where density was increasing. 

75

0.5 b 

O.4 AddedBSA 

2mg 

=0.3 

,_ El 3mg 

-'0.2 

- r--" 

o0.1 

'- N.D. N.D 

_ j = 1 

2 Yr Tr 1 Yr Tr Control 

2.3-- a 

_= _ 

o 

2.2-- 

2.1- 

{ Control 

1Yr 

z 

1 9- 

" 

Treatment 

Yr 

Tre;._tment 

1"77 I I I 

1991 1992 1993 

Year 

Figure 7.--Nitrogen and tannins contained in beech leaves the year following artificial defoliation, a: Total nitrogen, b: 

Weight of precipitated BSA by addition of 2 ml of 0.1% BSA solution (corresponding to addition of 2 mg of 

BSA) and 3 ml of 0.1% BSA solution (3 mg of B SA). N.D. indicates that no precipitation was detected.

t00 2 Yr 1 Yr 

o_ 80 Treatment Treatment . :_,," 

_- 60 

o 40 

o_ 20 

Control 

-- L_ 

0 20 40 60 80 0 20 40 60 80 0 20 40 60 80 

Days from Jun. 30 

Figure 8.--Survivorship curves of Q. punctatella larvae reared on beech which had suffered artificial defoliation. Each line 

began with an egg-mass containing about 50 eggs. X-axis is the number of days following egg attachment on June 

30th. 

500 9,_ 

400 I l 

--_ [----_ w 

300 

¢.... 

200 

; 291.5 345 392.9 Mean 

100 2 8 34 N 

_ ! 

450 

4O0 

o_ 

| 

350 [ I 

.....'_300 F ---_L--- ..... ] I 

250 

•-'= 200 

l ' n l 

"T_150 236.7 289.5 353.3 Mean 

3= 100 

50 

3 10 38 N 

! _I _1 

2 Yr Treatment 1 Yr Treatment Control 

Figure 9.--Body size of Q. pttnctatella larvae reared on beech which had suffered artificial defoliation. Weight of larvae at 

maturation is indicated for each sex. Bar means maximum and minimum value, and upper and lower sides of box are 

one standard error. 77

__ Maturation 500 o_ Maturation 

_ 600 E 

r-- I _ --- 300 

400 -'= 

447.1 507.5 504.7 Mean .=,,_200 339 7 425 6 405.7 Mean 

• • 

200 30 21 19 N _= 100 37 27 20 N 

0 

_t Pupation o_ Pupation 

600 500 

500 ]....._ I --- 400 

40o I Ij F-__] , 3oo I 1 w 

300 El, _ 200 ' 

._ 200 342.6 396.6 406.3 Mean 265.3 339.1 329.6 I_ean 

100 29 21 19 N _= 100 32 27 20 II J 

0 , , , 0 ' ' 

A B C A B C 

Locationof Q. punctatella Locationof Q. punctatella 

Figure lO.---Body size of three different regional Q. punctatella populations reared on beeches in the Tohoku Research 

Center of the Forestry and Forest Products Research Institute in Morioka. Weight of larvae at maturation and at 

pupation are indicated for each sex. Bar means maximum and minimum value, and upper and lower sides of box are 

one standard error. 

Tests of Two Hypotheses in Field Populations 

Larvae from all of the regions were reared on beech saplings in each of the respective regions. Plant quality influenced 

survivorship in each region. Although great differences in survivorship between regions were observed, little difference 

was found within a region (Table 1). Approximately half of the larvae reared on beech saplings in Sites B and C 

survived to maturity, whereas survivorship of all those reared on Site A beech was between 3.2% and 6.5%. The influence of 

Q, p,nctawlla population quality in each region on body size was clearly evident. Irrespective of the location of beech 

saplings on which the larvae were reared, the order of average body size among the regional populations was the same: B >__ 

C 2 A (Table 2). In particular, body size of A population larvae was significantly smaller than that of larvae from the other 

Table !.--Survivorship of Q. punctatella larvae from each region on beech in that region. 

78 

Q. punctatella 

A B C 

A 3.6% 3.2% 6.5% 

(6/165) (5/155) (10/154) 

E crenata B 48.9% 55.0% 53.9% 

(44/90) (33/60) (41/76) 

C 55.8% 52.9% 

(72/129) (36/68) 

(No. of matured larvae / No. of hatched larvae)

Table 2.--Body size of Q. purzctatella from each region on beech in that region. Weight of larvae at maturation is shown for 

each sex. 

Female Q. purzctatella 

A B C 

376.3+68.5 442.0+ 15.6 414.4+48.9 

A 388 453 489 

369 431 360 

3 2 5 

398.5+75.3 484.5+108.9 451.3_+58.8 

E crenata B 576 739 554 

309 333 340 

17 17 22 

392.9+26.3 455.0+68.5 

c 451 550 

351 342 

34 14 

Male Q. punctatella 

A B C 

288.3+60.5 369.7+ 15.6 336.8+41.3 

A 349 384 398 

228 353 298 

3 3 5 

294.5+63.6 384.9+36.99 335.3+45.2 

E crenata B 416 469 409 

204 332 254 

27 16 19 

353.3+31.9 396.0+-39.0 

C 412 450 

279 299 

38 22 

Mean+SD 

Max 

Min 

N 

two regional populations (p < 0.01 t-test). The influence of beech on body size was not as apparent as the population effect: 

female body size of larvae reared on Site B beech was close to that of larvae reared on Site C beech, while male body size of 

larvae reared on Site B beech was similar to that of those reared on Site A beech. 

Leaf properties were compared among the three sites (Fig. l 1), and weight per leaf among the three was almost the 

same. Leaves of Site A beech were significantly tougher than those of the other two sites, though they were much thinner 

than Site C leaves and almost the same as the Site B leaves. A remarkable point is that nitrogen content was significantly 

higher in leaves of Site A than in those of the other two sites (Fig. 12). Tannin content also increased greatly. Judging from 

the high N content of beeches in Site A, where severe defoliation had occurred three years before our experiment, trees had 

79

240 

__ 220 

O_ 

200 

180 

C 

= 0 160 

140 

4-- 

" 120 

.J 

100 ...... 

.. 

153.5 

.110 b 

__,_ 

_100 _____[ w 

90 88.3 89.7 101.7 

70 j ! I 

100 

a 

4-- 80 I 

,_J 

46.5 

,=- 4O 42.3 

-,-" °o,o 1 ....... • I 49 2 

.'=" 20 7----- 

.............. 1 i a 

A B C 

Site 

Figure l l._Leaf thickness and toughness of beech leaves in three different regions. Leaf thickness was represented by dry 

weight of 10 leaf disks (2 cm in diameter). The number is a mean value, bar means maximum and minimum value, 

and upper and lower sides of box are one standard error. 

80

06 b 

AddedBSA 

< 04 _ 2rag 

-'=' i:3 

-,03 

'F=,.0 2 

_-01 

_ N.D. 

,_ _ _ J 

3 a 

2.8 ...... _ 

2.84 

_2.6 i ............. 

=" 2.4 2.58 

...... ..... i------=-_- _ 

2.2 ] ! -- 2.36 w I iiiiii 

....... ,..... L_.__J"....... I _ _ H _, I _'_ " _fl 

A B C 

Sit,e 

Figure 12.--Nitrogen ahd tannins contained in beech leaves in three different regions, a: Total nitrogen (The number is a 

mean value, bar means maximum and minimum value, and upper and lower sides of box are one standard error), b: 

Weight of precipitated BSA when added at 2 mg and 3 mg BSA. N.D. indicates that no precipitation was detected. 

become vigorous. However, since the trees had also increased their defensive secondary compounds, they had also become 

less suitable for herbivores, and insect mortality was higher than in trees in the other two sites. 

DISCUSSION 

Self regulation by reducing adult body size and curtailing egg production in insects during an outbreak with resultant 

declines in the initial number of individuals in the next generation is thought to be one of the major factors governing the 

population dynamics of insects (Klomp 1966, Dempster and Pollard 1981). This has been shown to be a consequence, 

among other factors, of insufficient food supply due to complete defoliation of the host, as well as a density dependent 

outcome resulting from conditions experienced during larval development. However, this phenomenon is not restricted 

81

solely to the outbreak generation. Many insects have been shown to have body size changes associated with their population 

density dynamics. In other words, during a period when population density is increasing, body size also increases, and the 

reverse occurs when population density is declining, body size decreases also (Baltensweiler and Fischlin 1988, Myers 1990), 

In Q. punctatella, body size changed in the same way as the density. Changes in fecundity caused by these changes in body 

size were probably one factor causing the population fluctuation. 

Because the body size of Q. punctatella at Site A before the outbreak was approximately the same as that for the Site 

B population, the current small body size of the former is thought not to be genetically determined. Body size at Site A 

dropped to 70% of the norm the year following the outbreak and was 80% even two generations after the outbreak. This 

cannot not be explained solely by a lack of food during the outbreak. The results indicate that beech which has experienced 

de_i)liation the previous year is a poor host, causing both survivorship and body size of larvae to decline. Even though 3 

years had passed since the outbreak at Site A, larvae reared on Site A beech still had low survivorship. Corresponding to host 

quality, Q. punctatella quality was still bad in Site A. Even when Q. punctatella populations from this site were reared on 

beech from the other sites where density was increasing, the bodies of these larvae were still significantly smaller. High 

mortality, due to deterioration of food and small body size, and low egg production due to poor Q. punctatella quality, 

explains why its density continued to decline years after the outbreak. Several authors have demonstrated that poor nutritional 

status in one generation may deter normal development in the next. Examples include Lymantria dispar (Kovasevic 

1956), Malacosoma pluviale (Wellington 1965), and Hyphantria cunea (Morris 1967). Such intergenerational, cumulative 

effects are called "maternal effects" (Rossiter 1992). Rossiter (1992) states "maternal effects are the result of resource 

provisioning by one generation for the next... The resource based maternal effects are the product of gene-environment 

interactions experienced in the parental generation." Deterioration in Q. punctatella quality after the outbreak was the very 

result caused by maternal effects. 

Haukiojia and Neuvonen (1987) proved the hypothesis that poor food may select for individuals with superior ability 

to process low-quality diets. However, experiments conducted here did not confirm this: the survivorship of Site-A Q. 

punctatetla was significantly lower than that of Site C Q. punctate[la, and was almost the same as that of Site B Q. 

punctatella on poor-quality Site A beech. 

Concerning the food deterioration hypothesis, it is necessary to distinguish between induced changes in secondary 

compounds such as tannin, which are an active means of defense (induced defense hypothesis) and that of passive changes in 

secondary chemistry following episodes of defoliation (Myers 1988). The defense strategy of beech, in relation to severe 

defoliation, changed as time passed after an outbreak, though in each case it resulted in high mortality and small body size of 

Q. punctatella. In the year following defoliation, beech became less vigorous and the defense response was a passive one 

(food deterioration): tannins increased and nitrogen content decreased. Next, trees became more vigorous and the defensive 

strategy changed to a more active one; both nitrogen and tannin content increased. 

The results on Site B beech were unexplainable by this scenario; nitrogen content was lower than Site A beech and 

tannin content was almost the same, but the tree performance was comparable to Site C beech for rearing Q. punctatella. 

These results indicate that it is very dangerous to discuss the performance of plants by measuring only nitrogen or secondary 

compounds. Clancy (!99 l) also demonstrated that Douglas fir trees susceptible to the western spruce budworm had lower 

levels of foliar nitrogen and sugars than resistant trees, and that the susceptible trees had mineral/nitrogen ratios which were 

closer to optimal levels. Not only secondary compounds such as phenols and tannins but also minerals thus must be taken 

into consideration. 

Nitrogen content in untreated trees at Site A was nearly the same in all 3 years (1991-1993) and that in beeches 

following 1year of defoliation was the same for 2 years (1992-1993). It differed greatly among the three sites, but variance 

within a site was very minimal. Beech leaves at outbreak Sites A and B showed higher nitrogen content than at the nonoutbreak 

Site C. There is a possibility that nitrogen level is linked to the site dependent outbreak characteristics of Q. 

punctatella. 

The induced defense hypothesis advocates the following scenario. Due to the decline in food quality after an 

outbreak, succeeding generations decline, that is, insect growth is checked and population numbers decrease. With the 

recovery of plant quality, insect population density begins to increase again, thus creating the cyclical population dynmnic:s 

of these insects. Two different types of beech response were recognized in the decline in food quality as time passed afte:r an 

outbreak (Fig. 13). Beech becomes less vigorous soon after severe defoliation, and nitrogen content decreases but deferlsive 

82

Severe Defol iat ion Site-A 1993 

_ _ Tree vigour 

_ - 

j 

V i1 

i 

I 

V I 

_ ,, 

NitrogenContent 

Deiensive Secondary Compounds 

Food Quail ty 

oar 

Passive Defense -_-------_Active Defense 

Figure 13.--Schema of defense response of beech after severe defoliation. 

secondary compounds increase. This rather passive food deterioration causes high mortality and smaller body size in the Q. 

punctatella population. Thus, Q. punctatella density decreases and beech is released from severe herbivory pressure. Beech 

gradually becomes vigorous again and its defense strategy changes to a more active one; defensive compounds such as 

tannins increase, though nitrogen content also recovers. Such beeches are not suitable for herbivores, which results in high 

mortality of Q. punctatella. 

83

SUMMARY 

Body size of Q. punctatella at first increased with population density, but became very small the year following an 

outbreak. Its size was still small 3 years afterwards. Because adult size greatly influences fecundity, this was a factor in 

keeping Q. punctatella density low for several years after an outbreak. As for food quantity, severe defoliation caused a 

decrease in the amount of leaves the following year, and this limited the peak density. The defense strategy of beech in 

relation to severe defoliation changed as time passed. In the year following severe defoliation, nitrogen content decreased 

and tannins increased; this resulted in high mortality and small body size in Q. punctatella. Three years after an outbreak, 

nitrogen content recovered to the same level as prior to defoliation, but tannins were even higher. This active, induced 

defense kept larval survivorship low. Maternal effects were also recognized; the quality of Q. punctatella was poor even 3 

years after an outbreak. Thus, food deterioration and the correlated results of maternal effects suppressed Q. punctatella 

density for several years after an outbreak. 


We would like to sincerely thank Dr. Garry J. Piller (School of Environmental Earth Sciences, Hokkaido University) 

for his kind advice and discussion. 


BALTENSWEILER, W. and FISCHLIN, A. 1988. The larch bud moth in the Alps, p. 332-351. In Berryman A.A., ed. 

Dynamics of Forest Insect Populations. Plenum, New York-London. 

CLANCY, K.M. 1991. Douglas-fir nutrients and terpenes as potential factors influencing western spruce budworm defoliation, 

p. 124-134. In Baranchikov, Y.N., Mattson, W.J., Hain, F., and Payne, T.L., eds. Forest insect guilds: patterns of 

interaction with host trees. GTR-NE-153. Radnor, PA: U.S. Department of Agriculture, Forest Service. 

DEMPSTER, J.P. and POLLARD, E. 1981. Fluctuations in resource availability and insect populations. Oecologia 50: 412- 

416. 

EDWARDS, P.J. and WRATTEN, S.D. 1983. Wound induced defenses in plants and their consequences for patterns of 

insect grazing. Oecologia 59: 88-93. 

EDWARDS, P.J., WRATTEN, S.D., and GREENWOOD, S. 1986. Palatability of British trees to insects: constitutive and 

induced defenses. Oecologia 69:316-319. 

FEENY, P. 1970. Seasonal changes in oak leaf tannins and nutrients as a cause of spring feeding by winter moth caterpillars. 

Ecology 51:565-581. 

HAUKIOJIA, E. and NEUVONEN, S. 1987. Insect population dynamics and induction of plant resistance: the testing of 

hypotheses, p.411-432. In Barbosa, E and Shultz, J.C., eds. Insect Outbreaks. Academic Press, New York. 

IGARASHI, M. 1975. The beech caterpillar, Quadricalcarifera punctatella (Motschulsky)(Lep., Notodontidae), as an 

important defoliator of beech, Fagus crenata Biume. Monthly Report of Tohoku Research Center, Forestry and 

Forest Products Research Institute 162: I-4. l(in Japanese) 

KAMATA, N. and IGARASHI, Y. 1994. Problems in estimating the density of larval beech caterpillar, Quadricalcarifera 

punctatella (Motschulsky)(Lep., Notodontidae), using frass drops by modification of Southwood-Jepson's method. J. 

Appl. Ent. 118: 92-98. 

KAMATA, N. and IGARASHI, M. 1995a. Diurnal change of the adult behavior, daily oviposition, and influences of 

temperature on adult emergence and light trapping data of the beech caterpillar, Quadricalcarifera punctatella 

(Motschulsky) (Lep: Notodontidae). J. Appl. Ent. 119: 177-184. 

84

KAMATA, N. and IGARASHI, M. 1995b. Relationship betweeen temperature, number of larval instars, larval growth, body 

size, and adult fecundity of the beech caterpillar, QuadricaIcarifera punctatella (Motschulsky) (Lep., Notodontidae): 

As a cost-benefit trade-off. Environ. Entomol. 24: 648-656. 

KAMATA, N. and IGARASHI, Y. 1995c. An example of numerical response of the carabid beetle, Calosoma maximowiczi 

Morawitz (Col., Carabidae), to the beech caterpillar, Quadricalcarifera punctatella (Motschulsky) (Lep., 

Notodontidae). J. Appl. Ent. 119: 139-1142. 

KAMATA, N. and IGARASHI, Y. 1995d. Synchronous populations of the beech caterpillar, Quadricalcarifera punctatella 

(Motschulsky): Rainfall is the key, p. 452-473. In Hain, R.R., Salom, S., Ravlin, W., Payne, T., and Raffa, K., eds. 

Behavior, population dynamics, and control of forest insects. USDA Forest Service, Radnor, PA 19087. 

KAMATA, N. and YANBE, T. 1993. Frass production of last instar larvae of the beech caterpillar, Quadricalcarifera 

punctatella (Motschulsky)(Lep., Notodontidae), and method of estimating their density. J. Appl. Ent. 117:84-9 I. 

KLOMP, H. 1966. The dynamics of a field population of the pine looper (Lepi., Geom.). Adv. Ecol. Res. 3: 207-303. 

KOVASEVIC, Z. 1956. Die nahrungsswahl und das auftreten der pflanzenschadlinge. Anz. Schaedlingskd. 24: 97-101. 

MAKKAR, H.P.S., DAWRA, R.K., and SINGH, B. 1987. Protein precipitation assay for quantitation of tannins: determination 

of protein in tannin-protein complex. Anal. Biochem. 166: 435-439. 

MARTIN, J.S. and MARTIN, M.M. 1983. Tannin assays in ecological studies. -Precipitation of Ribulose-l,5-Bisphosphate 

Carboxylase/Oxygenase by tannic acid, quebracho, and oak foliage extracts. J. Chem. Ecol. 9: 285-294. 

MORRIS, R.F. 1967. Influence of parental food quality on the survival of Hyphantria cunea. Can. Entomol. 99: 24-33. 

MYERS, J.H. 1988. The induced defense hypothesis: Does it apply to population dynamics of insects? p. 345-365. In 

Spencer, K.C., ed. Chemical Mediation of Coevolution. Academic Press, San Diego. 

MYERS, J.H. 1990. Population cycles of western tent caterpillars: experimental introductions and synchrony of fluctuations. 

Ecology 71" 986-995. 

ROSSITER, M.C. 1992. The impact of resource variation on population quality in herbivorous insects: a critical aspect of 

population dynamics, p. 13-42. In Hunter, M.D., Ohgushi, T., and Price, RW., eds. Effects of Resource Distribution 

on Animal-Plant Interactions. Academic Press, San Diego. 

SATO, H., SHIMAZU, M., and KAMATA, N. 1994. Detection of Cordyceps militalis Link (Clavicipitales: Clavicipitaceae) 

by burying pupae of Quadricalcarifera punctatella (Motschulsky) (Lepidoptera: Notodontidae). Appl. Ent. Zool. 29: 

130-132. 

WELLINGTON, W.G. 1965. Some maternal influences on progeny quality in the western tent caterpillar, Malacosoma 

pluviale (Dyar.). Can. Entomol. 97: 1-14. 

85

THE IMPACT OF FLOWERING ON THE SUITABILITY OF BALSAM FIR 

FOR SPRUCE BUDWORM VARIES WITH LARVAL FEEDING BEHAVIOR 

1_.BAUCE and N. CARISEY 

D6partement des Sciences Foresti_res, Facult6 de Foresterie et de G6omatique 

Universit6 Laval, Ste-Foy (Qu6bec), Canada G 1K 7P4 

INTRODUCTION 

The production of staminate flowers by coniferous trees is known to affect the fitness of insect defoliators. For 

instance, spruce budworm, Choristoneurafumiferana Clem., larvae on flowering balsam fir, Abies balsamea (L.) Miller, had 

3 to 7 days shorter development times than those on nonflowering trees (Jaynes and Speers 1949, Blais 1952, Greenbank 

1963). Based on this effect, Blais (1952) hypothesized that flowering of balsam fir could trigger spruce budworm outbreaks. 

The impact of flowering on budworm survival and fecundity varies greatly from one study to another. Blais (1952) 

found no significant effect of flowering on spruce budworm survival, while Mattson et al. (1991) detected a 25% increase in 

larval survival when insects were reared on balsam fir branches bearing staminate flowers compared to those on nonflowering 

branches. Moreover, balsam fir flowering had no effect on the fecundity of spruce budworm (Blais 1952), whereas Pinus 

banksiana Lamb. flowering had a negative effect on the fecundity of the jack pine budworm, Choristoneura pinus. Although 

there is no agreement on whether flowering positively or negatively affects budworm mortality and fecundity, most authors 

believe that the impact of flowering results from either the high nutritive value of the pollen and/or the presence of staminate 

flower clusters that may provide a more favorable micro-habitat in term of heat and protection from natural enemies (Lejeune 

and Black 1950, Wellington 1950, Blais 1952, Greenbank 1963, Mattson etal. 1991). 

In regard to budworm dynamics, the production of staminate flowers can be followed, the next year, by either an 

increase (Graham 1935, Blais 1952, Volney 1988) or a decrease (Batzer and Jennings 1980, Bauce 1986) in budworm 

population density. In order to explain such opposite trends, we hypothesize and test here that the impact of flowering on the 

suitability of balsam fir for budworm larvae varies according to the feeding behavior of the larvae, which is affected by the 

density of budworm population (Blais 1952). 

METHODS 

Field Rearing Experiment 

The impact of flowering on spruce budworm growth, development, survival, and fecundity was determined by caging 

second instar larvae in the mid- and lower-crowns of flowering and nonflowering balsam fir trees. In 1992, a total of 10 

flowering and 10 nonflowering dominant balsam fir trees were randomly selected in a 60 yr-old balsam fir stand located in 

compartment 20 of the For_t Montmorency (47° 19' N, 79° 09' W), an experimental forest of Universit6 Laval. The stand fits 

Grandtner's (1966) description of the balsam fir-white birch association and was classified as site 1 quality with good 

drainage (B61anger et al. 1983), 15%slope, deep uncompacted glacial till, and 65% crown cover. 

A severe drought in 1991 triggered an intense production of staminate flowers the following spring. Trees located 

near streams, or not fully exposed to sunlight did not produce staminate flowers. On each sample trees, two 90 cm long 

branches, facing north-northwest, were selected at the mid and lower third sections of the crown. Staminate flowers were 

present only in the midcrown of flowering trees. Reproductive buds burst betbre vegetative buds and the pollen was dispersed 

during a period of 3 to 4 days. Each branch was enclosed with a fine-mesh cloth sleeve cage which served as an 

Mattson, W.J., NiemelS., R, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


86

enclosure %r 20 post-diapausing second instar larvae. Larvae were placed in the cages when lO0°C-days were attained, or 

approximately 2 weeks prior to opening of vegetative buds, 

Sample branches were cut when budworms reached pupal stage. Pupae were weighed, moths were melted, the 

number and the weight of eggs laid were recorded every day during the oviposition period. The progeny were placed at 2°C 

for a 29 week long diapause period, and the survival of the progeny was recorded at the end of diapause. On each branch, the 

weight of current-year foliage produced by the tree, and the weight of current-year and 1-year old foliage removed by the 

larvae were estimated. 

Data were analyzed using the Statistical Analysis System (Sas Institute 1988). Normality and variance homogeneity 

tests were performed before data were subjected to a nested-factorial, multivariate analysis of variance with individual trees 

nested within flowering class and crown sections crossed with flowering class. Percent survival was analyzed using Chisquare 

analysis. 

Laboratory Rearing Experiment 

The impact of spruce budworm feeding behavior on the effects of flowering on the insect growth and development 

was determined using laboratory rearing experiments with fresh food from the field. Also, this approach allowed us to 

separate the food effect from the microhabitat effect provided by the staminate flower clusters. 

A total of eight feeding scenarios were simulated under laboratory conditions (T = 12°C, RH = 65%, 18L:6D) using 

the methodology of Bauce et al. (1994). These conditions corresponded to the average conditions prevailing at the experimental 

field site during the period of spruce budworm feeding activity. The eight feeding scenarios were (1) larvae fed on 

staminate flowers during 4 days, and current-year foliage from the midcrown of flowering trees thereafter, (2) larvae mined 

1-year old needles until vegetative budbreak, and fed on current-year foliage from the lower crown section of flowering trees 

thereafter, (3) larvae fed on staminate flowers during 4 days, current-year foliage from the midcrown of flowering trees until 

they reached their larval sixth instar, and 1-year old foliage from the midcrown of flowering trees thereafter, (4) larvae fed on 

staminate flowers during 4 days, current-year foliage from the midcrown of flowering trees until they reached their larval 

sixth instar, and current-year foliage from the lower section of flowering trees thereafter, (5) larvae mined 1-year old needles 

until vegetative budbreak, and fed on current-year foliage from the midcrown of nonflowering trees thereafter, (6) larvae 

mined l-year old needles until vegetative budbreak and fed on current-year foliage from the lower crown section of nonflowering 

trees thereafter, (7) larvae mined l-year old needles until vegetative budbreak, fed on current-year foliage from the 

midcrown of nonflowering trees until they reached their larval sixth instar, and 1-year old foliage from the midcrown of 

nonflowering trees thereafter, and (8) larvae mined 1-year-old needles until vegetative budbreak, fed on current-year foliage 

from the mid-crown of nonflowering trees until they reached their larval sixth instar, and current-year foliage from the lower 

crown section of nonflowering trees thereafter. The scenarios 3, 4, 7, and 8 simulated a lack of current-year foliage when 

larvae reach sixth instar. This phenomenon usually occurs when larval density is high. Scenarios 3 and 7 simulated a 

backfeeding (feeding on 1-year old needles) while scenarios 4 and 8 simulated dispersal of the population from the midcrown 

to the lower crown section. Backfeeding and dispersal to the lower crown section were observed by Blais (1952) at high 

budworm population density. 

Two trees served as food source for two groups of 25 individually reared female larvae per feeding scenario. The 

food was replaced in the rearing containers at 2-d intervals and budworm growth and development were recorded for each 

instar separately. Insect development was monitored 24 hours a day every 6 hours and every 15 min during the moult. One 

hour after molting from one instar to the next, before feeding was resumed, each larvae was weighted along with he newly 

molted larval skin and then transferred to a new twig. Pupae were weighed 8 hours after pupation. Normality and variance 

homogeneity tests were performed before data were subjected to a two-stages nested multivariate analysis of variance with 

individual trees nested within feeding scenarios. 

87

RESULTS 

Field Rearing Experiment 

The production of staminate flox_ers significantly affected spruce budworm larval biology (9: Wilk's lambda F[2,171 

= 18.3, p = 0.0(_1, if: Wilk's lambda F[2,171 = 13. l, p = 0.0004) (Fig. 1, 2). Larval biolog 5 also significantly xaried 

according to crown sections (9: Wilk's lambda F[2,17] = 458.6, p = 0.0001 ; Of: Wilk's lambda F:[2,171= 331.8, p = 0.0001). 

However, the interaction between flo_vering class and crown section was significant (9: Wilk's lambda F[2,17] = 329.3, p = 

0.0(X)I o": Wilk's lambda F[2,17] = 549.9, p = 0.0001). This significant interaction results from the fact that floxvering 

affected budworm biology only in the midcrown of flowering trees where staminate flowers were available for consumption 

(Figs. 1, 2). 

F F NF NF 

o_ a a 

E loo ab 

,,,,,,.-1 

..c: 

.__ 80 a 

.c 

0"J 

a 

- 

) 

7 

a 

o 

.._ 

Ck 

60 

-'i 

[1.. 40 

2o 

o ,,,,, ,i =,.,. 

Figure 1.--Pupal weight of male and female spruce budworm reared in the midcrown and the lower sections (*) of flowering 

(F) and nonflowering (NF) balsam fir trees. Staminate flowers were present only in the midcrown of flowering trees 

(n = 10 trees). For a given sex, means followed by the same letter are not significantly different (p < 0.05). 

Both male and female budworms that had access to staminate flowers in the midcrown of flowering trees exhibited 5 

d shorter development time (9: F[1,18] = 32.7, p = 0.0001; _: F[1,18] = 27. l, p = 0.0001) and 9_'_(9) to 15% (o_) reduced 

pupal weight (9: F[1,18] = 12.3, p = 0.0026; if: F[1,1811= 4.9, p = 0.0397) compared to those located where staminate 

flowers were not available for consumption. However, no significant impact of flowering on budworm survival was detected 

(Chi-square = 75, df = 69, p = 0.28, n = 40) (Fig. 3). 

Insects reared on flowering branches had lower fecundity (F[ 1,18] = 10.8, p = 0.0041) than those reared on nonflowering 

branches (Fig. 4), but they produced individual eggs of similar weight (F[1,8] = 1.8, p = 0.21) (Fig. 5). At the end of 

the 29 week long diapause period, no significant difference was detected in terms of survival of the progeny maintained at 

2°C during their diapausing stage (Chi-square = 132, df = 129, p = 0.41, n = 46) (Fig. 6). 

Flowering branches produced significantly less current-year R_liagethan nonflowering branches (F[1,181 = 4.4, p = 

0.04) (Fig. 7). Although the insects reared on flowering branches removed approximately the same amount of current-year 

foliage as those reared on nonflowering branches (F[1,1811= 1.95, p = 0.18), they back-fed on 1-year old foliage, while the 

others did not (F[l, 18] = 23, p = 0.0(Y01) (Fig. 7). This indicates a possible reduction in the nutritive quality of the currentfoliage 

produced by flowering branches. 

88

2,500 

F 

000 

F NF NF 

b _7 a _;_ a _7 a 

E 

EL 

i • 

__0 

¢_1,000 

i 1 ; ; i 

• ! 

o i i , 

| , 

I , 

Figure 2.--Development time from post diapausing second instar to moth emergence of male and female spruce budworm 

reared in the midcrown and the lower sections (*) of flowering (F) and nonflowering (NF) balsam fir trees. Stami- 

nate flowers were present only in the midcrown of flowering trees (n = 10 trees). For a given sex, means followed by 

the same letter are not significantly different (p < 0.05). 

6o 

a 

F F NF NF 

t_ 

_o _ _ _ 

20 : 

lO [ EE 

o.....,....... _/_ ................... _, .....Z_ , I , b,- 

Figure 3.mPercent survival of spruce budworm reared in the midcrown and the lower sections (*) of flowering (F) and 

nonflowering (NF) balsam fir trees. Staminate flowers were present only in the midcrown of flowering trees (n = 10 

trees). Means followed by the same letter are not significantly different (p < 0.05). 

' i 

89

50+ + + 

"o a 

-_ 7 a 

° a _7 

°_ __ 

o b . i ; 

1 . , i 

: :" ! [ 

c I I 

50 i : I 

o , I 

o , / , ,t , J/ , .Z_ 

Figure 4.mFecundity of spruce budworm reared in the midcrown and ihe lower sections (*) of flowering (F) and nonflowering 

(NF) balsam fir trees. Staminate flowers were present onlyin the midcrown of flowering trees (n = 10 trees). 

Means followed by the same letter are not significantlydifferent (p < 0.05). 

E 

0.35 

F F NF NF 

A & & & 

U 

v a 

03 a a ./'---"-2 

0"} 0.25 L I 

_o_ / / 

._> 

"O t'- 

.m 

N--. 

o o.15 i 

..c:: ! 

.m03 

E 

(D E 

o.1 iI ¢) t 

0.05 f F 

J 

o , / , Z , _ , / 

Figure5.--Mean weightof individual eggslaid by sprucebudwormrearedin themidcrownandthelowersections(*) of 

flowering (F) and nonflowering (NF) balsam fir trees. Staminateflowers werepresent only in the midcrown of 

flowering trees (n = 10 trees). Means followed by the same letter are not significantly different (p < 0.05). 

90 

i

8O 

D 

09 i 

4,'.) 

F AS4L 

a 

Z___7 

2o ! 

i 

, 

.i J J 

Figure 6.--Percent survival at the end of the diapause period of the progeny of spruce budworm reared in the midcrown and 

the lower sections (*) of flowering (F) and nonflowering (NF) balsam fir trees. Staminate flowers were present only 

in the midcrown of flowering trees (n = I0 trees). Means followed by the same letter are not significantly different (p 

< 0.05). 

20 

.+.,a 

x2 a 

-cJ 

o? 

.N 

-6 b 

LL a 

5 a 

a 

I 

0 ! ! ! .... - ...... 

D defoliation 

D foliage production 

backfeeding 

Figure 7._Weight of current-year foliage produced by flowering (F) and nonflowering (NF) balsam fir trees at two crown 

levels (*), and weight of current-year and l-year old (backfeeding) foliage consumed by spruce budworm reared on 

flowering and nonflowering balsam fir trees at two crown levels (*). Staminate tlowers were present only in the 

midcrown of flowering trees (n = 10 trees). For a given parameter, means followed by the same letter are not 

significantly different (p < 0.05). 

I 

91

Laboratory Rearing Experiment 

P,esults from the laboratory rearing experiment indicated that the various feeding scenarios tested in this study 

significantly affected female spruce budwonn growth (Fig. 8) and development (Fig. 9) (Wilk's lambda F[49,15] = 4.7, p = 

().(}012). 

2OO 

st'5 [_ st'6 [_ pupa 

''O ti g 

D3 

E 

_ loo diE]7 t _ ;_ 

[i ........ _ j I ! 

5O 

_[ 

, 

,r 

i ! 

[ 

i 

' 

t 

; I 

[ 

i 

', 

] I 

l 

! 

a 

! 

a , ] a :/ a /_ I a / a / 

b _ _ b / b / t i ,/ / --"- 

' " I' // /" 

..__i/ i.J=. ,, / ,,a/,, IJa _a...g..,Z_i / 

Figure 8.---Female spruce budworm weight per instar according to various feeding scenarios using flesh food from flowering 

(F) and from nonflowering (NF) balsam fir trees. Two trees served as food sources for two groups of 25 individually 

reared larvae per feeding scenario (T = 12°C, RH = 65%, 18L:6D, + larvae fed on current-year foliage from the given 

crown section, - larvae fed on current-year foliage prior to reaching sixth instar and l-year old foliage thereafter, * 

larvae fed on current-year foliage from the midcrown prior to reaching sixth instar and current-year foliage from the 

lower"crown section thereafter). For a given instar, means followed by the same letter are not significantly different 

(p < 0.05). 

Insects fed on pollen during their early stages of development were smaller when they reached their sixth instar than 

those that did not have access to staminate flowers (Fig. 8). However, they had similar pupal weight as those that did not eat 

pollen but had access to current-year foliage during their whole development. The presence of pollen in the insect diet 

caused a significant 5 d reduction in development time prior to reaching sixth instar, and a significant 3 d reduction during 

pupal stage (Fig. 9). Although backfeeding during the insect sixth instar did not significantly affect development time, it 

reduced the pupal weight of larvae fed on food from flowering and nonflowering trees by 32% and 28% respectively (Fig. 8). 

On the other hand, dispersal from the midcrown to the lower crown section, when larvae reached sixth instar, did not significandy 

affect pupal weight. However, this feeding scenario resulted in a 6 d increase in the development time from sixth 

instar to moth emergence (Fig. 9). 

92 

I

2,500 _ st2-st4 El st5 0 st6 [_ pupa 

,ooo 

"-" b b b bc 

0.> --m 7 bc _ "I"--7 .----. 

:_ 1,5oo d d _ _..a..4 

@ 7 ...... -8,,'q /-'-7 _ .....-," ...... l".b',.,_ ...... a.," an',._ 

E ....._'1 ' "t '\,a\. : x ._ 

_9_O >: .,. _t ._ - - i ._-.I "

development time that they obtained from feeding pollen during their early stages of developrnent. On the other 

tland, when old larvae dispersed to the lower crown section, they avoided the negative effects of backfeeding, but they lost 

the advantage in development time that they obtained from feeding on pollen. Results from this study indicated that the 

production of staminate flowers by balsam fir trees could have opposite effects on spruce budworm population dynamics 

depending upon the insect population density when flowering occurs. 


We wish to thank M. Charest, M. Crdpin, D. Verge, G. Moreau, J.F. Demers, S. Perreault, C. Paquet, I. Cord, J. 

Forest, L. Roy, and M. Marin for technical assistance during insect rearings. Insects were supplied by the Insect Production 

Unit of the Forest Pest Management Institute, Sault Ste. Marie, Ontario, Canada. Financial support came from the Natural 

Sciences and Engineering Research Council of Canada. 


BATZER, H.O. and JENNtNG, DT. 1980. Numerical analysis of a jack pine budworm outbreak .in various densities of jack 

pine. Environ. Entomol. 9: 514-524. 

BAUCE, E. 1986. I_tude des variations de teneur en fibre brute du feuillage de sapin baumier, Abies bals'ames (L.) Mill., 

induites par diffErents facteurs de stress et leurs implications sur la tordeuse des bourgeons de l'dpinette, 

ChoristoHeurafumiferana (Clem.). M.Sc, thesis, Universit6 Laval, Ste-Foy (Quebec), Canada, 54 p. 

BAUCE, E., CREPIN, M., and CARISEY, N. 1994. Spruce budworm growth, development and food utilization on young 

and old balsam fir trees. Oecologia (in press). 

BLAIS, JR. 1952. The relationship of the spruce budworm to the flowering condition of balsam fir. Can. J. Zoot. 30:1-29. 

BELANGER, L., DUCRUC, J.R, and PINEAU, M 1983. Analyses et commentaires. Proposition d'uhne mdthodologie 

d'inventaire dcologique adaptde au territoire forestier pdriurbain. Nat. Can. 110: 459-476. 

GRAHAM, S.A. 1935. The spruce budworm on Michigan pine. University of Michigan, School of Forest Conservation, 

Bull No 6. 

GRANDTNER, MM. 1966. La vdgdtation fbrestiare du Qudbec mdridional. Universitd Laval, Ste-Foy (Qudbec), Canada. 

GREENBANK, D.O. 1963. S taminate flowers and the spruce budworm. Mem. Entomol. Soc. Can. 31: 202-218. 

JAYNES, H.A. and SPEERS, C.F. 1949. Biological and ecological studies of spruce budworm. J. Econ. Entomol. 42: 221- 

225. 

LEJEUNE, R.R. 1950. The effect of jack pine staminate flowers on the size of larvae of the jack pine budworm. Can. Ent. 

82: 34-43. 

LEJEUNE, R.R. and BLACK, W.E 1950. Populations of larvae of the jack pine budworm. For. Chron. 26: 152-156. 

MATTSON, W.J., HAACK, R.A., LAWRENCE, R.K., and SLOCUM, S.S. 1991. Considering the nutritional ecology of 

spruce budworm in its management. For. Ecol. Manage. 39: 183-2110. 

MILLER, C.A. 1957. A technique for estimating the fecundity of natural populations of spruce budworm. Can. J. Zool. 35: 

1-13. 

SAS INSTITUTE. 1988. SAS user's guide; statistics. 1988. SAS Institute Inc., Cary, NC. 

95

96 

SLANSKY, E, Jr. 1990. Insect nutritional ecology as a basis for studying host plant resistance. Florida Entomol. 73" 360- 

378. 

VOLNEY, W.J.A. 1988. Analysis of historic jack pine budworm in the Prairie province of Canada. Can. J. For. Res. 18: 

1152-1158. 

WELLINGTON, W.G. 1950. Effects of radiation on the temperatures of insectan habitats. Sci. Agr. 30: 209-234.

STAMINATE FLOWERING AND TREE PHENOLOGY AFFECT THE 

PERFORMANCE OF THE SPRUCE BUDWORM 

WILLIAM J. MATTSON, BRUCE A. BIRR, and ROBERT K. LAWRENCE 

USDA Forest Service, North Central Forest Experiment Station 

Pesticide Research Center, Michigan State University, E. Lansing, M148824, USA 

INTRODUCTION 

The important role of staminate flowers in the ecology and population dynamics of spruce budworm has been the 

subject of speculation for more than 50 years (Blais 1952, Greenbank 1963). This suspicion has been fueled by the fact that 

high populations of budworm have historically been associated with mature, flower-bearing trees (Mott 1963). Moreover, 

overwintering budworm larvae typically spin their silk shelter in the old staminate flower bracts, and spring emerging 2nd 

instars prefer to mine and feed in staminate flower buds rather than old needles. Budworms emerge 1 to 4 weeks before 

vegetative buds have begun to expand and are suitable for attack, whereas flower buds, when present, are always more 

advanced and thus better synchronized with the early, vernal budworms. Hence, second-stage larvae are invariably concentrated 

in the host's expanding flower buds. On the other hand, whenever budworm emergence is retarded relative to tree 

phenology, then larvae tend to go directly to expanding foliage rather than the staminate flower buds (Greenbank 1963). 

Obviously, then, the importance of flower buds to budworms may be linked to the phenology of the hosts' vegetative buds 

relative to the timing of spring emergence of the budworms. 

Larvae that feed within staminate flower clusters may actually grow faster than foliage-feeding larvae because such 

clusters are miniature greenhouses, trapping more radiant energy than comparable pure foliage environments. However, 

there is little other experimental evidence to support the hypothesis that staminate flowers can enhance budworm performance 

and thus contribute to outbreak development. For example, Jaynes and Speers (1949) and Blais (1952) reported that 

there is no difference in fecundity of budworms having fed on pollen or new foliage. 

Herms and Mattson (1992) hypothesized that plant reproduction can also indirectly affect herbivore performance 

through its strong impact on the allocation of plant nutrients and photosynthates. When plants produce large crops of pollen 

and/or seeds, the reproductive organs may preferentially receive scarce nutrients and energy that might otherwise have gone 

to plant defenses or even new growth on which phytophages depend. Hence, staminate flowering could indirectly impact 

budworms by reducing the concentrations of nutrients and perhaps even altering the levels of secondary metabolites in plant 

foliage. 

This study was undertaken to simulate and explore the consequences of abundant staminate flower production on the 

performance of pre-outbreak spruce budworm populations. We hypothesized that budworm larval performance would 

depend not just on the availability of staminate flower but also on the phenology of a tree's vegetative buds. We predicted 

that the presence of staminate flowers would be most important on very late flushing trees, and least important on very early 

flushing trees. 

METHODS 

In 1989 and 1991 we selected 20 half-sibling balsam fir, Abies balsamea, that were flowering at the Kellogg Experi- 

mental Forest (Michigan State University) near Augusta, Michigan. Trees were planted in 1970 and about 6-8 m tall at the 

time of the experiments. Before budbreak (April 21, 1989, and April 11, 1991), we selected per tree two flowering and two 

nonflowering branches near midcrown, which were then enclosed with a fine-mesh, cloth sleeve cage wherein we placed 

about 20 ready to emerge second instar budworm larvae still in their silken hibernacula on a gauze patch. Exactly 1 week 

Mattson, W.J., Niemel_i, P.,and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


97

later, we selected one more flowering and nonflowering branch on each tree which we once again enclosed with sleeve cages 

arid ir_;e_ted about 20 second-stage budworms. These two batches came from the same population, and they were randomly 

allocated to treatments. Hereafter, we respectively refer to them as cohort 1 and cohort 2. Approximately 2 weeks later we 

removed the gauze patches and counted the number of larvae which failed to emerge from their overwintering hibernacula. 

This gave us the initial number of insects in each sleeve cage. When most insects had reached the pupal stage, cages were 

removed by cutting each branch at its base. We then searched each bag for budworm larvae and pupae. Pupae were stored 

immediate[y in plastic vials and checked daily until all adults emerged. Moths were first frozen, then oven dried to constant 

weight and weighed and sexed. Because male and females responded in the same way to treatments, we pooled sexes by first 

converting males to female equivalents. Each male observation was multiplied by a constant (k) that was derived from the 

grand meaas (gdm) of the particular variable in question (e.g., weight, developmental time, developmental rate, etc): k_= gdm 

fcmate/gdm male, for each cohort and year. We calculated survival per sleeve cage by summing all larvae that achieved the 

final 6th larval stage (even though some were parasitized or preyed upon by stink bugs), pupae, and adults, and divided this 

grand total by the mm_berof larvae that actually emerged from their hibernacula. 

In 1991, we selected 32 (16 flowering, 16 nonflowering) white spruce trees, Picea glauca, that were growing at the 

edge of a provenance plantation (USDA North Central Forest Experiment Station) near Wellston, Michigan. Trees were 

planted in 1963, and were approximately 8-10 m tall at the time of the experiment. Just as above we selected (on May 3) two 

branci_es at midcrown level from each the flowering and the nonflowering trees for enclosure with sleeve cages and budworms. 

We placed only one cohort of budworm larvae on these trees, however. We collected and counted surviving budworms 

when most had reached the pupal stage. 

Aft trees were scored for their phenological development according to the protocol of Nienstaedt and King (l 970) at 

the time of placing insects on the trees. 

Difference between cohorts 1 and 2 in the balsam fir study were first tested via a 2 x 2 factorial, randomized block 

design where trees were treated as blocks and flowers and cohort timing were treated as treatments. Next each cohort was 

analyzed separately to test specifically for flower by tree phenology effects on budworm performance (weight gain, growth 

rate/day, days to complete development, and survival to the pupal stage). We subjected the data to univariate anova after 

appropriate transflmnations of the data (i.e., arcsin (%survival)). The balsam fir study was analyzed as a completely randomized, 

spiit,..piot design with phenology (3 classes) treated as the main plot and flowering (2 levels) as the subplot. The white 

spruce study was analyzed as a completely randomized 3 x 2 (phenology x flowering) factorial. 

Larval Survival 

RESULTS 

Balsam Fir: Flowering and Phenology Effects on Budworm 

First cohort survival was clearly affected by both flowering (F) and flowering x tree phenology (P) classes (Table 1). 

[n 1989, the significant F x P interaction was due largely to the enhancing effect of flowering, as expected, on the Sateflushing 

trees, where survival averaged 65.3% with, and 43.2% without flowers (Fig. 1). The smallest flowering effect 

occurred on early flushing tree (57.0% vs 53.6%, with and without flowers, respectively). In 1991, even though there were 

no significant main or interactio_ effects (Table 2),the largest positive impact of flowering was evident on the late flushing 

trees (as in 1989) where survival averaged 48.2% with, and 40.6% without flowers (Fig. 1). 

Second cohort survival in 1989 was significantly (p < 0.02) enhanced by flowering but not affected by phenology, nor 

by the F x P interaction (Table 1). In 1991, there were again significant flower, and F x Perfects (Table 2). This was due 

primarily to the huge difference in survival rates between flowering (54.1%) and nonflowering (32.0%) branches on late 

flushing trees (Fig. I). 

Testing the grand means for cohorts I and 2 in 11989and 1991 revealed in both cases that survival was higher for the 

first cohorts, but only statistically significant in 1989 (55% vs 43%). 

98

Table 1.--Mean spruce budworm performance (female equivalents) on balsam fir in 1989 on early, middle and late 

phenology trees (main plots) and flowering and nonflowering branches (subplots). Means varying significantly 

(p < 0.05) among main effects have different letters. For cohort one, mean survival was effected by flowering 

and by,the flowering x phenology interaction. For cohort two, mean survival was affected by flowering. 

Budworm performance Trees (main Elots)--Phenolo_ class Subplot--Flowerin i class 

variable Early Middle Late Yes No 

Cohort one 

Survival (larvae) 0.55 0.57 0.54 0.62a 0.49b 

Weight dwt (rag) 23.93 22.75 22.44 23.18 22.72 

Dev. time (days) 52.94 53.37 52.93 52.91 53.34 

Growth rate (mg/da) 0.45 0.43 0.43 0.44 0.43 

Cohort two 

Survival (larvae) 0.43 0.42 0.45 0.47a 0.39b 

Weight dwt (rag) 26.11 23.58 24.23 24.53 24.1 

Dev. time (days) 50.45 51.27 50.88 51. | 1 50.84 


0.70 _- 1989 

0.65 _ _,ow_.,[] N_,, 

0.60 

t Cohort 1 

Figure 1.---Survival rates of two "_ ,//_ 

cohorts of spruce budworm _- #/'A 

larvae in 1989 (upper panel) ._ 0.50 /)///_ 

and 1991 (lower panel) on ?' :.:;,:.:,:'.:il 

flowering and nonflowering _ 0.4.5 ',:,::'.,.::,:.:_ 

branches on 20 balsam fir trees ] 

classes: early, mid and late 0.35 

flushing at the Kellogg 

divided into three phenology 

Experimental Forest of 

o.ao f 

0.30 -- 

_i1 

Michigan State University in 

Augusta, Michigan. Flowering 0.55 l 

0.55f i] .. ,, oo o,, 

effects were statistically E2:I_,o_,,,,,_ s_,_ cohorta 

significant (p < 0.05) in one or 0.50 

both cohorts of both years 

(Tables 1 and 2). 

:_',_:!::: 

¢ Cohort 

and flowering x phenology t| 1981 .:,,:.:,',.':.':,_,_i! 

o.4o 

0.35 

0.30 __ __dIZ 

early- I mid- 1 late- 1 early-2 mid-2 late-2 

Tree Dhenolo_ty & budworm cohort classes 

L/// 

V// 

V'// 

99

Table 2.---Mean spruce budworm performance (female equivalents) on balsam fir in 1991 on early, middle and late phenology 

trees (main plots) and flowering and nonflowering branches (subplots). Means varying significantly (p < 0.05) 

among main effects have different letters. For cohort one, mean weight and growth rate were significantly affected 

by phenology, and mean development time by the f x p interaction. For cohort 2, only mean survival was significantly 

affected by flowering, and by the flowering x phenology interaction. 

Budworm performance ..... Trees, (main plots)--Phenology class SubplotmFlowerin_g "class 



Survival (larvae) 0.41 0.41 0.44 0.42 0.42 

Weight dwt (rag) 23.41b 26.47a 25.72a 24.79 25.00 

Dev. time (days) 51.29 50.97 51.05 51.05 51.23 

Growth rate (mg/da) 0.46b 0.52a 0.50a 0.49 0.49 

Cohort two 

Survival (larvae) 0.45 0.47 0.43 0.47a 0.4 l b 

Weight dwt (rag) 24.37 25.27 25.53 24.51 25.44 

Dev. time (days) 46.02 46.16 45.65 45.92 45.95 


Larval Growth 

There were no significant flower or F x P effects on weight gain (mg dwt) by the first or second budworm cohort in 

either 1989 or 1991 (Tables 1,2). However, there was a significant phenology effect on cohort 1 in 1991, where middle and 

late flushing trees yielded bigger adults than early flushers (26.5 vs 23.4 mg dwt). 

Likewise, there were no significant flowering, or F x P effects on budworm growth rates (mg/da) for either cohorts in 

1989, and 1991 (Table 1, 2). However, there was a significant phenology effect on the growth rate of the first 1991 cohort. 

Budworms on middle and late trees grew faster than those on early flushing trees (0.52 mg/da vs 0.46 mg/da). There was 

also a significant F x P effect on development time for cohort one (Fig. 2), but this may have been spurious. 

Comparing the grand means for cohorts 1 and 2 in 1989revealed that growth (22.95 vs 24.31 mg) was significantly 

higher and development shorter (53.13 vs 50.97 da) for the second than first cohort. By contrast, in 1991, there were no 

significant differences in overall growth between the two cohorts, but development time was, as before, much shorter for the 

second (51.14 vs 45.93 da), causing its growth rates to be significantly higher (0.49 vs 0.54 mg/da). 

Survival 

White Spruce: Flowering and Phenology Effects on Budworm 

Flowering and phenology had no significant effects on survival of larvae (Table 3), but the F x P interaction was 

nearly, significant (p < 0.10). There was an apparent trend for survival to increase with increasingly retarded phenology on 

nonflowering trees, and the opposite trend on flowering trees (Fig. 2). The fact that the two trends run counter to each 

explains why the main effects of flowering and phenology were insignificant. The grand means from flowering and nonflowering 

trees in each phenology class canceled one another out. 

Larval Growth 

There were strong, consistent phenology effects on both total weight gain and growth rates, the pattern being identical 

on both flowering and nonflowering trees (Table 3). Budworm mass and average growth per day increased with increasingly 

retarded tree phenology, Insects on the late flushing trees averaged 13.5% larger than those from the early flushing trees 

(32.5 mg vs 28.6 rag). Likewise their daily growth rates were 14.5% higher (80 mg/da vs 70 mg/da). Neither flowering, nor 

t00

52,00 

51,75 - 

51,50 - 

51.25 

50.75 _/// 

50.00 

0,70 

0.65 

0.60 

0.55 

OA5 

0.50I 

0.40 _- 

0.35 

38 

_!ow_rs _ None 

36 - _Z] _=lower__ No

Table 3..--Mean spruce budworm performance (female equivalents) on white spruce in 1991 on early, middle, and late 

phenology trees and flowering and nonflowering trees. Means varying significantly (p < 0.05) due to main effects 

have different letters. Only mean weight and development time were affected by phenology. There were no flowering 

effects. 

Budworm performance Tree phenol0_y class Tree flowering, class 



Survival (larvae) 0.57 0.57 0.55 0.57 0.57 

Weight dwt (rag) 28.61b 30.96a 32.48a 30.13 311.33 

Dev. time (days) 41.11 40.93 40.85 40.96 40.95 

Growth rate (mg/da) 0.70b 0.76a 0.80a 0.74 0.77 

F x P had any significant (p < 0.05) effects on growth, development time, or growth rates. Although, there was a hint (p < 

0.08) of a tendency for nonflowering trees to be superior for larval performance than flowering trees (Fig. 2). 

DISCUSSION and CONCLUSIONS 

Yearly Differences in Protocol and Weather Weaken the Tests 

The differences in the effects of flowering and phenology on spruce budworm performance between years is ahnost 

certainly due to the differences in host tree phenology when the studies were initiated. In 1989, we placed our first cohort of 

budworms on the trees at 133 degree days (dd), very close to the time native budworm populations would have been emerging 

(est. at 100 dd). The second cohort was placed out at 204 dd, 71 dd and 7 days later than the first. In 1991 the experiment 

was begun 10 calendar days earlier than in 1989 but at a phenologically later point, 210 dd, about the time of budbreak 

on the early flushing trees (Nienstaedt and King 1970). Hence, the first cohort in 1991 was more nearly equivalent to the 

second cohort in 1989. The second cohort on fir in 1991 was placed on the host plants at 262 dd thereby having no 1989 

equivalents. Finally, the single cohort on spruce in 1991 was placed on the trees at 280 degree days, about the time of 

budbreak for the later flushing spruce trees (Nienstaedt and King 1970). In addition, in 1991, just after placing the first 

cohort of second instars on the trees, the weather turned cool and wet keeping the young insects in their overwintering 

hibernaculum until nearly 1 week later when the next cohort was being placed on the trees. 

Although this study attempted to examine the combined effects of staminate flowering by host phenology on the 

performance of spruce budworm, we were unable to execute the experiment in perfect phenological duplication that would 

have allowed the most powerful tests of the hypotheses. 

Flower by Phenology Effects 

, The data clearly suggest that the survival of budworms is dependent on the flowering x phenology interaction of its 

host plants. In the case of fir, abundant flowering probably enhances survival most significantly on the later flushing trees, 

i.e., late relative to budworm emergence (they inevitably emerge before their hosts break bud). This was apparent especially 

in 1989. In that year all flowering branches supported higher survival than nonflowering branches regardless of tree phenology 

class, and regardless of cohort timing. In 1991 there was no apparent flowering effect on survival except for the late 

flushing trees, especially in the case of the second cohort. Likewise, on spruce there was a strong tendency, though only 

nearly significant (p

There was practically no evidence to suggest that flowering enhances the growth and development rate of budworms, 

not on fir, nor on spruce. Likewise, there seems to be no strong evidence that there is any kind of F x P effect on growth 

processes. To the contrary, the data on spruce, though not quite statistically significant (p

FOLIVORE FEEDING ON MALE CONIFER FLOWERS ° 

DEFENCE AVOIDANCE OR BET-HEDGING? 

THOMAS SECHER JENSEN 

Institute of Biological Sciences, Department of Zoology 

Aarhus University, DK-8000 Aarhus C, Denmark 

INTRODUCTION 

In the fifties and sixties a group of British researchers led by G.C.Varley made pioneering studies on the life history 

of the winter moth (Varley and Gradwell 1970). Among many other things, they concluded that the key to understanding the 

main mortality in this species was the timing between egg hatch and bud-burst. Larvae which hatch on trees, having buds 

firmly closed, fail to find food there. Instead such larvae disperse by spinning a thread of silk, and are blown away to an 

uncertain destiny. Larvae that hatch when buds have just burst, survive well. This first model was further developed by 

Feeny (1970) who in another pioneering study argued further that larvae which hatch after bud-burst face certain difficulties, 

this time in relation to nutritional conditions, mainly decreased nitrogen and water, and increased tannin concentrations. 

In summary these factors indicate that in Quercus robur, there is a "phenological window" represented by recently 

burst, highly nutritious and poorly-defended leaves. In oak many defoliators, especially lepidopteran species, take advantage 

of this window and often reach high densities, even outbreak levels, at that time. Also species richness is highest in the early 

spring. However, this situation in Quercus robur does not seem to apply in a number of other tree species. The period of 

shoot growth in oak is rather short, and Niemel/i and Haukioja (1982) demonstrated that in many other deciduous trees the 

period of shoot growth is much longer and seems to give defoliators a chance to find a much broader phenological window. 

In this paper the nutritional status of two coniferous genera, pine, Pinus, and spruce, Picea, are considered in relation 

to the phenological window, including the situation when alternative food, especially flowering buds are available. The 

insect species studied has mainly been the nun moth, Lymantria monacha, a close relative of the gypsy moth. The nun moth 

is a polyphagous species that often defoliates large areas of conifer forest in Central Europe (Jensen 1991). 

METHODS 

Rearings and experiments of nun moth larvae and other insects were performed in a laboratory at constant 20°C. 

Eggs were collected in the forest and brought to the laboratory during winter. In growth and survival experiments nun moths 

were kept in batches of ten in the first three instars; fourth and fifth instars were reared singly. Larvae were fed needles and 

flowers attached to excised twigs put in water. 

Chemical analysis were performed by means of micro-Kjeldahl analysis (nitrogen) and gas-chromatography (carbohydrates, 

cyclites and phenolic acids). For details see Jensen (I988). 

RESULTS 

Spruce 

The genus Picea comprises about 30 species in the northern hemisphere. In parts of Europe, Norway Spruce, Picea 

abies, is the native species, but it is also widely planted, along with the introduced North American species Sitka spruce, 

Picea sitchensis. The buds burst rather early in spring along with a lot of deciduous trees, and the new shoots and needles are 

Mattson, W.J., Niemel_i, R, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


104

soft for quite some time. It has 2-5 year classes of older needles. Nutritional content of the new needles is high: total N 

reaches 4.5% dwt, carbohydrate levels 250 nanomoles/g dwt (Jensen 1991). When needles get older, nitrogen levels drop to 

around 1%, whereas carbohydrate levels increase somewhat (Table 1). 

Table 1.--Concentrations of nutrients and secondary compounds in needles and male flowers of spruce, Picea abies. 

Total Amino Shikimic Quinic Pinitol Hexoses Inositol Cathecin Total 

Plant .part Nitrogen acids acid acid Phenolics 

% ................... n moles/mg dwt ................... 

Male flowers 5 _900 29 213 134 575 18 4 

Old top needles 1.2 _450 224 52 61 390 7 12 68 

New top needles 4 1,t20 108 625 133 248 17 3 35 

In Europe a substantial number of insect species are flush-feeders on spruce although some of them also consume 

older foliage, especially late in year or at outbreak densities (Table 2). This situation mainly applies to Lepidoptera although 

certain sawfly species also are flush-feeders. Another group feeds exclusively on older needles and is mainly comprised of 

Hymenoptera. In general, most spruce flush-feeders (e.g., L. monacha) perform substantially better when reared on the new, 

nutritious needles than on the old ones. Most newly hatched larvae die when put on old needles probably because their 

mandibles are not strong enough to cope with the hard needles (Jensen 1992). On the other hand, if third instar larvae are put 

on old needles, they survive and perform reasonably well. Preference indices for new needles (Table 3) decrease with instar 

number, indicating that larger larvae do not require young needles. 

Table 2.--Needle age preferences of insect larvae feeding on Norway spruce, Picea abies. 

Needle age class 

Insect species Newly flushed Current Old 

Lepidoptera 

Parasyndemis histrionana * (*) _' 

Eana argentata * * (*) 

Epinotia tedella (*) * 

Epinotia nanana * * (*) 

Epinotia pygmaea * 

Zeiraphera ratzeburgiana * 

Orgya antiqua * * (*) 

Lymantria monacha * * (*) 

Hymenoptera 

Cephalcia abietis * 

Cephalcia arvensis * * 

Gilpinia hercyniae * 

Pristiphora abietina * 

Pristiphora ambigua * 

Pristiphora saxeseni * (*) 

Pachynematus scutellatus * * (*) 

" * indicates unequivocal primary preferences, (*) indicates secondary preference, often by late instars. 

105

Table 3.--Lymantria monacha preference for new needles of spruce, Picea abies. 

Instar Preference index n P 

I 1.00 + 0.00 20 < 0.001 

II 1.00 + 0.00 17 < 0.001 

III 0.84 + 0.32 9 < 0.020 

IV 0.63 + 0.15 21 < 0.020 

V 0.54 + 0.26 21 n.s. 

However, as indicated in Table 2, some species actually prefer the old needles, and to such an extent that they will die 

if fed the tender, nutritious new needles. An example of this is the sawfly, Gilpinia hercyniae, known in North America as 

the European Spruce Sawfly. It is a former pest in Canada, but not in Europe. When offered only new foliage, this defoliato'r 

dies without eating (Jensen 1988). Experiments have shown that one possible deterring compound to G. hercyniae could be 

quinic acid, which together with shikimic acid is responsible for much of the acidity of these new needles. Quinic acid is 

found in huge amounts in the new needles, but in low concentrations in the old ones. When quinic acid was added to old 

(preferred) needles in an experiment (Jensen 1988), G. hercyniae refused to eat after concentrations were increased to the 

levels of new needles. 

With respect to the phenological window, insects like G. hercyniae, feeding on old needles have no problems. In 

contrast, the flush-feeders must find the phenological window. If they hatch very early, even dispersal will not help if the 

food resource is not yet present. The only way to survive is to find an alternative food. In the dense monocultures of 

European spruce forests it is not easy to find other host species. However, occasionally, i.e., with 2-7 year intervals flowering 

buds might offer this alternative. The pollen buds are prime food, being soft, with high nutritional content and rather low 

content of secondary compounds (Table 1). Laboratory experiments on newly emerged L. monacha showed that these larvae 

are very fond of male flowers. When they hatch early and are reared on branches with only old needles, they die. If male 

inflorescences are present they survive and grow well (Fig. 1). Once started on flowers, they can eat new and old needles. 

Later in the season the larvae may struggle for some time until vegetative buds burst. 

0.5 - Weight (g) 

0.3 - "" 

0.4 - o,,o 

0.2 - ,," __" 

0.1 - _o",, ,. " 

0.02 

/ 

0.04 0.00 

_/_/ ,," 

16. May 

o Male flowers ->New needles 

® Male flowers ->Old needles 

Old needles ->Old needles 

Figure 1.--Lymantria monacha growth on male flowers and needles of Norway spruce (Picea abies). 

106

Pine 

tn pine species the new needles emerge rather late in spring. Old needles are tough and of low nutritive value. This 

means that there is a long period with few feeding opportunities for herbivores, except for the existence of male inflorescences. 

In most places, pines set flower buds much more consistently than do spruce. 

New pine needles are less nutritive than spruce, nitrogen reaches only 2-2.5% of dwt. The toxins occurring in new 

needles are potent and it has been demonstrated (Ikeda et al. 1977) that they reduce growth rate, increase mortality, and 

extend larval period. Not surprisingly, most defoliator species feed on old needles or current year needles late in the year 

(Larsson and Tenow 1980, Table 4). 

Table 4.--Needle age preferences of insect larvae feeding on Scots pine, Pinus sylvestris. 

Insect species Needle age class 

Current Old 

Lepidoptera 

Bupalus pM[aria * * 

Cedestis spp * 

Cidaria firmata * 

Dendrolimus pini * * 

Ellopia fasciaria * 

Hyloicus pinastri * 

Panolis flammea * * 

Semiothisa liturata * 

Lymantria monacha * 

Hymenoptera 

Diprion pini * * 

Diprion ximilis * 

Gilpinia frutetorum * 

Microdiprion pallipes * * 

Neodiprion sertifer * 

Nun moth larvae normally hatch prior to bud-burst, leaving them with only old pine needles for food. Experiments in 

field enclosures and in the laboratory with first instar larvae on whole branches or twigs, showed almost 100% mortality. 

Only if the bagged branches included male flowers were larvae able to survive. In the laboratory, larvae survive and develop 

only if they are put on flowering branches in the first part of their life, irrespective of pine species (P contorta, P. sylvestris, 

P. mugo) (Table 5). Surviving larvae that later were fed new or old needles, performed equally well (Fig. 2). 

Table 5.--Percent survival of nun moth larvae on pines. T = 17 days, n = 85 per trial. 

P. silvestris P. mugo P. contorta 

Flowering 71.0 50.8 14.6 

Non-flowering 15.0 16.7 0.0 

107

500- 

400- 

i 300 

v 

¢ 

m 

a 200- 

I00 

Lymantrlamonacha 

on Pinusmugo 

Nonflowering 

0 2O 40 

Days 

Flo_ering,old 

Figure 2:--,L_ymantria monacha growth on flowering and nonflowering branches of mountain pine (Pinus mugo). On 

flowering branches larvae were later separated into feeding groups on new and old needles.

DISCUSSION 

Based on the above findings it seems necessary to add another set of curves to the Varley-Feeny phenological 

window, curves that represent the opening of male flowering buds (Fig. 3). These curves are fundamentally different in the 

pine and spruce. In spruce, where male flowers burst very early in spring, it can be noted that flush feeders simply get a 

broader phenological window in years with male flowers. This means that larvae which hatch early in such a year get a 

larger probability of finding highly nutritious food. However a strategy of early hatching is a risky business, because chances 

are high that there will be no flowers. On the other hand, if they hatch late the risk is small, but the nutritional quality is 

lower and hence development time and predation risk will be higher. 

,,.__../ 

4-_ 

___c2 

Spr'uce 

,_ s#.._ - ,-..... 

o" egg bud 

...... 

Lf) 

c_ Pi_e - ea_'l7 hatch 

M i 

egg o" bud 

,' /-X,"-"/ , \ 

Cxb 

_ Pine -late hatch m 

r.D 

o c; bud egg 

U9 

" I,'-', 

/'""..../qk 

0 

q _ tlltlfllk 

,' /', /\ lk 

,' ..: ',,/ "..lllltllllllllllk 

, .., / ,, NWIIIIIIIIIIIN 

_ .... / % III ItIi"N,ItlIIIIII IIII1_ 

,..."" ..... / _ IIITTtt,14LLIJIIIIIIII _ 

Figure 3._Modification of the Varley-Gradwell curves for the "phenological window": insect egg hatching, male flower and 

bud opening on spruce and pine. 

LID 

109

From the point of view of trees, the evolutionary response in terms of defense seems to be consistent with a concept 

k_own from deciduous species like oak or beech, as masting or mast seeding (Silvertown 1981). tn other words, in some 

years there will be high beech-mast or acorn production, but no seeds in the years in between. In the latter years, pest 

populations will decrease markedly. The strategy has probably evolved against seed predators but it could also work against 

_omedefoliators. 

The second line of defense is simply the variability of bud burst (Mitcherlich and Wellenstein 1942). Large variation 

in bud burst between provenances exist, but also large variation occurs within provenances. However, variability as a defense 

will mean that some larvae always will survive. 

The third line of defense against flush-feeders is a chemical one. Apparently, rather few spruce herbivores are 

deterred from feeding on newly burst needles to such an extent that they prefer old needles (Schopf 1986). 

In contrast, in pine it seems essential that flush feeders avoid plant defenses and here male flowers are of vitae 

importance. In pine, male flowers are a more predictable resource than in spruce. Larvae that hatch after the flowering 

period will die and hence the risk of early hatch is less because male flowers are most often present. However, the strategy 

might fail; they might hatch on a tree without flowers, or worse in a year without flowers in the whole stand. 

In the context of the phenological window it is actually necessary to consider the choices made by the female moth. 

Where does she deposit her eggs? Does she in fact spread the risk so that at least some of her larvae will survive? But at the 

same _ime,does she abandon an optimal strategy of maximizing short term fitness? Essential to all these considerations are 

_hedevelopment time of buds and eggs, both factors being related to spring temperatures, although probably not in the same 

fashioa. I hypothesize that the position of the eggs is a key factor in this synchronization simply because the temperature 

varies so much between microsites. 

1#mantria monacha females deposit their egg clusters mainly on the tree trunk under bark scales. Egg numbers and 

egg hatch differ consideraNy between north- and south-facing sides of the trunk (Raae 1979), and it might be that this is the 

mechanism the females employ when they spread the risk. This could be the way whereby females ensure that some of their 

progeny in this variable environment will hatch when the phenological window is open. Another bet-hedging mechanism 

couid be that after laying their first batch and becoming lighter, females fly to another tree and deposit the rest of their eggs. 

If the first tree is in the middle of a dense stand and the next is at the southern edge, the temperature regime is different and 

accordingly the bets change. 

litis interesting to observe that many flush feeders have wingless females or have reduced flight capability due to 

their heavy loads of eggs. In L monacha, females are winged but poor fliers and it is likely that the females lay their eggs 

close to if not directly on the tree where they pupated, thereby increasing the likelihood of synchronization between the egg 

i_atchand bud burst. 

In conclusion, avoiding defenses of new needles is absolutely necessary for pine defoliators but of minor importance 

_o spruce de_bliators. Bet hedging might prove to be the evolutionary response by the defoliators in order to find the host's 

phenological window. 


DEN BOER, RJ. 1968. Spreading of risk and stabilization of animal numbers. Acta Biotheoretica 18:165-194. 

FEEENY, P 1970. Seasonal changes in oak leaf tannins and nutrients as a cause of spring feeding by winter moth caterpillars. 

Ecology 51' 565-582. 

KEDA, T., MATSUMURA, E, and BENJAMIN, D.M. 1977. Chemical basis for feeding adaptation of pine sawflies, 

,_e_:dlpr_on rug(frorls and Neodiprion swainei. Science 197: 497-498. 

JENSEN, T.S. 1988. Variability of Norway spruce needles; performance of spruce sawflies (Gilpinia hercyniae). Oecologia 

77:313-32(t.

JENSEN, T.S. 199 la. Integrated pest management of the nun moth, Lymatltria monacha (Lepidoptera: Lymantriidae) in 

Denmark. For. Ecol. Mgmt. 39: 29-34. 

JENSEN, T.S. t99 lb. Patterns of nutrient utilization in the needle feeding guild. In Baranchikov et at., eds. Forest insect 

guilds: Patterns of interaction with host trees. GTR-NE-153. Radnor. PA: U.S. Department of Agriculture, Forest 

Service. 

JENSEN, T.S. 1992. The role of plant stress in the population dynamics of nun moths. IUFRO Conference Zakopane, 

Poland 1991. 

LARSSON, S. and TENOW, O. 1980. Needle-eating insects and grazing dynamics in a mature Scots pine forest in Central 

Sweden. In Persson, T., ed. Structure and function of northern coniferous forests: an ecosystem study. Ecol. Bull. 32: 

269-306. 

MITCHERLICH, H. and WELLENSTEIN, G. 1942. Die Nonne an Frtih- und Sp_ittreiberformen der Fichte. Mon. Z. ang. 

Ent. 15: 94-125. 

NIEMEL,_, R and HAUKtOJA, E. 11982. Seasonal patterns in species richness of herbivores: Macrolepidopteran larvae on 

Finnish deciduous trees. Ecol. Entomol. 7: 169-175. 

PHILIPPI, T. and SEGER, J. 1989. Hedging one's evolutionary bets, revisited. TREE 4:41:44. 

RAAE, K. 1979. The distribution of nun moth eggs in old spruce stands. Unpubl. rep. Agricultural Univ., Copenhagen. 

SCHONHERR, J. 1989. Outbreak characteristics of Lymantriids. In Wallner, W.E. and McManus, K.A., ed. Lymantriidae: 

A comparison of features of new and old world tussock moths. GTR-NE-123. Radnor, PA: U.S. Department of 

Agriculture, Forest Service. 

SCHOPE R. 1986. The effect of secondary needle compounds on the development of phytophagous insects. For. Ecol. 

Man. 15: 54-64. 

SEGER, J. and BROCKMANN, H.J. 1987. What is bet-hedging? Oxford Surv. Evol. Biol. 4: 182-211. 

SILWERTOWN, J.W. 1981. The evolutionary ecology of mast seeding in trees. Biol. J. Linn. Soc. 14: 235-240. 

VARLEY, G.C. and GRADWELL, G.R. 1970. Recent advances in insect population dynamics. Ann. Rev. Ent. 15: 1-24. 

111

THE BLACKMARGINED APHID AS A KEYSTONE SPECIES: 

A PREDATOR ATTRACTOR REDRESSING NATURAL ENEMY 

IMBALANCES IN PECAN SYSTEMS 

M.K. HARRIS and T. LI 

Department of Entomology, Texas A&M University, College Station, Texas 77843, USA 

INTRODUCTION 

Plants defend themselves against herbivores in many ways. Conceptualization of these defenses has resulted in many 

terms and contexts to describe them (see Harris and Frederiksen 1984). Snelling (1941) listed 15 categories. Painter (1951, 

Fig. 3) depicted 11 plant factors interacting with 8 insect factors mitigated by 8 environmental factors influencing 5 insectplant 

interaction factors as possible causes of resistance to insects, and then consolidated these factors into three resistance 

mechanisms: tolerance, antibiosis, and preference. These mechanisms were, by definition, heritable, effective in isolation, 

and particularly apt for usage in agriculture where producing a crop of good quality in the presence of the insect was paramount. 

Harris (1980) proposed a succinct and natural characterization of plant defense, listing escape in space and time, 

accommodation, confrontation, and biological associations as the primary mechanisms to consider. 

Biological Associations 

Of these, biological associations are the most complex and difficult to demonstrate as natural defense mechanisms, 

because both the genetics of the plant and the arthropod are involved. Perhaps the best known case of a natural biological 

association providing plants with a defense against herbivores is the ant-acacias. Howe and Westley (1988) review this and 

other lesser known ant-plant associations, as well as other biological associations including those important in dispersal, 

pollination, etc. Biological associations mediated by the genetics of the plant clearly play an important role in nature, yet 

they have seldom been deliberately exploited to a significant degree in agriculture. Exceptions could be corn leaf aphid, 

Rhopalosiphum maidis (Fitch) (Homoptera: Aphidae) on corn and sorghum, and apple rust mite Aculus schlechtendali 

(Nalepa) (Acarina: Eriophyidae) on apple, which are relatively innocuous at moderate densities and, if left alone, will often 

support densities of natural enemies that also suppress more pestiferous species of aphids and mites, respectively (Flint and 

van den Bosch 1977). The agricultural importance of corn leaf aphid and apple rust mite is clear, but the origin and role of 

these biological associations in natural systems are presently unknown. From an agricultural perspective, this area combines 

aspects of biological control and host plant resistancentwo subdisciplines of entomology that lack a common paradigm. 

BLACKMARGINED APHID AND PECAN SYSTEMS 

This paper proposes another example of a biological association, viz. how pecan, Carya illinoensis (Wang.) K. Koch 

(Juglandaceae), interacts with the blackmargined aphid, Monellia caryella (Fitch) (Homoptera: Aphidae), to influence other 

biological associations that may impact plant defense within the context of the matrices of interactions noted above. 

Pecan is a deciduous, monoecious, wind-pollinated, woody perennial that occupies alluvial soils from western Texas 

to the Mississippi Valley on the east and from southern Illinois into Mexico (Little 1971). Pecan trees can live more than 200 

years, grow to heights exceeding 35 m, and can constitute as much as 50% of the tree canopy in their natural habitat (Maggio 

et al. 1991). Reproduction is by seed (=200/kg) abundantly produced every 2-7 years (< 500 kg/ha) in natural populations, 

and each tree in nature is genetically distinct (Harris 1988). Leaves are compound and occur at a density of about 3 million 

leaves or roughly 50,000 m2/ha. 

Mattson, W.J., Niemel/i, E, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


112

Pecan domestication has been a gradual process, initially consisting of thinning riverine woodlands to pure stands of 

pre-existing pecan trees. Wild pecan nuts are commercially attractive and usually preferred by bakers and confectioners 

because they are typically priced about 25% less than selected varieties; they are about 40% smaller, allowing the whole 

halves to be pleasingly presented on small cakes, candies, etc.; and they are comparable, if not superior, in taste to selected 

varieties. More than 60% of current nut production in Texas comes from such trees with the remainder produced by vegetatively 

propagated varieties. Thus, the pecan landscape throughout the indigenous range consists of wild trees growing in 

mixed species woodlands, native pecan that has been thinned from woodlands, and vegetatively propagated orchards often 

found adjacent to one another. Management programs like fertilization, pesticide application, irrigation, and pruning range 

from being most intensive in orchards to virtually nonexistent in mixed species woodlands. This variety of habitats provides 

unique opportunities to examine many questions including effects of plant domestication on arthropods. Most investigations 

of the pecan arthropod complex have been spurred by economic considerations and have concentrated on species that pose 

economic threats or economic relief (Harris 1983). However, given the long, close association of the arthropod complex with 

the pecan and the physical proximity of the wild and cultivated varieties, the opportunity to interpret results of pecan arthropod 

interactions from other perspectives invites attention. The arthropod complex consists of more than a 100 phytophagous, 

predator, and parasite species that have been associated with pecan from the earliest of times (Harris 1983). 

The blackmargined aphid (BMA) is a monophagous, multivoltine, obligatorily alate, phloem feeder that overwinters 

as eggs placed in bark crevices and parthenogenetically infests pecan from shortly after budbreak until the fall some 7-8 

months later, when males are produced, mating occurs, and the overwintering egg population is established (Harris 1983). 

Seasonal phenology on wild trees typically consists of BMA densities initially below 1/leaf increasing to between 1-10 

aphids/leaf for a 2-3 week period during the summer, and then returning to densities below l/leaf for the remainder of the 

season. Orchard trees typically experience _>rivefold higher densities and exhibit BMA population increases earlier in the 

season that last for 3-4 weeks (Liao et al. 1984). 

BLACKMARGINED APHID AS A KEYSTONE SPECIES 

Effects of BMA on nut production appear to be negligible in wild trees (Liao and Harris I984), and with one exception 

(Wood et al. 1987), on orchard trees. However, routine use of insecticide in the early season typically results in epidemics 

of pecan aphids (PA) (Monelliopsispecanis Bissell), mites, and leafminers (Harris 1988, 1991). The pecan aphid complex 

consisting of BMA and PA has particularly been considered a primary threat to nut production (Dutcher and Htay 1985, 

Beshears 11988). Population densities of BMA are typically an order of magnitude lower than those of PA in such epidemics 

and, if the epidemic proceeds to defoliation, BMA is virtually absent during the latter stages (Bumroongsook and Harris 

1992). 

BMA outbreaks routinely occur once per season in nature, or can also be induced once following the use of broad 

spectrum insecticides. Cage studies show that BMA outbreaks occur 2-3 weeks after introducing BMA into natural enemy 

exclusion cages throughout the season on previously unexposed foliage (Edelson 1982, Liao and Harris 1984, Liao et al. 

1984, Liao et al. 1985, Bumroongsook and Harris 1992). These outbreaks naturally subside after 3-4 weeks, leaving intact 

photosynthetically active foliage (Bumroongsook and Harris 1992). BMA populations cannot reinfest the foliage for at least 

a month after the first induced outbreak (Liao et al. 1984). BMA outbreaks can be curtailed by opening the exclusion cages 

before the infestation has run its course or by introducing spiders or lady beetles into the cage at densities of about 1 per 10 

leaves (Liao et al. 1984). Lacewing, Chrysoperla rufilabris (Burmeister), oviposition increases as BMA densities increase. 

Lacewing egg densities are also higher on foliage exposed by opening cages containing an incipient outbreak of BMA (Liao 

et al. 1984). Spider densities increase as BMA densities increase, and spiders readily accept BMA as prey (Bumroongsook et 

al. 1992). Spider densities are also higher on foliage exposed by opening cages containing an incipient outbreak of BMA 

(Liao et al. 1984). M. pecanis is slower to outbreak on mature bearing trees (Bumroongsook and Harris 1992), and high 

densities result in defoliation whether the leaves have been conditioned by BMA or not, indicating these aphids are capable of 

causing more loss of foliage than BMA. 

These studies indicate that BMA may mitigate the impact of disruptions or natural declines of natural enemies in the 

pecan ecosystem by outbreaking and then attracting and reestablishing natural enemies to the system before more insidious 

phytophages increase in density (Bumroongsook and Harris 1992). The effects of such a role would not be limited to the 

pecan aphid complex, but would extend to all phytophagous species serving as prey to the polyphagous lacewings, lady 

beetles, spiders, etc. (Liao et al. 1984), that respond to M. caryella. 

113

Wood et al. (1987) report that M. caryella second and third instar nymphs excrete some fortyfold more honeydew 

than other pecan aphids (other life stage comparisons are three- to tenfold) and noted that some 95% of the 301 joules (J) of 

energy removed from pecan by one M. caryella aphid during its lifetime was due to honeydew excretion. M. cao'el/a energy 

consumption exceeds by 20% the largest consumption found in a review by Llewellyn (1987) for non-pecan aphids, and also 

has the lowest energy assimilation compared to other aphids. Honeydew is attractive to hundreds of species of insects 

(Klingauf 1987), including many predators and parasites. Tedders (1991) concluded pecan aphid honeydew "may be 

necessary for attracting and retaining large numbers of many beneficial species." Honeydew also improved the surrounding 

substrate for decomposers (Dixon 1985), enhancing the recycling of nutrients. These benefits must be balanced against the 

costs of M. cao'ella to pecan. 

Initial characterization of the costs of M. caryella requires accounting for energy lost to the aphid in relation to the 

overall photosynthate production capacity of the pecan. Wood et al. (1987) provide the former at 301.41 J for one M. 

caryella over a 19.3 d life span and calculate 11.42 J/d are removed by the average aphid in a population. Anderson (1991) 

reported pecan assimilates CO2 at a rate of 15.9 btmol/m2/s [Wood and Tedders (1986) report 11 g tool COJmZ/s]. Annual 

photosynthate productivity (PP)/ha, therefore, can be estimated by: 

Annual PP/ha = 2.65 bttool of glucose x 50,000 m2pecan foliage/ha x 6.25 x 106seconds/growing season x 

0.5 x 0.65 [(corrected for photosynthetic efficiency due to shading (0.5) and respiration (0.65)]. 

Where CO 2 has been converted to glucose (1 glucose = 6 COz), pecan foliage/ha was averaged from Wood et 

al. (1987) who reported 35,900 mZ/ha, Cutler (1976) who reported 52,580 m2/ha and Lozano (1982) who 

found 62,800 m2/ha; a growing season was defined as 217 d with 8 hr of sunlight each day; photosynthetic 

efficiency was estimated at 50% and respiration costs at 35%. 

This results in a PP/ha of 2.69 x 10s mols of glucose annually or 185.98 x 106 kcal/ha (1 tool of glucose = 691 kcal). 

Pecan foliage routinely is exposed to BMA at a level of a 100 to 500 or so aphid days per compound leaf each season 

(Tedders 1978, Flores 1981, Li 1990, Liao and Harris 1984, Tedders and Wood 1985, Bumroongsook 1986, Mansour and 

Harris 1988). Pecan leaf density estimates vary from 2.2 to 3.6 x l0 bleaves/ha (Cutler 1976, Lozano 1982, Wood et al. 

1987). Using an average of 3.0 x 106leaves/ha and 500 M. caryella aphid days/leaf/season results in an energy cost of 17.12 

x 10'_J/ha annually (I 1.42 J/aphid day x 3.0 x 106leaves/ha x 500 aphid days). Since 1 J = 0.24 cal, this equates to 4.1 x 10 6 

kcal as an average cost of BMA on each ha of pecan each season. 

Therefore, a 500 aphid day/leaf infestation of BMA each season would occur at a direct cost of about 2% (4.1 x 106 

+185.98 x 1106)of the photosynthate production of the pecan. This cost would be spread across all physiological needs of the 

plant with only a portion appearing as yield loss. Secondary effects of factors like saprophagous sooty mold growth and 

additional reductions in photosynthetic efficiency are not discussed because previous studies (Tedders and Smith 1976, Wood 

and Tedders 1986) indicate BMA densities of 500 aphid days/leatTseason are too low to cause measurable reductions via these 

secondary effects in mature bearing trees. 

This view of BMA is more sanguine than that presented by Wood et al. (1987). Resistance to pesticides in the pecan 

pest complex makes the prospect of maintaining BMA below its natural equilibrium level fraught with additional pesticide 

costs and other secondary pest problems like mites, leafminers, and other aphids (Harris 1991). Wood et al. (1987) report an 

average of 10 aphicide sprays ($500/ha at $50/spray)required to achieve this, and Dutcher (1991) reports 50 kg/ha rates of 

aldicarb ($330/ha at $6.60/kg) needed to achieve the same end for aphid control from mid-July onward. Although these 

approaches may protect the = 2% of photosynthate production at risk from BMA, that loss represents 20 kg from an estimated 

average yield of 1,000 kg/ha if the entire BMA energy removed resulted in nut loss. Most insecticides currently used in 

pecans are highly toxic to most natural enemies of aphids (Mizell 1991) and in danger of being rendered ineffective due to 

aphid resistance to them (Harris 1991). We believe that the pecan aphid complex rarely poses a direct threat to pecan 

production unless disrupted by pesticides, especially in the early season and that the judicious use of pesticides for key pests 

can usually avoid these disruptions. When the problem is "viewed from this perspective, pest managers may wish to reevaluate 

the risks posed by BMA compared to the benefits provided by sustaining natural enemies and enhancing decomposition. 

Further study is needed to determine whether a natural role of BMA in the pecan ecosystem is to re-establish waning 

densities of natural enemies. 

114

HARRIS, M.K. 1980. Arthropod-plant interactions related to agriculture, emphasizing host plant resistance, p. 23-51. In 

Biology and Breeding for Resistance. Dept. of Agric. Comm. TAES, Texas A&M University. MP-1451. 

HARRIS, M.K. 1983. Integrated pest management of pecans. Ann. Rev. Entomol. 28:291-318. 

HARRIS, M.K. and FREDERIKSEN, R.A. 1984. Concepts and methods regarding host plant resistance to avthropods and 

pathogens. Ann. Rev. Phytopath. 22: 247-272. 

HARRIS, M.K. 1988. Pecan domestication and pecan arthropods, p. 207-225. In The Entomology of Indigenous and 

Naturalized Systems in Agriculture. Westview Press Inc., Boulder, CO. 

HARRIS, M.K. 1991. Pecan arthropod management,, p. 6-15. In Wood, B.W. and Payne, J.A., eds. Pecan Husbandry: 

Challenges and Opportunities. First National Pecan Workshop Proceedings. ARS-96. U.S. Deptartment of Agriculture, 

Agriculture Research Service. 

HOWE, H.E, and WESTLEY, L.C. 1988. Ecological relationships of plants and animals. Oxford Univ. Press, Oxford. 

273 p. 

KLINGAUF, EA. 1987. Feeding, adaption and excretion, p. 225-253. In Minks, K.A. and Harrewign, E, eds. Aphids: Their 

Biology, Natural Enemies, and Control. Elsevier, New York. 

LI, T. 1990. Population growth and factors affecting the seasonal abundance Monellia caryella (Fitch) and Monelliopsis 

pecanis Bissell in Central Texas. Ph.D. Dissertation, Texas A&M University, College Station, TX. 181 p. 

LIAO, H. and HARRIS, M.K. 1984. Population growth of the blackmargined aphid on pecan in the field. Agric. Ecol. 

Environ. 12: 353-361. 

LIAO, H., HARRIS, M.K., GILSTRAP, EE., DEAN, D.A., AGNEW, C.W., MICHELS, G.J., and MANSOUR, F. 1984. 

Natural enemies and other factors affecting seasonal abundance of the blackmargined aphid on pecan. Southw. 

Entomol. 9: 404-420. 

LIAO, H., HARRIS, M.K., GILSTRAP, EE., and MANSOUR, E 1985. Impact of natural enemies on the blackmargined 

pecan aphid, Monellia caryella (Homoptera: Aphidae). Environ. Entomol. 14: 122-126. 

LITTLE, E.L. 1971. Atlas of United States trees. Vol. 1. Conifers and important hardwoods. Misc. Publ. 1146. Washington, 

DC: U.S. Department of Agriculture, Forest Service. 

LLEWELLYN, M. 1987. Aphid energy budgets, p. 109-117. In Minks, A.K. and Harrewign, E, eds. Aphids: Their Biology, 

Natural Enemies, and Control Elsevier, New York. 

LOZANO, R. 1982. A three-dimensional characterization of a bearing pecan tree. M.S. Thesis, Horticulture, Texas A&M 

University, College Station, TX. 71 p. 

MAGGIO, R.C., HARRIS, M.K., INGLE, S.J., and DAVIS, M.R. 1991. A summary of the location, abundance, distribution, 

and condition of Carya on the Brazos and Colorado River systems of Texas. Dept. of Agric. Comm. TAES, Texas 

A&M Univ. MP-1703. 

MANSOUR, E and HARRIS, M.K. 1988. Biology and phenology of the blackmargined aphid, Monellia caryella (Fitch), a 

new pest of pecan in Israel. Southw. Entomol. 13: 19-29. 

116

MIZELL, R.E 1991. Pesticides and beneficial insects: application of current knowledge and future needs, p. 47-54. In 

Wood, B.W. and Payne, J.A., eds. Pecan Husbandry: Challenges and Opportunities. First National Pecan Workshop 

Proceedings. ARS-96. U.S. Deptartment of Agriculture, Agriculture Research Service. 

PAINTER, R.H. 1951. Insect resistance in crop plants. Univ. Press of Kansas. Lawrence, KS. 520 p. 

SNELLING, R.O. 1941. Resistance of plants to insect attack. Bot. Rev. 7: 543-586. 

TEDDERS, W.L. and SMITH, J.W. 1976. Shading effect on pecan by sooty mold growth. J. Econ. Entomol. 69: 551-553. 

TEDDERS, W.L. 1978. Important biological and morphological characteristics of foliar feeding aphids of pecan. Tech. Bull. 

1529. Washington, DC: U.S. Department of Agriculture. 29 p. 

TEDDERS, W.L. and WOOD, B.W. 1985. Estimate of the influence of feeding by Monelliopsis pecanis and Monellia 

caryella (Homoptera: Aphididae) on the fruit, foliage, carbohydrate reserves, and tree productivity of mature 'Stuart' 

pecans. J. Econ. Entomol. 78: 642-646. 

TEDDERS, W.L. 1991. Alternative controls for pecan insects, p. 77-83. In Pecan Husbandry. ARS-96. U.S. Deptartment of 

Agriculture, Agriculture Research Service. 

WOOD, B.W. and TEDDERS, W.L. 1986. Reduced net photosynthesis of leaves from mature pecan trees by three species 

of pecan aphids. J. Entomol. Sci. 21: 355-360. 

WOOD, B.W., TEDDERS, W.L., and DUTCHER, J.D. 1987. Energy drain by three pecan aphid species (Homoptera: 

Aphididae) and their influence on in-shell pecan production. Environ. Entomol. 16(5): 1045-1056. 

117

PONDEROSA PINE RESPONSE TO NITROGEN FERTILIZATION AND 

DEFOLIATION BY THE PANDORA MOTH_ 

COLORADIA PANDORA BLAKE 

B.E. WICKMAN I, R.R. MASON z, and H.G. PAUL 2 

_Silviculture Laboratory, 1027 NW Trenton Avenue, Bend, OR 97701, USA 

2Forestry and Range Sciences Laboratory, 1401 Gekeler Lane, La Grande, OR 97850, USA 

INTRODUCTION 

The pandora moth, Coloradia pandora Blake (Lepidoptera: Saturniidae), is a well-known outbreak species on 

ponderosa, Pinus ponderosa Laws, and lodgepole pines, Pinus contorta Dougl., in the western United States (Mattson et al. 

1991). For example, it has periodically caused severe defoliation in the pineries of central Oregon. Historically, such 

outbreaks have been short-lived and caused little impact. But, in the 1920's, an outbreak in the Klamath Falls region did 

significant damage to the old-growth ponderosa pines so that they became highly susceptible to bark beetle attacks (Patterson 

1929). 

The insect has a 2-year life cycle (Carolin and Knopf 1968). Moths flights normally occur in June and July of evennumbered 

years (those near Bend, OR occur in odd-numbered years)(Patterson 1929, Carolin and Knopf 1968, Schmid et al. 

1982). Eggs are deposited in clusters on both needles and bark. Hatch takes place in August and the young larvae feed 

gregariously in small colonies on current-year needles until limited by cold weather in late fall. Feeding continues on warm 

winter days on the south side of trees and on trees highly exposed to sunshine. Larvae eventually disperse and feed solitari|y. 

Feeding intensity increases sharply with the onset of warm spring weather, and defoliation eventually becomes noticeable in 

early June as the caterpillars grow larger. By early July, fully grown larvae drop to the forest floor and pupate in the upper 

several centimeters of soil where they spend nearly a full year. 

The current outbreak in Central Oregon (first seen in 1988) has spread south of as well as north to the city of Bend 

because of massive moth flights in 1991 and 1993. In May 1992, we collected larvae in the original epicenter that were 

infected with a polyhedrosus virus. This usually portends the collapse of an outbreak, so 1994 could have been the year of a 

widespread virus epizootic. 

Because pandora moth outbreaks occur only every 20-30 years, the present infestation afforded an opportunity to test 

the effects of fertilization on the trees and the insects in the nutrient deficient soils in the central Oregon pineries. We 

hypothesized that increasing nutrient availability might improve canopy growth, and net photosynthesis and thereby lead to 

elevated carbon-based defenses against herbivory. Increased tree growth might also offset the effects of defoliation. 

In another forest-insect system, we found that fertilization may temporarily reduce the effects of defoliation (Mason 

et al. 1992, Wickman et al. 1992). During an outbreak of the western spruce budworm, Choristoneura occidentalis Freeman, 

both trees and insects were enhanced by fertilization, but trees more than budworms because they produced more new foliage 

than the budworm could eat. Consequently, fertilized stands suffered less growth impact from defoliation than did untreated 

stands. We tested this conclusion further in the ponderosa pine-pandora moth system. 

The exact test was to determine if a single treatment with nitrogen in the form of urea would significantly reduce the 

impact of pandora moth defoliation in a thinned second-growth ponderosa pine stand. A secondary objective was to determine 

the effect of the treatment on growth and feeding behavior of pandora moth larvae and evaluate the chemical composition 

of foliage and larval frass. 

Mattson, W.J., Niemel_i, P., and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


118

METHODS 

Study Site 

The study took place in an active outbreak of the pandora moth in the Deschutes National Forest about 20 km south 

of Bend, Oregon. The stand is a ponderosa pine/bitterbrush-manzanitJfescue plant community (Volland 1976). Lodgepole 

pine is the only other tree species present and is scarce. The original ponderosa pine was logged in the 1930's and the second 

growth stand thinned in the 1970's. The average age of the trees at the time of our study was about 55 years and spacing 

averaged 3.6 by 3.6 m, but the latter differed across the area (the study area averaged 165 trees per acre). 

Elevation of the study area is 1,250 in, slope varies from 0 to 5 percent, and aspect is slightly south. The soil mantle 

is 0.9-m-deep pumice, which resulted from the Mount Mazama volcanic eruption about 6,700 years ago. Bare pumice is 

common and there is only a thin organic layer under the trees. Soils are particularly deficient in nitrogen, phosphorus, and 

sulfur (Cochran 1978). Annual precipitation is approximately 51 cm and occurs mostly as snowfall from October through 

March. Precipitation at Bend was 23 cm in the 1988-1989 season, and first snowfall occurred within a week after we applied 

fertilizer. 

Experimental Design and Treatments 

In fall 1988, 10pairs of circular 1/10th of an acre plots (11.35-m radius) were located in a 4 by 5 grid with plot 

centers approximately 40 m apart (Fig. 1). All trees on the plots were marked with metal tags. On average, treated plots 

experiment near Sun River 

on the Deschutes National 

Figure l.--Schematic Forest. F = fertilized layout of and C 

= control. 

_ _ @ _ 

119

contained 17 trees and control plots 16 trees. The average tree diameter at breast height (dbh) was 29.5 cm. One plot in each 

pair was randomly selected for fertilization. Analysis of results was on a paired plot basis. In October, 27.2 kg of urea (47% 

N) was applied by hand to the fertilized plot in each of the 10 pairs. This simulated an application rate of 350 kg N ha -_. 

Insect and Tree Measurements 

Effects of the treatment were evaluated by measurements of tree buds, foliage, radial increment, and insect biomass 

on treated and untreated plots. Samples for foliage measurements and insect weights were collected periodically from tree 

crowns with pole pruners according to methods recommended by Schmid et al. (1982). Radial increment was evaluated at 

the end of the study from increment cores taken from two trees per plot. All sample trees were selected from the center of 

plots to minimize edge effect. The schedule for measurements and collections follows: 

Shoot and Insect Insect Radial 

Date foliage larvae pupae increment 

Fall 1989 X 

May 1990 X 

Fall 1990 X X 

Fall 1991 X 

May 1992 X 

Fall 1992 X X 

Larvae were sampled only once in 1990, using methods developed by Schmid et al. (1982). Sampling also was attempted in 

1992, but only a few larvae were found, indicating a virtual population collapse in the study area. Larvae were collected for 

weighing from 2 midcrown, 45-cm branch samples on each of 3 trees per plot. Larvae were preserved in 70% ethanol and 

oven dried at 45 °C to a constant weight. Mean dry weight per individual larva was calculated by dividing dry mass by 

density. Mean fresh weights of pupae were determined from a collection made from the soil of approximately 5 pupae per 

plot. 

Branch samples were collected each August or September from 1989 to 1992. Two 45-cm mid-crown branches per 

tree, from 3 trees per plot were cut:with pole pruners. In the laboratory, new buds were excised and current needles were 

stripped from the branches. These were oven dried at 45°C to a constant weight and expressed as foliage dry weight per 

centimeter of twig length. Mean foliage weight was summarized by plot and year. Buds were represented as weight per bud. 

In August 1992, at the end of the growing season, 2 cores per tree were collected at dbh on 2 trees per plot to measure radial 

increment. Cores were mounted, sanded, and measured to the nearest 0.01 mm on an incremental measuring instrument 

interfaced with a desk top computer, as described in Wickman et al. (1992). 

Insect frass was collected from canvas panels placed under trees on 5 treated and 5 untreated plots at the peak of 

larval feeding in June 1990. Frass was air dried and analyzed for total and available levels of N, P, K, and S at the University 

of Arizona, Tucson, in 1993. Dried foliage from 1989 and 1990 samples also was analyzed for N, P, K, and S and available 

nutrients at the same laboratory in 1993. Dried 1989 needles were analyzed for total reducing sugars and phenolics by Mr. 

B.A. Birr, North Central Forest Experiment Station, East Lansing, MI, in 1994. 

Statistical Analysis 

Data were analyzed using paired t-tests on the randomly selected paired plots for each year of measurement. Differences 

were judged to be significant when the probability was

the samples. Fertilization had substantial negative effects. Mean dry weights of individual larvae from fertilized plots were 

126.0 mg and 248.7 mg for 4th and 5th instars respectively, compared to 196.5 mg and 360.3 mg, respectively, tbr untreated 

plots. These were 29.6% (p = 0.036) and 32.9% (p = 0.006) reductions, respectively, due to fertilization. Mean fresh weights 

of treated and untreated pupae in September, 1990 were 290.1 mg and 30l .4 rag, respectively, but were not significantly 

different (p = 0.56). 

Foliage and Bud Weights 

Foliage weight was consistently greater on the control plots from 1989 through 1992, but this difference was significant 

(p = 0.03) only in 1990 (Fig. 2). The declining weights of all foliage from 1989 through 1992 was probably due to the 

combined effects of chronic drought and repeated de|bliation. Visual defoliation ratings made in fall, 1990, were similar for 

both fertilized and control trees. The trend of bud weight was similar to that found for foliage except in 1989 fertilized buds 

were heavier than the controls (Fig. 3). Buds from control plots were significantly heavier in 1991 and 1992 than fertilized 

buds (132.3 vs 120.8 mg, p = 0.053 and 69.5 vs 49.3 rag, p = 0.021, respectively). Again, the much smaller buds in 1992 

probably reflect combined effects of 2 defoliation episodes and 5 consecutive years of subnormal precipitation. 

60 

50 

"*" 40 

..c: ---7 

._ 

30 

(D 

03 : . . • 

._ 20 

0 

LL : 

1o / / 

0 -/ / 

1989 1990 1991 1992 

Year 

Figure 2.--Foliage per centimeter of shoot length. Each bar is an average of 10 plot means. 

Foliage and Frass Chemical Analyses 

Foliage from fertilized trees contained more total N in both 1989 and 1990 and more available N in 1989 (Table 1). 

Statistical analysis was not possible because the samples were pooled. Other nutrients (P, S, and K), apparently were similar 

in both fertilized and control foliage for both years (Table 1). 

Larval frass collected in June 1990 was also analyzed for nutrients (Table 2). Available N was over 6 times higher in 

frass from larvae feeding on fertilized trees. Total P and K in frass from fertilized trees was approximately half that from 

unfertilized trees. Because the samples were pooled, a statistical analysis was not possible, but these differences were 

striking. 

121

Figure 3.--Average weight per bud. Each bar is an average of 10 plot means. 

Table l.--Percent foliar nutrients in 1989 and 1990 needles (sample pooled from 10 fertilized plots and 10 control plots). 

122 

Total 

N P S K (ppm) 

1989 7,558 

Plots 1-10 (fertilized) 2.095 0.114 0.071 

11-20 (control) 1.568 0.127 0.075 7,471 

1990 8,641 

Plots 1-10 (fertilized) 2.012 0.t67 0.082 

11-20 (control) 1.750 0.185 0.088 9,995 

Available 

1989 

Plots l-t0 (fertilized) 0.0216 0.0690 0.0088 6,165 

11-20 (control) 0.0134 0.0581 0.0064 5,635 

1990 

Plots 1-10 (fertilized) 0.0206 0.1007 0.0039 7,082 

11-20 (control) 0.0216 0.0994 0.0052 7,739

Table 2.--Percent nutrient content of larval frass in 1990 (sample pooled from 5 fertilized plots and 5 control plots). 

Total Available 

N P S ppmK ppm N 

Fertilized 0.826 0.065 0.039 1,430 625 

Control 0.595 0.131 0.042 2,670 94 

There was no significant difference in the total reducing sugars (p = 0.36), or phenolics (p = 0.48) from fertilized 

trees and controls in 1989 foliage. 

Radial Growth 

Radial growth showed a steadily declining trend on both fertilized and control plots from 1988 through 1992, (Fig. 4). 

Fertilized trees, even though they produced less foliage and smaller buds during this period, had significantly (p

Fertilized and control trees all grew at essentially the same rates during the 16-year period prior to fertilization (Fig. 

5). The downward trend starting in 1986, probably induced by precipitation deficits, was sharply reversed in 1990 for 

fertilized trees. Those trees eventually resumed their downward trend, but the spurt of growth in 1990 gave them a noticeable 

advantage over the untreated controls. 

3 

2.5 

Control 

--Fertilized . 

_ ,, \ 

Treatment 

Application 

/ /" \ 

E / X... 

t -. 

E I ".. 

\ //\ 

0 \ _ s \\" \ 

,-- 1.5 ,'- 

(_ _ / " [] Precip.departures frommean \\ 

(b 

+10 cm 

1t,,.., 

09 

0 

< 

cu 1 

0.5 

'\\ 

-10cm 

,,o_,,o_,,,o_,,,o_,,,o_ ,o5,,o_,,,c_,,o5,o5_ 

0 -----T ] l l T I I I [ t I I I i i T ] T I 

1972 1976 1980 1984 1988 1992 

Year 

Figure 5._Trends of precipitation and radial growth of ponderosa pine at dbh. Plotted points are averages of 10 plot means. 

DISCUSSION 

All stages of the pandora moth are large and easily seen, and tree defoliation is spectacular during an outbreak. Such 

episodes, however, occur at intervals of 20 or 30 years, and tree mortality as a result of feeding is usually negligible. Even 

though the pandora moth makes an interesting test insect for herbivore studies, its irregular, boom or bust dynamics limit its 

consistent availability. Because of life history differences between western spruce budworm and pandora moth, comparisons 

of fertilizer effects are tenuous. The pandora moth has a 2-year life cycle; therefore, larvae, on alternate years, feed mostly 

on older needles rather than emerging new foliage so that the tree has a constant supply of new foliage. Conversely, the 

budworm has a l=year life cycle and feeds on new needles annually and can repeatedly remove new foliage for 7 to 10 years. 

We found that larvae responded negatively in 1990, at least in terms of individual weights, to the fertilizer treatment. 

This occurred even though available N measured in 1989 foliage was almost 50% higher in fertilized plots. The opposite 

effects on larval weights were reported for a similar study of western spruce budworm in Oregon (Mason et al. 1992, Waring 

et aL 1992). Available N in 1990 insect frass was over 6 times higher on fertilized plots. A possible explanation is that 

pandora moth larvae may be differentially absorbing proteins or amino acids from fertilized compared to unfertilized foliage 

and concentrating N in the frass. Why the increased nutrient resource did not result in larger individuals is an interesting 

question. Perhaps nutrients were imbalanced. 

124

Some tree responses to fertilization were also different than in other recent studies. Additional N provided by 

fertilization was expected to increase foliage production based on results of a study of fir infested with western spruce 

budworm (Wickman et al. 1992). We found fertilized trees produced less tbliage per centimeter of twig length for the period 

1989 through 1992 and buds also were smaller in all years except 1989. 

Radial growth usually is related to foliage quantity and quality. Fertilized trees had less foliage production but more 

radial growth than control trees, which suggested that resources may have been allocated to the tree bole at the expense of 

foliage. Fertilization resulted in greater individual tree growth similar to findings of Cochran (1978). He fertilized thinned 

ponderosa pine stands similar and close to our study area; however, his stands were not defoliated by pandora moth at that 

time. 

A study by Miller and Wagner (1989) of the effects of pandora moth defoliation on ponderosa pine growth in Arizona 

found greater radial growth in heavily defoliated trees l year after the last defoliation. This is not the usual growth response 

of trees defoliated by some other insects where growth is usually directly related to degree of defoliation and lags 1 year after 

defoliation (Wickman 1979, Wickman etal. in press) 

The study raised more questions than answers. Response of insects and foliage production of host trees was contrary 

to findings from a recent study of western spruce budworm (Mason et al. 1992, Waring et al. 1992, Wickman et hi. 1992). 

But radial growth seems to respond positively after fertilization on nutrient deficient soils irrespective of insect or host tree 

species, and perhaps from a site productivity perspective that is what really matters. 

It may be impossible to account for the smaller larvae from fertilized trees in this small study. Perhaps the added 

nutrient resources (N) were only partially allocated for growth of trees and some increased defensive chemistry production in 

fertilized trees resulted in smaller larvae. Maybe the larval growth was related to feeding efficiency determined by effects of 

foliage phenolics or other foliage attributes, like needle toughness, on the previous generation of larvae that carried over to 

the next generation. This explanation is difficult to pursue because we were able to measure larval weights from only one 

generation of the three that affected the host tree through the outbreak period. Perhaps the insect frass with its high levels of 

available nitrogen reacted synergistically with residues of N in the soil and this resulted in increased radial growth, even 

while the trees were being defoliated. Miller and Wagner (1989) suggest that the heavily defoliated pines ability to compensate 

for foliage loss involves certain compensatory growth mechanisms, perhaps accelerated nutrient cycling of insect frass. 

We did find that pandora moth frass is rich in nutrients, but we had no way of accounting for interactions of frass and our N 

treatments. 

This study, though small and short on chemical analysis, does point out the vagaries associated with interactions 

among fertilization, insects, defoliation, and the host tree. Our studies to date of budworm on fir and pandora moth on pine 

do not indicate that fertilization enhances plant defenses against herbivores through the production of secondary resistance 

compounds. One consistent result we have encountered is increased radial growth of fertilized trees. 

SUMMARY 

Responses of ponderosa pine and pandora moth to fertilization with N were studied for 4 years after treatment. 

Fertilization had a negative effect on larval weights. The 1990 generation of treated larvae were significantly smaller than 

control larvae, but there was no significant difference between pupal weights. Foliage and bud weights of fertilized trees 

were significantly lighter than controls. Available nitrogen in both foliage and insect frass was higher in fertilized plots. 

Radial growth at dbh of fertilized trees was significantly greater and almost double the growth of controls. The results 

indicate that effects of fertilization differ with species of host tree and insect herbivore, except for increased radial growth 

that has been consistently noted in all the recent fertilization studies in eastern Oregon. The highly complex interactions of 

increased nutrient cycling from herbivore feeding and artificial application of N may help explain variable results found in 

recent studies. 


Technical assistance was provided by Katie Bobowski, Ellen Stenard, and Wendy Sutton. We also thank Dr. 

Russel Mitchell for collecting insect frass and Dr. Arthur Tiedemann for arranging for chemical analysis of foliage and frass. 

125

Dr. William Mattson arranged for analysis of sugars and phenolics in foliage. Valuable insights which helped our 

interpretations were provided by Drs. William Mattson and Pekka Niemelfi during a visit to the study site in 1993. We 

appreciate helpful review comments from Dr. Arthur Tiedemann and Dr. Patrick Cochran, Pacific Northwest Research 

Station, and Dr. Pekka NiemelS., Finnish Forest Research Institute. 


CAROLIN, V.M. and KNOPE J.A.C. 1968. The pandora moth. U.S. Dept. Agric., For. Serv. Pest Leafl. 114. 

COCHRAN, RH. 1978. Response of a pole-size ponderosa pine stand to nitrogen, phosphorus, and sulfur. Res. Note PNW- 

319. Portland, OR: U.S. Department of Agriculture, Forest Service, Pacific Northwest Research Station. 8 p. 

MASON, R.R., WICKMAN, B.E., BECKWITH, R.C., and PAUL, H.G. 1992. Thinning and nitrogen fertilization in a grand 

fir stand infested with western spruce budworm. Part I: Insect response. For. Sci. 38(2):235-251. 

MATTSON, W.J., HERMS, D.A., WITTER, J.A., and ALLEN, D.C. 1991. Woody plant grazing systems: North American 

outbreak folivores and their host plants, p. 53-84. In Baranchikov, Y., Mattson, W.J., Hain, F.R, and Payne, T.L., eds. 

Forest insect guilds: patterns of interaction with host plants. GTR-NE-153. Radnor, PA: U.S. Department of 

Agriculture, Forest Service. 400 p. 

MILLER, K.K. and WAGNER, M.R. 1989. Effect of pandora moth (Lepidoptera:Saturniidae) defoliation on growth of 

ponderosa pine in Arizona. J. Econ. Entomol. 82(6): 1682-1686. 

PATTERSON, J.E. 1929. The pandora moth, a periodic pest of western pine forests. U.S. Dept. Agric. Tech. Bull. 137.20 p. 

SCHMID, J.M., BENNETT, D., YOUNG, R.W., MATA, S., ANDREWS, M., and MITCHELL, J. 1982. Sampling larval 

populations of the pandora moth. For. Res. Note RM-421. U.S. Department of Agriculture, Forest Service. 5 p. 

VOLLAND, L.A. 11976. Plant communities of the central Oregon pumice zone. R-6 Area Guide 4-2. U.S. Department of 

Agriculture, Forest Service. 113 p. 

WARING, R.H., SAVAGE, T., CROMACK, K., Jr., and ROSE, C. 1992. Thinning and nitrogen fertilization in a grand fir 

stand infested with western spruce budworm. Part IV: An ecosystem/pest management perspective. For Sci. 38(2): 

275-286. 

WICKMAN, B.E. 1979. How to estimate defoliation and predict tree damage. Douglas-fir tussock moth handbook. Agric. 

Handb. 550. Washington, DC: U.S. Deptartment of Agriculture. 15 p. 

WICKMAN, B.E., MASON, R.R., and PAUL, H.G. 1992. Thinning and nitrogen fertilization in a grand fir stand infested 

with western spruce budworm. Part II: Tree growth response. For. Sci. 38(2): 252-264 

WICKMAN, B.E., MASON, R.R., and SWETNAM, T.W. (In press). Searching for Long-term patterns of forest insect 

outbreaks. In Eather, S., Walters, K., Mills, N., and Watt, A., eds. Proc. Individuals, Populations, and Patterns in 

Ecology. L Intercept Press, United Kingdom. 

126

PHYTOCHEMICAL PROTECTION AND NATURAL ENEMIES IN THE 

REGULATION OF A ROOT AND STEM BORING INSECT 

A.V. WHIPPLE and D.R. STRONG 

Bodega Marine Laboratory, University of California, Box 247, Bodega Bay, California 94923, USA 

INTRODUCTION 

Much debate has centered on why herbivores in natural systems do not more often have a large negative effect on the 

plants on which they feed. Except for some aquatic trophic cascades in which plants are algae, (Carpenter et aL 1987, 

Carpenter and Kitchell 1988) and a few large mammalian grazers (McNaughton et al. 1988), herbivores have rarely been 

shown to substantially depress plant populations. Two main ideas why this is are top-down suppression by predators and 

bottom-up plant protection (Hunter and Price 1992). The natural enemy hypothesis suggests that populations are generally . 

controlled by their predators and/or pathogens (Hairston et al. 1960, Oksanen et al. 1981, Hairston and Hairston 1993). Thus, 

the plant feeders are kept too sparse to substantially depress plant populations. The endogenous plant protection hypothesis 

proposes that, in terrestrial and some aquatic systems, plant biomass is protected by noxious secondary chemicals (Ehrlich 

and Raven 1964, Hawkins 1992, Hay and Steinberg 1992). An adjunct to this is that higher plant biomass is also largely 

unavailable to most herbivores (Strong 1992, White 1993). This is because most herbivores cannot digest cellulose or lignin 

without the aid of microbial symbionts (Martin 1991). Because cellulose and lignin make up a large portion of the plant, 

many essential nutrients for herbivores are present only in low concentrations, which may limit the growth of herbivore 

populations so that they are not able to have substantial impact on plant populations. 

Of course, these hypotheses are not mutually exclusive, and in the system discussed here both plant defense and 

herbivore natural enemies may play a role in determining the level of herbivore effect on the plant population (Price 1992). 

The system of the bush lupine, Lupinus arboreus, and its stem and root boring ghost moth or "swift," Hepialus caI(fornicus, 

is ideal for examining the roles of these different factors because the level of herbivore attack varies over space with some 

areas exhibiting extensive plant mortality and others with little mortality (Strong et al. in prep.). 

Directly or indirectly, abiotic factors are probably involved in the differences between the sites in level of attack. The 

effect of an abiotic factor on the level of herbivory may act directly on the herbivore or be mediated through the herbivore's 

host plant, which would be bottom-up regulation, or through predators and pathogens, which would be top-down suppression. 

Herbivory 

and Resistance in the Bush Lupine 

The bush lupine is a dominant woody shrub of some coastal headland habitats in central California. At Point Reyes 

and on Bodega Head in Marin and Sonoma Counties, a major source of mortality to the bush lupine appears to be the swift, 

Hepialus californicus (Opler 1968, Davidson and Barbour 1977, Wagner 1986). The mid to late instars of the swift caterpillar 

bore in the roots and stem of the bush lupine. The early instars are thought to feed on litter and rootlets (Wagner 1986). 

The caterpillars reach 40-50 mm, and as many as 60 have been found in one bush. While those numbers are exceptional, it is : 

difficult to find a large bush with no evidence of swift caterpillar damage (pets. obs.). 

Lupines produce large quantities of quinolizidine alkaloids, which are thought to function in nutrient storage and in 

defense against herbivores and pathogens (Wink 1992). As little as 0.11%fresh weight of these alkaloids can be toxic to 

insect herbivores (Wink 1992). Bush lupine leaves contain high levels of these quinolizidine alkaloids in their foliage _ 

(Bentley and Johnson 1991). The bush lupine probably also has substantial levels of alkaloids in the stems and roots where 

the swift caterpillars feed (Wink and Witte 1984). The alkaloid is probably concentrated in the epidermis of the stem and root 

where exceptionally high levels of alkaloid have been found in other lupines (Wink 1992). 

Mattson, W.J., Niemel/i, R, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


127 

! 

ii

Whether these quinolizidine alkaloids defend the plant against attack by a particular herbivore species could depend 

on whether the herbivore in a specialist or a generalist. A specialist could be indifferent or prefer high levels of alkaloid, and 

a generalist would avoid or be harmed by high levels of alkaloid. Wink (1992):reviews the evidence for both of these cases; 

however, conclusive evidence for effects of alkaloids in nature has not yet been shown (ttartmann 1992). For swifts, the 

generalist-specialist distinction is ambiguous. The insect is polyphagous, having been found to feed on a wide variety of host 

plants over its geographical range. In addition, it is even more indiscriminantly polyphagous and fungivorous in early instars 

when feeding on detritus, before entering the root. However, as with many polyphagous species, local populations may be 

much more restricted in their host plant use than this implies (Fox and Morrow 1981, Singer and Parmesan 1993). The swifts 

at Bodega Marine Reserve seem to be feeding largely on lupine (though perhaps not entirely), despite the presence of hosts 

they have been found on in other locations (De Benedictis et al. 1990), and could show characteristics of specialist rather 

than generalist herbivores. Hence, it is not obvious what effect, if any, quinolizidine alkaloids will have on the interaction 

between the swifts and the lupine. The same can be said for another class of chemicals thought to function in defense that 

have been found in lupine roots, the isoflavonoids (Lane et al. 1987). 

Herbivory, and perhaps levels of endogenous resistance of bush lupine, vary over the reserve on Bodega Head. 

Dramatically more bushes died, and each contained many more caterpillars, in some areas than in others in 1992 and 1993 

(Strong et al. in prep.). In addition, the areas with lower bush lupine mortality in these 2 years have had a fairly consistent 

cover of lupines over the past 40 years. Conversely, areas with greater recent mortality show a historical pattern of fluctuating 

cover (Strong et al. in press). The plants in areas with different levels of herbivory may differ in their resistance to swift 

herbivory. Thus, in terms of endogenous plant protection, lower levels of herbivory by swift caterpillars could derive from 

lower nutritional value of the plant tissue to the herbivore, from higher secondary chemical levels, or both since in some 

cases they are inexorably intertwined. 

The bark and epidermis of the lupine may constitute another or different defense because the early instars may be 

prevented from feeding on the plant by some chemical or physical property of the outer bark. Thus, they would be more 

exposed to natural enemies for a longer period of time. 

Natural Enemies of Swift Caterpillars 

The early instars of Hepialus californicus are very small, with a starting length of approximately 1 ram, and could be 

quite vulnerable to predation because they feed externally on litter and roots. The high fecundity of the swift, followed by 

modest numbers of large caterpillars in lupine stems, indicates high early instar mortality. Potential predators in the soil 

around the roots include a geophilomorph centipede, which does eat early instars in laboratory trials (Beld unpubl, data). 

Wagner (1986) has reported cannibalism as a major source of mortality in laboratory colonies as well. 

As the caterpillars grow larger they are found feeding externally on lupine roots, at which time swift caterpillars 

killed by the new nematode species Heterorhabditis sp. have been found. Heterorhabditis sp. probably kills the early instars 

as well, but the tiny corpses are simply harder to locate. Larger caterpillars tunnel inside the lupine root and stem where they 

become relatively invulnerable to predation. A fungus, Beuvaria brogniartii, kills the medium to large swifts inside the stem 

and root (Williams 1905, Wagner 1986, Strong and Kaya unpubl.). 

The nematode has the potential to limit the density of swirl caterpillars in the stems. The prevalence of the nematodes 

in the soil at the base of lupine sterns, as measured by a GaUeria mellonella bioassay (Bedding and Ackhurst 1975), is 

quite variable between sites (Strong unpubl, data). 

Abiotic Factors 

Several abiotic factors may be of importance in this system either by depressing swirl populations directly or by 

affecting lupines or the nematode. Wind could affect swift moth populations at the different sites directly; ovipositing female 

swifts release eggs as they fly over suitable habitat. The prevailing winds come out of the north, correlated with low swirl 

caterpillar density in lupine roots. However, Wagner (1986) found that the moths do not usually fly if it is too windy and one 

of their flight times is predawn when winds are usually light. Also, desiccation may be a major cause of mortality in the early 

instar caterpillars. 

128

Indirect effects of abiotic factors on swift populations could be mediated either by their host plant or through predators 

and pathogens. Input of sea spray is higher at the upwind sites (Barbour et al. 1973). One consequence of higher salt 

spray loads could be higher soil phosphorous levels. A possible mechanism by which phosphorus could drastically alter 

allocation patterns in the plant has been demonstrated by Dinkelaker et al. (1989) in a congeneric species, Lupinus albus. 

Under conditions of low phosphorus, this plant secretes large quantities of citrate (up to 23% of its dry matter production) 

from proteoid roots. The citrate enhances the solubility of phosphorus, making it more available to the plant. Thus, plants 

that have more phosphorous available could have much more fixed carbon to allocate to growth rather than to acquiring 

nutrients since they would not need to secrete as much citrate to get their required phosphorous. If a similar mechanism is 

present in L. arboreus, this would almost certainly alter plant chemistry between sites with different levels of phosphorous in 

ways that could affect individual growth rates and survival of the swift caterpillars by changing the nutritional quality of their 

food. Differential g_:owth and survival rates could in turn determine the density of caterpillars in the stems and roots of 

plants. 

Abiotic factors, such as soil properties and moisture, could contribute to differences in the nematode populations 

between the two sites. Heterorhabditis sp. are probably vulnerable to desiccation (Kaya 1990), hence in the dry summer, 

slight differences in soil moisture between sites may be quite important to their survival. 

DISCUSSION 

The picture sketched by our preliminary information leads to several hypotheses concerning the differences in the 

impact of herbivore on its host plant among our different sites. Understanding the mechanism by which plants are "protected" 

from herbivory at sites with low herbivory as compared to sites with high herbivory will allow us to assess the 

relative importance of the top-down effects of natural enemies and the bottom-up effects of plant chemistry (including 

nutrition). The simplest hypotheses implicate a single abiotic factor or an abiotic factor mediated by either the nematode or 

the lupine. These include differences in oviposition due to wind, differences in levels of predatory nematodes due to differences 

in soil chemistry, and differences in the secondary chemistry of the lupine due to differences in nutrient levels. While 

each of these may play a role, it seems unlikely that any of them are sufficient alone. 

An additional hypothesis involves the interaction of plant chemistry and predation in explaining the observed pattern. 

The hypothesis is that the different nutrient levels at some sites allow plants to be better defended against herbivory, or to be 

more vigorous and better able to withstand herbivory (Price 1991), both of which would reduce the impact of the herbivore 

enough that the plants would be able to persist at that site. Then, since the plants persist at that site, rather then undergoing 

local extinction, there is a constant supply of the swift for the nematode to prey upon. At sites that experience local extinction 

of lupine, the nematode could have no host for a period of time and go extinct at that site. Since it is not a very mobile 

predator, it may not be able to recolonize rapidly when the lupine and the swift reappear. Thus, the nematode reinforces the 

pattern caused by the differences in herbivore "resistance" between the sites by reducing swift populations where the lupines 

usually persist, but not where they experience high mortality. 

Experimental manipulations in the field of natural enemies and conditions affecting plant chemistry will be needed to 

assess the relative roles of top-down control and bottom-up protection in controlling herbivore populations. The third 

alternative, as we have mentioned, is the action of abiotic factors directly on the herbivore populations; this possibility should 

also be examined by looking at egg deposition and the desiccation of early instars at the different sites. 

SUMMARY 

Mature bush lupine, Lupinus arboreus, die after heavy stem and root boring by caterpillars of the ghost moth or 

swift, Hepialus californicus. Mortality rates of these plants can be extremely high in some stands and quite low in other 

stands in close proximity. This system is ideal for testing for top-down versus bottom-up control of herbivore populations by 

examining the differences between sites where there is heavy herbivory and sites where there is not heavy herbivory. Nutrient 

levels that are likely to affect lupine plant allocation patterns probably vary over the area of interest. Population levels of 

an entomopathogenic nematode that prey upon the swift moth also vary over the reserve. It is hypothesized that both plant 

chemistry variation and the nematode natural enemy play an important role in determining the level of herbivore attack. 

129


We would like to thank John Maron for access to his data on lupine fluctuations, which is in manuscript. For 

financial support we would like to thank Barbara Bentley for a graduate student grant from the Bodega Field Conferences 

(AVW), the Sippl fund for travel support (AVW), and William Mattson and the IUFRO conference for aid with 

meeting fees (AVW). 


BARBOUR, M.G., CRAIG, R.B., DRYSDALE, F.R., and GHILSEN, M.T. 1973. Coastal Ecology: Bodega Head. University 

of California Press, Berkeley. 

BEDDING, R.A. and ACKHURST, R.J. 1975. A simple technique for the detection of insect parasitic rhapbditid nematodes 

in soil. Nematologica 21: 109-114. 

BENTLEY. B.L. and JOHNSON, N.D. 1991. Plants as tbod for herbivores: the roles of nitrogen fixation and carbon 

dioxide enrichment, p. 257-272. In Price, P.W., Lewinsohn, T.M., Fernandes, G.W., and Benson, W.W., eds. Plant- 

Animal Interactions. John Wiley & Sons, New York. 

CARPENTER, SR. and KITCHELL, J.F. 1988. Consumer control of lake productivity. Bioscience 38: 764-769. 

CARPENTER, S.R., KITCHELL, J.F., HODGSON, J.R., COCHRAN, P.A., ELSER, J.J., ELSER, M.M., LODGE, D.M., 

KRETCHER, D., HE, X., and VON ENDE, C.N. 1987. Regulation of Rakeprimary productivity by food web 

structure. Ecology 68: 1863-1876. 

DAVIDSON, ED. and BARBOUR, M.G. 1977. Germination, establishment, and demography of coastal bush lupine 

(Lupinus arboreus') at Bodega Head, California. Ecology 58: 592-600. 

DE BENEDICTIS, J.A., WAGNER, D.L., and WHITFIELD, J.B. 1990. Larval hosts of the microlepidoptera of the San 

Bruno Mountains, Calitbrnia. Atala 16: 14-35. 

DINKELAKER, B., ROMHELD, V., and MARSCHNER, H. 1989. Citric acid excretion and precipitation of calcium citrate 

in the rhizosphere of white lupine (Lupinus albus L.). Plant. Cell. Environ. 12: 285-292. 

EHRLICH, PR. and RAVEN, P.H. 1964. Butterflies and plants: a study in coevolution. Evolution 18: 586-608. 

FOX, LR. and MORROW, P.A. 1981. Specialization: species property or local phenomenon? Science 221: 887-893. 

HAIRSTON, N.G., JR and HAIRSTON, N.G., SR. 1993. Cause-effect relationships in energy flow, trophic structure, and 

interspecific interactions. Am. Nat. 142: 379-411. 

HAIRSTON, N.G., SMITtt, EE., and SLOBODKIN, L.B. 1960. Community structure, population control, and competition. 

Am. Nat. 94:421-424. 

HARTMANN, T. 1992. Alkaloids, p. 79-116. In Rosenthal, G.A. and Berenbaum, M.R., eds. Herbivores: Their Interactions 

with Secondary Plant Metabolites. Second edition, vol. I. Academic Press, San Diego. 

HAWKINS, B.A. 1992. Parasitoid-host food webs and donor control. Oikos 65: 159-162. 

HAY, M.E. and STEINBERG, ED. 1992. The chemical ecology of plant-herbivore interactions in marine versus terrestrial 

communities, p. 372-408. In Rosenthal, G.A. and Berenbaum, M.R., eds. Herbivores: Their Interactions with 

Secondary Plant Metabolites. Second edition, vol. I. Academic Press, San Diego. 

130

HUNTER, M.D. and PRICE, P.W. 1992. Playing chutes and ladders: heterogeneity and the relative roles of bottom-up and 

top-down forces in natural communities. Ecology 73: 724-732. 

KAYA, H.K. t990. Soil ecology, p. 93-11 I. In Gaugler, R. and Kaya, H.K., eds. Entomopathogenic Nematodes in Biological 

Control. CRC Press, Boca Raton. 

LANE, G.A., SUTHERLAND, O.R.W., and SKIPP, R.A. 1987. Isoflavonoids as insect feeding deterrents and antifungal 

components from root of Lupinus augustijblius. J. Chem. Ecol. 13" 771-783. 

MARTIN, M.M. 1991. The evolution of cellulose digestion in insects. Phil. Trans. R. Soc. Lond. B. 333: 281-288. 

MCNAUGHTON, S.J., RUESS, R.W., and SEAGLE, S.W., 1988. Large mammals and process dynamics in African ecosystems. 

Bio Science 38 794-800. 

OKSANEN, L., FRETWELL, S.D., ARRUDA, J., and NIEMELA, P. 1981. Exploitation ecosystems in gradients of primary 

productivity. Am. Nat. 118" 240-261. 

OPLER, RA. 1968. Unusual numbers of Hepialus sequoiolus Behrens in Sonoma County. Pan. Pac. Entomol. 44: 83. 

PRICE, RW. 1991. The plant vigor hypothesis and herbivore attack. Oikos 62:244-251. 

PRICE P.W. 1992. Plant resources as the mechanistic basis for insect herbivore population dynamics, p. 139-173. In Hunter, 

M.D., Ohgushi, T., and Price, RW., eds. Effects of Resource Distribution on Animal-Plant Interactions. Academic 

Press, San Deigo. 

SINGER, M.C. and PARMESAN, C. 1993. Sources of variations in patterns of plant-insect association. Nature 361:251- 

253. 

STRONG, D.R. 1992. Are trophic cascades all wet? Differentiation and donor-control in speciose ecosystems. Ecology 73: 

747-754. 

STRONG, D.R., MARON, J.L., and CONNORS RG. 1996. Phmt competition, endogenous defenses, and natural enemies in 

terrestrial food webs: an example from the bush lupine community. In Polls, G. and Winemiller, K., eds. Food Webs: 

integration of patterns and dynamics. Chapman and Hall, NY. 472 p. 

STRONG, D.R., MARON, J.L., HARRISON, S., CONNORS, RG., and JEFFRIES, R.L. 1995. High mortality, fluctuation 

in numbers, and heavy subterranean insect herbivory in bush lupine, Lupinus arboreus, ms to be submitted to 

Oecologia. 

WAGNER, D.L. 1986. The biosystematics of Hepialus F.s. lato with special emphasis on the californicus-hectoides species 

group. PhD. Thesis in Entomology, University of California, Berkeley. 

WILLIAMS, F.X. 1905. Notes on the life history of Hepialus sequoiolus Behrens. Entomol. News. 16: 283-287. 

WHITE, T.C.R. 1993. The Inadequate Environment: Nitrogen and the Abundance of Animals. Springer-Verlag, New York. 

WINK, M. 1992. The role of quinolizidine alkaloids in plant-insect interactions, p. 131-166. In Bernays, E., ed. Plant-Insect 

Interactions. vol. IV. CRC Press, Boca Raton, FI. 

WINK, M. and WITTE, L. 1984. Turnover and transport of quinolizidine alkaloids. Diurnal fluctuations of lupanine in the 

phloem sap, leaves, and fruits ofLupinus albus L. Planta 161: 519-524. 

131

RESISTANCE OF HYBRID AND PARENTAL WILLOWS TO HERBIVORES: 

HYPOTHESES AND VARIABLE HERBIVORE RESPONSES OVER 3 YEARS 

ROBERT S. FRITZ l, BERNADETTE M. ROCHE I, and COLIN M. ORIANS 2 

_Department of Biology, Vassar College, Poughkeepsie, New York 12601, USA 

2Department of Biology, Williams College, Williamstown, Massachusetts 01267, USA 

INTRODUCTION 

Analysis of herbivory on hybrid plants in natural populations has recently gained attention (Whitham 1989, Keim et 

al. 1989, Boecklen and Spellenberg 1990, Paige et al. 1991, Whitham et al. 1991, Aguilar and Boecklen 1992, Fritz et al. 

1994, Whitham et al. 1994), and it may be important for understanding plant-herbivore interactions (Whitham 1989). To 

assess the significance of interspecific hybridization of plants for plant-herbivore interactions, it will be important to know 

how interspecific hybridization of plants influences resistance of hybrid plants in relation to resistance of pure parents. It has 

been hypothesized that hybrid plants may act as an "ecological sink" for herbivorous insects (Whitham 1989). In mixed 

stands of parental and hybrid plants, this hypothesis predicts that a large fraction of the herbivore population would reside on 

hybrid individuals (Drake 1981, Whitham 1989, Floate et al. 1993). Whitham (1989) referred to highly susceptible hybrid 

poplars as being an ecological sink for a galling aphid, but in terms of metapopulation dynamics, hybrid plants could function 

as a "source" rather than a sink (e.g., Ericson et al. 1993). Hybrid plants have also been hypothesized to act as "evolutionary 

sinks" to pests (Whitham 1989). The "evolutionary sink" idea predicts that if herbivore fitness was higher on hybrids 

compared to their fitness on parental plants, there would be selection for specialization on the hybrids, perhaps preventing 

selection tbr herbivore virulence on the parental species. If so, then hybrid plant-herbivore interactions may be critical to 

understanding why plants seem to maintain pest resistance in the face of a greater evolutionary potential for virulence in 

pests. Finally, Floate and Whitham (1993) suggest that hybrid plants may facilitate host shifts of herbivores onto new plants, 

because they could act as a selective "bridge" favoring adaptation of an herbivore to previously non-host species. Tests of 

these hypotheses will require knowing the general patterns of herbivore responses to hybrid versus pure parental species. 

Hypotheses 

Fritz et al. (1994) proposed several alternative hypotheses of phytophage response to hybrid plants. Those hypotheses 

assume that pure parent species are being compared to F1 hybrids. The reason to focus on F1 hybrids is that this 

comparison could suggest: (1) how resistance mechanisms are inherited from parental species, and (2) how these traits may 

be expressed upon backcrossing. The Additive hypothesis (Fig. 1A) predicts that hybrids do not differ from the mean of the 

resistances of the two parents (i.e., the midparent value). Thus, F1 hybrids are intermediate between the resistances of the 

parental species, which suggests that hybrid resistances are due to the additive inheritance of resistance traits from both 

parents. 

The second hypothesis is the Dominance hypothesis (Fig. 1B). If hybrid resistance differs significantly from the 

mean resistance of both parents but does not differ significantly from that of one of the parents, the Dominance hypothesis 

would be supported. Hybrid resistance could resemble that of either the more resistant or the more susceptible parent. 

(Dominance in this context refers to phenotypic similarity between a parent and hybrids, not necessarily genetic dominance, 

though it may imply dominant inheritance of resistance traits.) Herbivore densities could be intermediate between the 

Additive and Dominance patterns, which would support an hypothesis of partial dominance. 

The third hypothesis is the Hybrid Susceptibility hypothesis (the Hybrids-as-Sinks hypothesis of Whitham 1989; see 

also Keim et al. 1989, Boecklen and Spellenberg 1990). This hypothesis predicts higher herbivore densities and/or higher 

herbivore pertbrmance on hybrids compared to parental taxa (Fig. 1C). The Hybrid Susceptibility hypothesis can be distinguished 

from the Dominance hypothesis in that the susceptibility of the hybrid must exceed that of the most susceptible 

parent. 

Mattson, W.J., Niemelfi, R, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


132

A° ADDITIVE C. HYBRID SUSCEPTIBILITY 

50 50 

)- 40 >- 40 

03 03 

Z Z 

w 30 w 

Q 0 30 

uJ LU 

O 20 O 20 

n- nw 

10 w 

I _: 10 

SPECIES 1 HYBRID SPECIES 2 SPECIES 1 HYBRID SPECIES 2 

B. DOMINANCE D. HYBRID RESISTANCE 

50 50 

>- 40 >'- 40 

F- __. 

03 

z z w 30 

w 30 o 

Q 

w LLI 

n-" n- 

O 2O O 20 

n- n- 

w 10 

w 

:z: 

10 

I 

0 "- 0 

SPECIES 1 HYBRID SPECIES 2 SPECIES 1 HYBRID SPECIES 2 

Figure I. Hypothesized patterns that would result from the Additive hypothesis (A), the Dominance hypothesis (B), the Hybrid 

Susceptibility hypothesis (C), and the Hybrid Resistance hypothesis (D). Imaginary densities (Susceptibility) are plot- 

ted on the Y-axis for each parent species and their FI interspecific hybrid. 

The final hypothesis, the Hybrid Resistance hypothesis, predicts that hybrid plants will be more resistant than either 

parent, resulting in lower herbivore densities on hybrid plants (Fig. ID). This hypothesis requires that hybrids be more 

resistant than the most resistant parent. A study by Boecklen and Spellenberg (1990) suggests support for this hypothesis for 

herbivore communities in two oak hybrid zones. Hybrids supported lower densities and diversities of herbivores than the 

mean of the parental species, but they did not distinguish between the fit of their data to the Hybrid Resistance hypothesis or 

the Dominance hypothesis. We present and discuss these hypotheses in genetic terms since support of any of the hypotheses 

may suggest probable underlying mechanisms of inheritance of resistance. 

It is not clear a priori why one of these hypotheses should be more likely than any others. If the Hybrid Susceptibility 

hypothesis is prevalent and herbivory reduces plant fitness, then herbivores could be important in maintaining distinct plant 

species by limiting the fitness of hybrids. Boecklen and Spellenberg (1990) found evidence for the Hybrid Resistance and for 

133 

m

our Additive hypothesis. Fritz et al. (1994) found support for the Additive, Dominance, and Hybrid Susceptibility hypotheses 

among herbivore and a pathogen species on hybrid and parental willows. The purposes of this paper are to investigate the 

responses of I 1 herbivore species from different guilds to parental and hybrid willows over three consecutive years, and to 

determine if herbivore responses to hybrid plants vary over time. This paper will show that different herbivores, even those 

in the same guild, vary dramatically in their response to hybrid versus parental taxa, but that most species, when considered 

over time, fit the Additive Hypothesis. 

METHODS AND MATERIALS 

The System 

This system has several advantages for studying herbivore-hybrid plant interactions. The two willow species that 

hybridize often co-occur in the same habitats with their hybrids, eliminating, to a certain extent, confounding environmental 

variation across naturally occurring hybrid zones with herbivore resistance variation among hybrids and parents. Willows 

have a number of herbivore species in different guilds, including gall-formers, leaf miners, leaf tiers, and chewers, that are 

welt characterized and that have been studied extensively on one of the willow species in this system. 

Salix sericea Marshall and Salix. eriocephala Michx. co-occur in swamps and along streams in central New York. S. 

sericea is a 0.5-4-m-high shrub that has lanceolate leaves with densely sericeous hairs on the lower leaf surface. Stipules are 

small, lanceolate, and usually absent from older leaf nodes (i.e., stipules are deciduous). This species occurs predominantly 

in swamps from Canada south through the northeastern U.S. and along the Appalachian range to Georgia (Argus 1986). S. 

eriocephala is a shrub that may reach 6 m in height, and its leaves are lanceolate to narrowly oblong and glabrous beneath. 

Stipules are large, persistent, half ovate, and half cordate at the base. It frequently occurs along streams, and its range 

extends from Canada south as far as Virginia and west to Missouri (Argus 1986). The ranges of the two species are broadly 

sympatric, and at the study site both species co-occur and intermingle. Species can be easily distinguished in the field based 

on leaf, stipule, and bud characteristics (Argus 1986). 

These willow species hybridize to form plants that are usually distinctive from each parent. Large, persistent, halfovate 

stipules (a S. eriocephala trait) and a sericeous (hairy) lower leaf surface (a S. sericea trait) serve as morphological 

markers to identify S. sericea x S. eriocephala hybrids in the field (Fritz et al. 1994). The natural hybrids at the study site do 

not appear to be extraordinary; S. eriocephala and S. sericea have been reported to hybridize throughout their range (Argus 

1974, t986; Mosseler and Papadopol 1989). Flowering phenology may explain the occurrence of hybridization between S. 

sericea and S. eriocephala (Argus 1974, Mosseler and Papadopol 1989), which differ by only a few days in the onset of 

flowering (foliage phenology is also similar) (Fritz, pets. obs.). 

Many herbivores attack the parental and hybrid willows. Each herbivore species attacks both parental willow species 

and hybrids. Gall-forming sawflies (Hymenoptera: Tenthredinidae) include Phyllocolpa nigrita, Phyllocolpa sp. nov., 

Phyllocolpa terminalis, and Pontania sp. (leaf galls). Gall-forming flies (Diptera: Cecidomyiidae) include the stem gallers 

Rabdophaga rigidae, the beaked willow gall; R. salicisbrassicoides, the willow cabbage gall; and a leaf galler, Iteomyia 

salic(folia. The other common leaf galler species is the gall mite Aculops tetanothrix (Acarina: Eriophyidae). There are two 

species of leaf miners, Phyllonorycter salicifoliella and Phyllocnistis sp. (Lepidoptera: Gracillariidae), that commonly form 

mines on willow leaves. 

There are two frequently seen leaf folder species (Lepidoptera) that have distinctive leaf folds, but have not yet been 

successfully reared or identified. One species (LF) folds over and sews the leaf margin to the lamina with silk. The other 

species (V) forms a tube of the leaf tip by sewing the edges of the leaf blade together. 

Methods 

These studies were performed from 1991 to 1993 at the Sosnowski site 3 km west of Milford, NY, along County 

Route 44. All plants are individually marked at the study site. Censuses of herbivores species on 14 S. eriocephala, 29 S. 

sericea, and 16 hybrids were conducted from late July to early August 1991. In 1992, we censused 20 S. eriocephala, 36 S. 

sericea, and 39 hybrids, and in 1993 we censused 40 S. eriocephala, 109 S. sericea, and 38 hybrids. Mostly hybrid plants 

known to be Fl-types (Fritz et al. 1994) were included in the hybrid category. Plants known to be backcrosses from RAPD 

were not included in this analysis. Some hybrid plants added to the census group after 1991 have not been examined for their 

RAPD genotype, and it is therefore possible that they are something other than Fl-types (Fritz et al. 1994). 

134

Galls, leaf miners, or leaf folds were counted on 50 shoots per plant for hybrids and S. eriocephala and on 300 shoots 

for S. sericea in 1991 only. Data are expressed as number of herbivores per 300 shoots. Because leaf folds of Phyllocolpa 

r_igritaand Phyllocolpa sp. nov. were not distinguished on S. sericea in 1991, we combined them for analyses in all years. 

All other species densities are considered separately. Species densities were considered statistically independent among 

plants within taxa for all years (Fritz et al. 1994), and species-wide significance tests rather than table-wide significance tests 

were applied. ANOVAs were performed for each year separately. We first performed a multivariate ANOVA, using loge 

transformed densities of alt herbivore species, and then we performed univariate ANOVAs for each species separately. 

Orthoganol contrasts were performed to test for the fit of each herbivore species to the hypotheses. We performed 

only two orthoganol contrasts to test the significance of herbivore densities on willow taxa (species) to avoid overparameterization 

of the analysis, since only three taxonomic groups were present (SAS Institute 1985). To test the Additive hypothesis, 

we tested the significance of the hybrid-midparent contrast (Contrast A) using sequential Bonferroni analysis with specieswide 

significance at P < 0.05/2 = 0.025 (Rice 1989). If this test was significant, we rejected the Additive hypothesis and then 

tested the contrast between herbivore density on hybrids and density on the numerically closest parent (Contrast B) (P < 0.05/ 

1 = 0.05). This contrast tested if there was a significant departure from the Dominance hypothesis. If the Dominance 

hypothesis was rejected, we inspected the means to determine if the Hybrid Susceptibility hypothesis (density on hybrids 

exceeded that of the highest parent) or the Hybrid Resistance hypothesis (density on hybrids was lower than the lowest 

parent) was supported, or if the density of hybrids was intermediate between that of one parent and the midparent value. This 

last possibility would support a partial Dominance hypothesis. 

RESULTS 

Multivariate ANOVAs showed highly significant contrasts between herbivore densities on hybrids and the mean 

herbivore densities of parent species for all three years (Table 1). This is a rejection of the Additive hypothesis overall. 

Further multivariate contrasts could not be made since the test of the Dominance hypothesis requires comparison to the 

numerically nearest parent, which would not be the same for all herbivores. In 1991, five herbivore species had significant 

hybrid vs. parents contrasts (Contrast A) at P < 0.025, and two herbivore species had marginally significant hybrid vs. parents 

contrasts at P < 0.05 (Table 1). In 1992, four species had significant hybrid vs. parents contrasts (Table 1). Three species that 

had significant contrasts in 1991 did not have significant contrasts in 1992. In 1993, five species had significant hybrid vs. 

parent contrasts (Table 1). Overall, then, there appear to be differences among the years in the species of herbivores that had 

significant departures from the Additive hypothesis. 

To illustrate the year-to-year differences in more detail and to draw conclusions about which hypotheses mentioned 

above are supported by each species, we present the separate results for each year for each species, which included all 

censused plants. 

Both contrasts for Phyllonoo, cter salicifoliella were significant in 1991 and 1993 (Table 1), and the means support 

for the Hybrid Susceptibility hypothesis (Table 2). In I992, neither contrast was significant, which supported the Additive 

hypothesis. The other leaf miner Phyllocnistis sp. also showed differences between years. In 1991 and 1993, the first 

contrast was not significant, supporting the Additive hypothesis (Tables 1 & 2). Fritz et al. (1994) concluded that this species 

supported the Hybrid Susceptibility hypothesis in 1991, because of the marginally significant contrast. In 1992 the Dominance 

hypothesis was supported with densities on S. eriocephala and hybrids being equal and lower than on S. sericea (Table 

2). 

Among species of the leaf galling guild, Phyllocolpa termMatis showed consistent results in each year (Tables I & 

2). In all cases the Hybrid Susceptibility hypothesis was supported. Phyllocolpa spp. supported the Additive hypothesis in 

1991 and 1993 (Tables 1 & 2). In each year, including 1992, densities on hybrids were intermediate between the two parental 

species. In 1992, the first contrast was significant and the second contrast was not significant, supporting the Dominance 

hypothesis (Tables t & 2). Pontania sp. supported the Additive hypothesis in 1991 and 1992, but the Dominance hypothesis 

was supported in 1993 (Tables 1 & 2). Iteomyia salicifolia supported the Hybrid Susceptibility hypothesis in 1991 and 1993 

(Tables 1 & 2). In each year, including 1992, densities on hybrids were greater than on either parental species. In 1992, the 

first contrast was not significant at P < 0.025 (although it was marginally significant at P < 0.05), thereby supporting the 

Additive hypothesis, even though the mean density on hybrids was more than twice as high as on S. eriocephala (Table 2). 

The last leaf galling species, A. tetanothrix, supported the Dominance hypothesis in 1991 and 1993 (Tables 1 & 2). In both of 

135

Table 1. --Summary of multivariate and univariate ANOVAs of herbivore densities on S. sericea, S. eriocephala, and hybrid 

plants for 1991, 1992, and 1993. F-values and degrees of freedom are shown for the univariate tests. Contrast A was 

considered significant if its P value was less than the Bonferroni criterion of P < 0.025. F-values for the multivariate 

ANOVAs are based on Wilks' lambda (other tests gave similar results). Herbivore species are grouped by feeding 

guild. 

YEAR 1991 1992 1993 

CONTRAST A B A B A B 

F(1,61) F(1,61) F(1,92) F(1,92) F(1,184) F(1,184) 

Multivariate 11.22? -- 11.28t -- 7.04? -- 

Leaf Miners 

Phyllonorycter 

salicifoliella 10.31"** 5.94* 2.06 0.68 7.27** 5.82*** 

Phyllocnistis sp. 4.24* 4.67* 13.02? 0.22 0.45 5.35* 

Leaf Gallers 

Phyllocolpa terminalis 18.35t 10.13"** 16.09? 10.26'** 36.91t 27.93? 

Phylocolpa spp. 2.18 0.31 6.13"* 0.03 1.66 11.75? 

Pontania sp. 0.87 8.60*** 4.28* 0.16 6.86** 1.82 

lteomyia salicifolia 13.04? 13.50t 4.32* 0.69 16.95? 11.93? 

Aculops tetanothrix 13.97t 2.27 30.23t 7.62*** 11.88t 2.34 

Leaf Folders 

Leaf folder-LF 4.49* 1.74 0.73 0.40 0.92 0.52 

Leaf folder -V 16.99? 9.30*** 0.44 2.36 0.40 0.47 

Stem Gallers 

Rabdophaga rigidae 1.99 0.21 0.50 0.47 2.13 0.22 

salicisbrassicoides 3.24 0.06 4.77* 0.11 0.13 0.44 

Total Herbivores 0.42 9.47*** 1.30 0.32 11.987 4.62* 

* - P < .05, ** P < .025, *** P < .01, # - P < .001 

these years, densities on S. sericea and hybrids were equal and lower than on S. eriocephala. In 1992, however, the Hybrid 

Resistance hypothesis was supported. Mite gall density on hybrid plants was significantly lower than on S. sericea. It should 

be noted, however, that in all three years the numbers of mite galls were lower on the hybrid plants than on the nearest parent. 

For the leaf folding guild, LF supported the Additive hypothesis in all years (Tables 1 & 2), but the first contrast was 

marginally significant in 1991 (Table 1). Species V supported the Hybrid Susceptibility hypothesis in 1991, but the Additive 

hypothesis was supported in 1992 and 1993 (Tables 1 & 2). 

In the stem galling guild, both R. rigidae and R. salicisbrassicoides supported the Additive hypothesis consistently in 

each year (Tables 1 & 2), but densities of R. salicisbrassicoides in 1992 were marginally significant for the first contrast, 

nearly supporting the Dominance hypothesis. 

Seven of the 11 herbivore species showed significant changes in the hypotheses that were supported between years. 

In five of the seven cases, densities of the herbivores had declined between the years of the change; in two cases densities had 

increased or there was little change. Lower herbivore densities, therefore, might have contributed to the changes in which 

hypotheses were supported. 

136

Table 2.--Summary of the hypotheses supported by each species based on the results in Table 1 and inspection of mean 

herbivore densities among the willow species in each year (Fritz et al. in prep.). When the Dominance hypothesis 

was supported the two equal taxa are indicated. In all cases they were more resistant than the other taxon. Herbivore 

species are grouped by feeding guild. 

Leaf Miners 

1991 1992 1993 

Phyllonorycter salicifoliella Susceptibility Additive Susceptibility 

Phyllocnistis sp. Additive Dominance (E=ES) Additive 

Leaf Gallers 

Phyllocolpa terminalis Susceptibility Susceptibility Susceptibility 

Phyllocolpa spp. Additive Dominance (ES=S) Additive 

Pontania sp. Additive Additive Dominance (E=ES) 

lteomyia salicifolia Susceptibility Additive Susceptibility 

Aculops tetanothrix Dominance (ES=S) Resistance Dominance (ES=S) 

Leaf Folders 

Leaf folder-LF Additive ,_dditive Additive 

Leaf folder -V Susceptibility Additive Additive 

Stem Gallers 

Rabdophaga rigidae Additive Additive Additive 

R. salicisbrassicoides Additive Additive Additive 

Total Herbivores Additive Additive Susceptibility 

To visualize the combined response of herbivores to hybrids vs. parents, we summed densities of herbivores on each 

plant and reanalyzed the responses. When mites were included in the analysis, there was no significant hybrid vs. parent 

contrast in 1991 and 1992 (Tables 1 & 2, Fig. 2), but the Hybrid Susceptibility hypothesis was supported in 1993. 

DISCUSSION 

The main conclusions of this study are that: (1) there is evidence that different herbivores support different hypotheses 

regarding the effects of interspecific hybridization on plant resistance; and (2) there is year-to-year variation in which 

hypothesis is supported by a given herbivore species. These results suggest the possibility that different resistance mechanisms 

are inherited differently in hybrid plants, or that different herbivores have different responses to the same resistance 

mechanism in the hybrid compared to the parent. The results also suggest the possibility that environmental variation affects 

the hybrid and parental willows differently between years, which alters susceptibility to herbivores. 

The hypothesized effects of interspecific hybridization on plant resistance have as an assumption that hybrids are 

Fl's. The plants that were censused were predominantly Fl's based on RAPD analysis (Fritz et al. 1994), but six plants were 

most closely identified as backcrosses to S. sericea. Known backcross plants were removed from this analysis, but it is 

possible that some backcrosses are included among the hybrid plants added to the censuses in 1992 and 1993. We have not 

yet analyzed these new hybrid plants using our RAPD technique to know their exact genetic status. Fritz et al. (1994) have 

shown that morphological characters are not always reliable for distinguishing S. eriocephala from hybrids, and F 1-type 

hybrids are indistinguishable from backcross plants using morphological traits (Fritz, unpublished data). With these comments 

in mind, we will proceed to interpret the results of this study in relation to the hypotheses (Fig. 1). 

137

450 

TOTAL HERBIVORES 

400 [] 1991-] 

[] 1992r 

350 ID993/ 

....... 

300 

250 

03 

F,, O. 

I 

IIEI 150 

100 

50 _,_,_ 

1/...-/,,, J 

O. _.',. ..... 

ERIOCEPHALA HYBRID SERICEA 

SPECIES 

Figure 2.--Density (+ 1 SE) of total herbivore numbers among S. eriocephala, S. sericea and their interspecific hybrids and 

among years. Table 2 indicates which hypotheses were supported by total herbivore numbers. 

The Additive hypothesis received the greatest support in each year (7 of 11 species, Table 2). Herbivore densities oa 

hybrids are often intermediate between the densities on the two parental species. This finding is consistent with an Additive 

model of inheritance of resistance traits. Other studies also provide support for the Additive hypothesis. Soetens et al. 

(1991) found for a species of Pontania on European willows and their hybrids that gall numbers were intermediate on 

hybrids (supporting the Additive hypothesis). Studies on sawflies on birches (Hanhim/iki et al. 1994), beetles on elms (Hall 

and Townsend 1987), and beetles on willows (Soetens et al. 1991, and Fritz unpublished), and various species on oaks 

(Aguilar and Boecklen 1992) also indicate that hybrids are often intermediate in suitability for herbivore growth. Hanhimfiki 

et al. (1994) have studied herbivore performance on hybrid and parental birches that grew in a common plantation in Finnish 

Lal_land. Of the 14 herbivore species studied (l 3 sawflies and a moth), I I species had intermediate growth rates on hybrid 

leaves compared to parental leaves. 

Monogenic and polygenic resistance traits where allele effects on traits are additive will have half the dosage of 

alleles, on average, in hybrids compared to parents, unless parents have the same defense mechanisms determined by the 

same loci. There is evidence from willows supporting additive inheritance of phenolic glycosides, tannins, and morphologi- 

cal defenses (Meier et al. 1989, Soetens et al. 1991, Nichols-Orians and Fritz, unpublished data). How the altered dosage of 

defensive traits influences herbivores will depend on the sensitivity of the herbivores to the resistance trait and the presence 

of other resistance traits inherited from species that might not normally be host plants of the herbivore. A consequence of 

additive inheritance of resistance is that backcrossing should result in a reconstitution of defense, rather than further break- 

down of resistance, as hypothesized by Whitham (1989) and Paige and Capman (1993). The resistances of backcross 

individuals should more closely resemble that of the recurrent parent as more parental genes are reincorporated into the 

genome. Whitham et al. (1994) found evidence supporting this hypothesis in their work on eucalypts. When F1 hybrids 

were compared with backcross plants and parents, often backcross plants were intermediate between FI hybrids and parents. 

The Dominance hypothesis received support from Aculops mites, Phyllocnistis sp. Phyllocolpa spp., and Pontania sp. 

in at least one year (Table 1). In this hypothesis, herbivore densities on hybrids resemble those of one parent but differ 

significantly from that on the other parental species. Hanhim_iki et al. (1994) found for three species, a moth and two 

sawflies, on the Finnish birches that performance on hybrids did not differ from that of one parent but did differ from the 

other parent, supporting the Dominance hypothesis. The Dominance hypothesis was proposed by Fritz et al. (1994), and it is 

particularly plausible because of the inheritance of defensive chemicals in hybrids of several crop and wild plants. For Lotus 

(O'Donoughue et al. 1990), Nicotiana (Huesing et al. 1989), and Papaver (Levy and Milo 1991), chemical defense mecha- 

nisms have been shown to be inherited as dominant traits in hybrids. Zangerl and Berenbaum (pers. comm.), in a review of 

138

the inheritance patterns of secondary chemicals, found that 13 of 32 compounds had dominant inheritance. Given dominant 

inheritance, if the parent possessing the high levels of defensive chemical is resistant, then F1 hybrids should also be resistant. 

Paige and Capman (1993) showed that hybrid breakdown in resistance to Pemphigus betae occurs only in backcrosses 

of Populus hybrids to the susceptible parent, and F1 hybrids are as resistant as the resistant P. fremomii. This illustrates the 

Dominance hypothesis in the cottonwood-aphid system. 

There was also considerable support for the Hybrid Susceptibility hypothesis. Three herbivore species consistently 

supported the Hybrid Susceptibility hypothesis, and over three years 9 of 33 cases supported this hypothesis. The Hybrid 

Susceptibility hypothesis was also supported by the analysis of total herbivore densities in 1993 (Table 1, Fig. 2). Mosseler 

(pers. comm.) reports that in his garden of F1 hybrids and parents of seven Salix species, hybrids were highly susceptible to 

Melampsora rust, but parents were immune. Fritz et al. (1994) has also found for Melampsora rust that hybrids were much 

more susceptible than either parental species. This observation shows that Fl's appear to show breakdown of resistance. 

The Hybrid Resistance hypothesis was supported only by the Aculops gall mite in 1992. Density of mites was lower 

on hybrids and on either parent. Boecklen and Spellenberg (1990) found some evidence for this hypothesis in their study of 

leaf mining and leaf galling herbivores on oak. A problem with the support for the Hybrid Resistance hypothesis from that 

study is that many herbivore species were combined into feeding guilds for analysis and they did not explicitly test for a 

significant departure from the Dominance hypothesis. 

The results of this study suggest that interspecific hybridization of plants has a variety of effects on resistance of 

hybrids and it does not lead only to a pattern of breakdown of resistance (i. e., support of the Hybrid Susceptibility hypothesis). 

The various herbivore responses suggest that different resistance traits of plants have different mechanisms of inheritance 

and/or that different resistance factors could affect herbivores differently. An important question is: How can the 

hypothesized herbivore responses be explained by the known patterns of inheritance of resistance traits? The Additive and 

Dominance hypotheses both have fairly straightforward explanations. If resistance is dosage-dependent, then an Additively 

inherited trait should lead to support for the Additive hypothesis. A dominantly inherited trait would be fully expressed in 

hybrids and should result in a pattern of herbivore resistance that supported the Dominance hypothesis. Backcrossing to the 

susceptible parent, as seems to have occurred in cottonwoods, could lead to the loss of one or more dominant resistance 

genes, thereby resulting in Hybrid Susceptibility of the backcross progeny. Paige and Capman (1993) suggest that there is 

more than one dominant gene involved in resistance to aphids in the cottonwood system. 

Additively inherited traits could also result in herbivore densities that support the Hybrid Susceptibility hypothesis. 

Herbivores that require a threshold amount of a chemical defense could have higher densities on hybrids (Hybrid Susceptibility) 

if the dosage in hybrids was less than the threshold amount. Hybrid Resistance could result from a mechanism of 

dominance inheritance. Hybrids with dominant inheritance of two different resistance traits from each parent would possess 

complete expression of two different resistance mechanisms. If particular herbivores are negatively affected by both resistance 

mechanisms, hybrids could be more resistant than either parent (i.e., Hybrid Resistance). It also seems possible that if 

two additively inherited resistance mechanisms interacted synergistically to affect herbivores Hybrid Resistance could also 

result. These alternative mechanisms suggest that various outcomes of interspecific hybridization on resistance are likely and 

that mechanisms of inheritance of resistance and their effect on herbivores need to be investigated in hybrid systems. 

The year-to-year variation in support for particular hypotheses was an unexpected result of this analysis. Seven of 

the 11 species had at least one change in which hypothesis was supported between adjacent years. These changes could be 

considered either: (1) random fluctuations in herbivore numbers on hybrids and parents or accidents of sampling, (2) 

consequences of temporarily lower population sizes that made differences difficult to detect, or (3) real year-to-year changes 

in the expression of resistance of hybrids and/or parent species due to variation in environmental factors. Point 1 is always 

possible, but it is impossible to discount using the census data. A number of shifts in herbivore response occurred between 

1991 and 1992 and then back to 1991 patterns in 1993. This may have something to do with the decreased abundance of 

herbivores in 1992 (unpublished results). Lower numbers could have made differences difficult to detect. While point 2 is 

possible, it would be wrong to attribute the shifts in support for hypotheses to this factor alone. Shifts in support for hypotheses 

may be real changes in the relative expression of resistance between hybrid and parental plants. Environmental variation 

(e. g., solar radiation, temperature, nutrients, etc.) is known to affect the expression of resistance among genotypes of many 

plant species. It therefbre seems likely that changes in environment could affect resistance of hybrid and parent species of 

plants in nature. There were some substantial shifts in the weather among the years of this study. The summer of 1992 was 

cool and rainy compared to normal in the northeast, whereas in 1993 there were some long periods of high temperatures and 

139

drought (pers. obs.). These factors might have contributed to the variation in resistance of hybrids. However, experiments 

wilt be required to test if tile relative resistance of hybrids shifts with manipulated changes in specific environmental factors. 

Guild membership was not a reliable indicator of similar responses of herbivores to hybrids and parents. Among the 

leaf galling guild, all four hypotheses were supported in at least 1 year. Leaf miners supported two different hypotheses 

(Additive and Hybrid Susceptibility) considering the 3 years together (Table 2). These results indicate that combining the 

densities of two or more species in the same guild for hybrid-parent comparisons will obscure the underlying patterns of 

herbivore densities on hybrids and parents. In this analysis, numbers of Phyllocolpa nigrita and Phyllocolpa sp. nov. were 

combined in 1992 and 1993, because densities of these species had not been distinguished on S. sericea of the willow species 

in 1991. Therefore, the conclusions drawn from these species should be considered with caution. 

SUMMARY 

We studied herbivory of two species of willows (Salix sericea and S. eriocephala) and their interspecific hybrids to 

test four alternative hypotheses concerning the effects of hybridization on plant resistance. Individually marked plants were 

identified using morphological traits in the field, and RAPD band analysis was used to verify the genetic status of some 

parental and hybrid plants. The densities of 11herbivore species were compared between 2 parents and their hybrids in the 

field. We found the most support for the Additive hypothesis and the Hybrid Susceptibility hypothesis over the three years. 

We found some evidence for the Dominance hypothesis, and one species in one year supported the Hybrid Resistance 

hypothesis. Guild membership was not a good predictor of similar responses of species to hybrid versus parental plants. 

This study demonstrates the diversity of responses of phytophages in response to interspecific hybridization, and indicates the 

presence of year-to-year variation, which might indicate the influence of environmental variation on the expression of hybrid 

resistance. 


We would like to thank Len and Ellie Sosnowski for allowing us to work on their property. We thank S. Kaufman, 

N. Murphy, P. Zee, M. Momot, A. Samsel, A. Dao, J. Hama, A. Samsel, and G. Sylvan for help in the field. This research 

was supported by the National Science Foundation grants BSR 89-17752 to R. S. Fritz and DEB 92-07363 to R. S. Fritz and 

C. M. Orians, and by the Vassar College General Research Fund. 


AGUILAR, J.M. and BOECKLEN, W.J. 1992. Patterns of herbivory in the Quercus grisea x Quercus gambelii species 

complex. Oikos 64: 498-504. 

ARGUS, G.W. 1974. An experimental study of hybridization and pollination in Salix (willows). Can. J. Bot. 52: 1613-1619. 

ARGUS, G.W. 1986. The genus Salix (Salicaceae) in the southeastern United States. Syst. Bot. Monogr. 9: 1-170. 

BOECKLEN, W.J.and SPELLENBERG, R. 1990. Structure of herbivore communities in two oak (Quercus spp.) hybrid 

zones. Oecologia 85: 92-100. 

DRAKE, D.W. 1981. Reproductive success of two Eucalyptus hybrid populations. II. Comparison of predispersal seed 

parameters. Aust. J. Bot. 29: 37-48. 

ERICSON, L., BURDON, J.J., and WENNSTROM, A. 1993. Inter-specific host hybrids and phalacrid beetles implicated in 

the local survival of smut pathogens. Oikos 68: 393-400. 

FLOATE, K.D., KEARSLEY, M.J.C., and WHITHAM, T.G. 1993. Elevated herbivory in plant hybrid zones: Chrysomela 

cot_uens, Populus and phenological sinks. Ecology 74: 2056-2065. 

140

FLOATE, K.D. and WHITHAM, T.G. 1993. The "hybrid bridge" hypothesis: host shifting via plant hybrid swarms. Am. 

Nat. 141 : 651-662. 

FRITZ, R.S., NICHOLS-ORIANS, C.M., and BRUNSFELD, S.J. 1994. Interspecific hybridization of plants and resistance 

to herbivores: Hypotheses, genetics, and variable responses in a diverse community. Oecologia 97:106-117. 

HALL, R.W. and TOWNSEND, A.M. 1987. Suitability of Ulmus wilsoniana, the 'Urban' elm, and their hybrids for the ehn 

leaf beetle, Xa_lthogaleruca luteola (Mtiller) (Coleoptera: Chrysomelidae). Environ. Entomol. 16: 1042-1044. 

HANHIM_.KI, S., SENN, J., and HAUKIOJA, E. 1994. Performance of insect herbivores on hybridising trees: the case of 

the subarctic birches. J. Anita. Ecol. in press. 

HUESING, J., JONES, D., DEVEERNA, J., MYERS, J., COLLINS, G., SEVERSON, R., and SISSON, V. 1989. Biochemical 

investigations of antibiosis material in leaf exudate of wild Nicotiana species and interspecific hybrids. J. Chem. 

Ecol. 15: 1203-1217. 

KEIM, R, PAIGE, K.N., WHITHAM, T.G., and LARK, K.G. 1989. Genetic analysis of an interspecific swarm of Populus: 

Occurrence of unidirectional introgression. Genetics 123: 557-565. 

LEVY, A. and MILO, J. 1991. Inheritance of morphological and chemical characters in interspecific hybrids between 

Pat)aver bracteatum and Papaverpseudo-orientale. Theor. Appl. Genet. 81: 537-540. 

MEIER, B., BETTSCHART, A., SHAO, Y., and LAUTENSCHLAGER, E. 1989. Einsatz der modernen HPLC for 

chemotaxonomische Utersuchungen morphologisch schwer zu differenzierender Salix-Hybriden. Planta Medica 55: 

213-214. 

MOSSELER, A. and PAPADOPOL, C.S. 1989. Seasonal isolation as a reproductive barrier among sympatric Salix species. 

Can. J. Bot. 67: 2563-2570. 

O'DONOUGHUE, L.S., RAELSON, J.V., and GRANT, W.E 1990. A morphological study of interspecific hybrids in the 

genus Lotus (Fabaceae). Can. J. Bot. 68: 803-812. 

PAIGE, K.N., CAPMAN, W.C., and JENNETTEN, R 1991. Mitochondrial inheritance patterns across a cottonwood hybrid 

zone: Cytonuclear disequilibria and hybrid zone dynamics. Evolution 45: 1360-1369. 

PAIGE, K.N. and CAPMAN, W.C. 1993. The effects of host-plant genotype, hybridization and environment on gall aphid 

attack and survival in cottonwood: The importance of genetic studies and the utility of RFLP's. Evolution 47: 36-45. 

RICE, W.R. 1989. Analyzing tables of statistical tests. Evolution 43: 223-225. 

SAS INSTITUTE. 1985. SAS user's guide: Statistics, Version 5 Edition. SAS Institute, Inc. Cary, NC. 

SOETENS, R, ROWELL-RAHIER, M., and PASTEELS, J.M. 1991. Influence of phenolglucosides and trichome density 

on the distribution of insects herbivores on willows. Entomol. Exp. Appl. 59: 175-187. 

WHITHAM, T.G. 1989. Plant hybrid zones as sinks for pests. Science 244: 1490-1493. 

WHITHAM, T.G., MORROW, RA., and POTTS, B.M. 1991. Conservation of hybrid plants. Science 254: 779-780. 

WHITHAM, T.G., MORROW, RA., and POTTS, B M. 1994. Plant hybrid zones as centers of biodiversity: The herbivore 

community of two endemic Tasmanian eucalypts. Oecologia, in press. 

141

F1 HYBRID SPRUCES INHERIT THE PHYTOPHAGOUS INSECTS 

OF THEIR PARENTS 

WILLIAM J. MATTSON, ROBERT K. LAWRENCE, and BRUCE A° BIRR 


B-7 Pesticide Research Center, Michigan State University, East Lansing, MI 48824, USA 

INTRODUCTION 

Interspecific hybridization in plants is a widespread and important natural phenomenon. In fact, Stace (1987) 

estimates that at least 50% of the extant angiosperm species have been derived via hybridization. Natural hybrids are also 

common among the gymnosperms (Wright 1955, Hanover and Wilkinson 1969, Schmidt-Vogt 1977, Zobel and Talbert 1984). 

In spite of the ecological and evolutionary importance of plant hybrid zones (Remington 1968, Barton and Hewitt 1985), 

little is yet known about how they affect the dynamics of plant/herbivore relationships. Whitham (1989) speculated that 

hybrids may be particularly prone to herbivory because usual resistance mechanisms derived from the parents are poorly 

expressed. For example, polygenic resistance traits might be intermediate in F1 hybrids, whereas monogenic and oligogenic 

traits might be fully expressed or not at all depending on the mode of gene action (e.g., dominant or recessive, etc.), and on 

whether there is genetic dysfunction preventing normal gene expression (Fritz et al. 1994, Whitham et al. 1994). Furthermore, 

to compound these problems, hybrids may attract specialist phytophages from both parent species, thereby being 

predisposed to a much larger pool of potential consumers than either of the pure parent types (Whitham et al. 1994). 

Insect Loading in Relation to a Plant's A/R Vectors and S/D Vectors 

To examine the issue of hybrid resistance to insects in a more mechanistic framework, we consider that there are at 

least two sets of plant traits affecting insect loading on plants: (1) the mix of properties which elicit insect attraction/ 

arrestmenffrepulsion (e.g., morphology, electromagnetic and biochemical cues, abiotic requirements, etc.) and thus govern the 

number of insect species that attempt to colonize it, and (2) the mix of properties that are sustaining/defensive for each of the 

would be colonizers as they attempt to feed, oviposit, and reproduce. These two sets of properties, which together set the 

potential for insect loading (total insect biomass or numbers), can be represented as two vectors: the species attraction/ 

repulsion (A/R) vector, [a_,a2,a3..an],and the suitability/defense (S/D) vector [Sl,S2,S3..s]. Each species of insect, Si, is 

represented by an element (a)in the species A/R vector, having a numerical value (0-1) that is proportional to the plant's 

average capacity to attract and arrest individuals of the particular insect species. Likewise, the S/D vector contains specific 

elements (s) that correspond to and represent the plant's average suitability, ranging from zero, where an insect colonizes a 

plant but cannot complete development on it (immunity), to 1 where the plant fulfills all of the insect's requirements, for each 

species in the species A/R vector. The product of the two vectors, [ayL+a2s2+..as ], represents the plant's overall potential 

"resistance expression", or potential insect loading per plant. 

To give an extreme example for one insect: the white pine weevil, Pissodes strobi, is highly attracted to Picea glauca 

shoots as evidenced by the great number of feeding scars it produces (i.e., a = 1), but it can only rarely reproduce in these 

shoots (i.e., s = 0), with the consequence that P glauca is considered "resistant" to the weevil (Lanier 1983). Actual insect 

loading depends, in addition, on the quantity and genetic diversity of insects (N) that continually attempt colonization. The 

total number of non-zero elements (all those a_> 0) in an individual plant's A/R or species vector (its species richness or 

species loading) is positively linked to the size and density of its population, the total areal distribution of the species, and 

negatively to its biochemical uniqueness wherein it occurs (for a review, see Tahvanainen and Niemela 1987). For many 

natural hybrids, their populations and geographical distributions are usually small (compared to their parents) thereby 

exposing them to fewer potential colonists, but their low biochemical uniqueness or high similarity to the two nearby parent 

species will likely have a counteracting influence making them highly susceptible to the reservoirs of consumers fiom two 

plant species (Whitham et al. 1994). 

Mattson, W.J., Niemel_, R, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


142 

;fl ¸_

Elements of the Suitability/Defense Vector in Hybrids: Expression of S/D Traits 

Most work on resistance expression (i.e., a_s_)in natural and artificial interspecies hybrids has focused on suitability/ 

defense (s) expression for particular insects, assuming that a plant's attractiveness (a) is high and invariant. Fritz et al. 

(1994) hypothesized four expression scenarios: (1) the "additive" case where the hybrid's S/D ranking (s_) against an insect 

is the simple mean of its parents, p_, and p: (sh_= (s_ + s2)/2), (2) the "dominance" case where the hybrid is consistently like 

one of the parents, e.g., p_(sh_= s_), (3) the "susceptibility" case where the hybrid has more insects than either of the parents 

(s_a< sh_> s2), and (4) the "resistance" case where the hybrid has fewer insects than either parent (s_ > sh_< s:). 

Most studies with trees have shown that hybrid S/D expression (s) against an insect is not cleanly predictable but 

varies with the particular insect and the host plant species (Boecklen and Larson 1994, Fritz et al. 1994). For example, 

Manley and Fowler (11969)and Osawa (1989) discovered that the severity of spruce budworm, Choristonenrafumiferana, 

defoliation of spruce trees was linked to the direction of their introgression between black and red spruces, P. maria_za and P. 

rubens. Hybrids closer to pure red were severely defoliated as were pure reds, whereas hybrids closer to pure black were 

only lightly defoliated as were pure blacks. Several hybrid elms, Ulmus, were consistently intermediate in resistance to the 

elm leafbeetle, Xanthogaleruca luteola, whenever one parent species was susceptible and the other was resistant (Hall and 

Townsend 1987, Hall et al. t987). Hybrids of P. sitchensis x P. glauca are usually resistant to Pissodes strobi, and so are 

such backcrosses to white spruce, the resistant parent whereas sitka spruce is highly susceptible (Mitchell et al. 1974, Copes 

and Beckwith 1977). Similarly, the Pinus coulteri x P jeffreyi hybrid carries resistance against the weevil, Cylirzrocopturus 

eatoni, that the coulteri parent possesses, but thejeffreyi parent does not (Miller 1950). Hybrids of several species of 

Japanese pines, Pinus, all apparently inherited their "resistance" against the scale, Matscoccus matsumurae, from the more 

susceptible parent species (McClure 1985). 

More recently, Floate et al. (1993) reported that Chrysomela confluens was invariably more abundant in a hybrid 

swarm of Populusfremontii x P angustifolia than in the parental populations. Whitham et al. (1994) reported that of some 

40 phytophagous taxa in a hybrid swarm of Eucalyptus risdonii x E. amygdalina only two were more abundant on FI hybrids 

than on the parents. Specialist insects were most abundant on hybrid backcrosses to their favored parent species, and 

generalist insects were generally more abundant on all hybrid types. Hanhimaki et al. (1994) reported that in the case of 14 

insects tested on hybrids of Betula pubescens ssp. tortuosa x B. nana, the hybrids were consistently equivalent to B. 

pubescens, but superior to B. nana. Fritz et al. (1994) tested 11 insects and one leaf rust on hybrids of Salix sericea x S. 

eriocephala, and concluded that 3 supported the additive, 2 the dominance, and 7 the susceptibility expression hypotheses. 

Boecklen and Larson (1994) measured densities of 8 species of galling wasps on hybrids of Quercus grisea x Q. gambelii, 

and reported that 3 supported the additive, 2 the dominance, 2 the susceptibility, and 1the resistance expression hypotheses. 

In the case of non-woody species, plant breeders have created innumerable interspecies hybrids which have become 

the cornerstone of breeding programs for the development of insect resistance in food and forage plants. Consequently, most 

resistance traits useful against native insects and pathogens have been derived from crosses to related but allopatric plant 

species (Harris 1975, 1980 1982; Bailey 1983). These studies reveal a surprisingly large number of cases for which insect 

resistance seems to be either monogenic or oligogenic, but polygenic inheritance is also important (Hanover 1980, Harris 

1982, Zobel and Talbert 1984, Singh 1986, Carson and Carson 1989, Geiger and Heun 1989). Moreover, the resistance genes 

are usually, but not always dominant in their expression (Harris 1982, Singh 1986), and evidence suggests that gene linkages 

and multiple resistance alleles are uncommon. Resistance genes appear to be insect specific, not conferring cross protection 

against many species (Harris 1982). 

Elements of the Species A/R Vector in Hybrids: Expression of A/R Traits 

General plant traits (e.g., crown form, height, growth rate, phenology, etc.) in hybrids are often intermediate between 

those of its parents (implying that S,_= (a_ + a2)/2) (Hanover and Wilkinson 1969, Bongarten and Hanover t982, Yeh and 

Arnott 1986) which suggests that interspecies hybrids will probably have the unique gestalt of properties to attract the all the 

specialist insects from both parents (Miller and Strickler 1984). In t_ct, Whitham et al. (1994) and Morrow et al. (1994) have 

found that this is true in a eucalypt hybrid swarm. They reported that the average hybrid tree supported 53% more phytophagous 

insect and fungal species than equivalent trees from the pure parental stands. 

143

Objectives 

Given the importance of a plant's herbivore loading, i.e., the number of herbivore species or non-zero elements in its 

species vector, this study tested the hypothesis that artificial white x blue spruce FI hybrids will inherit the phytophages from 

each of their parent tree species, with the consequence that the number of elements in the hybrid species vector will be the 

sum of the unique (specialists) plus the shared or common (generalists) elements of both parents. Because hybrid spruces do 

not occur naturally in Michigan, we only conducted tests in several test plantations where the trees are exposed to the natural 

phytophages of white spruce but not those of the other parent, blue spruce, a western North American tree species. However, 

one western spruce galling insect, Adelges cooleyi, is common in eastern North America on ornamental blue spruce. We also 

tested whether total insect loading on hybrids, i.e., the total number of individuals (pooled over all phytophage species) found 

per unit of foliage per plant, is equal to or greater than that found on each of the parent species. 

METHODS 

Hybrid spruces (P glauca x P. pungens) were sampled in 1991 and 1992 at two different locations in Michigan, along 

with populations of the parent species in order to assess differences in species, and total insect loading. One sample site 

occurred in south-central Michigan at the Michigan State University Kellogg Experimental Forest. The other occurred some 

450 km to the north in Michigan's Upper Peninsula at the Michigan State University Dunbar Experimental Forest. Kellogg 

(n = 40) and Dunbar (n = 24) sample trees were all F1 hybrids. Equivalent numbers of parent spruces were likewise sampled 

from adjacent or nearby (< 1 kin) plantations. 

At each site, trees were randomly selected, and examined twice per growing season (late May-early June, and late 

June-mid July). Trees were first examined visually to score for adelgid galls (Adelges abietis, A. cooleyi, and Pineus similis). 

Next they were sampled using standard beating methods whereby the apical half (45 cm length) of a midcrown branch on the 

west side was held over a specially designed stainless steel collection pan, and the branch was gently tapped with a wooden 

dowel for I0 seconds to encourage free feeding species to drop off into the pan. Such insects were gathered in vials and then 

stored in a freezer for later identification and counting. 

Data from both June and July samples (numbers of insect species, and total insect counts per tree) were analyzed 

separately using a completely randomized ANOVA, after log (X + 0.1) transformation. Means for each tree species were 

ranked and then separated using the SNK multiple range test. 

RESULTS 

Gall-forming Adelgid Specialists from Both Parent Spruces Successfully Attack Hybrids 

At the Kellogg experimental tbrest in southern Michigan, there were significant populations of both Adelges abietis, 

and A. cooleyi, largely specialists, respectively, on P. glauca, and P pungens and their near relatives (Furniss and Carolin 

1977). We found that the ihybrids contained substantial numbers of both kinds of adelgids, whereas the parent spruces had 

primarily their respective specialist adelgids (Table 1). One glauca individual had evidence of two poorly formed A. cooleyi 

galls, but in general glauca in eastern North America are not susceptible to the eastern A. cooleyi populations (personal 

observations) which live largely on ornamental blue and Engelmann spruces introduced from western North America. 

Similarly, there was only minor evidence for A. abietis on pungens: one tree, with one small, poorly formed gall. Although 

Table 1 shows very low populations ofA. cooleyi on pungens (one infested tree in a sample of 30-42), this was because the 

adelgid population had drastically declined in the preceding 2 years (1989, 1990) in the sampled pungens plantation which 

was about 1 kln from the hybrids. In 1988 and earlier there were much larger populations of A.cooleyi there: 76 of 100 

sample trees were infested (Mattson, unpublished data). Pineus simitis, another galling adelgid that is largely a specialist of 

glauca and near relatives, was not abundant enough to test hybrid susceptibility. 

At the Dunbar site in northern Michigan, A. cooleyi and P. similis populations were negligible, so we could not 

measure their colonization of the spruce hybrids. But, we did find there that A. abietis readily attacked hybrids, but not so 

readily pungens, as shown in the following tabulation of the percentage of trees attacked: white (84), hybrids (83), blue (15). 

144

Table l.--Comparing the mean numbers of insects and adelgid galls on blue, hybrid, and white spruces, using visual 

examination of two random midcrown branches per sample tree at the Kellogg Experimental Forest during late 

May-early June, and mid-late July. N = 30 - 42 trees. 

Insect Species Time period l: nos.! 2 branches Time period 2: nos./2 branches 

B.Spruce Hybrids W.Spruce B.Spruce Hybrids W.Spruce 

A. abietis galls 0.07b _ 1.14a 2.25a 0.00b 2.20a 2.27a 

A. cooleyi galls 0.03b 0.36a 0.00b 0.00b 0.33a 0.07b 

A. cooleyi alates 0.17b 24.76a 0.00b 0.00b 4.66a 0.00b 

Pineus similis galls O.OOa O.OOa O.19a O.OOa O.OOa O.OOa 

Means with different letters within time periods are significantly (p < 0.05) different. 

Species Loading, and Total Insect Loading Comparable Among Hybrids and White Spruce 

At Kellogg, the hybrid and white spruce herbivore species loadings were statistically equivalent (ca. 3 - 3.3 per 

branch) and significantly higher than those of blue (I.9) in the first sample, but not in the second, when all three spruces were 

statistically equivalent. At Dunbar, the hybrids, and white spruce also had equivalent herbivore loadings (ca. 3 - 3.8 species 

per branch) which were significantly higher than those (ca. 1.2 - 1.5 per branch) of blue spruce in both first and second 

samples (Table 2). 

Table 2.--Comparing mean numbers of phytophagous insect species, and total insect numbers occurring in beating samples 

from blue, hybrid, and white spruces at two study sites in Michigan. 

Study area and Mean nos. insect species per branch Mean nos. of insects per branch 

sample period 

B.Spruce Hybrids W.Spruce B.Spruce Hybrids W.Spruce 

Dunbar- 1 1.50b _ 2.96a 3.79a 2.42b 4.17a 6.71 a 

Dunbar-2 l. 17b 3.21 a 3.00a 1.25b 4.33a 4.71 a 

Kellogg-1 1.87b 3.31a 2.97a 1.13b 4.27a 4.23a 

Kellogg-2 2.97a 2.17a 2.90a 2.20a 2.12a 1.97a 

Means having different letters within an area and sample period category are significantly (p < 0.05) different. 

At Kellogg, there were 20 species of phytophagous insects on the hybrids, of which they shared 12 in common (60%) 

with white spruce (Fig. l). At Dunbar, we found twice as many (41 ) species of phytophagous insects on the hybrids, of 

which 33 (80.5%) they shared in common with white spruce (Fig. I). About 12-15% of the hybrid's phytophagous insects 

were "unique" at both sites, because they did not share these species with either white or blue spruce. About 49-60% of their 

total species pool, they shared with blue spruce. These data confirm the expectation that the phytophagous insects occurring 

on the hybrids are largely those commonly associated with white spruce. The same is true for blue spruce, which had a total 

of 23 species at Kellogg and 24 at Dunbar, of which 19 were shared in common with white spruce at both sites. 

145

Blue White White 

2 5% Blue 0% 

7.: 

Jnique 

" "_ "_''''''"'- Unique 12.2% 

W&B 15% W&B 

35% 41.5% 

Figure l.--The percent distribution of the phytophagous insect fauna of hybrid spruces that are shared with blue spruce only, 

white spruce only, both white and blue spruce, and unshared or unique at two sites: Kellogg (left side) in southern, 

and Dunbar (right side) in northern Michigan. At Kellogg there were 20 and at Dunbar there were 41 species of 

insects found in beating samples from the hybrids. 

DISCUSSION 

The data clearly suggest that white x blue spruce FI hybrids inherit most of the phytophagous insect species of their 

white spruce parents. Although we cannot claim that they similarly inherit most of the phytophagous insect species of their 

blue spruce parentage because the hybrids were not exposed to these insects in Michigan, we can never-the-less say that they 

readily inherited at least one of them, A. cooleyi. Likewise, total insect loading on the hybrids was comparable to that of 

white spruce. 

Thus, the data do not falsify the hypothesis that the number of phytophages (i.e., the number of non-zero elements) in 

the hybrid species vector will be the sum of the unique and shared species of its parents. We fully expect that if the hybrids 

were growing adjacent to large, natural populations of blue spruce, they would readily inherit its typical phytophages. 

Whitham et al. (1994) emphasized this very important point for Tasmanian eucalyptus hybrids, reporting that the average 

hybrid had 53% more phytophagous species than equivalent trees in pure parental stands. 

Neither do our data falsify the hypothesis that the total insect loading (i.e., the total number of all individuals of all 

insect species per unit foliage) on hybrids will be equal to or greater than the average loading of each of its parents. This may 

be true because the defensive traits (sh_)fbr any insect in the hybrids will probably never be more strongly expressed than in 

the less suitable or better defended of the two parent species (e.g., if sL_> s2_' then s2_> sh_), and in many if not most cases, it 

will be less poorly expressed (e.g., the additive, the susceptibility, and the dominant (high defense is recessive) expression 

scenarios of Fritz et al. 1994). At the level of the whole phytophagous insect community, the hybrid studies of Boecklen and 

Larson (1994), Fritz et al. (1994), Hanhimaki et al. (1994), and Whitham et al. (1994) unequivocally support this proposition. 

Hybrids may therefore generally sustain more herbivores and herbivore injury than either pure parent species. 

SUMMARY 

FI hybrids of the spruces Picea glauca x Picea pungens were studied at two Michigan locations to measure and 

compare their total phytophagous insect fauna (i.e., species and total loading) with that of the parent species in adjacent or 

nearby plantations. Hybrids had species compositions, and total insect loading nearly identical with that of Picea glauca. In 

addition, at least one adelgid specialist from P. pungens readily colonized and successfully reproduced on the hybrids. These 

data support the hypotheses that hybrids will (a) inherit the insect specialists from both parents, and (b) have an overall insect 

loading that equals or exceeds that of each parent. 

146 

i:i_ ¸J'"


The authors sincerely thank Drs. J. Kouki and J. Spence for their discussions about and their helpful critiques of 

earlier drafts of this manuscript. They also thank the personnel of the W.K. Kellogg Forest for their unstinting assistance. 


BAILEY, J.A. 1983. Biological perspectives of host-pathogen interactions, p. 1-32. In Bailey, J.A. and Deverall, B.J., eds. 

The Dynamics of Host Defense. Academic Press, N.Y. 

BAILEY, J.A. 1980. Arthropod-plant interactions related to agriculture, emphasizing host plant resistance, p. 23-51. In 

Harris, M.K., ed. Biology and Breeding for Resistance to Arthropods and Pathogens in Agricultural Plants. Texas 

A&M Univ., College Sta, TX. 

BAILEY, J.A. 1982. Genes for resistance to insects, emphasizing host-parasite interactions in Carya, p. 72-83. In 

Heybroek, H.M., tephan, B.R., and von Weissenberg, K., eds. Resistance to Diseases and Pests in Forest Trees. 

Pudoc, Wageningen, Netherlands. 

BARTON, N.H. and G.M. Hewitt, G.M. 1985. Analysis of hybrid zones. Ann. Rev. Ecol. Syst. 16: 1113-148. 

BOECKLEN, W.J. and SPELLENBERG, R. 1990. Structure of herbivore communities in two oak (Quercus spp.) hybrid 

zones. Oecologia 85: 92-100. 

BOECKLEN, W.J. and LARSON, K.C. 1994. Gall-forming wasps (Hymenoptera: Cynipidae) in an oak hybrid zone: testing 

hypotheses about hybrid susceptibility to herbivores, p. 110-120. In Price, P., Mattson, W.J., and Baranchikov, Y.N., 

eds. The ecology and evolution of gall-forming insects. Gen. Tech. Rep. NC-174. St. Paul, MN: U.S. Department of 


BONGARTEN, B.C. and J.W. HANOVER 1982. Hybridization among white, red, blue, and white x blue spruces. For. Sci. 

28: 129-134. 

CARSON, S.D. and CARSON, M.J. 1989. Breeding for resistance in forest trees--a quantitative genetic approach. Ann. 

Rev. Phytopath. 27: 373-395. 

COPES, D.L. and BECKWITH, R.C. 1977. Isoenzyme identification of Picea glauca, P sitchensis, and P. lutzii populations. 

Bot. Gaz. 138: 512-521. 

FLOATE, K.D., KEARSLEY, M.J.C., and WHITHAM, T.G. 1993. Elevated herbivory in plant hybrid zones: Chrysomela 

confluens, Populus and phenological sinks. Ecology 74: 2056-2065. 

FRITZ, R.S., NICHOLS-ORIANS, C.M, and BRUNSFELD, S.J. 1994. Interspecific hybridization of plants and resistance 

to herbivores: hypotheses, genetics, and variable responses in a diverse herbivore community. Oecologia 97: 106- 

117. 

FURNISS, R.L. and CAROLIN, V.M. 1977. Western forest insects. Misc. Publ. 1339. Washington, DC: U.S. Department 

of Agriculture, Forest Service. 654 p. 

GEIGER, H.H. and HEUN, M. 1989. Genetics of quantitative resistance to fungal diseases. Ann. Rev. Phytopath. 27: 317- 

341. 

HALL, R.W., TOWNSEND, A.M., and BARGER, J.H. 1987. Suitability of thirteen different host species for elm leaf 

beetle, Xanthogaleruca luteola (Coleoptera: Chrysomelidae). J. Environ. Hort. 5: 143-145. 

147

HALL, R.W. and TOWNSEND, A.M. 1987. Suitability of Ulmus wilsoniana, the 'urban' elm, and their hybrids fbr the elm 

leaf beetle, Xanthogaleruca luteola (Muller)(Coleoptera: Chrysomelidae). Environ. Ent. 16:1042-1044. 

HANHIMAKI, S., SENN, J., and HAUKIOJA, E. 1994. Performance of insect herbivores on hybridizing trees: the case of 

the subarctic birches. J. Animal Ecol. 63: 163-175. 

HANOVER, J.W. and WILKINSON, R.C. 1969. A new hybrid between blue spruce and white spruce. Can. J. Bot. 47: 

1693-1700. 

HANOVER, J.W. 1980. Breeding forest trees resistant to insects, p. 487-511. In Maxwell, F.G. and Jennings, ER., eds. 

Breeding Plants Resistant to Insects. John Wiley and Sons, NY. 

HARRIS, M.K. 1975. Allopatric resistance: searching for sources of insect resistance for use in agriculture. Environ. Ent. 4: 

661-669. 

KUDRAY, G.M. and HANOVER, J.W. 1980. A preliminary evaluation of the Spartan spruce (Picea glauca x P. pungens 

hybrid). Mich. State Univ. Agric. 

LANIER, G. 1983. Integration of visual stimuli, host odorants, and pheromones by bark beetles and weevils in locating and 

colonizing trees, p. 161-172. In Ahmad, S., ed. Herbivorous insects: host seeking behavior and mechanisms. Acad. 

Press, NY. 257 p. 

MANLEY, S.A.M. and FOWLER, D.E 1969. Spruce budworm defoliation in relation to introgression in red and black 

spruce. For. Sci. 15: 365-366. 

MCCLURE, M.S. 1985. Susceptibility of pure and hybrid stands of Pinus to attack by Matsucoccus matsumurae in Japan 

(Homoptera: Coccoidea: Margarodidae). Environ. Ent. 14: 535-538. 

MILLER, J.R. and STRICKLER, K.L. 1984. Finding and accepting host plants, p. 127-157. In Bell, W.J. and Carde, R.T., 

eds. Chemical Ecology of Insects. Sinauer Assoc., Sunderland, MA. 

MILLER, J.M. 1950. Resistance of pine hybrids to the pine reproduction weevil. For. Res. Notes 68. U.S. Department of 

Agriculture, Forest Service, Calif. For. Range Exper. Stn. 17 p. 

MITCHELL, R.G., JOHNSON, N.E., and WRIGHT, K.H. 1974. Susceptibility of 10 spruce species and hybrids to the white 

pine weevil (=Sitka spruce weevil) in the Pacific Northwest. PNW-225. U.S. Department of Agriculture, Forest 

Service.. 8 p. 

MORROW, EA., WHITHAM, T.G., POTTS, B.M., LADIGES, P., ASTON, D.H., and WILLIAMS, J.B. 1994. Gall-forming 

insects concentrate on hybrid phenotypes of Eucalyptus, p. 121-134. In Price, P.W., Mattson, W.J., and Baranchikov, 

Y.N., eds. The ecology and evolution of gall-forming insects. Gen. Tech. Rep. NC-174. U.S. Department Of Agriculture, 

Forest Service, N.C. For. Exp. Sta., St. Paul, MN 55108. 

OSAWA, A. 1989. Causality in mortality patterns of spruce trees during a spruce budworm outbreak. Can. J. For. Res. 19: 

632-638. 

REMINGTON, C.L. 1968. Suture-zones of hybrid interaction between recently joined biotas, p. 321-428. In Dobzhansky, 

T., Hecht, M.K., and Steere, W.C., eds. Evolutionary Biology, Vol. 2. Appleton-Century-Crofts, NY. 

ROSE, A.H. and LINDQUIST, O.H. 1977. Insects of eastern spruces, fir and hemlock. Can. For. Serv. For. Tech. Rept. 23. 

159 p. 

SCHMIDT-VOGT, H. 1977. Die Fichte. Verlag Paul Parey, Berlin. 647 p. 

SIMMS, E.L. and FRITZ, R.S. 1990. The ecology and evolution of host-plant resistance to insects. TREE 5: 356-360. 

148

SINGH, D.R 1986_ Breeding for resistance to disease and insect pests. Springer Verlag, NY. 222 p. 

STACE, C.A. 1987. Hybridization and the plant species, p. 115-127. In Urbanska, K.M., ed. Differentiation Patterns in 

Higher Plants. Academic Press, NY. 

TAHVANAINEN, J. and NIEMELA, R 1987. Biogeographical and evolutionary aspects of insect herbivory. Ann. Zool. 

Fenn. 24: 239-247. 

WHITHAM, T.G., MORROW, RA., and POTTS, B.M. 1994. Plant hybrid zones as centers of biodiversity: the herbivore 

community of two endemic Tasmanian eucalyptus. Oecologia 97: 481-490. 

WHITHAM, T.G. 1989. Plant hybrid zones as sinks for pests. Science 244: 1490-1493. 

WRIGHT, J.W. 1955. Species crossability in spruce in relation to distribution and taxonomy. For. Sci. 1: 319-349. 

YEH, EC. and ARNOTT, J.T. 1986. Electrophoretic and morphological differentiation of Picea sitchensis, Picea glauca, 

and their hybrids. Can. J. For. Res. 16:791-798. 

ZOBEL, B. and TALBER_I; J. t984. Applied forest tree improvement. John Wiley and Sons, NY. 505 p. 

149

RECENT ADVANCES IN RESEARCH ON WHITE PINE WEEVIL ATTACNG 

SPRUCES IN BRITISH COLUMBIA 

G.K. KISS I, A.D. YANCHUK 2, R.I. ALFARO a, J.E. CARLSON 4, and J.F. MANVILLE 3 

1British Columbia Forest Service, Kalamalka Forestry Centre, Vernon, British Columbia, Canada 

2British Columbia Forest Service, Research Branch, Victoria, British Columbia, Canada 

3Canadian Forest Service, Pacific Research Centre, Victoria, British Columbia, Canada 

4University of British Columbia, Biotechnology Laboratory, Vancouver, British Columbia, Canada 

INTRODUCTION 

Interior spruce (a collective term for white spruce, Picea gIauca (Moench) Voss; Engelmann spruce, Picea 

engelmanni Parry, and their varying degrees of hybrids) is a very important timber resource for the province of British 

Columbia. Over 100 million interior spruce seedlings are planted annually. An estimated 1.38 billion seedlings have been 

planted to date, and the total number of hectares planted in spruce is over 1.3 million. 

White pine weevil, Pissodes strobi Peck, is a major cause of Sitka spruce, Picea sitchensis (Bong.) Carr., plantation 

failure in coastal British Columbia, Washington, and Oregon (McMullen 1976; Furniss and Carolin 1977; MacSiurtain I981, 

Alfaro 1982, 1989). This pest is also becoming an important menace to interior spruce plantations, causing considerable 

damage to both yield and quality of product (Cozens 1983, Taylor et al. 1991). 

Adult weevils overwinter in the duff layer of the forest floor and begin laying their eggs in the bark of young spruce 

trees (generally 3-30 years old) near the tip of the previous ye.ar's shoot in early summer. Young weevil larvae feed on the 

phloem of the leader, moving downwards in the process (Silver 1968). Once a feeding ring is formed, the leader of the tree 

dies. In their downward movement, the larvae can destroy up to 4 years of growth. Literature sources refer to the successful 

colonization of the leader as weevil attack; this terminology will be maintained in this report. Repeated leader destruction 

causes loss of height growth and stem deformities, which leads to quality reduction (Alfaro 1989 and 1992). 

Selection of spruce varieties genetically resistant to weevil damage is a potential tool for the reduction of weevil 

damage in future plantations. Genetic resistance could manifest itself in several forms. Morphological and anatomical 

differences could make certain trees less susceptible to feeding, ovipositioning, and to the development of the larvae (Plank 

and Gerhold 1965, Stroh and Gerhold 1965). Another possible resistance mechanism consists of variations in the chemical 

composition of resistant trees that make them less attractive to weevils. Phagostimulants may be lacking in some trees, 

which can render them less attractive to weevils. Alternately, the presence of feeding repellents, deterrents, or toxic compounds 

may provide an effective defense mechanism (van Buijtenen and Santamour 1972; Bridgen et al. 1979; Alfaxo 1980; 

Alfaro et al. 1980, 1984; Alfaro and Borden 1982, 1985; Wilkinson 1985; Brooks et al. 1987 a, b). 

The impact of weevil damage could be mitigated by the ability of the trees to recover from the attack (Painter 1951 ). 

Tolerant trees are more likely to overcome the effect of attacks and will suffer less growth and quality loss. This depends 

largely on the ability of one of the lateral branches to gain apical dominance quickly after the leader has been killed. Trees 

lacking this ability will develop multiple leaders (forks) and often are stunted. 

Evaluation of various trials in British Columbia provided strong evidence for genetic variation in susceptibility for 

weevil damage in both Sitka spruce (Ying 1991, Alfaro and Ying 1990) and interior spruce (Kiss and Yanchuk 1991). 

Mattson, W.J., NiemelL P., and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. 

USDA For. Serv. Gen. Tech. Rep. NC-183, N.C. For. Exp. Sta., St. Paul, MN 55108. 

150

The objectives of this presentation are to (i) review measures taken by various agencies in British Columbia to 

elucidate the interrelationship between the weevil and its host, (ii) discuss studies in progress to take advantage of the genetic 

resistance of the host and (iii) present an overview of attempts to control one of the more destructive forest pests. The 

presentation will also review future plans. 

Prince George Progeny Trials 

WEEVIL RESEARCH 

Field Studies 

Due to the importance of interior spruce in British Columbia, a decision was made to embark on the genetic improvement 

of the species complex in the early 1960's. The senior author was hired to carry out this project in 1967. 

The basic approach to improving interior spruce was to designate geographically distinct areas (selection units or 

SU's) and select trees based on size and form in each selection unit (Kiss and Yeh 1988, Kiss and Yanchuk 1991). The first 

of these selection units was in the Prince George area. A total of 180 parent trees were selected. Cones and scions were 

collected from most selected parents. Scions were used to establish clone banks, and seeds were used to establish genetic 

tests. 

As part of the improvement program, genetic tests were established using open-pollinated progenies of individual 

parents selected in various selection units. The first set of these tests, representing the Prince George SU, was established in 

1972 and 1973 (Kiss and Yeh 1988, Kiss and Yanchuk 1991) using the open-pollinated progenies of 173 parent trees. 

The initial objective of the genetic tests was to identify well-performing families based on the growth performance of 

their progenies. This information aided in the construction of rogued first generation seed orchards and helped to identify 

candidates for advanced generation breeding. To this end, periodic height measurements were carried out and at age 15 

diameter measurements were also made. 

Following the 15-year height and diameter assessments, heavy infestation of white pine weevil was observed at some 

test sites. Cursory observation appeared to suggest a pattern to the infestation. Further studies confirmed this pattern and 

revealed the possibility of genetically controlled resistance mechanism(s) in a number of families. Results of this study are 

reported by Kiss and Yanchuk (1991). 

In summary, the results showed that: 

1. While there were different intensities of attack at the three individual test sites in the Prince George SU tests 

(i.e., 9% at Quesnel, 37% at Red Rock, and 63% at Aleza Lake), there were significant correlations between 

families attacked across sites (Figs. 1 and 2). 

2. The estimate of family-mean heritability for weevil resistance across sites was high (h2f=0.77 + 0.11). 

3. Families exhibiting greater vigor at any age were generally more resistant. Correlation between 6-year family 

height and incidence of weevil attack was negative and significant (Fig. 3, r=-0.42), as well as 15-year height 

and incidence of weevil attack (Fig. 4, r=-0.68). In fact, when families were ranked based on height growth at 

different ages, and average attack was calculated for the top and bottom 25% of families, the top families 

always had significantly lower attacks than those of the bottom families (Table 1). 

151

30 1 O0 . ._.--,,= 

A _ A 

r = 0.63 • A O_ r = 0.71 • • • • • • 2 • 

. m 80. - A _• _ _ • • 

3 • • N A 

• _iA'_A%A " 

• A A • A < _ 60. A• A_ AA} _AA &A •i A•A• A •• 

• A m A A •_ t 

"_ 15 AA A A _A • --_ A_ _A• A • 

© • • A A • • O_ • A AA•• _ A • 

>" 10 AA • • _ • AAA • A 

E 

A• A A AA_ _MkaAA A• -- A 

= 

A A •_ • A _i_ • ,4.._ 204 _ • 

c 5 •,_ ••_ AaA_ • =Aj 

• _A SA• AA A_A •A • • 

o o l 

0 20 40 60 80 0 20 40 60 80 

Mean family attack at Red Rock (%) Mean family attack at Red Rock {%) 

Figures 1 & 2.--Comparison of attack percentages at the Red Rock and Aleza Lake test sites and at the Red Rock and 

Quesnel test sites. 

8O 

• 80,. 

0 • A • • 

60 A A 2 •A d O0 a A • 

c:?-. • " A. 

'--- • _ • •_•_ #*A • _a > • •a_a AA • • 

• _A A 0.68 A a • •_2 A a 

. • _ r=- 

0 , r_ , • 

' ' ' 0 _ .......... A A A 

50 55 60 65 70 75 80 85 90 160 180 200 220 240 260 

Mean 6-year heights at Red Rock (%) o 300 32 

Mean 15-year heights at Red Rock (%) 

Figures 3 & 4.--Relationship between average family height measurements at 6 and 15 years of age and mean family weevil 

attacks at the Red Rock test site. 

This latter finding was especially gratifying as a literature survey indicated that white pine weevil favored more 

robust, longer leaders. While this still may be true at the phenotypic level, genetically superior families are less susceptible. 

We would have accepted susceptibility that was equal to the mean; the fact that faster growing families are less susceptible is 

a bonus. 

Quesnel Lakes Progeny Trials 

In the fall of t993 additional genetic test plantations were assessed for weevil attack. These plantations were 

established in 1983 to evaluate the genetic potential of I42 parent trees. The tests were established on three sites (KetchuE_ 

Creek, Little Benson Lake, and Clearwater) using randomized complete blocks with eight blocks per site and four seedlings 

per replication per family (total 32 seedlings per site per family for a grand total of 96 seedlings per family over the three 

sites). 

152

Table 1.--Average weevil attack on the top and bottom 25% of families at three test sites. Families were ranked by average 

heights at various ages and at 15-yr dbh (diameter at breast height). 

Test age 

Quesnel test Red Rock test Aleza Lake test 

Top Bot. Diff. Top Bot. Diff. Top Bot. Diff. 

Initial ht. 6.6 10.7 4.1 31.0 42.9 11.9 56.6 68.3 11.7 

3-year ht. 7.1 11.9 4.8 30.6 44.4 13.8 55.9 69.4 13.5 

6-year ht. 6.6 12.1 5.5 28.8 45.1 16.3 53.6 71.3 17.7 

10-year ht. 5.9 11.9 6.0 26.4 45.0 18.6 52.7 70.8 18.1 

15-year ht. 5.2 13.3 8.1 23.8 49.0 25.2 50.6 73.3 22.7 

DBH 6.3 11.7 5.4 28.3 44.1 15.8 55.6 70.1 14.5 

Moderate to heavy weevil infestation occurred on two of these plantations (Little Benson Lake - 17%, Clearwater - 

33%). As was the case with the earlier studies, the weevils again favored certain families and avoided others. Some individuals 

were attacked up to five times over the years, while more than 1,700 of the 2,700 intensively studied trees at 

Clearwater have never been attacked. There was high correlation between mean family attacks at the two sites (Fig. 5, r 

=0.66). 

A 

70 A 

60 A 

50 AA _AA • 

-_ O • AAI• 

AA • • 

23o _ • I i. ••_ 

_ 20 _, A• • • • 

r = 0.66 

10 • AA 

0 10 20 30 40 50 60 

Mean family attack at Little Benson(%) 

Figure 5.--Comparison of attack percentages at Little Benson and Clearwater. 

Once again the more vigorous, faster growing families were less susceptible to weevil attacks. Based on the overall 

ranking at all sites, the top 25% of families suffered substantially less damage than the bottom 25% (Table 2). 

Additional data collected on the Clearwater site in cooperation with Dr. Ren6 Alfaro relate to number of attacks per 

tree, recovery of attacked trees, and quality traits following recovery. These data are presently being evaluated and the 

results will be published elsewhere. 

A 

153

Table 2.--Average weevil attack on the top and bottom 25% of families at the Little Benson and Clearwater test sites. 

Families were ranked by average heights at various ages. 

Test age 

Comparative Trials 

Little Benson test Clearwater test 

Top Bot. Diff. Top Bot. Diff. 

Initial ht. 15.5 18.2 2.7 28.7 40.8 12.1 

3-year ht. 14.4 18.6 4.2 26.3 41.4 15.1 

6-year ht. 10.4 19.7 9.3 22.4 43.8 21.4 

10-year ht. 7.6 21.8 14.2 17.5 47.6 30.1 

In 1992 another pilot project plantation was evaluated for weevil damage. This plantation was established in 1976 to 

compare the relative performance of native British Columbian spruces with those originating from eastern Canada. Eastern 

spruces up to now perform well in B.C., and we are planning to incorporate some eastern material into our breeding program. 

This plantation is a replicated complete block design with 15 blocks, 10-tree row plots in each block from each of 21 

families. There are 9 families from eastern Canada (ENA) and 12 families from B.C. (four each from East Kootenay (EK), 

Prince George (PG), and Prince Rupert (PR)). 

Results indicate that eastern white spruce is very resistant to the white pine weevil. Of the 1,220 eastern trees, only 

67 were damaged by weevil, whereas 379 native spruces were damaged out of a total of 1,665 (Table 3). Interestingly, 25 of 

the 67 individuals damaged were progenies of the same parent. These results are very significant in light of the fact that these 

trees also grow and survive extremely well. 

Table 3.--Weevil attack differences among families of different geographic origin. 

Parents Number of trees 

Source Number Alive Attacked Percent 

ENA 9 1,220 67 5.5 

EK 4 545 154 28.3 

PG 4 562 85 15.1 

PR 4 558 140 25.1 

BC Total 12 1,665 379 22.8 

Further proof of the resistance is demonstrated by the eastern North American white spruce clones (about 75 clones) 

located in our breeding arboreta at the Kalamalka Forestry Centre. These clones are surrounded by native B.C. clones that 

are heavily attacked by weevil yet eastern clones are almost void of attacks. It is not possible to treat these data using 

statistical procedures due to the small number of ramets per clone (4), but there is no doubt about the trend. 

DISCUSSION 

Based on our observations, we must review weevil behavior on interior spruce. It has been postulated that in Sitka 

spruce, white pine weevil preferentially attacked the most vigorous, longest, and thickest leaders (VanderSar and Borden 

154

1977, Kline and Mitchell 1979, Wood and McMullen 1983). It was suggested that the same preference holds for interior 

spruce as well. Our investigation confirms beyond doubt that it is not the case. Susceptibility of the individual is the 

overriding principle that determines whether or not it will be attacked successfully. 

In field studies we observed that some seedlings of susceptible families, measuring 50 cm, were attacked by weevil. 

In some cases these young seedlings were attacked below the previous year's growth probably due to the inadequate thickness 

of the previous year's stem. 

To clarify weevil behavior, a number of studies are being initiated to elucidate the biology of white pine weevil in the 

interior and its relationship to its host. Verification of resistance in putatively resistant families will also be carried out. 

During the spring of 1993 a number of crosses were made using various putatively weevil resistant and susceptible 

parents. Judgement as to resistancy or susceptibility was made based on the progenies' weevil resistance performance. 

These crosses will form the basis of a number of experiments designed to verify resistancy and susceptibility and to study the 

mode and degree of inheritance of these traits. 

Future plans include continuation of the inventory of resistant families, provision of material for other agencies, and 

incorporation of the accumulated knowledge into the breeding program. 

Terpene Studies 

Laboratory Studies 

Our first intention following the identification of genetically controlled resistance was to try to identify the reason for 

resistance. To this aim, we enlisted the help of a team of biochemists lead by Dr. John Manville. They tested eight resistant 

and eight susceptible parental clones for terpenes that they judged to be of potential use in differentiating between them. The 

tests included both leaf and bark samples. 

They were able to classify 15 of the 16 parental clones as to susceptibility or resistance using six bark terpenes with 

multivariate technique (Manville et al. 1994). Two of the terpenes are identified (santene and citronellyl acetate) the other 

four are not yet identified. 

Leaf terpenes also provided similar results. Using six leaf terpenes (terpinolene and five unidentified terpenes), they 

were able to classify correctly 15 of 16 samples. 

Future plans include validation of the results (i.e., using a random sample of attacked and unattacked trees and 

applying the technique to them). They are also working on the identity of the unidentified terpenes. 

DNA Marker Studies 

We were also interested in investigating the potential differences at the gene level. To study these differences, we 

enlisted the help of Dr. John E. Carlson of the University of British Columbia (Carlson et al. 1994). 

The team investigated differences between resistant and susceptible interior spruce parents and their open-pollinated 

progenies. They constructed genetic linkage maps based on the recently developed Random Amplified Polymorphic DNA 

(RAPD) marker system, the first linkage maps available for any spruce species (Hong et al. 1993). Using a "Bulk Parental 

Analysis" (BPA) technique developed by the team, they pooled DNA from 12 resistant and 12 susceptible parents. This 

technique identified 20 DNA markers putatively associated with weevil resistance. 

The followup investigation was carried out on half-sib progenies of the original parents. Needles were collected from 

10 half-sib individuals of each parent of the previous study (a total of 240 half-sib progenies). 

155

Results of these studies showed that three of the markers were strongly correlated with resistance. None of the 

markers showed 100% association with resistance or susceptibility. The senior author has two potential explanations for this: 

1. The markers are only loosely linked to the gene(s) for resistance, so recombination between the marker and the 

gene can still take place. 

2. More than one gene may be involved in weevil resistance, and/or modifier genes are also necessary for a 

resistance gene to be expressed. 

These studies are also continuing. 

Micropropagation of Resistant Genotypes 

Dr. Ben Sutton and his team from the British Columbia Forest Biotechnology Centre in Vancouver are involved in 

utilizing elite full-sib crosses among resistant parents by multiplying them through embryogenesis (Roberts 1994). Their 

technique to produce large numbers of"emblings" has been developed to a point where they can mass produce these 

propagules. 

Embryogenic tissues can be stored in liquid nitrogen indefinitely. Once clones have been tested and proven to be 

resistant, they can be mass produced and used for operational plantations. 

Results of the studies described above will be utilized in developing an integrated weevil control system that will 

provide protection to newly planted spruce forests in British Columbia. There are a number of unanswered questions that 

will be addressed in future studies, but the results to date are encouraging. We are optimistic that our efforts will lead to 

success. 

SUMMARY 

Studies of interior spruce progeny trials revealed apparent genetic control of resistance to weevil damage in interior 

spruce in British Columbia. Correlation between mean family attacks across sites was high. 

Contrary to previous suggestions, weevil damage was more prevalent on less vigorous families. The difference 

between average damage of higher and lower ranked families (based on growth performance at any age) was always significant. 

Families that ranked highest for growth typically suffered less damage. 

Eastern North American white spruce appears significantly more resistant to weevil than are western sources. 

Western sources suffered more than fourfold the damage of eastern sources. 

Laboratory studies appeared to confirm differences between putatively resistant and susceptible spruces: 

I. Terpene analyses tentatively identified terpene content differences between the two strains. Fifteen of sixteen 

families could be distinguished by this technique; 

2. RAPD markers could also differentiate between resistant and susceptible families. While none of the markers 

had 100% association with resistance, the results are clearly pointing to markers loosely associated with 

resistance. 

Genotypes identified as resistant will be propagated using a technique referred to as somatic embryogenesis. A team 

at the British Columbia Research Incorporation has perfected the technique so that it can mass produce clones that are 

resistant to weevil. 

156

Further research is underway to clarify the many questions that are still remaining. Nevertheless, the resistance 

observed in our genetic trials will allow us to select and breed for this attribute in combination with growth and other 

desirable traits. 


The authors are indebted to many individuals who contributed to this manuscript. Technical support was provided by 

D. Wallden and G. Phillips over many years. The financial support of the British Columbia Forest Service was responsible 

for the establishment, maintenance, and evaluation of the trials that provided the database for the study. 


ALFARO, R.I. 1980. Host selection by Pissodes strobi Peck: Chemical interaction with the host plant. Ph.D. Thesis, Simon 

Fraser University, Vancouver, British Columbia. 

ALFARO, R.I. 1982. Fifty-year-old Sitka spruce plantations with a history of intensive weevil attack. J. Entomol. Soc. 

British Columbia 79: 62-65. 

ALFARO, R.I. 1989. Stem defects in Sitka spruce induced by Sitka spruce weevil, Pissodes strobi Peck, p. 177-185. In: 

Alfaro, R.I., and S. Glover, eds. Insects Affecting Reforestation: Biology and Damage. Proceedings of a IUFRO 

symposium held on July 3-9, 1988, in Vancouver, B.C., Canada, under the auspices of the XVIII International 

Congress of Entomology. Forestry Canada, Victoria, B.C. 

ALFARO, R.I., PIERCE H.D., Jr., BORDEN, J.H., and OEHLSCHLAGER, A.E. 1980. Role of volatile and nonvolatile 

components of Sitka spruce bark as feeding stimulants for Pissodes strobi Peck (Coleoptera: Curculionidae). Can J. 

Zool. 58: 626-632. 

ALFARO, R.I., BORDEN, J.H., HARRIS, L.J., NIJHOLT, W.W., and MCMULLEN, L.H. 1984. Pine oil, a feeding 

deterrent for the white pine weevil Pissodes strobi Peck (Coleoptera: Curculionidae). Can. Ent. 116: 41-44. 

ALFARO, R.I. and Borden, J.H. 1982. Host selection by the white pine weevil, Pissodes strobi Peck: feeding bioassays 

using host and nonhost plants. Can. J. For. Res. 12: 64-70. 

ALFARO, R.I. and BORDEN, J.H. 1985. Factors determining the feeding of the white pine weevil on its coastal British 

Columbia host, Sitka spruce. Proc. Ent. Soc. Ont. 116 (supplement): 63-66. 

ALFARO, R.I. and YING, C.C. 1990. Levels of Sitka spruce weevil, Pissodes strobi Peck, damage among Sitka spruce 

provenances and families near Sayward, British Columbia. Can. Ent. 122: 607-615. 

BRIDGEN, M.R., HANOVER, J.W., and WILKINSON, R.C. 1979. Oleoresin characteristics of eastern white pine seed 

sources and relationship to weevil resistance. For. Sci. 25: 175-183. 

BROOKS, J.E., BORDEN, J.H., PIERCE, H.D., JR., and LISTER, G.R. 1987a. Seasonal variation in foliar and bud 

monoterpenes in Sitka spruce. Can. J. Bot. 65: 1249-1252. 

BROOKS, J.E., BORDEN, J.H., and Pierce, H.D., Jr. 1987b. Foliar and cortical monoterpenes in Sitka spruce: potential 

indicators of resistance to the white pine weevil, Pissodes strobi Peck (Coleoptera: Curculionidae). Can J. For. Res. 

17: 740-745. 

CARLSON, J.E., HONG, YONG-PYO, and KISS, G. 1994. DNA markers associated with weevil resistance in interior 

spruce. Presented at the National white pine weevil workshop Richmond, British Columbia, January 19-21. In press. 

157

COZLN_, R.D. 1983. The spruce weevil, Pissodes strobi Peck (Coleoptera: Curculionidae). A review of its biology, 

damage and control techniques with reference to the Prince George Timber Supply area. B.C. Ministry of Forests 

Prince George. For. Serv. Intern. Rep. PM-PG-3. 

I::URNISS, R.L. and CAROLIN, V.M. 1977. Western forest insects. Misc. Pu. 1339. Washington, DC: U.S. Department of 


_IONG, YONG-PYO, GLAUBITZ J.C., PONOY, B., and CARLSON, J.E. 1992. DNA fingerprinting of conifer genomes 

using RAPD markers, p. 8-22. In Stomp, A.M., ed. Proc. Intl. Conifer Biotech Working Group Meeting, April 

23-28, 1992, Raleigh, North Carolina. 

KISS, G.K. and YANCHUK, A.D. 1991. Preliminary evaluation of genetic variation of weevil resistance in interior spruce 

in British Columbia. Can. J. For. Res. 21: 230-234. 

KISS_ G. and YEH, F.C. 1988. Heritability estimates for height for young interior spruce in British Columbia. Can. J. For. 

Res_ 18: 158-162. 

KHNE, L.N. and MITCHELL, R.G. 1979. Insects affecting twigs, terminals and buds. In Rudinsky, J.A., ed. Forest Insect 

Survey and Control. O.S.U. Book Stores, Inc. 

MACSIURTAIN, M.P. 1981. Distribution, management, variability and economics of Sitka spruce (Picea sitchensis 

(Bong.) Cart.) in coastal British Columbia. M.Sc thesis. University of B.C., Faculty of Forestry, Vancouver, B.C. 

MANVILLE, J.F., NAULT, J., VON RUDLOFF E., YANCHUK, A., and KISS, G.K. 1994. Spruce terpenes: Expression 

and weevil resistance. Presented at the National white pine weevil workshop Richmond, British Columbia, January 

19-21. In press. 

PAINTER, R.H. 1951. Insect resistance in crop plants. MacMillan Pub. Co. New York. 520 p. 

PLANK, G. H. and GERHOLD, H.D. 1965. Evaluating host resistance to the white pine weevil, Pissodes strobi, using 

feeding preference tests. Ann. Ent. Soc. Am. 58: 527-532. 

ROBERTS, D.R. 1994. Somatic embryogenesis for mass propagation of weevil resistant spruce. Presented at the National 

white pine weevil workshop Richmond, British Columbia, January 19-21. In press. 

SILVER, G.T. 1968. Studies on the Sitka spruce weevil, Pissodes sitchensis, in British Columbia. Can. Ent. 100: 93-110. 

STROH, R.C. and GERHOLD, H.D. 1965. Eastern white pine characteristics related to weevil feeding. Silvae Genetica 14: 

160-169. 

TAYLOR, S., ALFARO, R.I., and LEWIS, K. 1991. Factors affecting the incidence of white pine weevil damage to white 

spruce in the Prince George Region of British Columbia. J. Entomol. Soc. B.C. 88: 3-7. 

VAN BUIJTENEN, J.P. and SANTAMOUR, F.S. 1972. Resin crystallization related to weevil resistance in white pine 

(Pinus strobus). Can. Ent. 104: 215-219. 

VANDERSAR, T.J.D. and BORDEN, J. H. 1977. Visual orientation of Pissodes strobi Peck (Coleoptera: Curculionidae) in 

relation to host selection behaviour. Can. J. Zool. 55: 2042-2049. 

WILKINSON, R.C. 1985. Comparative white-pine weevil attack susceptibility and cortical monoterpene composition of 

western and eastern white pines. For. Sci. 31: 39-42. 

WOOD, R.O. and MCMULLEN, L.H. 1983. Spruce weevil in British Columbia. Can. For. Serv. Pac. For. Res. Cent. For. 

Pest Leafl. 2. 

YING, C.C. 1991. Genetic resistance to the white pine weevil in Sitka spruce. B.C. Min. of For. Res. Note 106. Victoria, 

B_C.,Canada. 17 p. 

158

A SUGI CLONE HIGHLY PALATABLE TO HARES: 

SCREENING FOR BIOAC_IIVE CHEMICA S 

HIROFUMI ttlRAKAWA, TADAKAZU NAKAStIIMA, YOSttlOKI HAYASHI, 

SEIJI OHARA, and TSUTOMU KUWAHATA 

Forestry and Forest Products Resealch Institute 

PO Box 16, Tsukuba-Norin, Ibaraki 305, Japan 

INTRODUCTION 

Herbivores often selectively feed on some plant individuals over other conspecifics. Such differences in palatability 

are usually chemically determined. Although typically confounded with environmental factors (Rousi et al. 1989, 1991, 

1993), genetic control of the differences in palatability observed among geographic origins (Matsuura and Maeda 1981; 

Radwan et al. 1982; Hansson et al. 1986; Rousi et al. 1989, 1991), among families (Rousi et al. 1989, 1991), and among 

clones (Radwan 1972, Hara 1973, Dimock et al. 1976, Kuwahata and Tanbara 1983) is obvious. 

Although mammalian herbivores can cause severe damage to trees, it is usually difficult to control such animals, 

because of economical, aesthetic, and ethical reasons. Breeding for resistance is one of the possible solutions to this problem. 

Therefore, efforts have been made to find the chemical agents responsible for the variation in palatability among trees. 

Atetsu 4 (A4), an elite sugi clone, Cryptomeria japonica, which originated in Atetsu, Okayama, Japan, is extremely 

palatable to hares. Its high palatability was first observed at a test planting site of 32 sugi clones, where A4 was highly 

damaged by hares just 1 year after planting. Later, an unusually high mortality rate was found in A4 at another test site of 37 

clones 5 years after planting, and damage by hares were considered to be the main cause of the mortality (Kuwahata and 

Tanbara 1983). 

This paper reconfirms the high palatability of A4 to hares, analyzes some of the chemical properties of the three sugi 

clones, and tests the palatability of crude chemical extracts and separated fractions using filter paper trials. 

MATERIALS AND METHODS 

Plant Materials and Hares 

The palatable sugi clone, A4, and two other unpalatable sugi clones, Uda 36 (U36) and Niimi 4 (N4), were used. The 

clones were transplanted to the nursery of the Forestry and Forest Products Research Institute in Ibaraki, and leaves were 

collected from trees at each feeding trial and, when stocked for use in chemical analysis, frozen at -43°C. Captive hares, 

Lepus brachyurus, kept in pens at the institute, were used for palatability tests. The hares were outdoors and maintained on 

commercial food pellets. They were either born in captivity or captured in the wild soon after birth. 

Palatability Test 

Feeding trials were conducted using two hares in separate 3.6 x 3.6 m pens. Fresh leaves or extracted residues of the 

clones were put on shallow mesh trays (24 x 32 x 1cm) and offered to the hares. Consumption was monitored by measuring 

the weight of the remnants over several days. Corrections were made for changes in leaf weight due to dehydration by 

measuring controls. 

Mattson, W.J., Niemel_i, P., and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. 


159

Filter paper tests were used to examine the palatability of fractions. Of six pieces of filter paper (5 x 8 cm, 1 g), two 

were treated with the extract of A4 and two with that of N4; the remaining two were left untreated as a control. The amount 

of fractions applied to each piece of filter paper was a one-gram dry-leaf equivalent. These were offered to hares kept in 

cages (80 x 80 x 80 cm) or pens (1.8 x 3.6 m), for one night and the amount (area) of filter paper left uneaten was measured. 

Commercial food pellets were always available during the trials, hence the intake of leaves or filter paper by hares 

was entirely voluntary. 

Seasonal Stability of Palatability and Solvent Extracts 

Seasonal stability of the palatability of fresh leaves of A4, U36, and N4 was investigated by feeding trials every 3 

months from May 1988 to May 1989. Leaves were taken from the trees and cut into approximately 5 to 10cm lengths. The 

leaves of each clone were put on two trays, which were set at opposite positions in a pen to prevent biased consumption due 

to tray position. 

For the same materials, solvent extracts were investigated: essential oils were taken by steam distillation for 10 h; 

dichloromethane and methanol extracts were consecutively obtained using a Soxhlet extractor. Leaves were cut into small 

pieces (< 3 mm in length) with pruning shears before the analysis. 

Effects of Extraction on Palatability 

Two series of consecutive extractions were made: (a) steam distillation for I0 h followed by methanol extraction; and 

(b) extraction by n-hexane and then methanol. The latter was done in a Soxhlet extractor. Leaves were cut into small pieces 

(ca. 3 mm in length) with pruning shears before extraction. 

Hot-water extraction was made using a common kitchen pot (capacity: 5 1) and a large steam-heated stainless pot for 

mass extraction (capacity: 50 1). Leaves of 10 to 20 cm in length were used. Leaves in the kitchen pot were boiled for 7 days 

(6 h per day) with water changed twice a day. In the steam-heated pot, leaves were boiled for 8, 16, or 24 h and once for 5 

days with water changed daily. 

The palatability of residual leaves were investigated by cafeteria test. Comparisons of palatability were made among 

the clones after each extraction. 

Effects of Methanol Extracts on Fresh and Residual Leaves 

The fresh leaves and residue after extraction of A4 and N4 were treated with the methanol extracts of A4 and N4 and 

feeding trials were made to compare the palatability before and after treatment. Methanol extracts were taken by soaking 

dried fresh leaves (600 g wet weight) first in 36 1 methanol for 13 days and then 18 1methanol for 8 days. The extracts were 

combined and concentrated in a vacuum evaporator to a small volume. The concentrate was then sprayed on fresh leaves or 

leaf residue with the amount that was originally included in the fresh leaves. 

Palatability of Extracts 

Fresh leaves of 10 to 20 cm in length (260 g wet weight) were put in large flask (capacity: 10 1)with methanol and 

heated in a hot water bath. The 2 h extraction was repeated twice for the same material each time with 7 1 methanol. The 

extracts were then combined (MW). 

The combined extract (MW) was added to 500 ml water and then evaporated to approximately 400 ml, and the waterinsoluble 

material was precipitated. Water-soluble (W) and -insoluble (M) fractions were separated by decanting. The water' 

soluble fraction was then consecutively extracted with n-hexane (Wh), ethyl acetate (We), and butanol (Wb) to leave a 

residual water fraction (Ww). The water-insoluble fraction was dissolved with methanol and then extracted with n-hexane 

(Mh) to leave a residual methanol fraction (Mm). 

The palatability of the various fractions was examined by filter paper tests. Each fraction was tested in pairs of A4 

and N4 .... 

160

RESULTS 

Seasonal Stability of Palatability and Solvent Extracts 

A4 showed highpalatabilitythroughout theyear andwas completelyconsumedin thefirst two nights with only one 

exception (Fig. 1). In contrast, U36 andN4 were not eatenmucheven afterA4 was depleted. In one instance,however, U36 

waseaten afterA4 was depleted although N4 did not show any such tendency. The two hares showed a notablesimilarityin 

their feeding. 

g 6O 

4O 

2 

Atetsu 4 Uda 36 Niimi 4 

May 1988 Aug. 1988 Nov. 1988 Feb. 1989 May 1989 

, J | i i 

g 60 "[... 

40 

20 

O • o-_¢--_o_ 

0 1 23 4 01 23 4 01 234 0 1 23 4 0 1 23 4 

Days of exposure 

Figure 1.--Palatability of fresh leaves of three sugi clones indifferent seasons. Vertical axis represents the weight of 

uneaten leaves remaining. Upper and lower graphs each represent the results for one hare. 

The mass of essential oils in A4 was much less than that in U36or N4 throughout the year (Fig. 2). Also, the amount 

of dichloromethane extract from A4 was much less than that from the other two unpalatable clones in all seasons. In contrast, 

the amount of methanol extract was not much different among the three clones. 


O O-- --,-- 

a: Essential oils b: Dichlorornethane c: Methanol extract (%) 

32 /__(ml_lOOgd'm') 108 306_ 4z 

extract (%) 40 20 _* 

0 May Aug.Nov. Feb. May May Aug.Nov. Feb. May May Aug.Nov. Feb. May 

Figure 2.--Seasonal fluctuation of essential oils and solvent extracts in three sugi clones. 

161

Effects of Extraction on Palatability 

When boiled in a common kitchen pot for 7 days, the residual leaves of A4 still retained high palatability. However, 

boiled in the steam-heated pot for 5 days, A4 lost its palatability to hares, although it retained high palatability after boiling 

for up to 24 h. The percent dry matter of the fresh leaves used in these trials was 35-39%; after boiling, residual leaves were 

22-25% of the original fresh weight. Thus, about one-third of the dry matter was extracted by boiling. However, when 

boiled in the steam-heated pot for 5 days, the residue of A4 was reduced in weight to 18%of the original, whereas N4 

remained at 24%. 

When essential oils were removed by steam distillation, A4 retained its palatability, but further extraction with 

methanol made it unpalatable (Fig. 3). Similarly, after extraction with n-hexane, A4 residue did not lose palatability but 

further extraction with methanol rendered it unpalatable. In contrast, the unpalatability of U36 and N4 seemed to diminish 

after each extraction. 


a: Fresh leaves b:Steam-distilled +MeOH-extracted c:Hexane-extracted+MeOH-extracted 

0 t , , .. , , 

_ _ I I 1 I _'I_.l_l, .ll[1._ * __l.J 

O I Z 3 0 I Z 3 0 I Z 3 0 I Z 3 0 I 2 3 

Days of exposure 

Figure 3.--Change of palatability of three clones after extraction. Vertical axis represents the weight of uneaten leaves 

or extracted residue leaves. Upper and lower graphs each represent the results for one hare. 

Effects of Methanol Extracts on Fresh and Residual Leaves 

Application of methanol extracts from dried A4 leaves did not affect the palatability of fresh leaves of A4 and N4 but 

made the residues after extraction palatable in both clones. In contrast, application of N4 extract made the fresh leaves of A4 

unpalatable, but showed no effect on the fresh leaves of N4 nor the residual leaves after extraction in both clones. 

Palatability Test of Fractions 

The filter paper test was not valid for some hares, because they showed little interest in filter paper and ate it only 

occasionally. One hare showed constant interest and gave suggestive results. Most of the fractions of A4 extracts showed 

high palatability, whereas some of the fractions of N4 extracts showed only weak palatability (Fig. 4). Little of the untreated 

filter paper was consumed. 

Of A4 extracts, both water-soluble (W) and water-insoluble (M) fractions were highly favored. The Wh fraction was 

less palatable than tile others. The Mh fraction also showed less palatability than did the Mm. For N4, although the MW 

extract was moderately consumed, the separate water-soluble and water-insoluble fractions showed the least consumption. 

The water-soluble fractions showed light (Wh) to fairly high palatability (Wb), whereas tile Mh was taken only in small 

amounts. 

162 

i

-

al. 1984, Tahvanainen et al. 1985, Clausen et al. 1986, Hansson et al. 1986, Snyder 1992). Although no specific stimulants 

have been reported to be involved in differences in palatability between genotypes, the possibility that A4 has some specific 

substance(s) that enhance browsing by hares has not been disproved. Our present results only suggest that volatile terpenes 

and water-soluble sugars do not have a decisive effect on the differences in the palatability between A4 and other clones. 

The A4 clone :isa genotype showing extremely high palatability compared to other plants of the same species. Other 

elite clones originating in Atetsu (Atetsu 1, 3, 5, 6) do not show higher palatability than clones from other areas, hence the 

palatability of A4 is difficult to attribute to its geographic origin (Kuwahata and Tanbara 1983). Although many studies 

describe large palatability variation within a species, A4 still seems exceptional. 

SUMMARY 

We examined the chemical factors determining the extremely high palatability of a sugi clone, A4, to hares. The 

palatability of A4 was fairly stable throughout the year. Although A4 had much less essential oil and dichloromethane 

extracts than the other clones, steam distillation (or boiling) and extraction by n-hexane did not reduce the palatability. 

Extraction by methanol made A4 unpalatable, whereas it seemed to make other unpalatable clones more palatable. Methanol 

extracts of A4 had a stimulating effect on feeding whereas that of N4 had a deterrent effect. A4 might contain specific 

feeding stimulants that are insoluble in n-hexane, soluble in methanol, and difficult to dissolve in water. 


BRYANT, J.P. 1981. Phytochemical deterrence of snowshoe hare browsing by adventitious shoots of four Alaskan trees. 

Science 213: 889-890. 

BRYANT, J.P., WIELAND, G.D., REICHARDT, P.B., LEWIS, V.E., and MCCARTHY, M.C. 1983. Pinosylvin methyl 

ether deters snowshoe hare feeding on green alder. Science 222: 1023-1025. 

CHtBA, S., OGAWA, A., NAGATA, Y., and TOMAKI, K. 1991. Effects of solvent extracts on resistance of Japanese larch 

and Kurile larch to voles. Proc. Hokkaido Branch, Japan Wood Res. Soc. 23:70-72 (in Japanese). 

CLAUSEN, T.P., REICHARDT, P.B., and BRYANT, P.B. 1986. Pinosylvin and pinosylvin methyl ether as feeding deterrents 

in green alder. J. Chem. Ecol. 12:2117-2131. 

DIMOCK, E.J., SILEN, R.E., and ALLEN, V.E. 1976. Genetic resistance in Douglas-fir to damage by snowshoe hare and 

black-tailed deer. For. Sci. 22: 106-121. 

FARENTINOS, R.C., CAPRETTA, P.J., KEPNER, R.E., and LITTLEFIELD, V.M. 1981. Selective herbivory in tasseleared 

squirrels: role of monoterpenes in ponderosa pines chosen as feeding trees. Science 213:1273-1275. 

HANSSON, L., GREE R., LUNDREN, L., and THEANDER, O. 1986. Susceptibility to vole attacks due to bark phenols 

and terpenes in Pinus contorta provenances introduced into Sweden. J. Chem. Ecol. 12: 1569-1578. 

HARA, M. 1973. The variation of hare damage among sugi clones. Proc. Tree Breed. Soc. 63-66 (in Japanese). 

HIRAKAWA, H. 1989. Screening of the preferential feeding substance(s) of the Japanese hare (Lepus brachyurus) included 

in the Japanese cedar (Sugi, Cryptomeriajaponica). Proc. Kanto Branch, Jap. Soc. For. 41:145-147 (in Japanese). 

KUWAHATA, T. and TANBARA, T. 1983. The variation of susceptibility to hare damage among sugi elite clones. Proc. 

Kansai Branch, Jap. For. Soc. 34:213-216 (in Japanese) 

MATSUURA, T: and MAEDA, M. 1981. Damage from red-backed vole (Clethrionomys rufocanus bedfordiae) to Saghalien 

fir (Abies sachalinensis) varies with each strain. For. Tree Breed. Hokkaido 24:15-20 (in Japanese). 

164 

il

OGAWA, A., NAGATA, Y., and SUKENO, S. 1992. Effects of solvent extraction on resistance of the hybrid between 

Japanese larch and Kurile larch to voles. Proc. Hokkaido Branch, Japan Wood Res. Soc. 24:61-64 (in Japanese). 

RADWAN, M.A. 1972. Differences between Douglas-fir genotypes in relation to browsing preference by black-tailed deer. 

Can. J. For. Res. 2: 250-255. 

RADWAN, M.A., CROUCH, G.L., HARRINGTON, C.A., and ELLIS, W.D. 1982. Terpenes of ponderosa pine and feeding 

preferences by pocket gophers. J. Chem. Ecol. 8:241-253. 

REICHARDT, RB., BRYANT, J.P., CLAUSEN, T.R, and WlELAND, G.D. 1984. Defence of winter-dormant Alaska paper 

birch against snowshoe hares. Oecologia 65: 58-69. 

ROUSI, M., TAHVANAINEN, J., and UOTILA, I. 1989. Inter- and intraspecific variation in the resistance of winterdormant 

birch (Betula spp.) against browsing by the mountain hare. ttolar. Ecol. 12: 187-1192. 

ROUSI, M., TAHVANAINEN, J., and UOTILA, I. 1991. A mechanism of resistance to hare browsing in winter-dormant 

European white birch (Betula pendula). Amer. Natur. 137: 64-82. 

ROUSI, M., TAHVANAINEN, J., and UOTILA, I. 1993. Effects of shading and fertilization on resistance of winter- 

dormant birch (Betula pendula)to voles and hares. Ecology 74: 30-38. 

SNYDER, M.A. 1992. Selective herbivory by Abert's squirrel mediated by chemical variability in Ponderosa pine. Ecology 

73: 1730-1741. 

TAHVANAINEN, J., HELLE, E., JULKUNEN-TIITTO, R., and LAVOLA, A. 1985. Phenolic compounds of willow bark as 

deterrents against feeding by mountain hare. Oecologia 65:319-323. 

165

GENETIC AND PHENOTYPIC VARIATION IN THE INDUCED REACTION OF 

SCOTS PINE TO LEPTOGRAPHIUM WINGFIELDII: 

REACTION ZONE LENGTH AND FUNGAL GROWTH 

F. LIEUTIER 1,B. LANGSTROM z, H. SOLHEIM 3 

C. HELLQVIST 2, and A. YART l 

_Station de Zoologie Foresti_re, Institut National de la Recherche Agronomique, Ardon, 45 160 Olivet, France 

2Division of Forest Entomology, Swedish University of Agricultural Sciences, S-77073 Garpenberg, Sweden 

3Department of Forest Ecology, Norwegian Forest Research Institute, As-NLH, N 1432, Norway. 

INTRODUCTION 

The induced defensive reaction of phloem generally plays a decisive role in the resistance of conifers to attack by 

bark beetles and their associated fungi. It is thought to be induced by fungi (Reid et al. 1967, Berryman 1972, Raffa and 

Berryman 1983, Christiansen and Horntvedt 1983, Matson and Hain 1985, Christiansen et al. 1987, Lfingstr6m et al. 1992), 

or by the boring activity of the beetle (Lieutier et al. 1988, Lieutier 1993). In most cases, by stimulating that energy demanding 

response, the fungi can lower the threshold of attack density above which the tree is overwhelmed and beetle attacks are 

successful. In all situations, the individual reaction is the basic mechanism of tree resistance. 

The between-tree variability in the threshold of attack density (Waring and Pitman 1980, Mulock and Christiansen 

1986), as well as the between-tree variability in the induced reaction itself (Shrimpton and Reid 1973, Peterman 1977, 

Lieutier and Ferrell 1989, Lieutier et al. 1993), has been linked to tree vigor. However, the genetic contribution to this 

variability has never been studied. 

In Scots pine, Pinus sylvestris L., an induced reaction is essential to resist attacks by lps sexdentatus Boern. and 

Tomicus piniperda L., and their associated fungi (Lieutier et al. 1988, 1989a; Solheim and L_ngstr6m, 1991; L_ngstr_m et al. 

1992). The blue-staln fungus, Leptographium wing_i'eldii Morelet, associated with T. piniperda, does not seem to play a role 

in the development of the induced reaction and in the success of attacks (Lieutier et al. 1988, 1989b). However, artificial 

inoculation of its sporulated cultures into Scots pine always incites strong defensive reactions (Lieutier et al. 1989a, Solheim 

and L_ngstr6m 1991) and the tree's response depends on the number of spores present in the inoculum (Lieutier et al. 1989b). 

Additionally, it may play a role in tree mortality, since weakened Scots pine can be killed by a high density of artificial 

inoculations (Solheim et al. 1993). 

The present study is part of a cooperative project with the following objectives: (1) to study the variation in the 

induced reactions (reaction zone formation, sapwood occlusion) between Scots pine provenances in two different climatic 

regions, (2) to study the performance of a French and a Swedish strain of L. wingfieldii in these provenances under different 

climatic conditions, and (3) to study the defensive chemistry involved in the induced reactions, particularly the roles of resin 

acids and phenolics, and to search for chemical markers for resistance among these groups of chemicals. This study addresses 

the first two issues. 

Mattson, W.J., Niemelfi, P., and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. 

USDA For. Serv. Gen. Tech. Rep. NC-I83, N.C. For. Exp. Sta., St. Paul, MN 55108. 

166


Field Work 

In early spring, 1991, provenances of Scots pine were selected in central France (Arboretum des Barres, Nogent-sur- 

Vernisson) and in southern Sweden (Remningstorp). Eleven (55 trees) and l0 (60 trees) provenances were chosen in France 

and Sweden respectively, so as to cover as wide a geographic distribution as possible, while taking into account the necessity 

of considering comparable provenances between the 2 countries (Table 1). In addition, the best lodgepole pine, Pinus 

contorfa, provenance was selected at Remningstorp (6 trees) for a separate comparison with the Scots pines. French trees had 

been planted between 1951 and 1957, except for 2 provenances planted in 1941 (Table 1), using 3-year-old seedlings. Their 

DBH ranged fiom 14-22.5 cm. Swedish trees had been planted in 1962 using 4-year-old seedlings and their DBH ranged 

from 10.5-18 cm. 

All trees were inoculated with a French and a Swedish strain of L. winglqeldii, originally isolated from T. piniperda in 

France (Lieutier et al. 1989a) and in Sweden (Solheim and LSngstr6m 1991) respectively. The inoculation materials were 2week-old 

cultures on malt agar. Holes were made to cambium depth with a 5-ram cork borer, 5-ram pieces of agar culture 

were placed in the holes, and the bark plugs were put back into the tree. In this study, we inoculated five rings spaced (ca. 30 

cm apart) on the lower stem (:from ca. 60 to 180 cm stern height) in six points spaced evenly around the circumference. Each 

of two spots at opposite sides of the stem received sterile agar (control), with French and Swedish strains of L. wing/i'e/dii. In 

order to avoid lesions growing into each other, inoculation points in subsequent rings were displaced by 30 degrees relative to 

the previous ring. The inoculations took place from 8 to 12April in France and on 26 April in Sweden. 

Three weeks after the inoculations, the reaction zones, i.e., the resin-soaked lenticular zones (often referred to as 

lesions) surrounding the points of inoculation, were sampled as follows: All 6 inoculation points of ring 2, and one of the two 

inoculation points of' each treatment in ring 4 (i.e., 3 inoculation points per treatment in total), were carefully exposed by 

removing the outer bark. The reaction zones were measured on the external face of the phloem upwards and downwards 

from each inoculation point. For re-isolation of the inoculated fungi, the right half of two reaction zones of each treatment 

was cut out, put individually in plastic bags, and cold-stored immediately at 2°C. The remaining reaction zones were put in 

paper bags or in vials, immediately frozen on dry ice, and stored for later chemical analyses. Two unwounded phloem 

samples were also collected in each tree and preserved the same way as the remaining reaction zones for chemical analyses. 

Six weeks after the inoculations, 3 other inoculation points per treatment, located in rings 3 and 4, were investigated 

in each tree and sampled as described above. Two new unwounded phloem samples were also taken. After sampling the 

reaction zones, 2 trees in France and 3 trees in Sweden were felled in each provenance, and discs were cut at rings 1,3 and 5. 

These discs were stored in a cold room (2°C) for later examination of sapwood reaction zones and re-isolation of the fungi. 

The total height of the felled trees and the lengths of their current leader shoots were measured. 

Twenty-one weeks (24 in Sweden) after the inoculations, the reaction zones were sampled (rings I and 5) from the 

remaining trees as described above. Trees were felled and discs were cut as described above also. 

Diameter and height of the trees differed significantly between provenances in both France and Sweden. In France, 

the provenances "Rud" and "Die" had the highest dimensions, while "Spe", "Seg" and "Fin" had the lowest. In Sweden, the 

native provenances "Vit", "Vfis" and "Kil" were the tallest and had the biggest diameter, while the provenances "Mal" and 

"Enz" were very small and thin. The provenance "Tab" was peculiar, as it had the lowest height although its diameter was 

comparable to other provenances. These trees were the last ones left at the plot after snow-break damage. On the contrary, 

lodgepole pine was taller, although its diameter did not differ from Scots pine provenances. 

In France, no between-provenance differences could be noticed in the length of the current leader shoots, except in 

September 1991, but damage by T. pit@erda did not permit any conclusions. In Sweden, significant provenance differences 

existed. Shoots of the foreign provenances were generally shorter than that of the local ones. Lodgepole pine also had long 

shoots. Six weeks after inoculations, the trees in France had completed ca. 2/3 of their leader growth, whereas trees in 

Sweden had only completed ca. 1/2 of theirs. 

167

Table 1.--Characteristics of the Scots pine provenances and of P. contorta studied in Sweden and in France. 

Designation origin latitude longitude altitude age 

(m) (years) 

Proven ances studied in France (Nogent-sur-Vernisson): 

Fin (Finland) 40 

Kal Kalmar (Sweden) 56° 30' 16° 20' 40 37 

Spe Spey valley (U.K., 57° 20' 03° 30'W 41 

Scotland) 

Rud Rudczany (Poland) 53° 40' 21° 29' 130 53 

Klo Klosterreichenbach 48° 30' 08° 25' 43 

(Germany, Schwarzwald) 

Niz Saint Nizier de Fornas 45° 25' 04° 05' 900 42 

(France, Loire) 

Die Saint Di6 (France, 48° 15' 07 ° 400 43 

Vosges) 

Mat La Matte des Angles 42" 36' 02° 07' 1520 53 

(France, Pyr_ndes) 

Ser (Serbie) 43 ° 21° 37 

Seg Segovie (Spain) 40° 52' 04°W 1400 41 

Can Canakkale (Turkey) 40° 10' 26° 25' 37 

Provenances studied in Sweden (Remningstorp): 

168 

Mal Malfi (Sweden) 65° 38' 21° 07' 75 33 

Sve Sveg (Sweden) 62" 03' 14° 19' 385 33 

Kil Kilafors (Sweden) 61° 08' 16° 31' 150 33 

Vat Varnhem (Sweden) 58° 22' [3 ° l l' 270 33 

V_is V_istervik (Sweden) 57° 46' 16° 39' 20 33 

Vit Vittsk6vle (Sweden) 55° 51' 14° 01' 30 33 

O1o Olonets (Russia, 61° 33° 1O0 33 

Carelia) 

]nv Inverness (U.K., 57° 10' 05°W 300 33 

Scotland) 

Tab Taborz (Poland) 53° 45' 20° 06' 130 33 

Enz Enzkl6sterle 48* 45' 08° 30' 700 33 

(German_, Schwarzwald) 

Con Stuart lake (Canada, 54* 30' 124° 15'W 600 33 

B.C.) = P. contorta 

i

Laboratory Procedures 

The week after collection, the samples were analyzed for fungus extension. From the reaction zones, phloem pieces 

were cut with a sterile razor blade at 5 mm intervals upwards and downwards from the inoculation points beyond the visible 

reaction zone. The sample pieces were placed on petri dishes with malt agar medium and incubated at 25°C for 1 week. 

Positive and negative records of L. wingfieldii were noted. Sapwood samples were handled the same way, except that 

sampling took place in the radial direction. The maximum extension of the sapwood reaction zone was also measured 

radially. 

Data Handling and Statistics 

All caiculations were done with the SAS statistical package. Preliminary analyses of vertical reaction zone lengths 

and fungal growth revealed no differences between the vertical extension of reaction or fungal zones upwards and downwards 

from the inoculation points regardless of the treatment (Student's t-test). Neither were there any differences between 

reaction or fungal zones from different inoculation rings of the same treatment. Hence, we pooled data within the same tree, 

and used tree-wise treatment means (based on 3 lesions or less in some cases with missing data) as calculation units. 

Treatment means were tested using one-way analysis of variance followed by Tukey's test for multiple comparisons. 

Two-way analysis of variance was also employed, as were pair-wise and standard t-tests, depending on the situation. Interdependency 

between different variables was explored by regression analyses. 

Reaction Zone Length 

Phloem 

Total length of the reaction zone varied significantly between provenances and between treatments (Fig. 1). In 

Sweden, local provenances generally exhibited longer reaction zones than foreign provenances at 6 and 24 weeks. The 

opposite tendency however, was visible in France at 6 weeks, where the longest reaction zones were observed in the southern 

foreign provenances, particularly the Spanish and Serbian provenances, while the shortest reaction zones belonged to the 

local provenances. At Remningstorp, lodgepole pine exhibited shorter reaction zones than all provenances of Scots pine after 

3 and6 weeks. Despite existing differences between provenances, no systematic trends (e.g., latitudinal) in reaction zone 

lengths could be detected. 

At Remningstorp, reactions to the two strains of L. wingfieldii were similar but, at Nogent-sur-Vernisson, the French 

strain caused a more extensive reaction zone than did the Swedish strain (Fig. 1 and 2). In both countries, reaction zones 

resulting from fungal inoculations were longer than reaction zones to the controls, but the latter were clearly longer in France 

than in Sweden. The correlation of the reaction zone length to the fungus increased more rapidly and to a higher level in 

Sweden than in France (Fig. 2). A plateau was generally reached after 6 weeks, except for the French fungus strain in France 

where the reaction zones still expanded after 6 weeks. A few provenances, however, did not fit with that description (Fig. 1). 

Few of the provenances studied occurred at both study sites. The pair "Kal" (France) and "Vas" (Sweden) originates 

from the same region and altitude in southeastern Sweden, and should hence be comparable. We can also assume that the 

Finnish "Fin" (France) and Carelian "O1o" (Sweden) as well as the pairs from Scotland, Poland, and Germany are fairly 

similar. However, a closer examination of these pairs of provenances does not disclose any deviations from the general 

results given above. 

Fungal Growth 

Total vertical fungal expansion varied significantly between provenances only for the 6 week sampling date (Fig. 3). 

At Nogent-sur-Vernisson, fungal growth was the most expanded in the southern, foreign provenance "Seg", and the least 

expanded in the French provenances, particularly the Pyrenean one. At Remningstorp, fungal growth was the most expanded 

169

Figure l.--Vertical extension of induced reaction zones in Scots pine phloem in response to artificial inoculations with a 

French (black) and Swedish (dotted) strain of Leptographium wingfieldii as well as a sterile control (white), in an 

experhnent in France (left) and Sweden (right) using different provenances of Scots pine and one of lodgepole pine 

(Sweden only). Samples were taken after 3 (A), 6 (B) and 21 (C) (France) or 24 (C) (Sweden) weeks; provenance 

abbreviations are explained in Table 1. Numbers of observations (trees) were 5 and 6 (France and Sweden, respectively) 

for the first two sampling dates, and 3 for last sampling date at both sites. Vertical bars indicate standard error. 

P-values refer to two-way-anovas for differences between treatments and provenances. 

17(i)

E 

120 

Nogent-sur-Vernisson Remningstorp 

c 

occ_ E. 80 11"_ i 

40 

E 20 

ot,,- 

E_ 

0 ......... _........... 

0 6 12 18 24 0 6 12 18 24 

30 ..... • ..... 

C_10 - _ 

C 

I1. 

0 0 6 12 18 24 0 6 12 18 24 

Weeks after inoculation Weeks after inoculation 

Figure 2.--Vertical progress in reaction zone size (upper row) and fungal growth (lower row) in Scots pine phloem in 

sterile control (triangle), in an experiment in France (left) and Sweden (right). Tree provenances were pooled. 

response to artificial inoculation with a French (circle) and Swedish (square) strain ofL. wingfieldii as well as a 

Samples were taken after 3, 6 and 21 (France) or 24 (Sweden) weeks. Numbers of observations (trees) were 5 and 6 

(France and Sweden, respectively) for the first two sampling dates, and 3 for last sampling date at both sites. Vertical 

bars indicate standard error. 

in the native provenance "Ki!". No significant differences were observed between treatments, except for the 21 week 

sampling date at Nogent-sur-Vernisson, where the French strain of L. wingfieldii grew further than the Swedish strain. 

However, in the Spanish provenance, that was also true for all sampling dates. In both locations, fungi expanded rapidly 

during the first 3 weeks, and thereafter stabilized (Fig. 2). At Nogent-sur-Vernisson, fungus expansion was higher for the 

French than for the Swedish strain after 21 weeks. The expansion of the Swedish strain even decreased at 21 weeks, except 

in the Pyrenean provenance (Fig. 3). 

17l

E 

Nogent-sur-Vernisson Remningstorp 

75 ..................................................................................................................................... 

E_ 50 

E 

i A Ptreatm= 0.1948 A Ptreatm = 0.6333 

r t t 

Pprov = 0.1211: Pprov = 0.0817 

0 .................................................... 

75 ............................................................................................................................................. 

------1 .............................. 

B Ptreatm= 0.1030 1 B Ptreatm= 0.9165 

i.+: 

o I t::i::i 

I 

E 50, Pprov = i0-0399_1 Pprov = 0.0099 

_'_2si , i 

It. 

]5 r......................................................................................................................................................................................... 

C Ptreatm= 0.0027 1 mal sve kil var vs vit olo inv tab enz con 

0.3669 

1 

E , I"Pr°V 

E , --I 

0 ...........L ........................................... 

] ]- 

fin kal spe rud klo niz die mat ser seg can 

Figure 3.---Vertical expansion of fungal growth after inoculation of a French and Swedish strain of L. wingfieldii in the 

172 

phloem of different pine provenances in an experiment in France (left) and Sweden (right). Samples were taken after 

3 (A), 6 (B) and 21 (C) (France only) weeks. Provenance abbreviations are explained in Table 1. Numbers of 

observations (trees) were 5 and 6 (France and Sweden, respectively) for the first two sampling dates, and 3 fk)r last 

sampling date (France only). Vertical bars indicate standard error. P-values refer to two-way-anovas for differences 

between treatments and provenances.

Interdependence Between Reaction Zone Length and Fungal Growth 

In all cases, fungat growth was much less expansive than reaction zone length (cf. Fig. l - 3). These two parameters 

were positively correlated with each other after 6 weeks (r=0.73 and 0.51 in France and Sweden, respectively, and after 21 

weeks (r=0.71, France only), but not after 3 weeks (r=0.38 and -0.27 in France and Sweden, respectively). 

Radial Occlusion and Fungal Growth Into Sapwood 

As sapwood occlusion was minimal (Fig. 4), data for the different provenances are not given in detail. After 6 weeks, 

the range in sapwood occlusion means was 0.4-2.1 mm and 0.2-3.1 mm in France and Sweden, respectively. After 21/24 

weeks, the corresponding ranges were 1.1-5.8 mm and 0.7-4.7 ram. Sapwood occlusion was always more developed after 

fungus inoculation than after control inoculation, and continued to expand between 6 and 21/24 weeks (Fig. 4). This was true 

for both localities. At Nogent-sur-Vernisson, the French strain of L. wingfieldii caused a larger sapwood occlusion than did 

the Swedish strain, 21 weeks after inoculation. 

,5 

4 

3 

2 

1 

OE 

4 

3 

Nogent-su r-Vernisson Remningstorp 

.._..t--& A 

o 0 6 12 18 24 0 6 12 18 24 

Weeksafter inoculation Weeksafter inoculation 

Figure 4._Radial occlusion of sapwood (upper row) and radial expansion of fungi (lower row) following inoculation with a 

French (circle) and Swedish (square) strain of L. wingfieldii as well as a sterile control (triangle), in different Scots 

pine provenances, in an experiment in France (left) and Sweden (right). Samples were taken after 6 and 21 weeks 

(France) or 6 and 24 weeks (Sweden). Provenance abbreviations are explained in Table 1. Numbers of observations 

(trees) were 2 and 3 (France and Sweden, respectively) for the 6 week sample, and 3 for the last sampling date (both 

sites). Vertical bars indicate standard error. 173 

T

As for radial occlusion, fungal growth in the sapwood was minimal (Fig. 4), and did not differ between provenances. 

Hence, data are not given in detail for the different provenances. After 6 weeks, the average fungal penetration into the 

sapwood ranged from 0.2-2.0 mm and from 0.2-6.0 mm in France and Sweden, respectively. After 21/24 weeks, the corresponding 

ranges were 0.0-1.6 mm and 0.0-9.0 ram. In both countries, fungi generally regressed or remained stable from 6 to 

21/24 weeks (Fig. 4). There were no differences between the 2 strains of fungi, but both fungi expanded further into the 

sapwood in Sweden than in France. 

DISCUSSION 

Comparison Between Provenances 

The observed differences between provenances in all variables studied (reaction zone length, sapwood occlusion, and 

fungal growth - vertically as well as radially) do not demonstrate a systematic pattern justifying general conclusions about 

provenance differences. In Sweden however, foreign provenances responded differently than native provenances to fungus 

infection. As this was not the case in France, no latitudinal or altitudinal trends can be detected. On the other hand, systematic 

differences in defense reactions may have been masked by the considerable inter-tree variation, possibly due to variations 

in micro-ecological conditions from one provenance plot to another, or even to genetic differences between trees of the same 

provenance. A larger number of replications would have been required. Even better would have been to work with clones 

instead of provenances. Thus in this experiment, reaction zone size alone did not provide enough intbrmation to evaluate 

possible genetic differences in host resistance to blue-stain fungi. 

In Sweden, lodgepole pine also responded in the same manner as foreign Scots pine provenances, which suggests it 

may be similarly susceptible to European bark beetles. 

Comparison Between Treatments 

As observed earlier, fungal inoculations produced larger reaction zones than the sterile zones (Wong and Berryman 

1977, Solheim 1988, Ross et al. 1992). For the variables measured, the two fungal strains caused larger differences in France 

than in Sweden, and this locality effect will be discussed below. The French strain of L. wingfieldii invariably caused larger 

reaction zones than the Swedish strain in France, and similar reaction zones in Sweden. As the pattern was similar for fungat 

growth, the French strain could be considered more aggressive than the Swedish one. The observed difference cannot be 

attributed to some environmental or host factor, as inoculations of the two strains occurred parallel in the same trees. Thus, 

some inherent difference between the fungal strains should cause this difference in aggressiveness. For sapwood, the French 

strain occluded more tissue than the Swedish strain after 21 weeks in France, hence its pathogenicity also could be higher. In 

all treatments however, sapwood occlusion was shallow, despite extensive reaction zone formation in the phloem (cf. 

Parmeter et al. 1992). This can be explained by the low density of inoculation, allowing the tree to contain the infection. 

Comparison Between Localities 

All variables measured indicated clear differences between localities. Reactions zones following fungal inoculations 

were larger in Sweden than in France, as was sapwood occlusion. That was also the case for fungal growth, although not so 

clearly tbr the phloem. In contrast, sterile inoculations caused larger reaction zones in France than in Sweden for the phloem, 

but not for the sapwood. 

As these differences were also true inside comparable pairs of provenances (e.g., "Kal" vs "Vas", "Fin" vs "Olo", 

Spe" vs "Inv", "Rud" vs "Tab" and "Klo" vs "Enz"), a provenance effect can be discarded. However, the finding that sterile 

inoculations resulted in larger reaction zones in France than in Sweden whereas fungal inoculations yielded the opposite 

pattern, is difficult to explain. This contradiction implies a higher sensitivity to mechanical wounding in France than in 

Sweden, but also a more efficient containment of the fungi in the former than in the latter case. Possibly, differences in 

prevailing temperature conditions may have favored fungal growth in Sweden, as L. wingfieldii is known to have a low 

temperature preference (Lieutier and Yart 1989). 

Although the experimental conditions were planned to be as similar as possible, trees may have differed in phenology 

between the localities. Judging from the relative shoot lengths 6 weeks after inoculation, Lorio's growth-differentiation 

174 

:i

alance hypothesis (Lorio 1986) cannot explain the locality effect on trees. Indeed, trees at Remningstorp were in a more 

intensive phase of growth than the trees at Nogent-sur-Vernisson, and should have had larger reaction zones in response to 

sterile inoculation. Age of the trees may be a more useful explanation, as trees were clearly older in France than in Sweden, 

and it has already been reported that older Scots pines have larger reaction zones than younger ones (Lieutier and Ferrell 

1989, Lieutier et hi. 1993). 

Although not demonstrated in our experiment, 

General 

genetic 

Conclusions 

variability in the tree's response to invading fungi (or beetles) 

probably exists in Scots pine. In addition, we have demonstrated that there is an inherent variability in strains of L. 

win_ieldii. However, phenotypic (geographic) variations concerning both tree and fungus, are superimposed on this pattern. 

This can be seen in the situation with larger reaction zones to fungus in Sweden than in France, whereas the opposite pattern 

occurredtions, for the control. One must keep in mind all these potential variations while studying tree/bark-beetle/fungus interac- 

SUMMARY 

The aggressiveness and pathogenicity of one French and one Swedish isolate of Leptographium wingfieldii were 

studied in different Scots pine provenances, both in France and Sweden. After inoculations, reaction zone length and fungal 

growth varied significantly between provenances, at least in the phloem, but no consistent trends could be detected. In 

France, fungal growth was more extended and the reaction zones were more developed for the French than for the Swedish 

strain, both in phloem and sapwood. In Sweden, fungal growth and reaction zone length were comparable for the two strains. 

Fungal growth and reaction zones induced by fungus inoculations were more extensive in Sweden than in France, but 

reaction zones induced by sterile wounds were larger in France. Fungus inoculations ' however, always induced larger 

reaction zones than sterile inoculations. 


This study is part of ongoing cooperative research between Institut National de la Recherche Agronomique (INRA), 

Swedish University of agricultural Sciences (SLU), and Norwegian Forest Research Institute (NISK), within the field of bark 

beetle/blue stain fungi/host tree interactions. A grant from the "Hildur and Sven Wingqvist's foundation for forest research" 

facilitated the Swedish part of the study, and is gratefully acknowledged. The CEMAGREF and the Arboretum des Barres at 

Nogent-sur-Vernisson are also acknowledged for providing experimental plots for the French part of the study. The authors 

express their gratitude to J. Garcia and P. Romary from INRA and to H. Knutsen, C. Nelleman and O. Olsen from NISK for 

their technical help. 


BERRYMAN, A.A. 1972. Resistance to bark beetle-fungi associations. BioScience 22: 598-602. 

CHRISTIANSEN, E. and HORNTVEDT, R. 1983. Combined Ips-Ceratocystis attack on Norway spruce, and defensive 

mechanisms of the trees. Z. angew. Entomol. 96:110-118. 

CHRISTIANSEN, E., WARING, R.H., and BERRYMAN, A.A. 1987. Resistance of conifers to bark beetle attack: searching 

for general relationships. For. Ecol. Manage. 22: 89-106. 

L]kNGSTROM, B., HELLQVIST, C., ERICSSON, A., and GREF, R. 1992. Induced defence reaction of Scots pine following 

stem attacks by Tomicus piniperda. Ecography 15:318-327. 

LIEUTIER, F. 1993. Induced defense reaction of conifers to bark beetles and their associated Ophiostoma species, p. 206- 

215. In Wingfield, M.J., Seifert, K.A., and Webber, J.E, eds. Ceratocystis and Ophiostoma: Taxonomy, Biology and 

Pathogenicity. Amer. Phytopathol. Soc. Press, St. Paul. 

175

LIEUTIER, E and FERRELL, G.T. 1989. Relationships between indexes of tree vigour and the induced defense reaction of 

Scots pine to a fungus associated with Ips sexdentatus (Coleoptera: Scolytidae), p. 163-178. In Payne, T.L. and 

Saarenmaa, H., eds. Integrated Control of Scolytid Bark Beetles. Virginia Polytech. Inst. St. Univ. 

LIEUTIER, E, YART, A., GARCIA, J. POUPINEL, B., and Lt_VIEUX, J. 1988. Do fungi influence the establishment of 

bark beetles in Scots pine? p. 317-330. In Mattson, W.J., Ldvieux, J., and Bernard-Dagan, C., eds. Mechanisms of 

Woody Plant Defenses Against Insects: Search for Pattern. Springer, New York. 

LIEUTIER, E, YART, A., GARCIA, J., HAM, M.C., MORELET, M., and Lt_VIEUX, J. 1989a. Champignons 

phytopathogbnes associds _ Ips sexdentatus Boern. et Tomicus piniperda L. (Coleoptera: Scolytidae) et dtude 

prdliminaire de leur agressivitd pour le pin sylvestre. Ann. Sc. For. 46: 201-216. 

LIEUTIER, F., CHENICLET, C., and GARCIA, J. 1989b. Comparison of the defense reaction of Pinus pinaster and Pinus 

sylvestris to attacks by two bark beetles (Coleoptera: Scolytidae) and their associated fungi. Environ. Entomol. 18: 

228-234. 

LIEUTIER, E, GARCIA, J., ROMARY, R, YART, A., JACTEL, H., and SAUVARD D. 1993. Inter-tree variability in the 

induced defense reaction of Scots pine to single inoculations by Ophiostoma brunneo-ciliatum, a bark beetle associated 

fungus. For. Ecol. Manage. 59: 257-270. 

LIEUTIER, F. and YART, A. 1989. Preferenda thermiques des champignons associds _ lps sexdentatus Boern. et Tomicus 

piniperda L. (Coleoptera: Scolytidae). Ann. Sc. For. 46:411-415. 

LORIO, Jr. RL. 1986. Growth-differentiation balance: a basis for understanding southern pine beetle-tree interactions. For. 

Ecol. Manage. 14: 259-273. 

MATSON, RA. and HAIN, F.R 1985. Host conifer defense strategies: a hypothesis, p. 33-42. In Safranyik, L. and 

Berryman, A.A. eds. The role of the host in the population dynamics of forest insects. Proc. IUFRO Conf. Banff 

Alberta Canada. 

MULOCK R and CHRISTIANSEN E. 1986. The threshold of successful attack by lps O,pographus on Picea abies: a field 

experiment. For. Ecol. Manage. 14: 125-132. 

PARMETER, J.R., SLAUGHTER, G.W., CHEN, M., and WOOD, D.L. 1992. Rate and depth of sapwood occlusion 

following inoculation of pines with blue stain fungi. For. Sci. 38: 34-44. 

PETERMAN, R.M. 1977. An evaluation of the fungal inoculation method of determining the resistance of lodgepole pine to 

mountain pine beetle (Coleoptera: Scolytidae) attacks. Can. Entomol. 109: 113-448. 

RAFFA, K.E and Berryman, A.A. 1983. The role of host resistance in the colonization behavior and ecology of bark beetles 

(Coleoptera: Scolytidae). Ecol. Monogr. 53: 27-49. 

REID, R.W., WHITNEY, H.S., and WATSON, J.A. 1967. Reactions of lodgepole pine to attack by Dendroctonus 

pseudotsugae Hopkins and blue stain fungi. Can. J. Bot. 45:I115-1126. 

ROSS, D.W., FENN, R, and STEPHEN, F.M. 1992. Growth of southern pine beetle associated fungi in relation to the 

induced wound response in loblolly pine. Can. J. For. Res. 22: 1851-1859. 

SOLHEIM, H. 1988. Pathogenicity of some lps typographus-associated blue-stain fungi to Norway Spruce. Medd. Nor. 

Inst. Skogforsk. 40:1 - l 1. 

SOLHEIM, H. and L]kNGSTROM, B. 199I. Blue-stain fungi associated with Tomicus piniperda in Sweden and preliminary 

observations on their pathogenicity. Ann. Sc. For. 48: 149-156. 

176

SOLHEIM, H., LANGSTROM, B., and HELLQVIST, C. 1993. Pathogenicity of the blue-stain fungi Leptographium 

wingfieldii and Ophiostoma minus to Scots pine: effect of tree pruning and inoculum density. Can. J. For. Res. 23" 

1438-1443. 

SHRIMPTON, D.M. and REID, R.W. 1973. Change in resistance of lodgepole pine to mountain pine beetle between 1965 

and 1973 Can. J. For. Res. 3:430-432 

• 

WARING, R.H. and PITMAN, G.B. 1980. A simple model of host resistance to bark beetles. Oreg. State Univ. for. Res. 

Lab. Res. Note 65. 2 p. 

WONG, B.L. and BERRYMAN, A.A. 1977. Host resistance to the fir engraver beetle. 3. Lesion development and contain 

ment of infection by resistant Abies grandis inoculated with Trichosporium symbioticum. Can. J. Bot. 55: 2358- 

2365. 

177

CAN PHLOEM PHENOLS BE USED AS MARKERS OF SCOTS PINE 

RESISTANCE TO BARK BEETLES? 

F. LIEUTIER _,F. BRIGNOLAS l, V. PICRON t, A. YART l, and C. BASTIEN 2 

tStation de Zoologie Foresti_re, Institut National de la Recherche Agronomique 

Ardon, 45 160 Olivet, France 

2Station d'Am61ioration des Arbres Forestiers, Institut National de la Recherche Agronomique 

Ardon, 45 160 Olivet, France 

INTRODUCTION 

Bark beetle damage is considerable in all coniferous forests in the temperate zone. In most cases, it results from 

beetle attacks surpassing the critical density (numbers/m 2)below which trees can successfully reject them. Above this 

threshold, natural mechanisms of tree resistance are overcome by the beetles and fungi which become successfully established. 

The level of this critical threshold of attack density is a fundamental expression of the strength or capacity of tree 

resistance (Berryman 1978, 1982; Christiansen et al. 1987). 

The basic phenomenon involved in the resistance is a phloem reaction induced by the invaders and localized around 

them. It consists of a rapid synthesis of chemicals, including secondary metabolites such as terpenes and phenols, which 

impregnate tree tissues (Reid et al. 1967, Berryman 1972, Shrimpton 1973, Christiansen and Horntvedt 1983, Christiansen et 

al. 1987, Lieutier et al. 1988, Lfmgstr6m et al. 1992, Lieutier 1993). The critical threshold of attack density depends on the 

quantity of energy the tree is able to rapidly mobilize for various syntheses in the reaction zones, and this quantity is necessarily 

limited (Christiansen et al. 1987). Thus, the threshold can be high if the lag time for stopping an aggressor at each 

point of attack is very short and hence the energy mobilized for that purpose in each reaction zone is low. Under these 

conditions of arrested attacks, a short reaction zone has been assumed to be more effective than an extended one (Lieutier et 

al. 1993). 

Markers of conifer resistance to bark beetles could be very useful; e.g., forecasting susceptibility to beetle attacks and 

beetle outbreaks, and aiding in selection of highly resistant trees. Because the concentrations of secondary metabolites vary 

considerably during tree responses to aggression, they merit a special interest. Preliminary studies on phloem phenolics have 

suggested dramatic changes in Scots pine, Pinus sylvestris L., in response to aggression (Lieutier et al. 1991a). This paper 

presents the results from two complementary approaches recently developed in France concerning the possible role of 

phenols in arresting bark beetle invasion and their use as biochemical markers of resistance. Detailed methods and other 

findings are presented elsewhere (Lieutier et al. 1995). 

METHODS 

Experimental Devices 

In our first experiment, six clones located in the INRA's nursery in Orlfians (France), each represented by one to three 

10-year-old trees, were used to test if a relationship existed between reaction zone length and phenolic composition. Trees 

were each inoculated with a malt agar sporulated culture of Ophiostoma brunneo-ciliatum Math.-K., a fungus previously 

isolated from lps sexdentatus Boern., and thought to play a role in the establishment of this beetle in Scots pine (Lieutier et 

al. 1988). Each tree received three inoculations. Three weeks later, the reaction zone around each inoculation point was 

measured and sampled. Three samples of unwounded phloem were also taken from each tree. 

Mattson, W.J., Niemel_i, P., and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. 


178

In a second experiment, three 30-year-old Scots pines located in the Orldans forest were inoculated with O. brunneo- 

ciliatum to study the kinetics of the trees' phenolic responses. Each tree received 33 inoculations distributed in 5 rings, each 

separated by 35 cm. After 3, 7, 14, 30 and 60 days, the reaction zones around respectively 15, 7, 5, 3 and 3 inoculation points 

Inoculations were always performed with 5 mm diameter discs of culture placed in cambium-deep, equally sized 

holes 

per tree 

made 

were 

with 

measured 

a cork borer, 

and sampled. 

and the 

At 

bark 

days 

plugs 

30 

were 

and 60, 

put 

samples 

back into 

of 

the 

unwounded 

holes, hnmediately 

phloem were 

after 

also 

being 

collected 

collected 

from 

from 

each 

the 

tree. 

trees, the phloem samples were frozen in dry-ice. In the laboratory, all samples were freeze-dried. 

second experiment, zone 

Phenols Extraction and Analysis 

3 

which were pooled all together (day 3) or by groups of 1 to 3 samples (day 7) to have sufficiently a high 

quantity of material for further chemical analyses. Extractions were performed at 4°C on ground samples. In order to 

remove resinous compounds (Alcubilla 1970), phloem powder was first washed with pentane before extraction of phenolics 

using 80% methanol. Pentane treatment was checked to have no significant effect on the compounds analyzed in our study. 

Gallic acid was used as internal standards for HPLC analyses. 

Analyses were performed with reversed HPLC (Waters 600E, Photodiode array detector 991), in a 250 mm column 

and 4 mm internal diameter. Stationary phase was a C-18 grafted silica (Merck, Lichrospher RRP18) with 0.005 mm porosity. 

Mobile phase was a mixture of a 1% acetic acid in ultra pure water (solvent A) with a solution of pure methanol, acetonitrile, 

acetic acid (49.5:49.5:1) (solvent B). The gradient is represented in Fig. 1. Readings were made at 310 nm. Results were 

expressed in internal standard equivalents per gram of freeze-dried powder. 

"" 100 

o,_ 

.¢ 

= _?IP.,l 

0.6 ....................... _ ..................................................... 

c_ 

ca ca gll 

ca ' 50 

0 4 ..... "_ = ........................ ° _I ................. ; ..... "'- s ................................ 

_ "_ 

• _I ca . o _ 

c_ = ...... -._ 

:,.7,, ; 6 = 

t •_ $" .." 

_' " 

= __ li !_ -r, 

.... :_ .......... i................. ! .................. i_ ................ _"' = ........ i........ u 

: 2 : "_ : "=, ! 

02-, . _:. 

,v,,,,,. " '" I .... I " " " " I " '-" ' r • _"i'" I " " " " I " " "'"" I ...... ; ' i'''' " " " " " " _ 

0 20 40 60 80 100 (ran) 

time (ram) 

Figure 1._Chromatogram of a reaction zone in Scots pine phloem and HPLC gradient. Solvent A = acetic acid, water 

(1/99); Solvent B = methyl alcohol, acetonitril, acetic acid (49.5/49.5/1). 

179

Characterization of compounds was made by using various methods: two-dimensional thin layer chromatography 

(TLC) on cellulose plates to observe fluorescence and to test specific chemical reagents, observation of spectral characteristics 

in U.V., cochromatography in HPLC with spots obtained from TLC and with various standards. The presence of glycosidic 

or ester links was tested with acid or alkaline hydrolysis, respectively. 


Comparison Between Unwounded and Inoculated Phloem 

Fourteen peaks were observed, of which eight were characterized at least at the chemical phenol family level (Fig. 1). 

Considerable differences existed between unwounded and inoculated phloem (Fig. 2). The most striking concerned the 

appearance of stilbenes, pinosylvin (Ps) and its monomethylether (Psme), and of one flavonoid, pinocembrin (Pc), in the 

reaction zone of all trees, while these compounds were always totally absent in the unwounded phloem. In all trees, the 

concentration of all compounds of the hydroxycinnamic acid group or its derivatives decreased during tree reaction. The 

concentration of flavonoids other than Pc increased or decreased depending on clone. 

Hydroxycinnamie 

Flavonoid Stilbene acid group or its 

group group derivatives 

(a) (a) (a) 

Figure 2.mVariations in the concentration of the phenolic compounds, between unwounded phloem and reactive phloem of 

different Scots pine clones, after inoculation with Ophiostoma brunneo-ciliatum. (a) = compound appearing in the 

reaction zone. One arrow means that the direction of variation was the same for all clones; several arrows mean that 

the direction of variation depended on the clone. Txf = taxifolin; Txfgl = taxifolin glycosid; Pc = pinocembrin; 

Ps = pinosylvin; Psme = pinosylvin monomethylether; Pcae = p-coumaric acid ester; Acphe = acetophenone glycoside. 

These results confirm the preliminary study by Lieutier et al. (1991a) in Scots pine. Phloem phenolic response to 

inoculations differed greatly according to compound, which caused dramatic modifications in the relative composition of this 

tissue. These results differ completely from the observations on terpenes in any conifer species (Russel and Berryman 1976; 

Raffa and Berryman 1982a,b; Delorme and Lieutier 1990; Lieutier et al. 1991b; Lfingstr6m et al. 1992), and makes certain 

phenols good candidates to play a role in arresting invaders and to be used as biochemical markers of tree resistance. 

Stilbenes and Pc, which were characteristic of phloem induced reactions, could play a particular role in reaction 

efficiency, and consequently in a tree's induced resistance to bark beetle and fungi attacks. However, similar results have 

been reported in needles after pollution by ozone (Rosemann et al. 1991), as well as in the sapwood of various pine species in 

response to several kinds of aggression (Ref. in Kuc and Shain 1977, Hart and Shrimpton 1979, Shain 1979, Hart 1981, 

Kemp and Burden 1986). In addition, similar results have also been obtained with a phloem wound without fungus (Lieutier 

and Yart unpubl.). Lieutier et al. (1991a) also mentioned the non-specificity of the phloem phenolic response after various 

aggressions. Tree response to bark beetles and their associated fungi are thus more a response to wounding than a response to 

particular aggressors themselves (Lieutier 1993). Mullick (1977) has suggested previously that a tree's response to aggres. 

sion is more concerned with tissue restoration than with defense. 

180

Considering biochemical pathways, it is interesting to note that the appearance of stilbenes and Pc in Scots pine 

corresponded to decreases in compounds of the hydroxycinnamic acid group, since this latter group is a precursor of both 

stilbenes and flavonoids (Ribereau-Gayon 1968, Gotham 1989). 

Between Clone Comparisons 

Clonal variability was first analyzed by multivariate analysis, in order to take into account all characterized com- 

pounds together. A canonical discriminant analysis was performed by taking reaction zones as Inain individuals and un- 

wounded phloem as complementary individuals. This procedure was necessary as the purpose was mainly to compare clone 

:reaction zones with each other, and only to place unwounded phloem relatively to them. 

Axes 1 2 and 3 explained 57, 23, and 15% of the variation, respectively. In the first two axes (Fig. 3), froln un- 

' 

wounded phloem to the reaction zone, phenolic composition seemed to evolve towards the same point of the plane, whatever 

the clone. That was confirmed in all significant axes (three), by calculating mean distances between points and their 

barycenter. Spread of the points relating to reaction zones was less than spread of the points relating to unwounded phloeln, 

in each clone and for all clones together (data not shown). Thus, each clone would react in a different way according to its 

constitutive composition, but always so that the composition of the reaction zone tended to be the same. 

The meaning of this phenolic modifications is not clear. It may be linked with protection against aggressors or with 

tissue restoration. Nevertheless, it is worthwhile to note that clone ...... c , whose reaction was the least efficient (reaction zone 

significantly longer than that of' all other clones), was the only one initially separated from the others by axis 2 (Fig. 3). If 

phenols effectively play a role in arresting beetles and their fungi, it could thus be possible to predict reaction efficiency of a 

clone on the basis of the whole phenolic composition of its constitutive phloem. 

8 

_'.%0 

eql 

-8 

. 

d 

""N& 

2 3, 

4 4 

/ 5 

b'b_za 6 "_so_ s 

I I I II 

-10 0 10 

Axis 1 (57%) 

Figure 3._Canonical discriminant analysis of phloem phenolic composition of different Scots pine clones, after inoculation 

with Ophiostoma brunneo-ciliatum. Reaction zones were considered as principal individuals and unwounded phloem 

as complementary individuals, a, b, c, d, e, f = unwounded phloem of the different clones (* represents the 

barycenters); 1, 2, 3, 4, 5, 6 = Corresponding reaction zones (o represent the barycenters). Arrows indicate the 

directions of variation from unwounded phloems to reactions zones. Clones a, b, c, d, e and f were respectively 

represented by 3, 2, 1, 2, 3 and 3 trees. 

t81

Clonal variability was also analyzed compound by compound, for the peaks characterized at least at the family level. 

To appreciate if certain compounds could be used as markers of efficiency of tree response, phenol concentration in each 

clone and for each compound, was compared to reaction zone length. The best correlation (r = 0.78) was for taxifolin 

glycoside in unwounded phloem (data not shown). The lowest concentrations corresponded to the least extended reaction 

zones, that is to the most effective response, and inversely. Concentration of taxifolin glycoside could thus be a marker of 

tree's response efficiency. Presence of taxifolin glycoside in the phloem prior to aggression could be favorable to the fungus, 

or could directly impede the development of the tree response. In the two cases, the result would be an increase in the lag 

time necessary to stop the fungus, thus causing a long reaction zone. 

Kinetics of the Phenolic Response 

A first approach to the kinetics was performed with multivariate analysis, the same way as in the previous experiment 

(Fig. 4). Axes 1 and 2 explained 88 and 9 percent of the variation, respectively. The composition of all unwounded samples 

was close to that of reaction zones at day 3, which suggest that phenolic composition of unwounded phloem did not vary 

3 

1 

4 u 

,,_,,, ..t 4 

2 

3 

4 

333 5 S 

2434 5 

At ed s 

-3 d 

[ 

-3 ' -i " i ' ; '.... i ' 

Axis 1 (88%) 

Figure 4.--Canonical discriminant analysis of the phenolic composition of Scots pine phloem collected at different dates after 

inoculations with Ophiostoma brunneo-ciliatum, in three different trees. Reaction zones were considered as principal 

individuals and unwounded phloem as complementary individuals. 1, 2, 3, 4, 5 = Reaction zones collected respectively 

3, 7, 14, 30 and 60 days after inoculations. A, B, C, D, E = corresponding barycenters, d, e = unwounded 

phloem collected at days 30 and 60. Arrows indicate the directions of variation during reaction development. 

during the whole experiment. Although between tree variability existed, the phenolic composition of the reaction zones at 

day 60 was clearly separated from that of all other samples along axis 1. Reaction zones at days 3 were separated from other 

samples along axis 2. Kinetics of phenolic response thus proceeded in two main phases. Concomitantly, growth of the 

reaction zone also proceeded in two main phases and was approximately done at day 30, indicating that the fungus had been 

stopped before this date (Fig. 5). One can thus hypothesize that the first phase, before day 30, is related to resistance and to 

interactions between tree and fungus, while the second more probably corresponds to wound healing processes. Considenng 

that, it is interesting to note that wound periderm began to be visible at day 30. 

Analysis, compound by compound, revealed three main kinds of kinetics. Ps, Psme and Pc increased in two phases, 

first early until day 7, then late after day 30 (Fig. 6a and 6b). Some other compounds, such as acetophenone glycoside, 

regularly decreased or remained constant (Fig. 6c), while others, such as taxifolin, increased only late, after day 30, SOmetimes 

after a decrease at the very beginning of development of the reaction (Fig. 6d). 

182 

5 

, i: i'i:i_i_'i!ill

- = 

120 

N /'_" 

=40 1/ 

_ 

j Days 

0 _ 

20 40 60 

0 ..___... tree A .-- m tree B ....... tree C 

Growth of the reaction zone 

Figure 5.--Mean reaction zone length at different dates after inoculation of three trees with Ophiostoma brunneo-ciliatum. 

Verticalbars represent standard deviations. 

.N 

k, 

_0 

0,5 

3' 

0,4 Pinocembrin t Pinosylvin monomcthylethcr 

-;--f_- o , J 

_ 0 _0 io 3'0 40 _0 60 0 _0 20 30 40 50 60 

° acctophcnone Taxifolin 

_ 15 glycoside . 4 

,o.., 1 

",L1s _l 

"''l "_I --__ i 2 

................ RZ ............................................ UP ] "_"" ....................... 17_--=='2"- Days 

0 .................................................................. 

" 0 " _i" _"_''_ 

o io io 3b 4b ._'o go 0 fo 2'0 _0 4'0 s'o 6'o 

Tree A Tree B ................... Tree C 

UP unwounded phloem RZ reaction zone 

Figure 6._Variations of the mean concentration of various phenolic compounds in reacting phloem and in unwounded phloem 

of three trees, after inoculation with Ophiostoma brunneo-ciliatum. Vertical bars represent standard deviations. 

183

Under these conditions, according to the above hypothesis, Ps, Psme and Pc would probably be involved in tree 

resistance, which could be related to the tree's ability to rapidly synthesize these compounds. After an early increase, their 

concentration remains constant until the fungus was stopped, which could indicate that they were metabolized by the fungus 

because of their toxic effect, as suggested by Lyr (1962). Their increase after day 30 could thus be due to the absence of 

fungus activity. This increase also suggests that they are involved in wound healing processes, as could be taxifolin. In Sitka 

Spruce as well as in hardwood trees, accumulation of phenolic compounds takes place parallel to suberin biosynthesis, a 

component of the wound periderm (Biggs 1985, Woodward and Pearce 1988). 

CONCLUSIONS 

Tree response could be more a generalized wound response than a pure defensive response. Both stilbenes and 

flavonoids could be involved in tree resistance. Taxifolin glycoside is a potential marker of tree resistance, all the more 

interesting as it belongs to constitutive phloem, but contrary to pinocembrin and stilbenes, its high concentration would 

indicate a low level of resistance. Whole phenolic composition of unwounded phloem could also be a marker of resistance. 

However, several conditions need to be verified before definitive conclusions are possible about their use as markers. Indeed, 

between clone differences could depend on tree physiological status or age, environmental conditions, locality or season. In 

addition, if genetic markers are looked for, the corresponding compounds must be genetically dependent and the corresponding 

genes must always be expressed. 

SUMMARY 

Scots pine phloem was investigated for phenolic markers of tree resistance to bark beetles. Our approach was based 

on the study of relations between the reaction efficiency and phenolic composition in six different clones, and on the study of 

the variations in phenolic composition during tree reaction development. Considerable modifications took place after 

aggression. Compounds lacking in unwounded phloem appeared in the reaction zone. During the development of the 

reaction zone, modifications of the phenolic composition clearly proceeded in two successive phases. Stilbenes and flavonoids 

seemed to be involved in the induced resistance. Taxifolin glycoside in unwounded phloem looked interesting as 

marker, as did the whole phenol composition of this tissue. 


The authors thank J. Garcia and P. Romary (INRA, Orleans, France) for their technical help and Dr. D. Sauvard 

(INRA, Orldans) for help with the statistics. They are grateful to Prof. H. Sandermann and Dr. W. Heller (G.S.F. Neuherberg, 

Deutschland) and to Dr. A. Scalbert (INRA Grignon, France) for providing pure compounds. They also thank Drs B. 

Monties, A. Scalbert and M.T. Tollier (INRA Grignon, France)/'or their help in characterization. These studies were granted 

by the Conseil R6gional de la Rdgion Centre (France) and by the French Ministry for Research and Technology. 


ALCUBILLA, M. 1970. Extraction, chromatographic separation and isolation of fungistatic substances from the inner bark 

of Norway Spruce. Z. Pflanz. Bodenk. 127: 64-74. 

BERRYMAN, A.A. 1972. Resistance of conifers to invasion by bark beetle fungus associations. BioScience 22:598-60 1. 

BERRYMAN, A.A. 1978. A synoptical model of the lodgepole pine/mountain pine beetle interaction and its implication in 

forest management, p. 98-105. In Berryman, A.A., Amman, G.D., Stark, R.W., and Kibbee, D.L., eds. Theory and 

Practice of Mountain Pine Beetle Management in Lodgepole Pine Forests. Coll. For. Res., Univ. Idaho, Moscow. 

BERRYMAN, A.A. 1982. Biological control, thresholds and pest outbreaks. Environ. Entomol. 11: 544-549. 

184

BIGGS, A.R. 1985. Suberized boundary zones and the chronology of wound response in tree bark. Phytopathol. 75:1191- 

1195. 

CHRISTIANSEN, E, and HORNTVEDT, R. 1983. Combined Ips/Ceratoo'stis attack on Norway spruce, and defensive 

mechanisms of the trees. Z. Angew. Entomol. 96:110-118. 

CHRISTIANSEN, E., WARING, R.H., and BERRYMAN, A.A. 1987. Resistance of conifers to bark beetle attack: Search- 

ing for general relationships. For. Ecol. Manage. 22: 89-106. 

DELORME, L. and LIEUTIER, F. 1990. Monoterpene composition of the preformed and induced resins of Scots pine, and 

their effect on bark beetles and associated fungi. Eur. J. For. Pathol. 20:304-316. 

GORHAM, J. 1989. Stilbenes and Phenanthrenes, p. 159-196. In Dey, M. and Harborne, J.B., eds. Methods in Plant 

Biochemistry. Vol. 1, Plants phenolics, acad. Press, MY. 

HART, J.H. 1981. Role of phytostilbenes in decay and desease resistance. Ann. Rev. Phytopathol. 19: 437-458. 

HART, J.H. and SHRIMPTON, D.M. 1979. Role of stilbenes in resistance of wood to decaY. Phytopathol. 69: 1138-1143. 

KEMR M.S. and BURDEN, R.S 1986. Phytoalexins and stress metabolites in the sapwood of trees. Dhytochemistry 25" 

1261-1269. 

KUC, J. and SHAIN,L. 1977. antifungal compounds associated with disease resistance in plants, p. 497-535. In Siegel, 

M.R. and Sisler, H.D., eds. antifungal Compounds. Vol. 2, Marcel Dekker, MY. 

LANGSTROM, B., HELLQVIST, C., ERICSSON, A., and GREF, R. 1992. Induced defence reaction in Scots pine follow- 

ing stem attacks by Tomicus piniperda. Ecography 15: 318-327. 

LIEUTIER, F. 1993. Induced defense reactions of conifers to bark beetles and their associated Ophiostoma, p. 206-215. In 

Wingfield, M.J., Seifert, K.A., and Webber, J.F., eds. Ceratocystis and Ophiostoma: Taxonomy, Biology and Pathogenicity. 

Am. Phytopathol. Soc., St Paul. 

LIEUTIER, F., YART, A., GARCIA, J., POUPINEL, B. and LEVIEUX, J. 1988 Do fungi influence the establishment of 

, 

bark beetles on Scots pine? p. 317-330. In Mattson, W.J., Levieux, J., and Bernard-dagan, C., eds. Mechanisms of 

Woody Plant Defenses Against insects, Search for Pattern. Springer, MY. 

LIEUTIER, F., YART, A., JAY-ALLEMAND, C., and DELORME, L. 1991a. Preliminary investigations on phenolics as a 

response of Scots pine phloem to attacks by bark beetles and their associated fungi. Eur. J. for. Pathol. 21: 354-364. 

LIEUTIER, F., BERRYMAN, A.A., and MILLSTEIN, J.A. 1991b. Preliminary study of the monoterpene response of three 

pines to Ophiosloma clavigerum (ascomycetes: Ophiostomatales)and two chemical elicitors. Ann. Sc. For. 48: 377- 

388. 

LIEUTIER, F., GARCIA, J., ROMARY, R, YART, A., JACTEL, H., and SAUVARD, D. 1993. Inter-tree variability in the 

induced defense reaction of Scots pine to single inoculations by Ophiostoma brunneo-ciliaturn, a bark beetle associated 

fungus. For. Ecol. Manage. 59: 257-270. 

LIEUTIER, F., PICRON, V., BRIGNOLAS, F., YART, A., SAUVARD, D., BASTIEN, C., and JAY-ALLEMAND, C. 1995. 

Changes in soluble phenolic metabolites of Scots pine phloem in response to challenge with Ophkvstoma brunneociliatum, 

a bark beetle associated fungus. Eur. J. for Pathol. in press. 

LYR, H. 1962. Detoxification of heartwood toxins and chlorophenols by higher fungi. Nature 194: 289-290• 

MULLICK, D.B. 1977. The non-specific nature of defence in bark and wood during wounding insect and pathogen attack, 

Recent Adv. Phytochem., Vol. 2, Plenum, New York: 395-441. 

185

RAFFA, K.E and BERRYMAN, A.A. 1982a. Accumulation of monoterpenes and associated volatiles following inoculation 

of grand fir with a fungus transmitted by the fir engraver, Scolytus ventralis (Coleoptera: Scolytidae). Can. EntomoI, 

114:797-810. 

RAFFA, K.F. and BERRYMAN, A.A. 1982b. Physiological differences between lodgepole pines resistant and susceptible to 

the mountain pine beetle and associated microorganisms. Environ. Entomol. 2: 486-492. 


ponderosae Hopk. and blue stain fungi. Can. J. Bot. 45:1 1115-1126. 

RIBEREAU-GAYON, R 1968. Les compos6s ph6noliques des v6g6taux. Dunod ed., Paris. 253 p. 

ROSEMANN, D., HELLER, W., and SANDERMANN, H. 1991. Biochemical plant response to ozone. II. Induction of 

stilbene biosynthesis in Scots pine (Pinus sylvestris L.) seedlings. Plant Physiol. 97: 1280-1286. 

RUSSEL, C.E. and BERRYMAN, A.A. 1976. Host resistance to the fir engraver beetle: 1 - Monoterpene composition of 

Abies grandis pitch blisters and fungus-infected wounds. Can. J. Bot. 54: 14-18. 

SHAIN, L. 1979. Dynamic responses of differentiated sapwood to injury and infection. Phytopathol. 69:1143-1147. 

SHRIMPTON, D.M. 1973. Extractives associated with the wound response of lodgepole pine attacked by the mountain pin 

beetle and associated microorganisms. Can. J. Bot. 51: 527-534. 

WOODWARD, S. and PEARCE, R.B. 1988. Wound-associated response in Sitka Spruce root bark challenged with 

Phaeolus schweinitzii. Physiol. Molecular Plant Pathol. 33:151-162. 

186

DIFFERENTIAL SUSCEPTIBILITY OF WHITE FIR PROVENANCES 

TO THE FIR ENGRAVER AND ITS FUNGAL SYMBIONT 

IN NORTHERN CAHFORNIA 

G. T. FERRELL _and W. J. OTROSINA 2 

_USDA Forest Service, Pacific Southwest Research Station, 2600 Washington Avenue, Redding, California 96001, USA 

2USDA Forest Service, Southeastern Forest Experiment Station, 320 Green Street, Athens, Georgia 30602, USA 

INTRODUCTION 

The fir engraver, Scolytus ventralis LeC., attacks white fir, Abies concolor (Gord. and Glend.) Lindl., and other true 

firs, Abies spp., in western North America. The biology, attack behavior, and ecology of this bark beetle were recently 

summarized by Berryman and Ferrell (1988). During the summer flight season, the attacking beetles bore into the cambial 

zone of fir boles, introducing a pathogenic brown-staining fungus, Trichosporium symbioticum Wright. Resistant firs react to 

the invasion by forming a resinot_s necrotic wound in the phloem and outer sapwood, which contains the spread of the fungus 

and repels or kills the beetles. In such interactions, the tree usually survives. This reaction is less intense or absent in 

susceptible firs, resulting in reproductive success of the beetles, and severe damage or even death of the fir. Sporadic 

outbreaks of the fir engraver, associated primarily with droughts, have caused widespread mortality of true firs in nearly 

every decade of this century in western North America. 

In the drought years of 1987-88, four 26-year-old white fir provenance test plantations at Camino, California, located 

at 1028 m elevation in the central Sierra Nevada, sustained considerable levels of mortality. Subcortical examination of a 

subsample of the dead firs indicated that all had reproductively successful gallery systems of the fir engraver. Drought 

continued until winter 1992-1993. Surveys beginning in fall 1988 revealed pronounced differences in this mortality, among 

both plantations and provenances, and also within plantations. 

This paper describes the observed patterns of susceptibility and resistance in relation to known patterns of geographic 

variation in white fir in western North America and discusses studies underway to understand the mechanisms responsible for 

these patterns. 

METHODS 

The adjacent provenance test plantations represent two studies. The "geographic range" plantations (1 and 3) consist 

of 39 provenances from throughout most of the western portion of white fir's natural geographic range in western North 

America. Each provenance was originally represented by three replications, each of three seedlings. One replication was 

removed by thinning in 1970. The "elevational transect" plantations (2 and 4) contain only four provenances representing a 

west-east elevational transect across the Sierra Nevada at the latitude of Camino. Each of these provenances is represented 

by 10 half-sib families, each originally with nine trees, since thinned to six. In all four plantations, provenances were planted 

in an interlocked randomized non-contiguous plot layout designed to minimize effects of microsite variation (Libby and 

Cockerham 1980). 

Beginning in August 1988, trees in these plantations were surveyed annually in summer or early fall for mortality and 

susceptibility to fir engravers. Surveys were conducted after trees attacked and killed by fir engravers the previous year had faded 

crowns, but before trees killed by current attack had faded crowns. During these surveys, trees topkilled by the previous year's 

attack were also noted. 

Mattson, W.J., Niemel/i, P., and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. 

USDA For. Serv. Gen. Tech. Rep. NC-183, N.C. For. Exp. Sta., St. Paul, MN 55108.

Patterns of tree resistance were surveyed in December 1989; numbers of pitch streamers indicating unsuccessful fir 

engraver attacks were classified as 0, < 5, 5-10, and > 10 on stems of all live trees. 

Distributions of trees classified by status (live, dead) and number of pitch streamers were compared among provenances 

by Chi-square Contingency Test (SAS Institute Inc. 1988). Data from the two replicates of each type of plantation 

(geographic range, elevational transect) were similar and were pooled for analysis. In the geographic range plantations, 

provenances were combined according to Hamrick and Libby (1972) as modified by Libby et al. (1980). The latter authors 

divided white fir in the western part of its range into five major geographic groups based primarily on needle morphology: 

(1)"Northern"--central Oregon and northwestern California; (2) "Central"-- south-central Oregon, central and northeastern 

California; (3) "Southern California"; (4) "Interior South"m Arizona; and (5) "Interior North"-- eastern Nevada and western 

Utah. 

RESULTS 

Of the 1,127 firs alive in all four plantations in 1987, 393 (35%) were killed by the fir engraver by 1993. Most (253) 

of the mortality resulted from the attack season of 1987, with an additional 121 resulting from the 1988 attack season. 

Relatively few (29) were killed by attack in 1989-1993. Only a few (10) trees were topkilled, two of which were killed in 

subsequent years. Firs on the outside edges of the plantations were infrequently killed (4/97 or 4%). 

Through 1993, cumulative mortality in the elevational transect plantations (256 of 578 total trees, or 44%) was almost 

twice that in the geographic range plantations (137/549, or 25%). The associated Chi-square (adjusted for continuity) was 

45.505, p < 0.0001 at 1 dr. 

In the geographic range plantations, mortality differed among major geographic groups of Hamrick and Libby (1972) 

and Libby et al. (1980). In both plantations, provenances of the Northern and Central groups had sustained the highest 

mortality, while very few of the firs of the Interior South and Interior North groups were killed (Table 1). Low levels (1 1 

percent) of mortality occurred in southern California firs. The Chi-square associated with among-group variation was 

40.631, p < 0.0001 at 4 df. Among individual provenances, the one from nearest Camino (AK; Omo Ranch; 32 km distant) 

had the highest mortality in both plantations. 

Table 1.--Mortality of white fir provenances caused by fir engravers in the 

geographic range plantations t at Camino, CA I987-1993 

Group 2 Plantation 1 Plantation 3 

Total Dead Pctdead Total Dead Pctdead 

Northern 55 12 22 84 33 39 

Central 106 22 21 124 55 44 

Southern Calif. 40 3 8 47 7 15 

Interior South 28 0 0 31 4 13 

Interior North !4 0 0 20 1 5 

1Each plantation contained 39 provenances from the western portion of white fir's 

geographic range. 

ZProvenances combined by major morphological groups of Hamrick and Libby 

(1972). 

In the elevational transect plantations, mortality in the two lower- elevation westside provenances nearest Camino 

(AK, AL) averaged more than one-third greater than that in the two upper-elevation eastside provenances (AM, AN), which 

are only about 50 km more distant (Table 2). The among-provenance Chi-square was 24.757, p < 0.0001 at 3 df. 

Analysis of the distribution of pitch streamers on boles of surviving firs revealed that in the elevational transect 

plantations, virtually all surviving firs had been attacked, and many heavily so (>10 streamers). In the geographic range 

188

Table 2.--Mortality of white fir provenances caused by fir engravers in the 

elevational transect plantations _at Camino, CA 1987-1993 

Provenance Plantation 2 Plantation 4 

Total Dead Pct dead Total Dead Pctdead 

AK 76 37 49 95 59 62 

AL 72 30 42 72 42 58 

AN 65 19 29 63 27 43 

AM 61 9 15 74 33 45 

_West-east transect across Sierra Nevada at latitude of Camino. AK, AL are 

lower elevation westside; AN, AM are upper elevation eastside. 

plantations, however, more than 38% of Interior South and Interior North firs had no pitch streamers, and fewer than 11% 

had more than l0 pitch streamers (Table 3). In contrast, among Northern and Central firs, fewer than 6% had no pitch 

streamers while more than 24% had over 10. Distribution of pitch streamers on Southern California firs was intermediate 

between these two patterns. The among-group Chi-square was 128.424, p < 0.0001 at 12 df. The Central provenances from 

the Sierra Nevada where Camino is located had the lowest percentage of firs with no pitch streamers (ca. 2%) and the highest 

percentage with more than 10 streamers (ca. 41%). 

Table 3.--Incidence of pitch streamers caused by fir engravers on boles of white fir 

provenances in the geographic range plantations, _Camino, 1989 

Group 2 Number of pitch streamers 

0 10 

Northern 5 34 31 23 

Central 3 22 61 60 

Southern Calif. 11 16 33 16 

Interior South 27 14 8 6 

Interior North 13 12 8 1 

JPlantations 1 and 3, combined. 

2provenances combined by major morphological groups of Hamrick and Libby (1972). 

DISCUSSION 

Results indicated that Northern (Oregon) and Central (Northern California) provenances were susceptible to the fir 

engraver and its fungal symbiont in the Camino plantations, while Interior South (Arizona), and Interior North (Nevada, 

Utah) provenances were virtually nonsusceptible. Southern California provenances demonstrated low susceptibility. These 

patterns agree closely with known geographic patterns of morphological and chemical (cortical monoterpenes) variation in 

white fir over its natural range in western North America (Hamrick and Libby 1972, Zavarin et al. 1975). Susceptible 

provenances were those characterized as green-foliaged, with needle morphology suggesting only moderate drought resistance, 

and low in camphene and 3-carene, consisting of California white fir (var. lowiana (Gord.) Lemm.) from northern 

California and intermediates with grand fir, A. grandis (Doug.), from south-central Oregon. Provenances evidencing virtually 

no susceptibility were blue-green foliaged with needle morphology suggesting higher drought resistance, relatively high 

in camphene and 3-carene, and of the Rocky Mountain variety (vat. concolor (Gord. & Glend.) Lindl.). Southern California 

provenances evidencing low susceptibility were characterized as blue-green foliaged, relatively high in 3-carene but nearly 

lacking in camphene and thus intermediate between the Northern and Interior South groups. 

189

In all plantations, susceptibility was always highest in provenances nearest Camino. This was evident not only on a 

large scale in the geographic range plantations where Central (Northern California) provenances were the most susceptible, 

but also on a fine scale in the elevational transect plantations where provenances 40 km distant from Camino were more 

susceptible than those from 80 kin distant. In loblolly pine plantations in South Carolina, Powers et al. (1992) also found that 

local provenances were more susceptible to southern pine beetle, Dendroctonusfrontalis, than distant provenances. 

Mechanisms responsible for the observed differential susceptibility among provenances are under investigation. 

Their basis, however, is undoubtedly primarily genetic as the plantations are all adjacent with provenances planted intermixed 

to minimize chances that microsite would differentially affect provenances. 

The very low susceptibility of trees on the outside edges of the plantations compared with those in the interior was 

probably primarily determined by stand factors such as differences in soil moisture or beetle pheromone dispersal rather than 

tree genetics because susceptibility occurred without regard to provenance. Mechanisms underlying this difference are under 

investigation. 

We have two studies underway testing the hypothesis that observed differential susceptibility is attributable to 

differential moisture stress among provenances, and between firs on plantation edge versus interior. White firs with highly 

negative water potentials (pre-dawn in August during summer dry, and fir engraver flight, seasons) are known to be susceptible 

to the fir engraver (Ferrell 1978). Results to date failed to find appreciable differences among provenances in 1989, 

1990, or 1993. "Edge" firs did average higher moisture stress than "interior" firs in 1992 and 1993. Both studies will 

continue for one more post-drought year. Results are necessarily conditioned by the fact that moisture stress can be studied 

only in surviving firs. Thus, parallel studies are underway to compare radial growth patterns in surviving versus killed firs, 

under the hypothsis that differential moisture stress should be evident in radial growth pattterns. 

Differences in fir engraver attack preference and success are being investigated by reciprocally caging "local" and 

"exotic" populations of beetles with bolts cut from "local" and "exotic" provenances of white fir. These tests are designed to 

reduce or eliminate differential resistance caused primarily by differences in environmental factors such as moisture stress 

and to isolate for analysis differential resistance caused by factors that are primarily genetically determined such as constitutive 

bark chemicals. Preliminary results indicate that beetles from the vicinity of Camino initiate far fewer attacks in Arizona 

bolts than in "local" bolts. No such preference was evident in tests with Nevada beetles caged with bolts of firs from Nevada 

and the vicinity of Camino. These tests are being repeated with beetles and bolts of other provenances. Parallel studies are 

also under way using various geographic isolates of the Trichosporium fungus inoculated into stems of surviving firs in the 

plantations. Among-provenance variations in inoculation wound size and monoterpene composition are being studied. Thus 

far consistent differences in virulence have been tbund between isolates regardless of provenance of the fungus or the fir. 

Studies continue to explore seasonal and yearly variations and to analyze monoterpene composition. 

Results indicate that local populations of host conifers can be more susceptible to local populations of bark beetles 

and their fungal symbionts while exotic host populations can be less susceptible. If generally true, this phenomenon may 

have to be taken into account in tree improvement programs, most of which prefer to utilize local genetic material as planting 

stock because this material is preadapted to the growing site. If results at Camino are any indication, however, exotic seed 

sources may be less susceptible to local bark beetle populations provided that these trees are adequately adapted to the 

physiographic factors of the growing site. At Camino, many of the nonsusceptible exotic provenances had grown fully as tall 

as the best-growing local provenances, indicating that they were thus far as welt adapted to the growing site. 

Another important implication of the Camino results is that maintenance of genetic diversity in conifer populations 

may be an important safeguard in protecting conifer stands against bark beetles. The much greater mortality experienced in 

the elevational transect plantations containing only four provenances may be an expression of this, although it was evidently 

also influenced by the local origin of all four of these provenances. 

t90

Results from the geographic range plantations suggest that planting well-adapted exotic and local provenances in 

mixture may be a useful strategy for avoiding problems from bark beetles and their fungal symbionts. Any use of exotic 

provenances should probably be limited to high-value trees, particularly those grown on stressful sites, because more widespread 

planting could lead to loss of resistance through local pest populations becoming adapted to them. 

SUMMARY 

During a drought-associated fir engraver outbreak in California, local white fir provenances were more susceptible 

than exotic provenances, and, doubtless partly in consequence, less genetically diverse plantations were more susceptible 

than plantations with greater genetic diversity. 

Mechanisms underlying the observed differential susceptibility remain unknown but are the subject of continuing 

investigations. 

Results suggest that planting a mixture of well-adapted exotic, as well as local, provenances for maintenance of high 

genetic diversity may be an important strategy for protecting conifer hosts against bark beetles and their symbiotic fungi. 


We thank Dr. Tom Conkle and staff of the Pacific Southwest Research Station's Institute of Forest Genetics and Prof. 

Bill Libby of the cooperating University of California College of Environmental Science, Berkeley, for access to plantations 

and for encouragement. 


BERRYMAN, A.A. and FERRELL, G.T. 1988. The fir engraver in western states, p. 555-577. In Berryman, A.A., ed. 

Dynamics of Forest Insect Populations. Plenum, New York. 

FERRELL, G.T. 1978. Moisture stress threshold of susceptibility to fir engraver beetles in pole-size white firs. For. Sci. 24: 

85-92. 

HAMRICK, J.L. and LIBBY, W.J. 1972. Variation and selection in western montane species. I. White fir. Silvae Genetica 

21" 29-35. 

LIBBY, W.J. and COCKERHAM, C.C. 1980. Random non-contiguous plots in interlocking field layouts. Silvae Genetica 

29: 183-190. 

LIBBY, W.J., ISIK, KANI, and KING, James P. 1980. Variation in flushing time among white fir population samples. 

Annales Forestales, Jugoslavenska Akademija Znanosti 8: 123-138. 

POWERS, H.R., JR., BELANGER, R.P., PEPPER, W.D., and HASTINGS, F.L. 1992. Loblolly pine seed sources differ in 

susceptibility to the southern pine beetle in South Carolina. South. J. Appl. For. 16:169-174. 

SAS INSTITUTE INC. 1988. SAS/STAT User's Guide, Release 6.03 Edition. SAS Institute Inc. Cary, North Carolina. 

ZAVARIN, E., SNAJBERK, K., and FISHER, J. 1975. Geographic variability of monoterpenes from the cortex of Abies 

concolor. Biochem. System. and Ecol. 3: 191-203. 

191

MILD DROUGHT ENHANCES THE RESISTANCE OF N©RWAY SPRUCE TO 

A BARK BEETLE-TRANSMITTED BLUE-STAIN FUNGUS 

ERIK CHRISTIANSEN and ANNIE MARIE GLOSLI 

Norwegian Forest Research Institute, N-1432 A.s, Norway 

INTRODUCTION 

Outbreaks of the spruce bark beetle, Ips typographus (L.), one of the most serious pests of the Eurasian spruce 

forests, are frequently triggered by large-scale windfelling. Although correlative evidence indicates that long-lasting drought 

incites and aggravates such epidemics (cf. Christiansen and Bakke 1988), very little experimental work has been carried out 

to test this assumption. 

Ideally, the defensive capacity of the trees should be tested using prescribed numbers of bark beetle attacks. In Pinus 

contorta this has been done successfully by screening the branch-free part of the stem, and inducing a defined number of 

Dendroctonus ponderosae attacks under the screen (Raffa and Berryman 1983). Without screening, additional attacks by 

"wild" beetles in the neighbourhood will occur if the beetles possess aggregation pheromones. 

Unlike mountain pine beetle attacks on lodgepole pine, I. typographus attacks often extend far into the living crown 

of Norway spruce, Picea abies, making screening impractical. For this reason we chose to assay host resistance using 

prescribed loads of artificial mass-inoculation with the blue-stain fungus, Ophiostoma polonicum Siem., a close associate of 

the spruce bark beetle. 

Spores of O. polonicum are transmitted both externally and internally by L typographus (Furniss et al. 1990), and the 

fungus is consistently isolated from the advancing front of blue-stain in successfully attacked trees (Solheim 1986, 11992). O. 

polonicum can kill healthy Norway spruce when artificially inoculated under the bark (Horntvedt et al. 1983, Christiansen 

1985a, Solheim 1988). Similarly, other spruce species and even Pseudotsuga menziesii may succumb to mass-inoculation 

with this fungus (Christiansen and Solheim 1990). Parallel to the "threshold of successful attack" for bark beetles 

(Thalenhorst 1958), a "threshold of successful infection" exists for mass-inoculation with this fungus (Christiansen 1985b). 

Drought is generally accompanied by hot weather, and the two factors are not easily separated when one attempts to 

find the causes for large-scale infestations. We hypothesize that (1) elevated temperatures act directly on the various life 

stages of the beetles to favour population build-up, and/or that (2) drought affects the physiology of the trees, making them 

more susceptible to attack or more suitable as food. Here, we report an experiment designed to shed light on aspects of 

Hypothesis 2: Norway spruce trees were artificially drought stressed, and their defensive capability was assayed and compared 

with unstressed, control trees. 

In ordinary forest stands, trees will vary considerably in their susceptibility to beetle/fungus attack. This necessitates 

large numbers of experimental trees to obtain statistically significant results. In the present experiment, however, we utilized 

clones of spruce trees and thus permined work with relatively few trees. 

METHODS 

Experimental Trees and Their Treatment 

The Norway spruce clones used in this study grew in a multi-clonal stand at Hogsmark Experimental Farm in ,As, 

Akershus, Norway, and had been produced as follows: Seeds from selected trees were sown in 1951 and cuttings from the 

Mattson, W.J., Nieme1_i, P., and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. 

USDA For. Serv. Gen. Tech. Rep. NC-183, N.C. For. Exp. Sta., St:.Paul, MN 55108. 

192

esulting seedlings were then rooted. In spring, 1970, rooted ramets of each clone were planted using 2x2 m spacing. The 

planting was done in a west-facing, gently sloping agricultural field where underground drain pipes had been installed and the 

clayey soil tilled before planting. The homogeneous soil and the exact planting array reduced phenotypical variation in the 

stand to a minimum. 

From each of two clones, #194 and #582, twelve ramets were selected. Trees of Clone 194 had been used in an 

earlier mass-inoculation experiment with O. polonicum, and had proved relatively resistant to the infection (Christiansen, 

unpubl.). Nothing was known about the susceptibility of Clone 582 as to its susceptibility. 

The continuous rows of clones followed the slope of the field. The row of trees between Clone 194 and Clone 582 

had been cut in early October 1992, and so had the two rows on their other side. Thus both rows had a distance of 4 m to 

their nearest neighbours. Trees at both ends of the two rows were discarded as phenotypically aberrant, having much larger 

crowns than trees inside the stand. 

In late October, 1992, all branches below ca. 2.5 m were removed from all 24 trees. These branches were all dead 

due to shading. The ground on both sides of the upper six trees in each row was then covered with plastic sheets to prevent 

autumn and winter precipitation from percolating into the ground. Instead, this water was led down to the lower six trees in 

the two rows. On 20 April t993 the plastic ground cover was removed. Ropes were strung at 2 m above ground between the 

upper six trees of the rows, and non-transparent plastic tarpaulins were suspended from these ropes down to drains fastened 

to the stumps of the tree rows that had been removed. The drains collected the rain falling through the canopy, and channelled 

it down to the lower six trees in the rows. 

This way the six upper trees in a row were deprived of most of the precipitation falling from October till the end of 

the experiment. The lower six trees of the two rows served as unstressed controls. Because the early summer of 1993 was 

very dry, these control trees were given an extra 47 mm of water, applied under the canopy by means of a garden sprinkler. 

Monitoring of Drought Stress 

Starting on 12 May 1993, drought stress was monitored by repeated measurements of pre-dawn xylem water potentials 

(hereafter termed "WP"), using a pressure chamber. Freshly excised shoots taken 4-5 m above ground were used. To 

detect a possible natural WP gradient along the slope, trees at the upper and lower ends of the rows were measured. By mid- 

June WPs of the experimental trees were clearly different from those of the controls, and after a monitoring of all 24 trees on 

23 June we decided that they had reached a satisfactory level of stress for inoculation with the fungus. 

Fungal Inoculation 

Prior to inoculation a 80 cm long template was attached around the stern between 1.2 and 2.0 m above ground. In the 

template evenly spaced holes had been punched at a density of 4 per dm 2,and the points of inoculation were marked through 

these holes. The choice of this particular dose was based on earlier experience with the susceptibility to infection by trees of 

Clone 194 and of two other clones growing in the same stand. Inoculation occurred on 25 June. Bark plugs were removed 

with a 5 mm cork borer. The inoculum, actively growing O. polonicum mycelium on malt agar, was placed in the cambiumholes, 

whereupon the bark plug was returned to its original position (cf. Horntvedt et al. 1983). 

Further Field and Laboratory Procedures 

Around the time of inoculation, exudation of constitutive resin was estimated twice (cf. Christiansen and Horntvedt 

1983): on each tree 10 thin plastic tubes were inserted into holes cut with a cork borer about 2.1 m above ground. After 24 

hours the length of the resin column in the tube was measured. 

On l August, about 5 weeks after inoculation, exudation of resin from the point of inoculation was recorded on a 

subjective scale from 0 to 5, and used as an estimate of induced resinosis. On 14 September 1993, 81 days after inoculation 

all trees were felled, and tree height and height of the lowest green whorl of branches were measured. The inoculated stem 

section was cut out, and brought to the laboratory. 

193

in the laboratory two ca. 5 mm thick cross-sectional discs were cut 20 cm inside each end of the section. The 

he_._wo_>d-+sapwood border and blue-stained areas were marked out on these two discs. Two separate measuren_ents for 

fung_t sc_ccess were used: (I) percent of the sapwood which had become blue-stained, and (2) percent of the disc circurn ferer_ce 

where the cambial area had been killed. For both measurements, the disc having the most ft._ngal proliferation was used 

t(_ represen_ the tree. 

Above the lower of these thin discs a thicker one was cut. This ca. 3 cm disc was used for re-isolation of' fungi from 

Ihe wood. Three small wood chips were taken along one radius of the disc where staining occurred, one at the front of the 

advancing blue=s'tain, one near the carnbium, and one in between. 

The data were analyzed using the Minitab statistical system (Ryan et al. 1992). 

RESUUFS 

Dt.Jrin_4the period late October 1992 to late June 1993 precipitation at As amounted to ca. 400 ram, and a further 200 

mm f:cll fron_ late June (inoculation) till inid-September (felling). Although small quantities of water must have leaked 

{hrot_gh the ground cover and later the roof, the experimental trees were still deprived of most of this precipitation. The 

spring of 1993 was drier than normal; only 124 mm felt during the period 1 March - 30 June, as opposed to a normal of 2 t 5 

mm_ Normal annual precipitation is ca. 800 ram. 

Measurement of pre-dawn xylem water potentials in late June demonstrated significant differences between stressed 

and control trees in both clones (Fig. 1). In Clone 194 the WP values of stressed trees ranged between 0.75 and 0.95 MPa as 

compared to 0.35..o0.50 MPa in the control trees. No stressed tree had as high a WP as any of the controls. Clone 582 

exhibited les_ difference in WP between the two categories of trees, and one sheltered tree (#6) was no more stressed than 

three of the control trees. Trees of the same clones at the top and bottorn of the slope showed no clear differences in WP 

(rest_It nol shown). 

On 29 June measurements indicated that stressed trees of Clone 194 yielded less constitutive resin in 24 hours than 

controls (p=0.0()2), but on 6 July no difference could be detected. In Clone 582 such differences vere absent on both occa- 

sions. Subjective estimates of surfTxce exudation of resin from the points of inoculation gave a rating that was about three 

lime higher in stressed trees of both clones than in corresponding controls (p=0,0(X)). 

LIpon felling on 14 September, no tree had lost its foliage. Several trees of Clone 194 had a yellow tint to their 

fi_liage, but this occurred both in stressed and control trees. 

The cross-sectional discs showed sapwood blue-staining varying from 0 to 100 percent (Fig. 2). Staining was much 

more pronounced in controls than in drought stressed trees of Clone 194. The stressed tree that had the most fungal invasion 

had mucf_ tess s{ain than the least affected control (p=O.000). In Clone 582, on the other hand, there was no difference 

between stressed and control trees. This measure of fungal success coincided with percentage of the disc periphery where the 

cambial area had been killed (Fig, 2). There was no overlap between stressed trees and controls in Clone 194 (p=0.()0()). [n 

Chine 582 conm_t _rees proved to be somewhat more affected than the stressed ones (p=0.033). 

O. poloruc_m was reisolated from 10 of the 12 trees of Clone 194 and 8 out of 12 in Clone 582. Because almost 12 

weeks elapsed from i_oculation to felling, other fungi had also invaded the wood (cf. Solheim 1992). These were mainly 

Nectria spp. (f:o_.md in 14 trees), but another Ophiostoma species, O. piceae (Mtinch) H. and R Syd., occurred together wiih 

O. I;,o!

-0.2 

0 

5 82. 

_j -0,4 e a a a 

D4 

E 

I ® e ® 

-0,6 ® 

_4 e 

.4 4.3 

C; -0.8 

o 

.# 

o 

_4 _ .... t I 0 O 0 | . l 

(1) 

194 

-0 .2 

- 0.4 _ 

if, 

3: 

o 

zx 

a 

zx 

_I 0 

-o. 6 

04 o o 

- o. 8 " 

@ 

e 

-1 0 | I I I I I I I ....I I | 

1 2 3 4 5 6 7 8 9 i0 ll 12 

Tree number 

Figure 1.--Pre-dawn xylem water potential of Norway spruce trees on 23 June 1993. Trees 1-6 of both clones (194 and 582) 

are artificially drought stressed; trees 7-12 are controls. 

An earlier experiment in Norway spruce addressed possible after-effects of drought on the resistance to O. polonicum 

infection (Christiansen 1992). In that case, three summers of severe artificial drought did not affect the resistance of the trees 

during a fourth season of normal precipitation. Thus neither that experiment nor the present one supports a general hypothesis 

that drought stress makes coniferous trees more susceptible to attack by bark beetles and their associated blue-stain fungi 

(cf. Christiansen et al. 1987). 

The drought stress was relatively mild in the present study, though possibly adequate to have influenced important 

physiological processes in the trees (cf. Hsiao 1973), even including gas exchange (cf. Havranek and Benecke 1978). Such 

changes may be thought to reduce resistance, but vascular wilts are characteristically found to be more severe in wet than in 

dry soils (Schoeneweiss 1986). 

A 

195

109 ICO 

I 9 4 _ _7 _ _ I00- 582 _ 92 

25 30 

._°_+m+,|_+ ...... ++.+,5_+_ .,_ o, +- , 

1 2 3 4 5 6 7 8 9 I0 II 12 1 2 3 4 5 6 7 8 9 I0 II 12 

E I00 I00 94 

._ _ _y 194 

_ 88 _ 90 89_ 582 _ 86 88 

40 _'+' + i .... 2; _! ;+ 

+_3o i+ + R > + 2+ ++{ +o ++l+ .?LI ++ .... + 

°" o J[Jt ......... +...... o .... m+_m...... 

1 2 3 4 5 6 7 8 9 i0 11 12 1 2 3 4 5 6 7 8 9 I0 II ]2 13 

Tree nunbe_< Tree ntmber 

Figure 2.--Fungal proliferation in Norway spruce trees inoculated with Ophiostoma polonicum. Trees 1-6 of both clones 

(194 and 582) are artificially drought stressed; trees 7-12 are controls. Upper graphs: Percent blue staining of the 

sapwood cross-sectional area. Lower graphs: Percent cambial killing. 

In a field experiment in young Pinus taeda, long-lasting water deprivation resulted in pre-dawn xylem water potentials 

of the same magnitude as in the present study. Although resin flow was significantly reduced in the stressed pines, they 

were also less attacked by Dendroctonusfrontalis than the irrigated controls. The conclusion was that drought of this extent 

does not necessarily lower resistance (Dunn and Lorio 1993). 

Like several other inoculation experiments in Norway spruce, the present one demonstrated that the invasive success 

of O. polonicum is negatively correlated with the quantities of resin that accumulate in "reaction zones" around the points of 

infection and ooze out from these sites (Christiansen and Horntvedt 1983, Christiansen 1985a, Christiansen and Ericsson 

1986, Horntvedt 1988, Horntvedt and Solheim 1991). As in other conifers, this "secondary" resinosis is assumed to be due 

mainly to localized, induced defence reactions (Berryman 1972). 

During the early stages of invasion in conifers, the bark beetles encounter a first line of defence, i.e., constitutive or 

primary resin oozing from severed ducts in the phloem. In our study there was no unequivocal difference between stressed 

and control trees in this character, although the first of two measurements gave higher resin yields in control trees of Clone 

194 than in stressed ones. However, in the end, these controls turned out to be less resistant to infection than the stressed 

trees. In Norway spruce, the quantities of primary resin appear to be highly variable (Christiansen 1991). Despite its 

apparent importance for the defence of certain pines (Cates and Alexander 1982), this constitutive resin appears to be less 

important for others such as Pinus sylvestris (Schroeder 1990). It has been suggested that tree species that are repeatedly 

exposed to attacks by multiple generations of beetles per year may rely more on a constitutive defences than those which 

generally have only one critical period of attack (Matson and Hain 1983). 

For some pine bark beetles (e.g., D. frontalis and D. brevicomis), studies suggest that drought has a negative effect on 

resistance, and that this effect is cumulative both on a shorter and a longer time scale (Craighead 1925, Miller and Keen 

196

1960, Kalkstein 1976). Increased susceptibility may be due to a reduced yield of constitutive resins. However, this explanation 

may not hold when drought is only mild (cf. Dunn and Lorio 1993). 

It seems likely that the degree of stress to which trees are exposed will influence their resistance. Assuming that a 

higher level of stress would have made the trees more susceptible, it is relevant to ask how often it occurs in the field under 

natural conditions. In this study the sheltered trees were deprived (depending on their root systems) of as much as half a 

year's precipitation (c. 400 mm) prior to inoculation, and a further 200 rnm were lost thereafter. Such an abrupt loss of water 

does not seem likely to occur under field conditions in Norway. At the norrnat time of mass flight by I. typographus (i.e., in 

May), soil water reserves are not likely to be seriously depleted. 

In the present case it can be hypothesized that the relativley mild drought may have triggered a mobilization of some 

defence mechanism in Clone 194; it may have induced tile accumulation of defensive compounds or have elicited the 

development of defence-related cell types (V.R. Franceschi, pets. comm.). A more severe drought might have prevented the 

mobilization of such defences by critically reducing the availability of some important resources, but this remains to be 

proved. 

SUMMARY 

Young Norway spruce trees of two different clones were artificially deprived of precipitation water over a period of 8 

months, starting in late October. In late June, pre-dawn water potentials of the experimental trees were significantly lower 

than those of unstressed control trees, but the stress was still tlairly mild. At this time both categories of trees were inoculated 

with a standard dose of the blue-stain fungus Ophiostoma polonicum, a usual associate of the spruce bark beetle Ips 

typographus. The stressed trees of one clone proved to be significantly more resistant to infection than unstressed controls of 

the same clone. The other clone showed a similar but much less pronounced trend. It is suggested that the mild stress has 

triggered the mobilization of some defence mechanism in the former clone. 


The study was supported by a grant from the Agricultural Research Council of Norway (Contract no. 205030). 

Thanks are due to Tore SkrOppa and colleagues for providing experimental trees; to Line Hamar, Heidi Knutsen, Olaug 

Olsen, and Torfinn Smther for field assistance; to Olaug Olsen for fungal reisolation; to Olaug Olsen and Halvor Solheim for 

fungal identification; and to Halvor Solheim and Ye Hui for critical examination of the manuscript. 


BERRYMAN, A.A. 1972. Resistance of conifers to invasion by bark beetle-fungus associations. BioScience 22: 598-602. 

CATES, R.G. and ALEXANDER, It. 1982. Host Resistance and Susceptibility, p. 212-263. In Mitton, J.B. and Sturgeon, 

K.B., eds. Bark Beetles in North American Conifers. University of Texas Press, Austin. 

CHRISTIANSEN, E. 1985a. Ceratocystis polonica inoculated in Norway spruce: Blue-staining in relation to inoculum 

density, resinosis, and tree growth. Eur. J. For. Pathol. 15' 160-167. 

CHRISTIANSEN, E. 1985b. lps/Ceratocystis-infection of Norway spruce: What is a deadly dosage? Z. Angew. Entomol. 

99:6-11 

CHRISTIANSEN, E. 199 I. Ips typographus and Ophios'toma polonicum versus Norway spruce: Joint attack and host 

defence, p. 321-324. In Baranchikov, Y., Mattson, W.J., Hain, F., and Payne, T., eds. Forest insect guilds: patterns of 

interaction with host trees. Gen. Tech. Rep. NE-153. Radnor, PA: U.S. Department of Agriculture, Forest Service: 

321-334. 

197

CHRISTIANSEN, E. 1992. After-effects of drought did not predispose young Picea abies to infection by the bark beetletransmitted 

blue-stain fungus Ophiostoma polonicum. Scand. J. For. Res. 7: 557-569. 

CHRISTIANSEN, E. and BAKKE, A. 1988. The spruce bark beetle of Eurasia, p. 479-503. In Berryman, A.A., ed. Dynamics 

of Forest Insect Populations. New York and London: Plenum Publishing Corporation. 

CHRISTIANSEN, E. and ERICSSON, A. 1986. Starch reserves in Picea abies in relation to defence reaction against a bark 

beetle transmitted blue-stain fungus, Ceratocystispolonica. Can. J. For. Res. 16: 78-83. 

CHRISTIANSEN, E. and HORNTVEDT, R. 1983. Combined Ips/Ceratocystis attack on Norway spruce, and defensive 

mechanisms of the trees. Z. Angew. Entomol. 96:110-118. 

CHRISTIANSEN, E. and SOLHEIM, H. 1990. The bark beetle-associated blue-stain fungus, Ophiostoma polonicum, can 

kill various spruces and Douglas fir. Eur. J. For.Pathol. 20: 436-446. 

CHRISTIANSEN, E., WARING, R.H., and BERRYMAN, A.A. 1987. Resistance of conifers to bark beetle attack: Searching 

for general relationships. For. Ecol. Manage. 22: 89-106. 

CRAIGHEAD, F.C. 1928. Interrelation of tree-killing bark beetles (Dendroctonus) and blue stains. J. Forestry 26: 886- 

887. 

DUNN, J.R and LORIO, EL. 1993. Modified water regimes affect photosynthesis, xylem water potential, cambial growth, 

and resistance of juvenile Pinus taeda L. to Dendroctonusfrontalis (Coleoptera: Scolytidae). Environ. Entomol. 22: 

948-957. 

FURNISS, M.M., SOLHEIM, H., and CHRISTIANSEN, E. 1990. Transmission of blue-stain fungi by Ips typographus 

(Coleoptera: ScoOytidae) in Norway spruce. Ann. Entomol.Soc. Am. 83: 712-716. 

HAVRANEK, W.M. and BENECKE, U. 1978. The influence of soil moisture on water potential, transpiration and 

photosythesis of conifer seedlings. Plant and Soil 49:91-103. 

HORNTVEDT, R. 11988. Resistance of Picea abies to Ips typographus: Tree response to monthly inoculations with 

Ophiostoma polonicum, a beetle transmitted blue-stain fungus. Scan. J. For. Res. 3" 107-114. 

HORNTVEDT, R., CHRISTIANSEN, E., SOLHEIM, H., and WANG, S. 1983. Artificial inoculation with lps typographusassociated 

blue-stain fungi can kill healthy Norway spruce trees. Medd. Nor. inst. skogforsk. 38(4): 1-20. 

HORNTVEDT, R. and SOLHEIM, H. 1991. Pathogenicity of Ophiostoma polonicum to Norway spruce: The effect of 

isolate age and inoculum dose. Medd. Skogforsk.44(4): 1-11. 

HSIAO, T.C. 1973. Plant responses to water stress. Ann. Rev.Plant Physiol. 24: 519-570. 

KALKSTEIN, L.S. 1976. Effects of climatic stress upon outbreaks of the southern pine beetle. Environ. Entomol.5: 653- 

658. 

MATSON, RA. and HAIN, F.R 1985. Host conifer defence strategies, p. 33-42. In Safranyik, L., ed. The Role of the Host 

in the Population Dynamics of Forest Insects. Proc. IUFRO Conf., Banff, Alberta, Canada, 4-7 Sept. 1983. Pacific 

For. Res. Centre, Can. For. Serv., Victoria B.C. 

MILLER, J.M. and KEEN, F.R 1960. Biology and control of thewestern pine beetle. U.S. Department of Agriculture, 

381 p. 

RAFFA, K.F. and BERRYMAN, A.A. 1983. The role of host plant resistance in the colonization behavior and ecology of 

bark beetles (Coleoptera: Scolytidae). Ecol. Monogr. 53: 27-49. 

198

RYAN, B.F., JOtNER, B.L., and RYAN, T.A., Jr. t992. Minitab Handbook. Second edition. Boston: PWS-KENT Publ. 

Comp. 

SCHOENEWEISS, D.F. t986. Water stress predisposition to disease - an overview, p. 157-174. In Ayres, EG. and Boddy, 

L., eds. Water, fungi and plants. Cambridge University Press. 

SCHROEDER, M. 1990. Duct resin flow in Scots pine in relation to attack of the bark beetle, Tomicus piniperda (L.) (Col., 

Scolytidae). J. Appl. Ent. 109:t05-112. 

SOLHEIM, H. 1986. Species of Ophiostornataceae isolated from Picea abies infested by the bark beetle, Ips O'pographus. 

Nord. J. Bot. 6: 19%207. 

SOLHEIM, H. 1988. Pathogenicity of some Ips 07)ographt_s-associated blue-stain fungi to Norway spruce. Medd. Nor. 

inst. skogforsk. 40(14): 1-11. 

SOLHEIM, H. t992. The early stages of fungal invasion in Norway spruce infested by the bark beetle, Ips O'pographus. 

Can. J. Bot. 70: 1-5. 

THALENHORST, W. t958. Grundztige der Populationdynamik desgrossen Fichtenborkenk_ifers Ips typographus L. 

SchrReiheforstl. Fak. Univ. G6ttingen 21: 1-126. 

199

A[ PROACHES TO STUDYING ENVIRONMENTAL EFFECTS ON 

RESISTANCE OF PINUS 7AEDA L. 

TO DENDROCTONUS FRONTALIS Z IMM.ERMANN 

..... ._ , 

PETER L. LORIO, JR. 

USDA Forest Service, Southern Forest Experiment Station, 2500 Shreveport Highway, Pineville, Louisiana 7 t360, USA 

INTRODUCTION 

E_vh>n_nental conditions and the genetic potential of loblolly pines, Pinus taeda L., affect or determine tree resista_ce 

to ;:mackby the southern pine beetle, DendroctonusJ?ontalis Zimmermann, by operating on physiological processes 

(Fig_ _). Er_vironment and genetic potential must operate through physiological processes to determine the quantity and 

quality of growdu as well as to express resistance to invasion by pathogens and bark beetles (Kramer t986). Only in this way 

can eiflaer e,_vironment or genetics affect growth and development at the cell or the whole tree level. Itere, I would like to 

focus primarily cm one major aspect of the environment that commonly affects the growth and development of lobtolly pines 

gradtheir _etative resistance to attack by the southern pine beetle; that is, the water regimes under which they may grow. It is 

important _:_consider the eft_ectsof environment across a range of time fi'ames, from very short (diurnal or even hourly) to 

very h:m_(litk:time). For example, trees growing on wet sites and in humid environments may grow rapidly and reach large 

size over king time frames because of prolonged wet conditions; however, in short time frames of days or weeks they may be 

su}:_icctedto severe water deficits not evident when data are summed over long periods. 

ROLE OF PHYSIOLOGY IN FORESTRY 

GENETIC POTENTIAL ENVIRONMENTAL FACTORS 

Tree Improvement Atmospheric, soil, 

programs and biotic 

Silvicultural treatments 

PHYSIOLOGICAL PROCESSES 

At whole plant level 

At cellular level 

QUANTITY AND QUALITY 

OF GROWTH 

(Usually far below the possible maximum) 

Figure I.o--Diagram illustrating the role of physiology in forestry. Genetic potential and environment operate through physiological 

processes in determining the quantity and quality of growth. Expression of resistance mechanisms of conifers 

to invasion by patho,,ense. _ and bark beetles is likewise governed by, these relationships (From Kramer, 1986). 

Ma{tson, W,J., Niemel_i, E, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest tbr pattern and principle. USDA 

For Serv. Gen. Tech. Rep. NC-t83, N.C. For. Exp. Sta., St. Paul, MN 55108. 

2{X}

It is seldom feasible to control water regime in the practice of tbrest management as one can control the density and 

spacing of trees, or the sites on which trees are grown. They may be planted in various ways, on dry or wet sites, thinned to 

reduce competition among individuals for light, water, and nutrients, and fertilized or pruned if desired, but in the southeastern 

United States there is little one can do about the great variation in rainfall through growing seasons and from year to year. 

Varying water regime can be a major factor in tree growth and development, as well as in tree resistance to southern pine 

beetle attack. 

One may approach the study of tree responses to water regimes in a variety of ways. Commonly, a manipulative 

approach is chosen to establish a range of conditions that may produce measurable responses considered important to the 

purpose. Although it is difficult, if not impossible, to modify one aspect of the environment without altering one or more 

other factors, manipulative studies can be very useh_l. However, it helps considerably to consider the possible effects of 

prevaili ng environmental conditions and the ontogenetic stage of trees on responses to treatment. 

Observational studies can be very helpful, if not essential, to the development of the most useful manipulative 

studies. By conducting such studies, one can develop some understanding that can help in the design of better studies in 

which treatments are imposed. The two approaches can be complementary. For example, (Chou 1982) emphasized the 

importance of seasonal predisposition of host trees in planning inoculation trials for any purpose. He found that Pinus 

radiata susceptibility to Diplodia pinea varied with season of the year, being high in spring-summer but low in autumnwinter. 

Knowledge of structural and physiological changes associated with ontogeny can be invaluable to interpretation of 

results of fungal inoculation studies, as indicated in studies by (Paine 1984) and (Owen et al. 1987), as well as with pinewood 

nematode inoculations (Myers 1986). Similarly, physiological changes that occur during ontogeny of trees may alter their 

susceptibility to forest pests (Kozlowski 1969). 

We have conducted a number of observational studies toward discovering tree and site characteristics correlated with 

the initiation and recurrence of southern pine beetle infestations (Lorio 1966, 1968; Lorio and Hodges 1968b, 1971; Hodges 

and Lorio 1971; Lorio et al. 1972; Lorio and Bennett 1974). These studies led to practical applications, such as the development 

of stand risk rating for the southern pine beetle (Lorio 1978, 1980; Lorio and Sommers 1981a, 1981b; Lorio et al. 1982; 

Zamoch et al. 1984; Hedden and Lorio 1985). A compendium on the southern pine beetle was based to a large extent on 

observational studies (Thatcher et al. t980). 

Here, I would like to review some studies that we have conducted with loblolly pine and the southern pine beetle in 

attempts to determine the effect of water regime on tree resistance to beetle attack. They include both manipulative and 

observational studies. Contrary to our early assumption that stress in the form of water deficit is bad for trees and good for 

bark beetles, a more sound assumption is that "it depends." It depends on a number of factors, such as tree size, growth stage 

or age, the timing of the deficit during the growing season, how severe the deficit becomes, how long it lasts, site conditions, 

tree root distribution and condition, and perhaps many other factors. It is especially important to remember that any such 

factors affect trees by operating through physiological processes (Kramer 1986). 

METHODS 

Early on we conducted both manipulative and observational studies to assess the effects of environmental conditions, 

principally soil water supply, on loblolly pine physiology and susceptibility to southern pine beetle attack. One of our 

manipulations was to dig trenches around groups of trees or individuals, and line them with polyethylene sheeting to restrict 

lateral movement of water and prevent roots of study trees from tapping soil water beyond the trench (Lorio and Hodges 

1968b, 1977; Hodges and Lorio 1969). We constructed various types of shelters that limited soil water recharge from 

rainfall and allowed us access to the trees for measurements. Other methods included continuously flooding tree root 

systems for long periods of time by maintaining ponded water on the site, and applying collars of solid CO 2 around the lower 

tree bole to inhibit the movement of water through sapwood to induce accute water deficit (Moeck et al. 198t, Stephen et al. 

1988). In more recent times we have girdled trees (Dunn and Lorio 1992), sheltered root systems without trenching, and 

irrigated in attempts to provide wen-watered conditions to compare with the effects of sheltering (Dunn and Lorio 1993). 

This and other studies included artificially inducing southern pine beetle attacks (Lorio and Hodges 1977; Dunn and Lorio 

1992, 1993), but in some studies, volunteer attack was relied upon (Lorio and Hodges 1968b, Hodges and Lorio 1969). 

201

Observational studies included examining site and stand conditions associated with southern pine beetle infested 

stands compared with noninfested stands (Lorio 1968), measuring soil water regimes, depth to free water in perforated 

tubing, and tree diameter growth (Lorio and Hodges 1971); determining tree rooting characteristics and distribution, and 

oleoresin exudation pressures in relation to microrelief (Lorio and Hodges 1968a, Lorio et al. 1972); and comparing methods 

of measuring field techniques for assessing the water status of trees (Hodges and Lorio 1971). One study, conducted over 

several years, included calculation of soil water balance (Zahner and Stage 1966), measurement of cambial growth by 

periodically extracting xylem plugs from tree boles and counting the number of tracheids produced in the current annual ring, 

counting vertical resin ducts in the current annual ring and calculating their density, and periodically sampling resin flow 

from wounds through the years (Lorio et al. 1990). 

RESULTS 

Manipulative studies to establish levels of water deficit that would severely weaken trees relative to nontreated 

controls (Lorio and Hodges 1968b, 1977; Hodges and Lorio 11969,1975) demonstrated that severe water deficits, aside from 

reducing growth, caused physical and chemical changes apparently associated with reduced tree resistance to beetle attack. 

However, it was not demonstrated that the southern pine beetle preferentially attacked trees subjected to severe water deficit, 

or that they required such weakened hosts tbr successful attack. On the contrary, it was clear that vigorously growing hosts 

could be overcome readily by this bark beetle species, and that such hosts could produce abundant brood. 

An observational study of relationships among xylem water potential of twigs (Scholander et al. 1965), relative water 

content of needles (Weatherley 1950), oleoresin exudation pressure in the lower main bole (Vit6 1961), and short-term flow 

of resin (Mason 1969, 1971) showed good relationships between resin pressure and the conventional measures of water 

status, but no relationship between resin pressure and flow (Hodges and Lorio 1971). The question of a direct relationship, or 

lack thereof, between oleoresin exudation pressure and oleoresin flow from bark wounds is discussed elsewhere (Lorio in 

press). 

Recent attempts to evaluate the effects of environmental conditions on loblolly pine resistance to the southern pine 

beetle provide some insight about the importance of environment to tree resistance. Results from over a 2-year period (1984- 

1985) of an observational study (Lorio et al. 1990) indicated substantial differences between years in calculated soil water 

storage, cumulative soil water deficits, cambial growth, density of resin ducts in the current year's annual ring, and the 

patterns of resin yields through the growing seasons. The results for 1986 and 1987 are reported here (Fig. 2). Clearly 

different water regimes prevailed from one year to the next, as did the apparent effects on cambial growth, resin duct densities, 

and resin yields. No treatments were applied and beetle attacks were not induced, and the apparent responses are 

correlative. However, the pattern of resin yields in 1987 closely mimics a conceptual model of seasonal changes in pine 

resistance to beetle attack in years having soil water balance patterns similar to the long-term average (Fig. 3). Resin flow 

from small bark wounds is considered indicative of the relative resistance of southern pines to southern pine beetle attack 

(Hodges et al. 1979). 

A manipulative study carried out within the same stand in the spring of 1986 and the late summer of 1987 indicated a 

strong environmental effect preceding and during the study (Stephen et al. 1988). Another manuscript that focuses on 

important differences in environmental conditions, growth and development of the host trees, and southern pine beetle 

population levels is pending. 

Girdling of bark to the face of the xylem prior to the development of latewood, and under well-watered conditions, 

produced considerably different responses to fungal inoculations or induced southern pine beetle attacks at 2 weeks compared 

to 8 weeks after girdling (Dunn and Lorio 1992). Two weeks after girdling, carbohydrate status, cambial growth, resin flow, 

phloem moisture content, lesion formation, and beetle colonization differed little above or below girdles. After 8 weeks the 

carbohydrate status below girdles was greatly reduced, as was cambial growth (no latewood formed). Although not statisti- 

cally significant, resin flow was reduced almost 50% below compared to above girdles. Phloem moisture was significantly 

higher below than above girdles, and beetles oviposited three times as many e_,,,s and constructed three times as much eee 

gallery below than above girdles. These results were obtained during a period of almost continually declining soil moisture 

and steadily accumulating calculated soil water deficits (Fig. 4). 

202

15 15 

u. 10 |O 

O 

N 

tal 

5 5 

g ao 

o 4o 40 

T 

I11 

X 

:0 20 20 

Z 

e_EoaO M 

z40 4O 

w 

a 

_20 20 

LATEWOOD 

_RLYWOOD 

EARLYWOOD ..,o./" , , __ , J , A 

0 _.__ .............. I ........ I _ ! l, m 1 ! I I _ 0 1 ' : : _' : : : _ ' , I I ! 

.-- g H 

.= 10 10 

N 

_. B B 

E at 6 

N 

4 4 

2 2 

I I I I I I I I I I I I _ __! I II I II I 1 ......L ........... L__ _ _ 

0 30 60 |10 t20 150 180 210 240 270 300 330 380 0 30 80 90 120 150 180 2t0 _N,0 270 300 330 380 

DAY OF'YEAR DAY OFYEAR 

Figure 2.--Graphs of water regimes, the course of xylem growth and development, vertical resin duct density in the 

current annual ring, and resin yield from standard bark wounds in 1986 (A, B, C, D) and 1987 (E, F, G, H). 

Daily soil water storage (solid line) and cumulative daily water deficits (dotted line), (A, E). Tree bole cambial 

growth and development expressed as the number of tracheids formed, time of transition from earlywood to 

latewood, and the total amount of earlywood and latewood formed (B, F). Vertical resin duct densities in the 

current annual ring (C, G). Oleoresin yield over 24-hour periods (D, H). Vertical bars are standard errors, n=l 1 

for 1986, n--13 for 1987. Water regimes, growth and development, and resin yields differ dramatically between 

years. Resin yields for 1987 closely approximate the seasonal changes in resistance suggested in the conceptual 

model shown in Figure 3 for resistance to southern pine beetle attack in years that have soil water regimes 203 

similar to that of the long-term average. Severely dry conditions in the summer of 1986 apparently not only 

limited carbon partitioning not only to growth, but also to secondary metabolism; e.g., oleoresin synthesis. 

4;

= *- _ *'- '_MODERATEWATEROEFICtTS -* 

Z 

"< 

_o 

_,a.# 

= 

EABLYWOOB LATEWO00 

FORMATION _- TRANSITION _ FORMATION 

-- VERTICALDUCTFORMATION- -- -" 

J F M A M J J A S 0 N D 

MONTHOF THEYEAR 

Figure 3.-,.-_,,,_A conceptual model of seasonal changes in pine resistance to southern pine beetle attack for years that have soil 

water balance patterns similar to that of the long-term average. Resistance to the earliest attacking beetles is considered 

to be highly dependent on the potential flow of oleoresin at the wound site. (From Lorio eta/. l,., 990). 

V 

=° /-'% 

_ \ ! \ INOCULATIONS-, .. 

_1 \ , SPB ATTACKj * -'" 

16 TREES CUT - x . 

X """" 

o 12 _, 

_ '_ .-" CDEF 

b-- 

8 

Z ." X 

I,a,! 

I 

150 180 21 0 240 270 

DAY OF YEAR 

Figure 4.--Daily precipitation (vertical bars), calculated daily soil water storage and cumulative daily water deficits (CDEF) 

tbr June-September, 1989, near Alexandria, LA. The two-level soil water storage program of Zahner and Stage 

21)4 (11966)was used, _suming as sandy clay loam soil holding :20.32 cm of available water in the tree rooting zone. 

Dates of girdling, inoculations, and induced southern pine beetle attack are indicated. (M(_ified from Dunn and Lorio 

1992).

Only recently have we conducted studies wilh juvenile pines {trees no_ yet into reproductive growth). One ,_ct_ study 

involved sheltering root systems to liinit water supply ai_d ir_igating to supplement rai__fal{ (l)uni_ am] I.,

O 

C 

Lower activity of apical meristems of the top 

relative to photosynthetic capacity 

I 1 ° 

Excess photosynthate in the top E I1 

Greater translocation to other active meristems" _ _o 

root apical and lateral cambia _E 

(D 

and c 

,,,--- 

.13 0 

-._ Greater differentiation of cells and tissues 

£:}_ 

1. protective secondary metabolites = 

2. higher soluble and insoluble storage compounds _, 

3. flower induction 

Figure 5.--Carbon partitioning among growth and developmental processes in response to changing nutrient and water 

supplies influence the competition and fluctuating balances between growth and differentiation processes, e.g., the 

various cambial meristems versus protective secondary metabolites; as well as among aboveground and belowground 

systems within a tree. (After Gordon and Smith 1987). 

Voluminous literature in the area of physiological ecology and plant/herbivore interactions in general appears to be 

related closely to the growth-differentiation balance concept (Mooney and Chu 1974, Mooney et al. 1983, Bryant et al. 1983, 

Coley et al. 1985, Chapin et al. 1987, and others). Although the paths of research in those areas did not include consideration 

of Loomis's (1932) concept, I find that it provides considerable support for developments in those fields. For example, it 

seems to me to be essentially in agreement with the carbon-nutrient balance hypothesis (Bryant et al. 1983, Tuomi et al. 

1988), but provides a broader framework, primarily by providing explicit consideration for water. Several papers include 

consideration of the plant growth-differentiation balance concept in forming theories of evolutionary development of plant 

defenses against insects (Tuomi et al. 1990, Herms and Mattson 1992, Tuomi 1992). 

Here, I have indicated some of the research results from both early and recent studies. The more recent studies were 

carried out with consideration of plant growth-differentiation balance relationships in mind, and with considerable stimula- 

tion from advances in physiological ecology and studies of plant/herbivore interactions in general. We are continuing 

research in the same vein. The study with juvenile pines (Dunn and Lorio 1993) has been extended to an older stand, with 

some improvements in techniques, and with results indicating strongly that water deficits have nonlinear effects on tree 

206 

o 

cO 

"O c 

,,t

esistance to beetle attack. Another study is in progress, once again with juvenile pines, in a plantation subjected to thinning 

and fertilization. We will be continuing work in that direction. At this time preliminary results indicate that the short-term 

effects of thinning and fertilization are to enhance carbon partitioning to growth and to reduce carbon committed to secondary 

metabolism; e.g., resin synthesis. 

SUMMARY 

There are a number of ways to approach the problem of assessing the effects of environmental conditions, such as 

water regime, on tree physiological responses and resistance to bark beetle attack. It helps to keep in mind that environmental 

factors operate through physiological processes (Fig. 1, and Kramer 1986), and that there are concepts, such as plant 

growth-differentiation balance (Loomis 1932) and carbon-nutrient balance (Bryant et al. 1983; Tuomi 1992; Tuomi et al. 

1988, 1990), that can provide bases for forming testable hypotheses. It is especially important to know as much as possible 

about the host tree and the specific bark beetle of interest. Observational studies are particularly important because they can 

provide a baseline for interpreting the results of manipulative studies. Whenever feasible, it is especially helpful in manipulative 

studies to characterize environmental conditions before and during a study. Because physiological changes that occur 

during ontogeny of trees can alter their susceptibility to herbivores, knowledge of tree stage of growth and development can 

be especially helpful. 


Cooperative Agreements (19-86-036 and 19-87-20) between the Mississippi Agricultural and Forestry Experiment 

Station and the USDA Forest Service, Southern Forest Experiment Station, contributed in part to results reported here. 

' LITERATURE CITED 

BRYANT, J.P., CHAPIN, F.S., III, and KLEIN, D.R. 1983. Carbon/nutrient balance of boreal plants in relation to vertebrate 


CHAPIN, ES., III, BLOOM, A.J., FIELD, C.B., and WARING, R.H. 1987. Plant responses to multiple environmental 

factors. BioScience 37: 49-57. 

CHOU, C.K.S. 1982. Susceptibility of Pinus radiata seedlings to infection by Diplodia pinea as affected by pre-inoculation 

conditions. New Zealand J. For. Sci. 438-441: 

COLEY, P.D., BRYANT, J.P., and CHAPIN, F.S., III. 1985. Resource availability and plant antiherbivore defense. Science 

230: 895-899. 

DUNN, J.P. and LORIO, P.L., Jr. 1992. Effects of bark girdling on carbohydrate supply and resistance of loblolly pine to 

southern pine beetle (Dendroctonusfrontalis Zimm.) attack. For. Ecol. and Manage. 50:317-330. 

DUNN, J.P. and LORIO, P.L., Jr. 1993. Modified water regimes affect photosynthesis, xylem water potential, cambial 

growth, and resistance of juvenile Pinus taeda L. to Dendroctonusfrontalis (Coleoptera: Scolytidae). Environ. 

Entomol. 22(5): 948-957. 

GORDON, J.C. and SMITH, W.H. 1987. Tree roots and microbes: a high priority for high technology, p. 423-432. In 

Boersma, L.L., et al., eds. Future Developments in Soil Science Research. Soil Science Society of America, Madison, 

WI. 

HEDDEN, R.L. and LORIO, RL., Jr. 1985. Rating stand susceptibility to southern pine beetle attack on national forests in 

the Gulf Coastal Plain. Res. Pap. SO-221. Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern 

Research Station. 5 p. 

207

HERMS, D.A. AND MATTSON, W.J. 1992. The dilemma of plants: to grow or defend. Quart. Rev. Biol. 67: 283-335. 

HODGES, J.D. and LORIO, EL., Jr. 1969. Carbohydrate and nitrogen fractions of the inner bark of loblolly pines under 

moisture stress. Can. J. Bot. 47: 1651-1657. 

HODGES, J.D. and LORIO, EL., Jr. I971. Comparison of field techniques for measuring moisture stress in large loblolly 

pines. For. Sci. 17: 220-223. 

HODGES, J.D. and LORIO, EL., Jr. 1975. Moisture stress and composition of xylem oleoresin in loblolly pine. For. Sci. 22: 

283-290. 

HODGES, J.D., ELAM, W.W., WATSON, W.E, and NEBEKER, T.E. 1979. Oleoresin characteristics and susceptibility of 

four southern pines to southern pine beetle (Coleoptera: Scolytidae) attacks. Can. Entomol. 111:8 89-896. 

KLEBS, G. 1903. WiIlkurliche Entwickelungsanderungen bei Pflanzen. G. Fisher, Jena. 

KLEBS, G. 1910. Alterations in the development and forms of plants as a result of environment. Proceedings of the Royal 

Society of London 82B: 547-548. 

KOZLOWSKI, T.T. 1969. Tree physiology and forest pests. J. For. 67:118-123. 

KRAMER, P.J. 1986. The role of physiology in forestry. Tree Physiol. 2: 1-16. 

LOOMIS, W.E. 1932. Growth-differentiation balance vs carbohydrate-nitrogen ratio. Proceedings of the American Society 

for Horticultural Science 29: 240-245. 

LOOMIS, W.E. 1953. Growth correlation, p. 197-217. In Loomis, W.E., ed. Growth and Differentiation in Plants. Iowa 

State College Press, Ames, IA. 

LOOMIS, R.S., LUO, Y., and KOOMAN, E 1990. Integration of activity in the higher plant, p. 105-124. In Rabbinge, R. et 

al., eds. Theoretical Production Ecology: Reflections and Perspectives. Pudoc, Wageningen, The Netherlands. 

LORIO, EL., Jr. 1966. Phytophthora cinnamomi and Pythium species associated with loblolly pine decline in Louisiana. 

Plant Dis. Rep. 50: 596-597. 

LORIO, EL., Jr. 1968. Soil and stand conditions related to southern pine beetle activity in Hardin County, Texas. J. Econ. 

Entomol. 61: 565-566. 

LORIO, EL., Jr. 1978. Developing stand risk classes for the southern pine beetle. Res. Pap. SO-144. Asheville, NC: U.S. 

Department of Agriculture, Forest Service, Southern Research Station. 9 p. 

LORIO, EL., Jr. 1980. Loblolly pine stocking levels affect potential for southern pine beetle infestation. South. J. Appl. For. 

4: 259-265. 

LORIO, EL., Jr. 1986. Growth-differentiation balance: a basis for understanding southern pine beetle-tree interactions. For. 

Ecol. and Manage. 14: 259-273. 

LORIO, EL., Jr. 1993. The relationship of oleoresin exudation pressure (or lack thereof) to flow from wounds. J. Sustain. 

For. (In Press). 

LORIO, EL., Jr. and BENNETT, W.H. 1974. Recurring southern pine beetle infestations near Oakdale, Louisiana. Res. 

Pap. SO-95. Asheville, NC: U.S. Department of Agriculture, Forest Service, Southern Research Station. 6 p. 

LORIO, EL., Jr. and HODGES, J.D. I968a. Microsite effects on oleoresin exudation pressure of large loblolly pines. 

Ecology 49(6): 1207-1210. 

2O8

LORIO, EL., Jr'. and HODGES, J.D. 1968b. Oleoresin exudation pressure and relative water content of inner bark as 

indicators of moisture stress in loblolly pines. For. Sci. 14(4): 393-398. 

LORIO, RL., Jr. and HODGES, J.D. 1971. Microrelief, soil water regime, and loblolly pine growth on a wet, mounded site. 

Soil Science Society of America Proceedings 35: 795--800. 

LORIO, RL., Jr. and HODGES, J.D. 1977. Tree water status affects induced southern pine beetle attack and brood production. 

Res. Pap. SO-135. New Orleans, LA: U.S. Department of Agriculture, Forest Service, Southern Forest Experiment 

Station. 7 p. 

LORIO, RL., Jr. and HODGES, J.D. 1985. Theories of interactions among bark beetles, associated microorganisms, and 

host trees, p. 485-492. In Shoulders, E., ed. Proceedings of the 3rd Biennial Southern Silvicultural Research Conference. 

Gen. Tech. Rep. SO-54. New Orleans, LA: U.S. Department of Agriculture, Forest Service, Southern Forest 

Experiment Station. 

LORIO, RL., Jr. and SOMMERS, R.A. 1981a. Central Louisiana, p. 23-39. In Coster, J.E. and Searcy, J.L., eds. Site, Stand, 

and Host Characteristics of Southern Pine Beetle Infestations. Tech. Bull 1612. Pineville, LA: U.S. Department of 

Agriculture, Forest Service. Combined Forest Pest Research and Development Program. 

LORIO, RL., Jr. and SOMMERS, R.A. 1981b. Use of available resource data to rate stands :for southern pine beetle risk, p. 

75-78. In Hedden, R.L., Barras, S.J., and Coster, J.E., tech. coords. Proceedings of Symposium Hazard-Rating 

Systems in Forest Pest Management. Gen. Tech. Rep. WO-27. New Oreleans, LA: U.S. Department of Agriculture, 

Forest Service, Southern Forest Experiment Station. 

LORIO, RL., Jr., HOWE, V.K., and MARTIN, C.N. 1972. Loblolly pine rooting varies with microrelief on wet sites. 

Ecology 53:1134-1140. 

LORIO, P L., Jr., MASON, G.N., and AUTRY, G.L. 1982. Stand risk rating for the southern pine beetle: integrating pest 

management with forest management. J. For. 80:212-214. 

LORIO, RL., Jr., SOMMERS, R.A., BLANCHE, C.A., HODGES, J.D., and NEBEKER, T.E. 1990. Modeling pine resistance 

to bark beetles based on growth and differentiation balance principles, p. 402-409. In Dixon, R.K., Meldahl, 

R.S., Ruark, G.A., and Warren, W.G., eds. Process Modeling of Forest Growth Responses to Environmental Stress. 

Timber Press, Portland, OR. 

MASON, R.R. 1969. A simple technique for measuring oleoresin exudation flow in pine. For. Sci. 15: 56-57. 

MASON, R.R. 1971. Soil moisture and stand density affect oleoresin exudation flow in a loblolly pine plantation. For. Sci. 

17: 170-177. 

MOECK, H.A., WOOD, D.L,. and LINDAHL, K.Q., Jr. 1981. Host selection behavior of bark beetles (Coleoptera: 

Scolytidae) attacking Pinus ponderosa, with special emphasis on the western pine beetle, Dendroctonus brevicomis. 

J. Chem. Ecol. 7(1): 49-83. 

MOONEY, H.A. and CHU, C. 1974. Seasonal carbon allocation in Heteromeles arbutifolia, a California evergreen shrub. 

Oecologia 14: 295-306. 

MOONEY, H.A., GULMON, S.L., and JOHNSON, N.D. 1983. Physiological constraints on plant chemical defenses, p. 21- 

36. In Hedlin, RA., ed. Plant Resistance to Insects. American Chemical Society, Washington, D.C. 

MYERS, R.F. 1986. Cambium destruction in conifers causedby pinewood nematodes. J. Nematol. 18: 398-402. 

OWEN, D.R., LINDAHL, K.Q., WOOD, D.L., and PARMETER, J.R. 1987. Pathogenicity of fungi isolated from 

Dendroctonus valens, D. brevicomis, and D. ponderosae to ponderosa pine seedlings. Phytopathology 77: 631-636. 

209

PAINE, "I.D. 1984. Seasonal response of ponderosa pine to inoculation of the mycangial fungi isolated from the western 

pine beetle. Can. J. Bot. 62:551-555. 

SCHOLANDER, RE, HAMMEL, H.T., BRADSTREET, E.D., and HEMMINGSEN, E.A. 1965. Sap pressure in vascular 

plants. Science 148: 339-346. 

STEPHEN, F.M., LIH M.R, PAINE, T.D., and WALLIS G.W. 1988. Using acute stress to modify tree resistance: impact on 

within-tree southern pine beetle populations, p. 105-119. In Payne, T.L. and Saarenmaa, H., eds. Integrated Control 

of Scolytid Bark Beetles. Proceedings of the IUFRO Working Party and XVII International Congress of Entomology 

Symposium, Vancouver, B.C., Canada, July 4, 1988. Virginia Polytechnic Institute and State University. 

THATCHER, R.C., SEARCY, J.L., COSTER, J.E., and HERTER, G.D. (eds.). 1980. The southern pine beetle. Tech. Bull. 

1631. U.S. Department of Agriculture, Forest Service, and Science and Education Administration. 267 p. 

TREWAVAS, A. 1985. A pivotal role for nitrate and leaf growth in plant development, p. 77-91. In Baker, N.K., et al., eds. 

Control of Leaf Growth. Cambridge University Press, Cambridge. 

TUOMI, J. 1992. Toward integration of plant defence theories. Trends in Ecol. and Evol. 7: 365-367. 

TUOMI, J., NIEMELA, R, and SIREN, S. 1990. The Panglossian paradigm and delayed inducible accumulation of foliar 

phenolics in mountain birch. Oikos 59: 399-410. 

TUOMI, J., NIEMELA, R, CHAPIN, F.S., III, BRYANT, J.R, and SIREN, S. 1988. Defensive responses of trees in relation 

to their carbon/nutrient balance, p. 57-72. In Mattson, W.J., Levieux, J., and Bernard-Dagan, C., eds. Mechanisms of 

Woody Plant Defenses Against Insects. Springer-Verlag, New York. 

VITE, J.R 1961. The influence of water supply on oleoresin exudation pressure and resistance to bark beetle attack in Pinus 

ponderosa. Boyce Thompson Inst Contrib 21 : 37-66. 

WEATHERLEY, RE. 1950. Studies in the water relations of the cotton plant. I. The field measurement of water deficits in 

leaves. New Phytologist 49: 81-97. 

ZAHNER, R. and STAGE, A.R. 1966. A procedure for calculating daily moisture stress and its utility in regressions of tree 

growth on weather. Ecology 47: 64-74. 

ZARNOCH, S.J., LORIO, RL., Jr., and SOMMERS, R.A. 1984. A logistic model for southern pine beetle stand risk rating 

in central Louisiana. J. Georgia Entomol. Soc. 19: 168-175. 

210

EFFECTS OF ROOT INHABITING INSECT-FUNGAL COMPLEXES 

ON ASPECTS OF TREE RESISTANCE TO BARK BEETLES 

KENNETH E RAFFA and KIER D. KLEPZIG _ 

Department of Entomology, University of Wisconsin, Madison, Wisconsin 53706, USA 

INTRODUCTION 

Forest declines have proven difficult to address, because they involve complex interactions in which the key processes 

are only poorly understood (Houston 1981, Witcosky and Hansen 1985, Witcosky et al. 1986, Klein and Perkins 1988, 

Mueller-Dombois 1988, Huettl et al. 1990, Manion 1991, Saxe 1993, Woodcock et al. 1993). Without a basic understanding 

of the underlying mechanisms, it is difficult to relate field observations to various management practices, anthropogenic or 

environmental stresses, or natural successional events. One of our major information gaps concerns the interactions among 

different insect and fungal feeding guilds that both respond and contribute to plant stress physiology. We have been addressing 

this question in declining red pine, Pinus resinosa, stands in the Great Lakes region of North America. Our objective is to 

link field correlations with underlying physiological mechanisms. 

Insects and Pathogens Associated With Declining Red Pine Stands 

Red Pine Decline Disease is characterized by a progressive deterioration and death of mature plantation trees. 

Affected portions of stands are highly localized, and mortality progresses radially from an epicenter (Klepzig et al. 1991). 

No living trees remain in this epicenter, and the resulting gap is colonized by herbaceous and early succession vegetation. 

Trees killed during the current growing season ring this zone. Neighboring trees show reduced radial and crown growth, 

whereas trees farther into the stand show no symptoms. Each year, these zones expand. The killed trees become windblown, 

the low vigor trees are killed, and additional trees show signs of stress. This syndrome can be characterized at the betweenstand, 

within-stand, and within-tree levels. In a 3-year evaluation of 19 sites at 13 locations in Wisconsin, Klepzig et al. 

(1991) found that the most consistent between-stand correlate with Red Pine Decline was the abundance of root- and lowerstem 

infesting beetles (Table 1). These included three curculionid and two scolytid species. Several other scolytids and 

curculionids were also present in high numbers, but their densities did not consistently differ between declining and asymptomatic 

plantations of the same age category. Likewise, there were no strong differences associated with abiotic soil factors. 

The beetles associated with Red Pine Decline differ in their within-stand spatial distributions relative to the margin of 

tree mortality. For example, four beetles occur at higher densities in declining than asymptomatic stands. However, two of 

these show no significant within-stand distribution patterns, one occurs more commonly near the zone of stressed trees, and 

one occurs at higher densities in regions of the stand not yet showing visible symptoms (Table 1). Interestingly, the red 

turpentine beetle, D. valens Le Conte, occurs at higher densities in the asymptomatic than symptomatic portions of declining 

stands, yet does not show statistically significant between-stand differences. 

The subcortical belowground beetle guild inhabiting Wisconsin pines shows a high degree of niche partitioning. This 

separation among closely related beetles is based on host tissue tree vigor, olfactory behavior, and symbiotic relationships 

(Raffa and Hunt 1989, Rieske and Raffa 1990, 1991, Raffa and Klepzig 1992, Hunt et al. 1993). For example, H. porculus 

1Currentaddress: Department of Plant and Soil Science, Southern University and A & M College, Baton Rouge, Louisiana 70813 

USA. 



211

Table 1.--Root- and lower-stem infesting beetles associated with Red Pine Decline Disease (From Klepzig et al. 1991). 

A. Species occurring at higher densities in declining than asymptomatic stands 

Hylobius radicis (Curculionidae) 

Hylobius pales (Curculionidae) 

Pachylobius picivorus (Curculionidae) 

Hylastes porculus (Scolytidae) 

B. Species occurring at higher densities in asymptomatic portions of declining stands 

Dendpvctonus vatens (Curculionidae) 

tlylastes porculus (Scolytidae) 

C. Species occurring at higher densities in symptomatic portions of declining stands 

Pachylobius picivorus (Curculion idae) 

and H. rhyzophagus colonize lateral roots; H. pales, H. radicis, and P.picivorus colonize the root collar region; and D. vaIens 

enter trees just above the soil line. Moreover, beetle behavioral responses to volatiles associated with stress (Kimmerer and 

Kozlowski 1982) reflect differences in preferred host physiology. For example, H. pales and P. picivorus differ in their 

relative responses to various ratios of terpene-ethanol mixtures (Rieske and Raffa 1991). Each of the beetles in Table 1 is 

closely associated with Ophiotomatales fungi. The scolytids are primarily associated with Leptographium terebrantis Barras 

and Perry, and the curculionids are primarily associated with Leptographium procerum (Kendrick) Wingfield (Klepzig et al. 

1991). These fungi are moderately pathogenic to P. resinosa but cannot kill healthy trees (Raffa and Smalley 1988a, Klepzig 

et al. 1994a). The rates of association between these beetles and fungi average around 70%, and the vector efficiencies are 

around 50% (Klepzig et al. 1994b). 

Colonization by the beetles shown in Table 1 and their associated fungi does not kill host trees. A survey of living red 

pines revealed that a high percentage of the trees peripheral to the margin of tree mortality were infested with D. valens, H. 

radicis, or both (H. pomulus was not sampled). Such trees can remain alive for several years. In most cases, however, these 

trees are subsequently killed by the stem colonizing bark beetle lps pini (Say), and its associated fungus Ophiostoma ips 

(Rumb.) Nannf (Fig. 1). All of the dead, but none of the living, trees were infested with I. pini. Thus, the I. pini - O. ips 

complex appears to be the ultimate cause of tree death, but root infect:ion appears to be a predisposing factor. I. pini appears. 

to be highly opportunistic. Despite its strong association with individual tree death, between-stand population densities were 

not associated with stand condition. 

The above view is supported by our observations of a declining stand over a 5-year period (Fig. 2). The percentage 

of trees killed by I. pini that showed prior colonization by H. radicis averaged 86.8 + 9.7(SD%). The percentage of trees 

killed by I. pini that showed prior colonization by D. valens averaged 19.4 + 12.4 (SD%). Most of the trees infested by one 

of these root colonizing species were also infested by the other. Interestingly, the lowest proportion of trees killed by L pini 

that had previously been infested by root insects occurred during 1988, a year of severe drought (H. radicis: 72.7%; D. 

valens: 8.2%). Belowground excavations of five declining and control stands revealed a similar, but even more striking, 

trend (Klepzig et al. 1991). In declining stands, fungal staining and root deterioration proceeded in advance of the insect 

vectors and any aboveground symptoms. Significant root infection occurs 10-15m beyond the margin of killed trees. In 

control stands, and in the portions of symptomatic stands far beyond the stressed trees, root infection levels are negligible. 

Leptographium fungi progress through root grafts, which are extensive throughout mature P resinosa plantations. The extent 

of grafting is so common in these plantations that the phloem tissue of stumps can remain viable for several years (Klepzig et 

af. 1994a). The exact mechanism by which Leptographium fungi progress through root grafts into healthy tissue remains 

unknown, however. 

212

OCCURRENCE OF ROOT AND STEM SUBCORTICAL 

BEETLES JN LmVINGRED PINES 

NONE i 

D. VALENS 

HYLOBIUS 

I, PIN_ N = 

184 

_----T------__----T------ 

0 10 20 30 40 50 60 

% iNFESTED 

OCCURRENCE OF ROOT AND STEM SUBCORTICAL 

BEETLES IN KILLED RED PINES 

NONE 

,,_ 

D. VALENS _ 

HYLOBIUS 

/ N = 229 

I. PINi i 

..... "" l • I " I "'" .... _ ............ " .......... i 

0 20 40 60 80 100 

% INFESTED 

Figure 1.--Association of root- and stem-feeding subcortical beetles with living (above) and dead (below) red pines in 

declining plantations. 

213

.,_.o,4e,., HQ,,N,.,_ 

tn 8 O- H. radicis 

UJ I00- __ K,,.,._ 

IJl.! ",Xx,, " 

_" 60- xx _ 1987 

t_ 

I.U ........X ........ 1992 

t.t) 40- 

o_ 20- 

__ 

(D O tt') It) 

V v.- £Xl (Xl 

, i A 

t,.O 'r'y-,,. 

100 - x ......... -x_ 

a.... .----_',::,,.. . . 

"_,:., los pm/ 

t,_ 80" --'-"-O - 1987 

" ", .... 0 .... 1989 

(/") .... 4_.... 1990 

m 40" ,,. \ x • "... 

" N \ \""-.:"-.. ---o---199_ 

z N-, \ 'u. "'-"" 

'- .,,.. -...::-,.,., 

"---.-,,,,-..:.,:... 

0 _ ......... "_'_"-- 

(D 

V 

0 

'rtO 

¢q 

, 

LO 

CXl 

A 

CO "-- 

,,I-m 

DISTANCE FROM 1986 MARGIN OF TREE MORTALITY (m) 

Figure 2.--Progression of infestation in a declining red pine stand. Infestation rates of the pine root collar weevil, Hylobius 

radicis (above) and the pine engraver, lois pini (below) relative to the originally recorded margin of killed trees are 

shown. 

214

]Predisposition of Root-infected Trees to Stem Colonizing Insect-fungal Complexes 

Based on these observations, we hypothesized that prior infection by root- and lower-stem colonizing insects and 

fungi reduces P. resinosa resistance against L pini - 0. ips complexes. Testing this hypothesis requires both an understanding 

of red pine defense mechanisms against subcortical invasion, and comparisons of the key parameters between root-infected 

and non-root-infected trees. 

The response of P resinosa to invasion by biotic agents involves the secretion of preformed allelochemicals from 

severed resin ducts, and the rapid formation of a necrotic lesion around, and accumulation of additional allelochemicals at, 

the entrance site (Raffa and Smalley 1988b, 1994). These allelochemicals are comprised primarily of monoterpenes and 

phenolics. This is a rather generalized response that can be elicited by multiple agents and is characterized more by quantitative 

than qualitative allelochemica[ changes. It resembles those reported for other conifer-scolytid-Ophiostoma systems (e.g., 

Berryman 1972; Hodges et al. 1979; Cates and Alexander 1982; Raffa and Berryman 1982a,b; Christensen and Horntvedt 

1983; Lorio t986, 1993; Paine and Stephen 1987a,b; Lieutier and Berryman 1988; Nebeker et al. 1993; Raffa et al. 1993). 

Both constitutive and reaction allelochemicals are important to tree defense. However, allelochemical concentrations are 

higher in reaction tissue, and they have higher inhibitory effects on fungal germination and growth and on insect feeding and 

survival (Klepzig et al. 1994d). For example, I. pini adults strongly prefer media amended with extracts or simulated 

monoterpene doses from constitutive over reaction phloem tissue. Likewise, the monoterpene levels present in constitutive 

stem phloem tissue can kilt 60% of adult L pini within 2 days, but monoterpene concentrations present only 3 days after 

initiation of an induced response are sufficient to kill 90% of the beetles (Fig. 3) (Raffa and Smalley 1994). Without effective 

induced responses, the surviving beetles would probably attract enough additional beetles to deplete host resins and overwhelm 

tree resistance; without high constitutive levels, trees would be unable to delay aggregation long enough for induced 

reactions to be mobilized (Raffa and Berryman 1982a, 1983a,b, 1987; Raffa et al. 1993). 

E _00 100 

c._ f,q 

c:_ 60 80 >- 

._J 

L) ._ 

z 60 60 __ 

O 

40 40 

o 

2: 

Ill TOXICITY 

__J MONOTERPENE CONO 

Z LU 

D_ 

o: 

UJ 

I-- 

0 

z 

C) 

20 

0 

20 

0 

D 

LU 

0:3 

3 

DAYS AFTER ATTACK 

Figure 3.--Effects of constitutive and induced monoterpene concentrations on survival of Ips pini adults. Monoterpene 

concentrations in mature red pines from days 0 (pre inoculation) to 3 following inoculation with O. ips are shown in 

foreground. Mortality rate to adult I. pini for each concentration of synthetic monoterpene in laboratory assay is 

shown in background. 

215

We conducted a series of comparisons between root-infected and asymptomatic trees for a range of parameters 

associated with host susceptibility to I. pini - O. ips complexes. These comparisons included field assays at the whole tree or 

controlled bait level, and laboratory assays involving application of controlled treatments. Both groups of assays involved 

behavioral tests with I. pini, and tree physiological responses to simulated I. pini - O. ips attack. The results are summarized 

in Table 2. 

Behavioral assays evaluated three phases of scolytid orientation to host trees: initial landing, host entry, and gallery 

formation (Wood 1972, Elkinton and Wood 1980, Raffa and Berryman 1982c, Raffa 1988). Arrival rates to baited traps did 

not differ between stem phloem strips from root-infected or healthy trees, stems artificially infected with various fungi 

associated with Red Pine Decline, or volatiles associated with various intensities of stress (Table 2). When male I. pini were 

caged directly onto trees, however, the entry rates were 2.4 X higher on trees adjacent to the margin of dead trees than on 

trees farther into the stand. Prior excavations had demonstrated high levels of root disease in the former group (Klepzig et al. 

1991). Three conclusions can be drawn from these results: (1) beetles can detect changes in stem phloem physiology and 

Table 2.wComponents of I. pini colonization of P. resinosa affected by prior infestation with root-colonizing insects and 

fungi 

Parameter Treatment Comparison Result 

a. Initial landing Host stem tissue Rootinfested vs. healthy trees ns 

Host stem tissue Fungal-infected vs. non-infected ns 

stems (Oi, On, Lt) 

Volatiles etOH, ct-pinene, EtOH & tx-pinene, ns 

vs. controls 

b. Host entry Caging onto whole trees Root infected vs. healthy trees Prefer diseased trees: 7.X 

c. Gallery formation Phloem strips Root weevil infected vs. healthy Prefer infested logs: 2.4X 

logs 

Amended phloem Nonpolai" extracts from constitu- Prefer healthy trees: 10X 

tive stem phloem of rootinfected 

vs. healthy trees 

Amended phloem Nonpolar extracts from reaction Prefer diseased trees: 3.0X 

stem phloem of root-infected 

vs. healthy trees 

Amended phloem Monoterpene contents present Prefer diseased trees: 3.5X 

in reaction stem phloem of 

root-infected vs. healthy trees 

d. Constitutive resin Resin flow Stem phloem of root-infected ns 

vs. healthy trees: 1.5m height 

Stem phloem of root-infected Higher in healthy trees: 4X 

vs. healthy trees: 10m height 

e. Reaction resin a-pinene Root diseased vs. healthy trees Higher in healthy trees: 1.3X 

216 

Data from Klepzig et al. (in prep.), Raffa and Smalley (in prep.).

chemistry associated with root colonization:; (2) there is no evidence that pre-landing behavior is influenced by factors 

associated with root colonization; (3) the pattern of tree mortality, in which trees near the "pocket" are more likely to be 

killed by I. pini than are trees distant from the margin of killed trees (Fig. 2), is not just a function of proximity to emerging 

brood. Even when the factor of beetle dispersal is removed, higher entry rates are observed on neighboring trees. 

Beetle preference for root-diseased over healthy trees persists under laboratory conditions. L pini excavated more 

extensive galleries in phloem strips of logs that had been artificially infested with H. radicis than in control logs. Likewise, 

extracts from tlhe stem phloem tissue of root-diseased and healthy trees affect L pini behavior differently (Klepzig et al. 

1994c). When these extracts are applied to ground denatured phloem incorporated into agar, male L pini show strong 

preferences. Both the extent of tunnelling and the location of beetles are influenced by the source of host extracts under twoway 

choice conditions. Interestingly, the beetles preferred extracts from the constitutive tissue of healthy over diseased trees. 

This suggests that low levels of monoterpenes function as feeding incitants. However, when trees are challenge inoculated 

with O. ips, the pattern is reversed. The reaction tissue of healthy trees is much more repellent than the reaction tissue of 

diseased trees (Table 2). The same results can be achieved by amending the media with synthetic monoterpene concentrations 

determined to be present in these various extracts. 

Once entered by beetles, root-infected trees are physiologically less able than uninfected trees to resist attack by/. 

pini and its associated fungus O. ips. Differences occur in both primary and induced resin (Table 3). Interestingly, we did not 

observe differences in primary resin flow rates between root-infected and healthy trees when the wounds were applied at 

1.5m height. However, flow rates at the base of the canopy yielded only 25% of the resin volume in root-diseased as in 

healthy trees. Thus, the earliest symptoms of root infection may arise relatively high up in the stem. 

Based on these results, we hypothesized the following sequence: (1) root- and lower stem- infecting beetles colonize 

selected trees and vector Leptographium fungi into the root system; (2) root fungi progress through root grafts into healthy 

trees. These organisms do not kill mature red pines, but reduce their resistance to the stem-colonizing I. pini and its associated 

fungus O. ips; (3) colonization by this stem inhabiting complex is lethal to host trees; (4) killed trees provide a food base 

for secondary root-feeding insects, which introduce additional inoculum into the epicenter. 

Table 3.wEnvironmentat factors affecting host suitability to insects and fungi associated with Red Pine Decline Disease 

Factor Effect 

Reduced lights Increased lesion length follwoing inoculation with L. terebrantis: 2.5X 

Defoliation Increased lesion length following inoculation with O. ips: 1.2X 

Defoliation Decreased resin flow following mechanical wound: 0.45X 

Defoliation Preferential feeding by H. pales: 2.3X 

Nitrification Increased feeding by H. pales: 1.4X 

Nitrification Decreased development time by H. pales females: 0.89X 

Nitrification Decreased development time by H. pales males: 0.92X 

Nitrification Increased growth by H. pales females: 1.2X 

Nitrification Increased growth by H. pales males: 1.1X 

Nitrification incrased survival by H. pales: 2.1X 

Data from Hunt et al. (1993), Klepzig et al. (in 1994c.),Krause and Raffa (in prep.) 

217

Environmental Predisposition to Subcortical Insects and Fungi Associated with Declining Red Pine Stands 

Although the above description may partially describe within-stand dynamics, it cannot explain the onset of decline. 

To address this question, we have conducted a variety of experiments on the effects of environmental stress on host suitability 

to insects and pathogens associated with Red Pine Decline. The results suggest that a relatively broad range of biotic and 

abiotic factors can improve host suitability to these organisms (Table 3). For example, defoliation decreases resin flow, 

reduces tree ability to confine O. ips, and increases feeding preference to H. pales (Krause and Raffa 1994). Likewise, 

nitrification increases feeding, development, and survival by H. pales (Hunt et al. 1993). Reduced light availability also 

decreases host ability to respond to L. terebrantis (Klepzig et al. 1994c). Both defoliation and reduced light availability 

reduce the overall pool of photosynthate for carbon-based defenses such as terpenes and phenolics (Wright et al. 1979, 

Ericsson et al. 1985, Lorio and Summers 1986, Miller and Berryman 1986, Peet and Christensen 1987, Waring 1987, Dunn et 

al. 1990, Reich et al. 1992, Krause et al. 1993, Krause and Raffa 1994). These stresses can themselves be initiated by biotic 

agents. For example, outbreaks by insect (or microbial) defoliators are often followed by mortality due to bark beetles 

(Wright et al. 1984, Paine and Baker 1993), densely crowded stands more commonly experience bark beetle outbreaks 

(Nebeker et al. 1993), and suppressed trees are less able to resist attack (Safranyik et al. 1975, Raffa and Berryman 1982b, 

Mitchell et al. 1983, Waring and Pitman 1985, Paine and Baker 1993, Preisler and Mitchell 1993). 

Our current understanding of the biotic and abiotic interactions contributing to Red Pine Decline is depicted in Figure 

4. Each of the major assumptions underlying this model has been supported, although some of the mechanisms are only 

poorly understood. In particular, our knowledge of the relative importance, impact, and tolerance ranges of the various 

inciting factors lags behind our knowledge of within-stand spread. Our current understanding is now sufficient, however, to 

test putative natural and anthropogenic stress agents on specific tree physiological, insect behavioral and developmental, and 

microbial developmental parameters, and to project these impacts to the stand level. This model can also serve as a framework 

for expanding our basic understanding of interactions among feeding guilds, the interaction of constitutive and inducible 

allelochemicals in host plant resistance, and microbial mediation of plant-insect interactions. 

.,__,_._ _:_ • ..., _,_...,_...,, :- . :. .. 

..... g g g ..... ..... . 

Figure 4._Sequence of events in the infection, decline, and mortality of plantation red pines. 

218

Implications to Natural Resource Management and Forest Ecosystem Health 

The dynamics of insect and microbial colonization also suggests some specific management tactics that can prevent 

the onset and spread of Red Pine Decline. For example, mixed-block rather than uniform-species planting would reduce the 

transmission of Leptographium fungi through root grafts. This would contain root-diseased trees within relatively small 

groups rather than entire stands. Where economic incentives for red pine make large-scale planting of"alternate species 

impractical, planting buffer strips of alternate species might serve the same purpose. Pinus strobus would probably be a 

better candidate than Pinus banksiana, because the former is less preferred by the root weevils in our region (Hunt et al. 

1993). Judicious thinning would also reduce stress and thereby reduce stand susceptibility. However, the timing is quite 

important. Open light penetration can favor site suitability to root weevils by warming and drying the soil (Wilson and 

Millers 1983). Moreover, thinning should be accompanied by stump removal to avoid leaving a food base for the root- and 

lower-stem feeding species (Witcosky and Hansen 1985, Witcosky et al. 1986). Our model also suggests that severing root 

grafts in advance (@ 15m) of killed trees could halt the transmission of Leptographium. We have begun tests on this idea in 

Wisconsin. We further recommend sanitation removal of dead and living trees (including stumps) within the radius of root 

severing. 

Although the insects and fungi described here are typically viewed as forest pests, these moderately pathogenic 

species might serve as useful bio-indicators. Reproductive success by these organisms is contingent on their ability to detect 

trees that are sufficiently stressed to have impaired defensive capability, but not so severely stressed as to be available to 

saprophytic competitors. As such, the population densities and behavior of moderately virulent endoparasites may be 

indicative of reduced ecosystem vigor associated with anthropogenic disturbances or injudicious forestry practices. Our 

monitoring data and bioassay show this idea is feasible for at least one form of forest decline. That is, populations of some 

root- and lower stem-feeding insects were higher in declining than in asymptomatic stands. Moreover, different species vary 

in how their densities relate to the progression of decline and in how they respond to tree stress physiology. We propose that 

the most useful bioindicator candidates might be those that show increased numbers or behavioral alteration before any 

symptoms become apparent, but are not by themselves major outbreak pests. 

SUMMARY 

Interactions between feeding guilds that affect different plant parts can strongly affect the dynamics of forest declines. 

Environmental stresses can mediate these outcomes, especially against highly generalized plant defense systems. Such 

interactions among insects and fungi are primarily responsible for Red Pine Decline, a syndrome affecting mature plantation 

trees in the Great Lakes region of North America. Root- and lower stem-colonizing insect-fungal complexes reduce tree 

resistance against stem-feeding bark beetle-fungal complexes. Between-stand transmission of moderately pathogenic 

Ophiostomatales fungi occurs by insect vectoring, but within-stand progression also occurs through extensive root grafts in 

plantations. Subterranean scolytids and curculionids show a high degree of niche partitioning based on host tissue and 

condition, chemical ecology, microbial associates, and spatial and temporal distributions. 

Healthy trees can resist colonization by rapidly accumulating terpenes and phenolics to concentrations that repel or 

kill I. pini and inhibit O. ips spore germination and mycelial growth. However, tree constitutive and inducible defenses are 

compromised by root infection. Behavioral assays are generally the most sensitive to the early stages of root injury. Among 

these, post landing assays of I. pini acceptance behavior are most sensitive, especially those in which beetles are exposed to 

host reaction tissue formed in response to challenge inoculation with fungi. Host susceptibility to the root colonizing agents 

can be increased by environmental stresses such as reduced light, defoliation, and nitrification. 

An understanding of the complex interactions underlying forest declines can improve our ability to relate various 

syndromes to management practices, anthropogenic stress, and natural succession. These systems can also serve as useful 

models to improve our overall understanding of plant-insect and insect-fungal interactions. The underlying mechanisms also 

suggest some specific tactics for preventing or ameliorating Red Pine Decline. Marginally pathogenic agents such as the 

organisms described here may also serve as useful indicators of forest ecosystem health. Because their existence is contingent 

on detecting trees under sufficient stress to fully resist attack but not yet available to saprophytic competitors, the 

population densities, behavior, and developmental success of these insects and fungi n-fight be good indicators of incipient 

stress at the stand or tree level. 

219


This work was supported by USDA CSRS Competitive Grant 90-3725P-5581, the University of Wisconsin Graduate 

School, the University of Wisconsin College of Agricultural and Life Sciences, Mclntire Stennis, and Hatch Regional Project 

W-187. We would also like to thank our colleagues L. Rieske, S. Krause, and K. Kleiner )'or input on the ideas discussed in 

this paper. 

REFERENCES CITED 

BERRYMAN, A.A. 1972. Resistance of conifers to invasion by bark beetle-fungus associations. Bio Science 22: 598-602. 

CATES, R.G. and ALEXANDER, H. 1982. Host resistance and susceptibility, p. 212-263. In Mitton, J.B. and Sturgeon, 

K.B., eds. Bark Beetles in North American Conifers: A System for the Study of Evolutionary Biology. University 

Texas Press, Austin. 

CHRISTENSEN, E. and ERICSSON, A. 1986. Starch reserves in Picea abies in relation to defense reaction against a bark 

beetle transmitted blue-stain fungus, Ceratocuptispolonica. Can. J. For. Res. 16: 78-83. 

CHRISTENSEN, E. and HORNTVEDT, R. 1983. Combined lps Ceratocystis attack on Norway spruce and defensive 

mechanisms of the trees. Zeit. angew. Entomol. 96:110-118. 

DUNN, J.E, POTTER, D.A., and KIMMERER, T.W. 1990. Carbohydrate reserves, radial growth, and mechanisms of 

resistance of oak trees to phloem-boring insects. Oecologia 83: 458-468. 

ELKINTON, J.W. and WOOD, D.L. 1980. Feeding and boring behavior of the bark beetle lps paraconfusus (Coleoptera 

Scolytidae) on the bark of a host and non-host tree species. Can. Entomol. 112: 797-809. 

ERICSSON, A., HELLQUIST, C., LANGSTROM, B., LARSSON, S., and TENNOW, O. 1985. Effects on growth of 

simulated and induced shoot pruning by Tomicus piniperda as related to carbohydrate and nitrogen dynamics in Scots 

pine. J. Applied Ecol. 22: 105-124. 

HODGES, J.D., ELAM, W.W., WATSON, W.E, and NEBEKER, T.E. 1979. Oleoresin characteristics and susceptibility for 

four southern pines to southern pine beetle attacks. Can. Entomol. 11: 889-896. 

HOUSTON, D.R. 1981. Stress triggered tree diseases_the diebacks and declines. NE-INF-41-81. Radnor, PA: U.S. 

Department of Agriculture, Forest Service, Northeastern Forest Experiment Station: 4-7. 

HUETTL, R.E, FINK, S., LUTZ, H-J., POTH, M., and WISMEWSKI, J. 1990. Forest decline, nutrient supply and diagnostic 

fertilization in Southwestern Germany and in Southern California. For. Ecol. and Manage. 30: 341-350. 

HUNT, D.W.A., LINTEREUR, G., SALOM, S., and RAFFA, K.E 1993. Performance and preference of Hylobius radicis 

Buchanan, and H. pales (Herbst) (Coleoptera:Curculionidae)on various Pinus species. Can. Entomol. 125:1003- 

1010. 

KIMMERER, T.W. and KOZLOWSKI, T.T. 1982. Ethylene, ethane, acetyl aldehyde, and ethanol production by plants 

under stress. Plant Physiol. 69: 840-847. 

KLEIN, R.M. and PERKINS, T.D. 1988. Primary and secondary causes and consequences of contemporary forest decline. 

The Bot. Rev. 54: 1-43. 

KLEPZIG, K.D., RAFFA, K.E, and SMALLEY, E.B. 1991. Association of insect-fungal complexes with Red Pine Decline 

in Wisconsin. For. Sci. 37: 1119-1139. 

220

KLEPZIG, K.D., RAFFA, K.F., and SMALLEY, E.B. 1994a. Reaction of red pine to inoculation with isolates of the fungi 

associated with Red Pine Decline Disease. In prep_ 

KLEPZIG, K.D., RAFFA, K.F., and SMALLEY, E.B. 1994b. Dendroctonus valens and Hylastes porculus: vectors of 

pathogenic fungi (Ophiostomatales) associated with Red Pine Decline Disease. Subm. to Environ. Entomol. 

KLEPZlG, K.D., KRUGER, E.L., SMALLEY, E.B., and RAFFA, K.F. 1994c. Monoterpenes and phenolics associated with 

the response of healthy and diseased red pine to bark beetle vectored fungi. In prep. 

KLEPZlG, K.D., RIBA, J., SMALLEY, E.B., and RAFFA, K.F. 1994d. Effects of red pine phenolics and monoterpenes on a 

subcortical insect - fungal complex associated with a decline disease. In prep. 

KRAUSE, S. and RAFFA, K.F. 1994. Effects of long-term defoliation on insect herbivores of red pine. In prep. 

KRAUSE, S.C., RAFFA, K.F., and WAGNER, M.R. 1993. Tree response to stress: a role in sawfly outbreaks? p. 211-227. 

In Wagner, M.R. and Raffa, K.F., eds. Sawfly Life History Adaptations to Woody Plants. Academic Press, New 


LIEUTIER, F. and BERRYMAN, A.A. 1988. Elicitation of defensive reactions in conifers, p. 313-320. In Mattson, W.J., 

Levieux, J., and Bernard-Dagan, C., eds. Mechanisms of Woody Plant Defenses Against Insects: Search for Pattern. 

Springer-Verlag, New York. 

LORIO, P.L., JR. 1986. Growth-differentiation balance: a basis for understanding southern pine beetle-tree interactions. 

For. Ecol. & Manage. 14: 259-273. 

LORIO, RL., JR. 1993. Environmental stress and whole-tree physiology, p. 81-101. In Schowalter, R.D. and Filip, G.M., 

eds. Beetle-Pathogen Interactions in Conifer Forests. Academic Press, New York. 

LORIO, RL., JR. and SUMMERS, R.A. 1986. Evidence of competition for photosynthates between growth processes and 

oleoresin synthesis in Pinus taeda. Tree Physiol. 2: 301-306. 

MANION, RD. 1991. Decline diseases of complex biotic and abiotic origin, p. 324-339. In Manion, RD., ed. Tree Disease 

Concepts. Prentice Hall, Englewood Cliffs, NJ. 

MILLER, R.H. and BERRYMAN, A.A. 1986. Carbohydrate allocation and mountain pine beetle attack in girdled lodgepole 

pines. Can. J. For. Res. 16: 1036-1040. 

MILLER, R.H., WHITNEY, H.S., and BERRYMAN, A.A. 1986. Effects of induced translocation stress and bark beetle 

attack (Dendroctonus ponderosae) on heat pulse velocity and the dynamic wound response of lodgepole pine (Pinus 

contorta var. latifolia). Can. J. Bot. 64: 2669-2674. 

MITCHELL, R.G., WARING, R.H., and PITMAN, G.B. 1983. Thinning lodgepole pine increases tree vigor and resistance 

to mountain pine beetle. For. Sci. 29: 204-211. 

MUELLER-DOMBOIS, D. 1988. Forest decline and die backmA global ecological problem. Trends in Ecol. and Evol. 3: 

310-312. 

NEBEKER, T.E., HODGES, J.D., and BLANCHE, C.A. 1993. Host response to bark beetle and pathogen colonization, p. 

157-178. In Schowalter, R.D. and Filip, G.M., eds. Beetle-Pathogen Interactions in Conifer Forests. Academic 

Press, New York. 

PAINE, T.D., and BAKER, F.A. 1993. Abiotic and biotic predisposition, p. 61-79. In Schowalter, R.D. and Filip, G.M., eds. 

Beetle-Pathogen Interactions in Conifer Forests. Academic Press, New York. 

221

PAINE, T.D. and STEPHEN, F.M.. 1987a. The relationship of tree height and crown class to the induced plant defenses of 

loblolly pine. Can. J. Bot. 65: 2090-2092. 

PAINE, T.D. and STEPHEN, F.M. 1987b. Influence of tree stresses and site quality on the induced defensive system of 

loblolly pine. Can. J. For. Res. 17: 569-571. 

PEET, R.K. and CHRISTENSEN, N.L. 1987. Competition and tree death. BioScience 37::587-595. 

PREISLER, H.K. and MITCHELL, R.G. 1993. Colonization patterns of the mountain pine beetle in thinned and unthinned 

lodgepole pine stands. For. Sci. 39 528-545. 

RAFFA, K.F. 1988. Host orientation behavior of Dendroctonusponderosae: Integration of token stimuli and host defenses, 

p. 369-390. In Mattson, W.J., Levieux, J., and Bernard-Dagan, C., eds. Mechanisms of Woody Plant Defenses 

Against Insects: Search for Pattern. Springer-Verlag, New York. 

RAFFA, K.F. and BERRYMAN, A.A. 1982a. Physiological differences between lodgepole pines resistant and susceptible 

to the mountain pine beetle and associated microorganisms. Environ. Entomol. 11: 486-492. 

RAFFA, K.F. and BERRYMAN, A.A. 1982b. Accumulation of monoterpenes and associated volatiles following fungal 

inoculation of grand fir with a fungus transmitted by the fir engraver, Scolytus ventralis (Coleoptera: Scolytidae). 

Can. Entomol. 114:797-810. 

RAFFA, K.F. and BERRYMAN, A.A. 1982c. Gu_tatory cues in the orientation of Dendroctonus ponderosae 

(Coleoptera:Scolytidae) to host trees. Can. Entomol. 114: 97-104. 

RAFFA, K.F. and BERRYMAN, A.A. 1983a. Physiological aspects of lodgepole pine wound responses to a fungal symbiont 

of the mountain pine beetle. Can. Entomol. ! 15: 723-734. 

RAFFA, K.F. and BERRYMAN, A.A. 1983b. The role of host plant resistance in the colonization behavior and ecology of 

bark beetles. Ecol. Monogr. 53: 27-49. 

RAFFA, K.F. and BERRYMAN, A.A. 1987. Interacting selective pressures in conifer-bark beetle systems" A basis for 

reciprocal adaptations? Am. Nat. 129: 234-262. 

RAFFA, K.F. and HUNT, D.W. 1989. Microsite and interspecific interactions affecting emergence of root-infesting pine 

weevils (Coleoptera: Curculionidae) in Wisconsin. Ann. Entomoi. Soc. Am. 82: 438-445. 

RAFFA, K.F. and KLEPZIG, K.D. 1992. Tree defense mechanisms against insect-vectored fungi, p. 354-390. In 

Blanchette, R.A. and Biggs, A.R., eds. Defense Mechanisms of Woody Plants against Fungi. Springer-Verlag, New 


RAFFA, K.F. and SMALLEY, E.B. 1988a. Host resistance to invasion by lower stem and root infesting insects of pine: 

Response to controlled inoculations with the fungal associate Leptographium terebrantis. Can. J. For. Res. 18" 675- 

681. 

RAFFA, K.F. and SMALLEY, E.B. 1988b. Response of red and jack pines to inoculation with microbial associates of the 

pine engraver, lps pini. Can. J. For. Res. 18:581-586. 

RAFFA, K.F. and SMALLEY, E.B. 1988c. Seasonal and long-term responses of host trees to microbial associates of the 

pine engraver, Ips pini. Can. J. For. Res. 18: 1624-1634. 

RAFFA, K.F. and SMALLEY, E.B. 1994. Induction of red pine subcortical monoterpenes to insecticidal levels by fungal 

associates of bark beetles. Subm. to Oecologia. 

222

RAFFA, K.F., PHILLIPS, T., and SALOM, S. 1993. Strategies and mechanisms of host colonization by bark beetles, p. 103- 

128. In Schowalter, T.O. and Filip, G., eds. Beetle-Pathogen Interactions in Conifer Forests. Academic Press, New 


REICH, RB., WALTERS, M.B., KRAUSE, S.C., VANDERKLEIN, D.W., RAFFA, K.F., and TIBONE, T. 1992. Gas 

exchange and growth of red pine seedlings and trees one year after defoliation. Trees - Structure and Function. 7: 67- 

77. 

RIESKE, L.K. and RAFFA, K.F. 1990. Dispersal patterns and mark-recapture estimates of two pine root weevil species 

Hylobius pales and Pachylobius picivorus (Coleoptera: Curculionidae), in Christmas tree plantations. Environ. 

Entomol. 19: 1829-1836. 

RIESKE, L.K. and RAFFA, K.F. 1991. Effect of varying ethanol and turpentine levels on attraction of two pine root weevil 

species Hylobius pales and Pachylobius picivorus (Coleoptera: Curculionidae). Environ. Entomol. 20: 48-52. 

SAFRANYIK, L., SHRIMPTON, D.M., and WHITNEY, H.S. 1975. An interpretation of the interaction between lodgepole 

pine, the mountain pine beetle and its associated blue stain fungi in western Canada, p. 406-428. In Baumgartner, 

D.M., ed. Management of Lodgepole Pine Ecosystems. WSU Coop. Ext. Serv. 

SAXE, H. 1993. Triggering and predisposing factors in the "Red" decline syndrome of Norway spruce (Picea abies). Trees 

8: 39-48. 

WARING, R.H. 1987. Characteristics of trees predisposed to die. BioScience 37: 569-574. 

WARING, R.H. and PITMAN, G.B. 1985. Modifying lodgepole pine stands to change susceptibility to mountain pine beetle 

attacks. Ecology 66: 889-897. 

WILSON, L.F. and MILLERS, I. 1983. Pine root collar weevil: Its ecology and management. Tech. Bull. 1675. Washington, 

DC: U.S. Department of Agriculture, Forest Service. 

WITCOSKY, J.J. and HANSEN, E.M. 1985. Root-colonizing insects associated with Douglas-fir in various stages of 

decline due to black-stain root disease. Phytopathol. 75: 399-402. 

WlTCOSKY, J.J., SCHOWALTER, T.D., and HANSEN, E.M. 1986. The influence of time of precommercial thinning on 

the colonization success of Douglas-fir by three species of root-colonizing insects. Can. J. For. Res. 16: 745-749. 

WOOD, D.L. 1972. Selection and colonization of ponderosa pine by bark beetles. Symposium of the Royal Entomol. Soc. 

London. 6:110-117. 

WOODCOCK, H., PATTERSON, W.A. III, and DAVIES, EM., JR. 1993. The relationship between site factors and white 

ask (Fraximus americana L.) decline in Massachusetts. For. Ecol. & Manag. 60:271-290. 

WRIGHT, L.E., BERRYMAN, A.A., and GURUSIDDIAH, S. 1979. Host resistance to the fir engraver beetle, Scolytus 

ventralis (Coleoptera:Scolytidae). 4. Effect of defoliation on wound monoterpenes and inner bark carbohydrate 

concentrations. Can. Entomol. 111: 1255-1261. 

WRIGHT, L.A., BERRYMAN, A.A., and WICKMAN, B.E. 1984. Abundance of the fir engraver, Scolytus ventralis, and 

the Douglas-fir beetle, Dendroctonus pseudotsugae, following tree defoliation by the Douglas-fir tussock moth, 

Orgyria pseudotsugata. Can. Entomol. 116: 293-305. 

223

DOUGLAS-FIR AND WESTERN LARCH DEFENSIVE REACTIONS TO 

LEPTOGRAPHIUM ABIETINUM AND OPHIOSTOMA PSEUDOTSUGAE 

DARRELL W. ROSS 1and HALVOR SOLHEIM 2 

_Department of Forest Science, Oregon State University, Corvallis, Oregon, 97331-7501, USA 

2Norwegian Forest Research Institute, Section of Forest Ecology, Division of Forest Pathology, N-1423 As-NLH, Norway 

INTRODUCTION 

The Douglas-fir beetle, Dendroctonus pseudotsugae Hopkins, is one of the most important insects associated with 

Douglas-fir, Psuedotsugae menziesii (Mirb.) Franco, in western North America (Furniss and Carolin 1977). At high population 

densities, the beetle is capable of causing significant tree mortality and impacting forest resource values (Cornelius 1955, 

Rudinsky 1966, Johnson and Belluschi 1969, Furniss and Orr 1978, Furniss et al. 1979). The Douglas-fir beetle usually 

breeds in recently dead or live Douglas-fir trees. The only other tree species in which the beetle has been reported to successfully 

breed is western larch (Ross 1967). However, the beetle is apparently able to breed only in dead larch (Furniss et al. 

1981). Live larch may be attacked, but brood production from live trees has never been observed. Reed et al. (1986) 

compared various chemical and physical properties of Douglas-fir and western larch to identify possible reasons that brood 

fail to survive in live larch trees. Compared with Douglas-fir, larch had a higher content of 3-carene, thinner phloem, higher 

phloem moisture content, larger diameter vertical resin ducts, and lower oleoresin exudation pressure. The authors speculated 

that the high content of 3-carene in live larch may be responsible for the failure of Douglas-fir beetle brood in those 

trees. 

One or more species of ophiostomatoid fungi are invariably associated with each species of conifer-infesting bark 

beetle (Whitney 1982, Harrington 1988). Although the ecological relationships among the fungi and bark beetles are not 

completely understood in many cases, there is evidence that some of these fungi help the beetles to overcome the natural 

defenses of mass attacked trees. Conifer response to invasion by bark beetles and their associated fungi involves two 

different systems: (1) the flow of preformed resin from severed resin ducts and (2) the induced wound response or resistant 

reaction (Reid et al. 1967, Berryman 1972, Christiansen and Horntvedt 1983). The dimensions of the induced response in 

host tree tissues are generally correlated to the growth of the fungi or the extent of mechanical damage associated with beetle 

tunneling (Ross et al. 1992, Lieutier 1993, Solheim 1993). If the fungi associated with the Douglas-fir beetle are involved in 

overcoming tree resistance, then a possible reason for the failure of brood development in larch may be the inability of the 

fungi to survive and grow in larch. A significant variation in the pathogenicity of the fungi to Douglas-fir and western larch 

should he reflected by differences in the induced response following artificial inoculations. 

Despite a considerable amount of past research related to the biology and management of the Douglas-fir beetle, little 

is known about its funsal associates. The two species of ophiostomatoid fungi most commonly associated with the Douglasfir 

beetle are Ophiostoma pseudotsugae (Rumb.) yon Arx and Leptographium abietinum (Peck) Wingf. (Rumbold 1936, 

Harrington 1988, Solheim unpubl.). There are no published reports on the pathogenicity of these fungi or their ecological 

relationships to the beetle. In 1993, we installed a study to evaluate the pathogenicity of these fungi to Douglas-fir. The final 

results of that study will be published elsewhere in the near future. At the same time, we installed a small test to assess the 

response of Douglas-fir and western larch, Larix occidentalis Nutt., to artificial inoculations with these fungi. The objective 

of our study was to measure the length of the induced response in the phloem of Douglas-fir and western larch following 

artificial inoculation with O. pseudotsugae and L. abietinum. This paper reports the results of this test. 

Mattson, W.J., Niemela, E, and Rouse, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 

For. Serv. Gen. Tech. Rep. NC-183, N.C. For. Exp. Sta., St. Paul, MN 55108. i 

224

METHODS 

The study was installed in a mixed conifer stand (elevation - 1460 m) in northeastern Oregon. On June 29, 1993, 

seven Douglas-fir and seven western larch were selected from an area of less than 1 ha. The trees were chosen based on 

similarities in size and lack of any visible injuries or disease. Three inoculation sites were established equidistant around the 

circumference of each tree at dbh (1.4 m height). The three treatments that were randomly assigned to the inoculation sites 

were: (1) sterile wound, (2) L. abietinum, and (3) O. pseudotsugae. At each inoculation site, a 1.2-cm diameter hole was 

drilled to the sapwood surface with a sterile bit. For the sterile wound, a cotton plug was placed in the hole that was then 

covered with duct tape. For the fungal inoculations, a small agar plug containing the appropriate fungus was placed in the 

hole that was then similarly covered. The fungi used in the study had been isolated from Douglas-fir beetles collected in 

British Columbia, Canada (Solheim unpubl.). 

On August 25, 1993, the outer bark covering each inoculation site was removed and the length of the necrotic lesion 

on the phloem surface was measured. 

The mean dbh for trees of each species were compared by a t-test. Lesion length was subjected to analysis of 

variance for a split-plot design with tree species representing whole plots and fungal inoculation representing subplots. 

RESULTS 

There was no significant difference (P < 0.19) in dbh between Douglas-fir (mean = 51.0 cm) and western larch (mean 

= 48.8 cm). The tree species x treatment interaction was significant (F = 3.29; P < 0.05) for lesion length. For both tree 

species, inoculation with either fungus resulted in a larger lesion than that produced by a sterile wound (Fig. 1). The length 

of the lesions induced by a sterile wound or inoculation with O. pseudotsugae did not differ between tree species, but the 

lesion induced by inoculation with L. abietinum was significantly shorter in western larch than in Douglas-fir (Fig. 1). 

DISCUSSION 

There were no obvious differences in the general appearance of lesions (i.e., intensity of resin accumulation, radial 

growth, etc.) between tree species, so lesion length should provide a relatively accurate indication of fungal growth within the 

phloem of the host trees. The smaller lesions produced by western larch in response to L. abietinum may reflect a more rapid 

and effective defensive reaction to the fungus. These results suggest that L. abietinum is not as well adapted to utilize 

western larch as a host compared with Douglas-fir. If L. abietinum is critical to the survival of Douglas-fir beetle brood in 

live host trees, then this may help to explain the complete mortality of brood in western larch. In contrast, there was no 

difference in the induced response to O. pseudotsugae between the two tree species. Our results must be interpreted cautiously 

since we have only measured lesion length in phloem tissue. Lesion length is not always directly correlated to 

pathogenicity (Solheim 1988). Fungal growth and tree response may be quite different in the xylem compared with the 

phloem. However, the difference in response to fungal inoculations that we observed suggests that further research on the 

host tree-fungus interactions in this system would be worthwhile. The significance of these results will only become apparent 

when we learn more about the role that these fungi play in the population dynamics of the Douglas-fir beetle. 

SUMMARY 

There was no difference in lesion length at the phloem surface between Douglas-fir and western larch following 

sterile wounds or inoculation with O. pseudotsugae. However, lesion length following inoculation with L. abietinurn was 

significantly shorter in western larch compared with Douglas-fir. These results may help to explain the lack of Douglas-fir 

beetle brood survival in western larch. 

225

5O 

40- 

Douglas-fir a 

a 

Western larch a 

E 

E b 

30- 

J:E 

E 

"-- C 

E 20o 

(_ 

_J 

10- 

0 

Wound L. ab/et/num O. pseudotsugae 

Treatment 

Figure 1.--Mean length of the induced lesion at the phloem surface following sterile wounding or fungal inoculation in 

Douglas-fir and western larch (n=7). Bars with the same letter are not significantly different (P = 0.05). 


BERRYMAN, A.A. 1972. Resistance of conifers to invasion of bark beetle-fungus associations. BioScience 22: 598-602. 

CHRISTIANSEN, E. and HORNTVEDT, R. 1983. Combined Ips/Ceratocystis attack on Norway spruce, and defensive 

mechanisms of the trees. Z. Ang. Entomol. 96:110-118. 

CORNELIUS, R.O. 1955. How forest pests upset management plans in the Douglas-fir region. J. For. 53:71 I-713. 

FURNISS, M.M. and ORR, P.W. 1978. Douglas-fir beetle. For. Insect & Dis. Leafl. 5. Washington, DC" U.S. Department 

of Agriculture, Forest Service. 

FURNISS, M.M., MCGREGOR, M.D., FOILES, M.W., and PARTRIDGE, A.D. 1979. Chronology and characteristics of a 

Douglas-fir beetle outbreak in northern Idaho. Gen. Tech. Rep. INT-59. Ogden UT: U.S. Department of Agriculture, 

Forest Service. 

226 

: ii

FURNISS, M.M., LIVINGSTON, R.L., and MCGREGOR, M.D. 1981. Development of a stand susceptibility classification 

for Douglas-fir beetle, p. 115-128. In Hedden, R.L., Barras, S.J., and Coster, J.E., tech. coords. Hazard-Rating 

Systems in Forest Insect Pest Management: Symposium Proceedings. Gen. Tech. Rep. WO-27. Washington, DC: 

U.S. Department of Agriculture, Forest Service. 

FURNISS, R.L. and CAROLIN, V.M. 1977. Western forest insects. Misc. Publ. 1339. Washington, DC: U.S. Department 


HARRINGTON, T.C. 1988. Leptograhium species, their distributions, hosts and insect vectors, p. 1-40. In Harrington, T.C. 

and F.W. Cobb, Jr., eds. Leptographium Root Diseases on Conifers. American Phytopathological Society Press, St. 

Paul, MN. 

JOHNSON, N.E. and BELLUSCHI, RG. 1969. ttost-finding behavior of the Douglas-fir beetle. J. For. 67: 290-295. 

LIEUTIER, E 1993. Induced defence reaction of conifers to bark beetles and their associated Ophiostoma species, p. 225- 

234. In Wingfield, M.J., Seifert, K.A., and Webber, J.E, eds. Ceratocystis and Ophiostoma Taxonomy, Ecology, and 

Pathogenicity. American Phytopathological Society Press, St. Paul, MN. 

REED, A.N., HANOVER, J.W., and FURNISS, M.M. 1986. Douglas-fir and western larch: chemical and physical properties 

in relation to Douglas-fir bark beetle attack. Tree Physiol. 1: 277-287. 


ponderosae Hopkins and blue stain fungi. Can. J. Bot. 45:1115-1126. 

ROSS, D.A. 1967. Wood- and bark-feeding Coleoptera of felled western larch in British Columbia. J. Entomol. Soc. Brit .... 

Columbia 64: 23-24. 

ROSS, D.W., FENN, R, and STEPHEN, F.M. 1992. Growth of southern pine beetle associated fungi in relation to the 

induced wound response in loblolly pine. Can. J. For. Res. 22:1851-1859. 

RUDINSKY, J.A. 1966. Host selection and invasion by the Douglas-fir beetle, Dendroctonus pseudotsugae Hopkins, in 

coastal Douglas-fir forests. Can. Entomol. 98: 98-111. 

RUMBOLD, C.T. 1936. Three blue-staining fungi, including two new species, associated with bark beetles. J. Agric. Res. 

52: 419-437. 

SOLHEIM, H. 1988. Pathogenicity of some Ips typographus-associated blue-stain fungi to Norway spruce. Medd. Nor. 

inst. skogforsk. 40(14): 1-11. 

SOLHEIM, H. 1993. Ecological aspects of fungi associated with the spruce beetle Ips typographus in Norway, p. 235-242. 

In Wingfield, M.J., Seifert, K.A., and Webber, J.F., eds. Ceratocystis and Ophiostoma Taxonomy, Ecology, and 

Pathogenicity. American Phytopathological Society Press, St. Paul, MN. 

WHITNEY, H.S. 1982. Relationships between bark beetles and symbiotic organisms, p. 183-211. In Mitton, J.B. and 

Sturgeon, K.B., eds. Bark Beetles in North American Conifers. University of Texas Press, Austin, TX. 

227

INTERRUPTION OF BARK BEETLE AGGREGATION BY A 

VIGOR-DEPENDENT PINUS HOST COMPOUND 

KENNETH R. HOBSON 

Utah State University, Department of Forest Resources and USDA Forest Service, Forestry Sciences Laboratroy, 

Logan, Utah 84322, USA 

INTRODUCTION 

The attraction of bark beetles to pheromones has been abundantly demonstrated in the last quarter century (Wood 

1982, Borden 1985). Much less is known about the response of bark beetles to host volatiles. This paper reports preliminary 

evidence that one common host compound of North American conifers is a strong repellent or interruptant of aggregation for 

several species of bark beetles. It also reviews work that suggests that the level of this compound may be an indicator of host 

tree stress typically associated with bark beetle infestation in several species of pines. This compound, methyl chavicol, 

fulfills the two essential requirements for a kairomonal cue that can explain the selective infestation of susceptible trees by 

bark beetles, lit is correlated to tree susceptibility, and bark beetles respond to it. 

The search for a biochemical indicator of host tree stress that might serve as an olfactory cue for bark beetles is not 

new (Rudinsky 1962 and references therein, Cobb et al. 1968). Miller et al. (1968) examined the resin from smog-damaged 

ponderosa pines, Pinus ponderosa, in the San Bernardino Mountains east of Los Angeles. These trees were being selectively 

infested by mountain pine beetle, Dendroctonus ponderosae, and western pine beetle, Dendroctonus brevicomis. Of the six 

monoterpenes examined, none were shown to be significantly different between damaged and healthy trees. However, Cobb 

et al. (1972) found a strong, significant difference in the level of one compound, methyl chavicol, in the foliage of smogdamaged 

ponderosa pines and healthy trees. Smog-damaged trees had 71% less methyl chavicol in their foliage - by far the 

sharpest phytochemical difference found. 

More recently Nebeker et al. (1995) measured 18 host volatiles from healthy lodgepole, Pinus contorta, and those 

diseased by dwarf mistletoe, Arceuthobium americanum; armillaria root disease, Armillaria mellea; or comandra blister rust, 

Cronartium comandrae. The latter two diseases are reported to be the most common predisposing agents of lodgepole to 

mountain pine beetle attack in the Intermountain West (Tkacz and Schmitz 1986). In the case of both predisposing diseases 

of lodgepole, there was a large and significant difference in the level of 4-allylanisole (a synonym for methyl chavicol) 

between healthy and diseased trees. This was not true for dwarf mistletoe, which was not shown to be associated with bark 

beetle infestation (Tkacz and Schmitz 1986). For both comandra and armillaria, the difference in the level of methyl chavicol 

was one of the greatest phytochemical differences found between diseased and healthy trees. Trees infected with comandra 

had 43.6% less methyl chavicol than healthy trees; trees infected with armillaria had 63% less methyl chavicol than healthy 

trees. Cobb et al. and Nebeker et al. both found the strongest biochemical differences between diseased and healthy trees in 

one compound. 

These results begged the question, "What is the response of bark beetles to methyl chavicol?" Electroantennograms 

(EAG) show a strong response to methyl chavicol. In 1989, Peter White found that estragole (methyl chavicol) produced the 

third highest response of all 11 host compounds tested on the red turpentine beetle, Dendroctonus valens (White and Hobson 

1993). In 1993 Armand Whitehead showed a strong EAG response of mountain pine beetle to methyl chavicol (Hobson et al. 

in prep). A large EAG response does not indicate that a compound will be a strong attractant or repellent. However, ecologically 

relevant compounds that produce a large EAG are likely to have behavioral significance (Masson and Mustaparta 

1990). These data encouraged us to conduct behavioral field studies. 



228

METHODS 

In the spring and summer of 1993, we tested the response of five species of bark beetles to methyl chavicol in 

California and Idaho. In April and May, we tested western pine beetle, red turpentine beetle, and two species of Ips, L 

paraconfusus and I. pini, at the University of California's Blodgett Forest Research Station in the central Sierra Nevada. 

Tests were conducted as described previously (Hobson 1992, Hobson et al. 1993). For each beetle species, attractive lures 

were placed on lindgren traps. In each test there were four treatments" (1) attractant, (2) attractant and methyl chavicol, (3) 

methyl chavicol alone, and (4) control blank. 

Beetles were collected daily and treatments were randomized after each collection. All four treatments were replicated 

in 4-10 blocks. Methyl chavicol was released from four open 1.5-ml ependorf tubes at a rate of 0.5 ml/day. The 

attractant for D. valens was a set with four 1.5-ml ependorf tubes of S(-)[3-pinene. For the western pine beetle, I. pini and I. 

paraconfusus commercial lures provided by Phero Tech (Delta, BC) were used. Western pine beetle lures contained exobrevicomin, 

frontalin and myrcene. I. pini lures contained ipsdienol (3%+/97%-). I. paraconfusus lures were ipsenol (50%+/ 

50%-), ipsdienol (97%+/3%-), and cis-verbenol. 

In July and August of 1993, we tested mountain pine beetle and I. pini in the Sawtooth National Recreation Area in 

south central Idaho. The experimental design was as above. Attractant lures for mountain pine beetle (trans-verbenol, exobrevicomin 

and myrcene) were obtained from Phero Tech. Attractant lures for I. pini were as above. 

RESULTS 

Methyl chavicol significantly reduced attraction of western pine beetle and mountain pine beetle to their respective 

aggregation pheromones (Table 1). Attraction of D. valens to [3-pinene was not significantly interrupted by methyl chavicol. 

L pini catch was reduced by 29%, but this was not significant. I. paraconfusus flew too late in California for us to collect 

adequate data for statistical analysis; however, in five of six blocks where I. paraconfusus was collected, traps baited with 

methyl chavicol and the bait had fewer beetles than traps with bait alone. 

Table 1.--Bark beetle response to Methyl chavicol 

Pheromone Pheromone & Percent 

baifi SD methyl chavicol _ SD reduction 

Western pine beetle 53.5 a2 69.7 21.6b 28.0 60 

Mountain pine beetle 27.4c 47.5 7.9d 12.4 71 

Red turpenti ne beetle 10.6 6.28 9.06 6.8 14 

Ips pini 20.5 19.7 14.5 12.2 29 

_Meancatch. 

2Numbersin a row followed by different letters are significantly different o_= 0.05 ANOVA followed by Wilcoxon test. 

229

Where beetle response to methyl chavicol was compared by sex, mountain pine beetle males and females were both 

significantly less attracted to lures with methyl chavicol. The data for western pine beetle were not sufficient to statistically 

test gender-specific response, but methyl chavicol reduced the catch of males by 53% and females by 66%. Methyl chavicol 

reduced the catch of D. valens females by 21% and of males by 11%, but neither of these reductions was significant. 

DISCUSSION 

The aggregation of the two most aggressive bark beetles in this study, the mountain and western pine beetles, was 

strongly interrupted. The less aggressive species, D. valens, L pini, and I. paraconfusus were not so strongly affected. 

Subsequent testing of the L paraconfusus with methyl chavicol and a larger sample size produced a consistent 40% 

interruption of aggregation (Storer, pers. comm.). The aggregation of two other aggressive bark beetles in Alaska, D. 

rufipennis and D. simplex, were interrupted by 82% and 73%, respectively (Werner 1995). Haack and Lawrence have tested 

methyl chavicol in traps with attractants of Tomicus piniperda (Haack, pers. comm.). Subsequent to the presentation of this 

paper, Hayes and co-workers published similar results with Dendroctonusfrontalis (Hayes et al. 1994 a, b) showing 37% and 

56% reduction in catch with methyl chavicol (4-allylanisole). Hayes and Strom (1994) also confirmed our results with 

mountain pine beetle, obtaining a 77% reduction in catch with methyl chavicol. Their work differed from this work in that 

they could not obtain significant interruption of attraction for western pine beetle, and they did find interruption of attraction 

at the 43% level for a Wisconsin population of I. pini. 

Methyl Chavicol 

Methyl chavicol, an aromatic ether or phenylpropanoid, is also commonly known as 4-allylanisole, tarragon, or 

estragol. (Other chemical synonyms include: isoanethole, p-allylmethoxybenzene, 4-allyl-l-methoxybenzene, chavicol 

methyl ether, esdragon and 1-methoxy-4-2(2-propenyl)benzene) (Aldrich Chemical - Material Data Safety Sheet). It is 

widespread in the oleoresin of new world pines in the subgenus Pinus, occurring in P ponderosa, P. taeda, P palustris, P. 

elliottii, P patula, P jeffreyi, P. tenufolia, P. hartwegii, P michoacana P. lumholtzii (Mirov1961 and references therein), and 

P. caribaea (Smith, R.M. 1975). Among old world pines it occurs in P. sylvestris and P nigra (Bardysev et al. 1970). It is 

abundant in the foliage of ponderosa pine (23%) and present in the foliage of lodgepole and digger pines (Mirov 1961). It is 

also found in 12 or more species of spruce (Zavarin, pers. comm.). 

Methyl chavicol is the principal volatile ingredient or a major volatile of several strongly scented herbs: tarragon, 

Artemisia dracunculus; fennel, Foeniculum vulgare; star anise, Illicium verum; basil, Ocium basilicum; and cloves, Szygium 

aromaticum (Duke 1985). Methyl chavicol is also known from the leaves of three rutaceous plants, west african zigua, 

Clausena anisata, where it is the major component of the oil of leaves used to repel mosquitoes (Okunade and Olaifa 1987); 

7_nthoxylum spp., where it is toxic to Dacus eggs (Marr and Tang 1992); and southeast asian wood apple, Feronia limonia 

(Ahmad et al. 1989). In addition, it is found in oil palm, Elaeis guineensis, where it is attractive to the weevil Elaeidobius 

kamerunicus (Hussein et al. 1990). In recent plant herbivore studies, it is most well known from the work of Metcalf and his 

co-workers. They have found it in the flowers of Cucurbita maxima, where it is attractive to Diabrotica spp., corn rootworms. 

This association is the basis of a productive research effort in biorational management of Diabrotica (Metcalf and 

Lampman 1989). 

Biosynthetically methyl chavicol is derived from the shikimic acid pathway (Zavarin et al. 1971). This pathway is 

disrupted by the herbicide glyphosate (Roundup, Monsanto) (Borden, pers. comm.). Interestingly, when Bergvinson and 

Borden (1991) applied glyphosphate to lodgepole pine, they found that treated trees were readily colonized by mountain pine 

beetles. They concluded that the herbicide had inhibited the treated trees' secondary defense response against the beetle's 

symbiotic fungi. This is consistent with the hypothesis that the success of mountain pine beetles in glyphosate-treated trees 

is, in part, due to a drop in methyl chavicol. Bridges (1987) found methyl chavicol was the most inhibitory host compound to 

the three symbiotic fungi associated with D. frontalis, suggesting that it may be an important defensive mechanism of loblolly 

pine against southern pine beetle and its vectored fungi. Hayes et al. (1994a) found that methyl chavicol was strongly 

reduced in southern pines following wounding and treatment with metham-sodium and dimethyl sulfoxide. Treated trees 

were attacked by D. frontalis after the level of natural methyl chavicol dropped. Methyl chavicol tested subsequently in field 

bioassays strongly reduced attraction of southern pine beetle to pheromone-baited traps. 

230

The importance of a compound as a repellent is somewhat novel in bark beetle host selection research. Limonene 

and other terpenes that kill bark beetles and their vectored fungi have been thought to have a mainly short-range effect as 

feeding deterrents (Smith, R.H. 1975, Coyne and Lott 1976). The majority of work in this field, however, has looked for 

compounds that increase with stress and are attractive. Few studies have looked for compounds that decrease with stress and 

are repellent. 

SUMMARY 

The significance of these results rests in three principal areas" 

1) The large difference in methyl chavicol between healthy and diseased trees provides a clear olfactory sign of stress 

for bark beetles, far clearer than has been demonstrated for any monoterpene to date. 

2) The stresses that produce the large drop in methyl chavicol are the same stresses that are associated with bark 

beetle infestation, suggesting that methyl chavicol is one of the biochemical links between ecology and beetle 

behavior. 

3) Methyl chavicol reduces the arrestment and landing of bark beetles in the field sufficiently strongly to suggest 

management applications for several of our most important bark beetle species with a naturally occurring, abundant 

host compound. 

Methyl chavicol is the best candidate so far for a compound that responds to stress and to which beetles respond. It 

may serve as an indicator and reveal, via biochemical linkage, the elements of host defenses that most strongly affect bark 

beetle success. Further investigation of other host species applying stress of the sort known to favor beetle colonization may 

determine some new insights for our understanding of the biochemical basis of conifer defenses. Ultimately, we may 

understand how environmental stresses that favor beetle colonization produce the characteristics of susceptibility which 

beetles favor. We may be able to provide the biochemical linkage between ecological conditions and host selection. 


AHMAD, A., MISRA, L.N., and THAKUR, R.S. 1989. Composition of the volatile oil from Feronia limonia leaves. 

Planta-Medica 55" 199-200. 

BARDYSEV, I.I., PAPANOV, G.J., and PERCOVSKIJ, A.L. 1970. Chemical composition of the balsams ofPinus sylvestris 

and P. nigra growing in Bulgaria. Dokl. AN BSSR 14: 539-40. 

BERGVINSON, D.J. and BORDEN, J.H. 1991. Glyphosate-induced changes in the attack success and development of the 

mountain pine beetle and impact of its natural enemies. Entomologia Experimentalis et Applicata. 60:203-212. 

BORDEN, J.H. 1985. Aggregation pheromones. In Kerkut, G.A. and Gilbert, L.I., eds. Comprehensive Insect Physiology, 

Biochemistry & Pharmacology. Pergamon Press, Oxford. 

BRIDGES, J.B. 1987. Effects of terpenoid compounds on growth of symbiotic fungi associated with the southern pine 

beetle. Phytopathol. 77: 83-85. 

COBB, EW. Jr., WOOD, D.L., STARK, R.W., and MILLER, P.R. 1968. Theory on the relationships between oxidant injury 

and bark beetle infestation. Hilgardia 39: 141-152. 

COBB, EW. Jr., ZAVARIN, E., and BERGOT, J. 1972. Effect of air pollution on the volatile oil from leaves of Pinus 

ponderosa. Phytochemistry 11: 1815-1818. 

COYNE, J.F. and LOTT, L.H. 1976. Toxicity of substances in Pine oleoresin to southern pine beetles. J. Georgia Entomol. 

Soc. 11: 301-305. 

231

DUKE, J.A. 1985. CRC Handbook of Medicinal Herbs. CRC Press, Boca Raton, FL. 677 p. 

HAYES, J.L., INGRAM, L.L. Jr., STROM, B.L., ROTON, L.M., BOYETTE, M.W., and WALSH, M.T. 1994a. Identification 

of a host compound and its practical application: 4-allylanisole as a bark beetle repellent. Proceedings of the 4th 

Southern Station Chemical Sciences Meeting; Feb. 1-2, 1994. Gen. Tech. Rep. SO-104. Asheville, NC: U.S. 

Department of Agriculture, Forest Service. 

HAYES, J.L., STROM, B.L., ROTON, L., and INGRAM, L.L. Jr. 1994b. Repellent properties of the host compound 4allylanisole 

to the southern pine beetle. J .Chem. Ecol. p. 1595-1615. 

HAYES, J.L. and STROM, B.L. 1994. 4-Allylanisole as an inhibitor of bark beetle (Coleoptera: Scolytidae) aggregation. J. 

Econ. Entomol. 87: 1586-1594. 

HOBSON, K.R. 1992. Studies on host selection and biology of Dendroctonus valens and fungal host interactions of 

Dendroctonus brevicomis. Ph.D. Dissertation, University of California, Berkeley. 

HOBSON, K.R,, WOOD, D.L., COOL, L.G., WHITE, RR., OHTSUKA, T., KUBO, I., and ZAVARIN, E. 1993. Chiral 

specificity :inresponses by the bark beetle D. valens to host kairomones. J. Chem. Ecol. 19: 1837-1846. 

HUSSEIN, M.Y., LABS, N.H., and ALI, J.H. 1990. Biological and chemical factors associated with the successful introduction 

of Elaeidobius kamerunicus Faust, the oil palm pollinator in Malaysia. In van Heemert, C. and de Ruijter, A., 

eds., The Sixth Int'l Syrup. on Pollination. 

MARR, K.L. and TANG, C.S. 1992. Volatile insecticidal compounds and chemical variability of Hawaiian Zanthoxylum 

(Rutaceae) species. Biochem. System. and Ecol. 20: 20%217. 

MASSON, C. and MUSTAPARTA, H. 1990. Chemical information processing in the olfactory system of insects. Physiol. 

Rev. 70: 199-245. 

METCALF, R.L. and LAMPMAN, R.L. 1989. Estragole analogues as attractants for Diabrotica species (Coleoptera: 

Chrysomelidae) corn rootworms. J. Econ. Entomol. 82: 123-129. 

MILLER, RR., COBB, EW., Jr., and ZAVARIN, E. 1968. Effect of injury upon oleoresin composition, phloem carbohydrates, 

and phloem pH. Hilgardia 39: 135-139. 

MIRO'v; N.T. 1961. Composition of gum turpentines of pines. Tech. Bull. 1239. Washington, De: U.S. Department of 

Agriculture. 

NEBEKER, T.E., SCHMITZ, R.E, TISDALE, R.A., and HOBSON, K.R. 1995. Chemical and nutritional status of dwarf 

mistletoe, armillaria root rot, and comandra blister rust infected trees that may influence tree susceptibility to bark 

beetle attack. Can. J. Bot. (accepted 11/94, in press) 

OKUNADE, A.L. and OLAIFA, J.I. 1987. Estragole: an acute toxic principle from the volatile oil of the leaves of Clausena 

anisata. J. Nat. Prod. 50: 990-991. 

RUDINSKY, J.A. 1962. Ecology of Scolytidae. Ann. Rev. Entomol. 7: 327-348. 

SMITH, R.H. 1975. Formula for describing effect of insect and host tree factors on resistance to western pine beetle attack. 

J. Econ. Entomol. 68(6): 841-844. 

SMITH, R.M. 1975. Note on gum turpentine of Pinus caribaea from Fiji. N. Zealand J. of Sci. 18: 547-548. 

TKACZ, B.M. and SCHMITZ, R.E 1986. Association of an endemic mountain pine beetle population with lodgepole pine 

infected by ArmiUaria root disease in Utah. Res. Note INT-353. Ogden, LIT: U.S. Department of Agriculture, Forest 

Service. 

232

WERNER, R.A. 1995. Toxicity and repellency of 4-allylanisole and monoterpenes from white spruce and tamarack to the 

spruce beetle and eastern larch beetle (Coleoptera: Scolytidae) Environ. Entomol. 24: 151-158. 

WHITE, RR. and HOBSON, K.R. t993. Stereospecific antennal response by the red turpentine beetle, Dendroctonus valens 

to chiral monoterpenes from ponderosa pine resin. J. Chem. Ecol. t9:2193-2202. 

WOOD, D.L. 1982. The role of pheromones and atlomones in the host selection and colonization behavior of bark beetles. 

Ann. Rev. Entomol. 27:411-446. 

ZAVARIN, E., COBB, F.W., Jr., BERGOT, J. and BARBER, H.W. 1971. Variation of the Pinus ponderosa needle oil with 

season and needle age. Phytochemistry 10:3107-3114. 

233

WILL GLOBAL WARMING ALTER PAPER BIRCH 

SUSCEPTIBILITY TO BRONZE BIRCH BORER ATTACK? 

ROBERT A. HAACK 


1407 S. Harrison Road, East Lansing, M148823, USA 

INTRODUCTION 

According to several general circulation models, a rapid climatic warming of 1-5°C in the annual global-mean surface 

air temperature is predicted by the year 2050 (Shands and Hoffman 1987, MacCracken et al. 1990, Karl et al. 1991, 

Schlesinger and Jiang 1991, IPCC 1992, Peters 1992, Bengtsson 1994). Some forest entomologists predict an increase in 

outbreak frequency of certain forest insects as a result of global warming (Kellomaki et al. 1988, Hedden 1989, Franklin et 

al. 1992, Fleming and Volney 1995), although the actual response may be difficult to predict due to the many interacting 

biotic and abiotic factors (Hedden 1989, Porter et al. 1991, Cammell and Knight 1992, Dewar and Watt 1992, Ayres 1993, 

Scriber and Gage 1995, Williams and Liebhold 1995). 

Is there a way to test whether climate change will alter tree susceptibility to insect attack? One method is to look for 

patterns of differential insect attack within genetic test plantations where trees of known genetic background grow together at 

the same test site. Frequently, the primary objective of genetic test plantings is to identify either specific families or geographic 

seed sources (provenances) that performed well at a particular location, especially with respect to growth and form 

characteristics. The trees in genetic test plantings, particularly provenance tests, commonly originate from several distant 

locations, often representing several contrasting climates. Because tree populations become genetically adapted to their local 

conditions, one major factor that affects their performance when moved to a new location is the difference in climate between 

the original and the new location (Wright 1976, Kozlowski et al. 1991, Schmidtling 1994). Data from genetic test plantings 

have only recently been examined to elucidate how some tree species might respond to global warming (Matyas 1994, 

Schmidtling 1994). Similarly unexplored has been the use of provenance plantation data to predict how global warming will 

affect tree susceptibility to insect attack. 

In this paper, I explore how global warming might alter susceptibility of paper birch, Betula papyrifera Marsh., to the 

bronze birch borer, Agrilus anxius Gory (Coleoptera: Buprestidae). This particular tree-insect combination was selected 

because (1) paper birch appears highly sensitive to environmental stress given that several large scale declines of paper birch 

have been reported this century in eastern North America, and (2) the bronze birch borer has usually been the ultimate 

mortality agent of stressed birch trees (Slingerland 1906, Swaine 1918, Spaulding and MacAloney 1931, Balch and Prebble 

1940, Hawboldt 1947, Nash et al. 1951, Barter 1957, Redmond 1957, Clark and Barter 1958, Haack and Mattson 1989, Jones 

et al. 1993, Braathe 1995). Birch mortality as a result of bronze birch borer attack has followed stress events such as drought, 

extreme cold winter temperatures that follow closely behind a winter thaw, elevated growing season temperatures, high soil 

water tables, years of heavy seed crops, and years of severe insect defoliation (Redmond 1955, 1957; Clark and Barter 1958; 

Herms 1991; Herms and Mattson 1991; Auclair et al. 1992; Jones et al. 1993; Auclair et al. 1995; Braathe 1995). 

Paper birch is a boreal species, and in Michigan it reaches its southernmost range near the center of Michigan's lower 

peninsula. The ecotone between boreal and temperate forests closely follows the 47°F (8.3"C) average annual mean temperature 

isotherm in Michigan (Fig. 1). Typically, annual mean temperature decreases steadily with increasing latitude, but in 

Michigan, the Great Lakes cause a semi-marine type climate, causing isotherms to closely follow the lakes' shorelines (Fig. 

1). This feature of moderately vertical isotherms in portions of Michigan allows for great variation in climatic conditions 

within a rather narrow latitudinal band. Michigan is relatively flat, ranging in elevation from 174 to 604 m (572-1,980 ft) 

above sea level. 

Mattson, W.J., Niemel/i, R, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


234

41 40 

42 \ 4t 

42 41 

42 ",,,\ 

42 40 40 

41 / 42 

40 43 

4O 

41,1 44 

2 

43 

44 

42 45 ,,. 

45 

46 / 

47 ,_ 46 

48 I 1 

49 _ 49 

.,, / 

49 49 50 

Figure 1.--Average annual mean temperature isotherms for Michigan in °E based on the period 1940-1969 (MDA 1974). 

Inset shows Michigan, with county-level resolution, the surrounding Great Lakes (shaded area), and outlines of 

several adjacent states and Canada. Approximate °C equivalent values are: 40°F = 4.4°C, 41 °F = 5.0°C, 42°F = 5.60C, 

43°F = 6.1 °C, 44°F = 6.7°C, 45*F = 7.2°C, 46°F = 7.7°C, and 47°F = 8.3°C 

Models of global warming in the northern hemisphere predict that a species range will change most dramatically 

along the southern limits of its range (Cannell et al. 1989, Botkin and Nisbet 1992, Woodward 1992, Matyas 1994, Sykes and 

Prentice 1995). Fortunately, in southern Michigan there is a paper birch progeny test (Fig. 2) that consists of over 200 halfsib 

(i.e., where the identity of the maternal parent is known) families that were collected from throughout the natural range of 

birch in Michigan. In addition, it is coincidental that the mean annual temperature varies by about 10°F (5.6°C) between the 

southern and northern extremes of Michigan (Fig. 1). Therefore the paper birch trees growing in the genetic test plantation in 

southern Michigan are theoretically experiencing an increase in mean annual temperature of 1° to 5°C. Given that the bronze 

birch borer preferentially attacks weakened birch trees, it seems plausible that borer attack could vary among seedlots of 

varying geographic origins if they were experiencing differential levels of stress. 

The bronze birch borer, like many other buprestids in the genus Agrilus, usually attacks host trees that are stressed by 

drought or repeated insect defoliation (Anderson 1944; Clark and Barter 1958; Barter 1965; Carlson and Knight 1969; Haack 

and Benjamin 1982; Haack and Slansky 1987; Mattson and Haack 1987a, 1987b; Haack and Mattson 1989; Herms 1991; 

Herms and Mattson 1991; Wargo and Haack 1991; Haack 1992; Haack and Acciavatti 1992; Solomon 1995). The biology 

and life history of the bronze birch borer has been reported by several authors (Chittenden 1898; Balch and Prebble 1940; 

Anderson 1944; Barter 1957; Ball and Simmons 1980; Loerch and Cameron 1983, 1984; Akers and Nielsen 1990; Wilson 

and Haack 1990). Briefly, the bronze birch borer is univoltine. Adult beetles emerge in early summer from within their host 235 

50 

..,- 47

47°N - / 

8 1 

45ON _ 

44ON _ 

43ON_ 

42°N 

Figure 2.--State of Michigan, showing the 26 locations where paper birch seed was collected in 1975. The * in southwestern 

Michigan indicates the location of the plantation site where the seedlings were planted in 1976. Inset shows the 

conterminous 48 states of the US, with Michigan shaded. 

trees, and then search for suitable trees to attack. Adults often reattack the same tree until it dies, but from year to year they 

attack only those portions that are still living at the time of oviposition. It often takes 2-3 years of successive attack before a 

given branch or trunk section dies. Adults are believed to be attracted primarily by host odors given that no pheromone has 

yet been discovered for a buprestid (Haack and Slansky 1987, Haack 1992). After mating, females oviposit on the bark 

surface of the trunk and branches. After hatching, larvae tunnel through the bark and then feed in the cambial region, scoring 

both the inner bark and outer sapwood. Larvae construct meandering galleries throughout the summer months, passing 

through four larval instars. In late summer or fall, larvae construct individual pupal chambers in the outer sapwood and there 

pass the winter. Pupation occurs the following spring, and soon the adults emerge to renew the cycle. 


As described in Miller et al. (1991), seed was collected from 218 open-pollinated paper birch seedlots (half-sib 

families) in 1975, representing 26 Michigan populations; 123 trees from 13 populations in Michigan's Upper Peninsula and 

95 trees from 13 populations in Michigan's Lower Peninsula (Fig. 2, Table 1). All parents were growing in natural forests 

between latitudes 43.3°N and 47.4°N, and ranged from 5 to 22 m in height, and 9 to 46 cm in diameter (Table 1). The seed 

from the 218 half-sib families was planted in February 1976 and grown in the greenhouse for 5 months, held outdoors for 2 

236

Table 1.--Summary data for the 26 populations of Michigan paper birch trees, representing 218 trees, from which seed was 

collected in 1975, including for each population the number of trees sampled, location data, climatic data, and mean 

height and diameter at breast height (DBH) of the parental trees. 

Seed source I Location 2 Climatic data3 Height (m) 4 DBH (cm) 4 

Pop No °N Lat Area County Temp DD FFD Mean (range) Mean (range) 

t 7 46.8 UP Ontonagon 42 1600 I10 14.0 (10-21) 25.3 (9-46) 

2 10 47.4 UP Keweenaw 40 1600 140 11.8 (5-19) 24.0 (10-45) 

3 10 46.7 UP Baraga 40 1600 130 16.2 (9-21) 21.0 (15-29) 

4 7 46.3 UP Gogebic 40 1800 60 15.0 (12-19) 22.6 (21-24) 

5 10 46.1 UP Iron 41 1800 100 15.1 (13-16) 19.6 (15-25) 

6 10 45.2 UP Menominee 42 1600 140 14.6 (12-16) 18.6 (14-23) 

7 10 46.1 UP Delta 41 1800 70 13.4 (9-16) 20.4 (13-31) 

8 10 46.6 UP Marquette 42 1600 150 14.7 (8-21) 23.5 (10-40) 

9 10 46.5 UP Alger 41 1400 140 12.0 (11-15) 16.5 (13-20) 

10 11 46_0 UP Schoolcraft 42 1600 110 16.3 (12-22) 24.6 (14-32) 

11 9 46.7 UP Chippewa 41 1400 150 11.1 (11-13) 17.6 (12-28) 

12 10 46.3 UP Chippewa 41 1400 130 13.5 (11-16) 22.6 (17-30) 

13 9 46.0 UP Mackinac 41 1600 120 13.2 (8-16) 18.3 (15-24) 

14 10 45.6 LP Emmet 42 1800 150 13.7 (7-16) 25.4 (15-32) 

15 5 44.9 LP Antrim 44 2000 140 15.5 (14-16) 24.5 (21-26) 

16 10 44.9 LP Montmorency 43 1800 70 13.7 (11-16) 19.8 (13-25) 

17 10 45.5 LP Presque Isle 44 1800 140 12.3 (9-16) 17.7 (12-28) 

18 10 44.8 LP Alcona 43 1800 130 16.3 (15-18) 25.8 (19-33) 

19 4 43.4 LP Montcalm 46 2400 130 15.2 (11-18) 21.5 (14-27) 

20 5 43.9 LP Lake, Mason 46 2200 120 14.6 (11-16) 21.7 (19-24) 

21 8 44.3 LP Manistee 46 2200 150 15.8 (11-19) 20.5 (16-27) 

22 9 44.4 LP Missaukee 43 1800 100 14.5 (11-19) 22.6 (14-38) 

23 9 44.2 LP Gladwin 44 2000 100 16.2 (11-18) 20.4 (15-28) 

24 9 44.0 LP Arenac 45 2000 130 13.8 (11-16) 17.5 (11-25) 

25 5 43.6 LP Sanilac 46 2400 130 12.4 (11-15) 15.6 (13-17) 

26 1 43.3 LP Saginaw 47 2600 150 12.1 n.a. 18.2 n.a. 

Seed source: Pop - seed source population number as given in Fig. 1; No = number of parental trees from which seed 

was collected in each particular population. 

2 Location: Area - Michigan seed source locations where UP refers to Upper Peninsula and LP refers to Lower 

Peninsula; County = Michigan county from which trees were sampled. 

3Climatic data for each population of sampled trees using isopleth data from various Michigan climate maps: Temp = 

average annual mean temperature (°F; 1940-1969; MDA 1974); DD - seasonal (March - October) growing degree days 

(base 50°F; 1951-1980; Eichenlaub et al. 1990); FFD - average number of frost free days (i.e., number of days between 

last spring 32°F reading and first fall 320F reading; 1940-1969; data drawn from MDA 1971). 

4 Height and diameter at breast height (DBH; 1.3 m) of sampled trees by population. 

weeks, and then planted in a plowed and disked field in June, 1976. The plantation is located near 42°N, at the Fred Russ 

Experimental Forest, Michigan State University, in southwestern Michigan (Fig. 2). Trees were planted in a completely 

randomized design on 2.4 x 2.4 m square spacing with an average of 5 trees per seedlot. Overall, 1,088 trees were planted, 

with 636 trees from Upper Peninsula seed sources and 452 trees from Lower Peninsula seed sources. 

Measurements of this plant material began in 1976 and continued through 1992. Height, diameter, bark color, insect 

damage, and tree survival were measured at varying intervals over 15 years (see also Miller et al. 1991). Height was measured 

in years 2, 6, 7, 8, and 10 post-establishment. Bronze birch borer damage was evaluated in years 10, 12, and 15 on a 

scale of 0 to 3, where 0 = tree apparently healthy with no evidence of borer infestation; 1 = less than half of the crown 

237

anches dead or dying with borer-related bark ridges obvious on the bark surface; 2 = more than half of the crown branches 

dead or dying with bark ridges obvious and borer exit holes often present; and 3 = tree dead, usually with borer-related bark 

ridges and exit holes present. Bark ridges (callus tissue) occur on branch and trunk sections in response to larval feeding 

(Anderson 1944, Barter 1957, Wilson and Haack 1990, Herms 1991). The presence of borer-related bark ridges was scored 

in years 12 and 15. Completion of larval development was also scored in years 12 and 15 by noting the presence of borer exil 

holes. Tree survival was computed for each of the 218 families in years 2, 6, 8, 10, 12, 15, and 16. 

Historical temperature data were obtained from the Michigan Department of Agriculture, Climatological Division 

(MDA 1974). Average annual mean temperature was considered the most relevant climatic variable to use when considering 

global warming. The average annual mean temperature assigned to each seed source was obtained from a map showing 

annual mean temperatures for Michigan averaged over the 30-year period 1940-1969 (Fig. 1, Table 1, MDA 1974). This was 

done by locating all seed collection sites on the map and then assigning a particular annual mean temperature value to each 

seed source based on its location relative to the map isotherms. Although actual temperature data for each of the 26 collections 

sites would have been preferred, this was not possible because many collection sites had no nearby weather stations. 

Additional climatic data for each of the 26 collection sites were obtained in similar fashion from other maps (Table 1). 

Using the map shown in Figure 1, the annual mean temperature values assigned to the 218 seed sources ranged from 

40° to 47°F (4.4*-8.3°C; Table 1). Actual average annual mean temperature values from several official weather stations for 

this same area and time period ranged from 39.7* to 47.1"F (4.30-8.4"C; MDA 1971). The average annual mean temperature 

of the plantation site was about 49.00F (9.4°C; MDA 1971). Overall, assuming that the birch seed sources were genetically 

acclimated to the average annual mean temperature of their original sites, then the birch trees growing at the plantation site 

were theoretically experiencing levels of climatic warming equivalent to 1.9° to 9.3*F (1.1 ° to 5.2°C; Table 2). For ease of 

discussion, I will use the term climatic differential (CD) to represent the difference (in °C) between the average annual mean 

temperatures of the seedlot origin and the plantation site. 

Table 2.--Average annual mean temperature data and percent of paper birch trees (both live and dead) that had bronze 

birch borer-induced ridges and/or borer exit holes along their lower trunk in years 12 (1987) and 15 (1990) postestablishment 

by climatic differential treatment. 

Mean temperature 

difference between 

Corresponding origin and Percent of trees with Percent of trees with 

isotherms 2 plantation site 3 borer-induced ridges 4 borer exit holes 4 

CD j °F °C *F °C Yr 12 Yr 15 Yr 12 Yr 15 

CD t ° 46, 47 7.7, 8.3 2 1.1 30ab 32b 10b 13b 

CD2 ° 44, 45 6.7, 7.2 4 2.2 24b 33b 8b 18b 

CD3 ° 42, 43 5.6, 6.1 6 3.3 39ab 46b 12b 25b 

CD4 ° 40, 41 4.4, 5.0 8 4.4 50a 60a 23a 40a 

CD = Climatic differential, which is a term that reflects the difference between the average annual mean temperature 

of the seed sources within a particular treatment group and the average annual mean temperature of the plantation 

site, which was 49°F (9.4"C). The value 1° in CD1 °, for example, reflects the annual mean temperature difference 

between the origin and the plantation site, e.g., 1.I*C (see footnote 3 below). 

2The isotherms of average annual mean temperature in *F (and the approximate °C equivalent), as given in Fig. 1, 

that correspond to each CD grouping. 

3For each CD grouping, the mean temperature difference between the origin and the plantation site was calculated 

as follows. For CDI*, the corresponding isotherms were 46* and 47°F. However, these isotherms span the temperature 

range from 46* to 47.9°F, with a mean of about 47"E Thus, the temperature difference between the origin and the : 

plantation site in this case is 2*F (49°-47 *) or about 1.I*C. These values represent the theoretical level of climatic 

warming that the corresponding trees were experiencing while growing at the plantation site in southern Michigan. 

4 Percent of trees with borer-induced ridges and/or borer exit holes along the lower trunk; percent values followed by 

the same letter within a column are not significantly different 

followed by the Dunn's test for multiple comparisons). 

238 

(p

One weakness of using data from genetic tests plantings to study the impact of climatic warming, however, is that 

other variables besides temperature change when moving seed from its site of origin to the test location. Photoperiod, for 

example, is one important variable for which many tree species, including paper birch, are genetically adapted (Vaartaja 

1959, Wright 1976, Kozlowski et al. 1991, Farmer 1993). In the northern hemisphere, typically, northern sources tend to 

flush earlier and terminate shoot growth earlier when moved southward compared with more local sources (Wright 1976, 

Kozlowski et al. 1991). In the present study, birch seed was collected between Lat. 43.3°N and 47.4°N, and then planted at 

42.0°N. On the longest day of the year there is about 15.2 hours of light at 42.0°N, 15.5 hours at 43.3°N, and 16.0 hours at 

47.4°N. Similarly, on the shortest day of the year, they experience about 9.1 hours of light at 42.0°N, 9.0 hours at 43.3°N, and 

8.4 hours at 47.4°N. As a means of minimizing any influence of photoperiod, linear regression analyses were also conducted 

on two subsets of the original 26 populations (collection sites) in which the latitudinal spread among the seed sources was 

further reduced. In the first case, only the 13populations from the Lower Peninsula of Michigan were tested (populations 14- 

26, Fig. 2, Table 1), representing collections within a 2° span of latitude and a 5°F (2.8°C) span in annual mean temperature. 

In the second case, 12 birch populations were tested (populations 4-7, 10, 12-18; see Fig. 2, Table 1), representing collections 

within a 1.5° span in latitude and a 4°F (2.2°C) span in annual mean temperature. 

Analyses were conducted using SAS and SigmaStat. Height and borer data were analyzed on an individual tree basis, 

whereas tree survival was calculated on a family basis (n = 218), using only those birch families that had at least one tree 

surviving through the second growing season (1977). Each tree was assigned to one of four "treatments" based on the 

average annual mean temperature of its origin. The four groupings were isotherms 40 ° and 41°F, 42 ° and 43°F, 44 ° and 45°F, 

and 46 oand 47°F (Table 2). The climatic differential term assigned to each grouping was CD1 ° (46-47°F), CD2°(44°-45°F), 

CD3 ° (42°-43°F), and CD4 ° (40°-41°F; Table 2). These terms, CDI°-CD4 °, reflect the average climatic differential in °C fbr 

each treatment group, e.g., for CD1 ° the midpoint climatic differential value is actually 1.I°C and represents the average 

degree of climatic warming that the trees in isotherms 460F and 47°F were theoretically experiencing (Table 2). Analysis of 

variance (ANOVA) was used to test for differences among treatments; ap level of 0.05 was used for significance. Tree 

height and survival data were analyzed using the GLM procedure of SAS, followed by the Student-Newman-Keuls means 

separation test. The arcsin square root transformation was performed on the survival data prior to analysis. Borer data were 

analyzed using the Kruskal-Wallis nonparametric test of SigmaStat, which performs an ANOVA on ranks; Dunn's test was 

used to make multiple comparisons among treatment means. 

RESULTS 

Paper birch growth, survival, and susceptibility to bronze birch borer attack were significantly affected by the 

climatic differential between the origin of the seed source and the plantation site (Fig. 3). Overall, height growth decreased 

as the climatic differential increased (Fig. 3A). Differences in height growth were already evident by the end of the second 

growing season in 1977 (p < 0.0001; Fig. 3A). By the end of the sixth growing season (1985), CDI ° trees were the tallest, 

with tree height becoming successively shorter for the CD2 °, CD3 °, and CD4 ° trees (p < 0.0001 ; Fig. 3A). 

As the trees aged, tree survival tended to decrease as the climatic differential increased (Fig. 3B). From year 2 (1977) 

to year 8 (1983) post-establishment, little tree mortality occurred, but thereafter tree mortality accelerated, especially among 

the trees experiencing the greatest climatic differential (CD4°; Fig. 3B). Treatment survival values first became significantly 

different in year 12 post-establishment (1987; p < 0.006), when survival for the CD4 ° trees was lower than that for trees from 

the other three groups. A similar pattern of tree survival continued into years 15 (1990; p < 0.001) and 16 (1991; p < 0.0008) 

post-establishment. 

As the trees became older, tree susceptibility to the bronze birch borer increased with increasing climatic differential 

(Fig. 3C). Of the 146 paper birch trees that died between 1985 and 1991, 143 (98%) of them had signs of borer attack along 

their lower trunk. A similar pattern emerged during each of the 3 years when borer damage was assessed (1985, 1987, 1990; 

p < 0.001 for each), i.e., CD4 ° trees experienced the highest levels of borer attack, while CDI ° trees experienced the lowest 

levels (Fig. 3C). As would be expected from the above discussion, the occurrence of borer-related bark ridges and exit holes 

was more common on CD4 ° trees than on CD 1° trees (Table 2). 

Regarding the two subsets of birch populations in which the latitudinal spread was reduced, paper birch growth, 

survival, and susceptibility to bronze birch borer were all significantly affected by the climatic differential between the origin 

of the seed source and the plantation site. In the first case, linear regression analysis of the 13 populations from Michigan's 

239

- A 

7 A 

_" 

--l-- 

6 

5 

C 

c_ -,.-- CD1 o 

"1- 

U_I 

3 _ CD2 ° 

I'-- r'r" 

i"Ti 

2 

4 jc 

_ CD3 o 

1 I_B +- CU4 ° 

__1 

! r 1 _ _ _ .... f i' _ ...... _ T ' i ' E 1 

2 4 6 8 10 12 14 16 

100 : ........... A A 

0_ _ 70 ",_ ",,,_AA.. 

,,u4,6o 

¢7" B B 

!--'- 50 

40 , ," , .......... 

2 4 6 8 10 12 14 16 

"_0_"/ , .... .., ...., ..,B , ...., , ., , , 

2 4 6 8 10 12 14 16 

TREE AGE (Years) 

Figure 3.kPaper birch mean height growth (A), mean survival (B), and mean susceptibility to borer attack (C) over time 

(years post-establishment) in a provenance plantation in southern Michigan. Survival was based on only those trees 

that lived for the first two years. The 0-3 borer rating scale is presented in the Methods section. Trees are divided 

into four treatment groups based on the climatic differential (CD) between the average annual mean temperature of 

the origin of each seed source and the annual mean temperature of the plantation site. On average, trees assigned to 

treatment CD 1° had a 1.1°C difference between the origin and the test site, CD2 ° a 2.2°C difference, CD3 ° a 3.3°C 

difference, and CD4 ° a 4.4*C difference (see Table 2). For each parameter, means followed by the same letter within 

a column (i.e., year) are not significantly different (p

Lower Peninsula (populations 14-26) demonstrated that as the climatic differential between the origin and the test site 

increased, tree height decreased (p < 0.006, r2 = 0.51, n = 13 birch populations), tree survival decreased (p < 0.012, r2 = 0.49, 

n = 13), and tree susceptibility to borer attack increased (p < 0.006, r2= 0.51, n = 13). Similarly, in the second case, which 

included 12 birch populations (populations 4-7, 10, 12-18), linear regression analysis indicated that as the climatic differential 

between origin and test site increased, tree height decreased (p < 0.02, r2= 0.43, n = 12 birch populations), tree survival 

decreased (p < 0.038, r2= 0.36, n = 12), and tree susceptibility to borer attack increased (p < 0.012, r2 = 0.45, n = 12). 

DISCUSSION 

Cannell et al. (1989) state that most tree species will benefit from an increase in annual mean temperature of I°C, but 

genetic increases will often cause growth reductions or mortality. Similarly, Matyas (1994) states that an increase in temperature 

will positively affect tree growth only within each species' limits of physiological and ecological tolerance. Paper birch 

appears to have narrow limits of heat tolerance. In the present study, as the climatic differential increased for individual seed 

sources, birch trees grew slower, experienced higher mortality rates, and sustained higher rates of borer attack. Such results 

strongly suggest that paper birch will experience dramatic decline along its southern range in response to climatic warming, 

and that the bronze birch borer will likely be the principal mortality agent involved in future birch decline. Similar predictions 

for widespread decline of paper birch in response to global warming have been made by Pastor and Post (1988), 

Overpeck et al. (1991), Botkin and Nisbet (1992), Reed and Desanker (1992), Reed et al. (1992b), Solomon and Bartlein 

(1992), and Jones et al. (1994). Several of the general circulation models predict reduced rainfall or increased rates of 

evapotranspiration in association with increased air temperatures (Karl et al. 1991, IPCC 1992). Thus, if higher temperatures 

are combined with reduced rainfall, then many tree species will likely be even more susceptible to attack by trunk-boring 

insects (Mattson and Haack 1987a, 1987b). 

As mentioned above, differences in sensitivity to photoperiod between northern and southern seed sources could 

partly explain the relatively poorer performance exhibited by the more northerly birch families in the present study. However, 

when I restricted the birch data set to seed sources that spanned as little as 1.5 ° and 2° latitude, the same patterns of 

reduced paper birch growth, survival, and resistance to bronze birch borer persisted. In a Picea abies provenance study, 

Vaartaja (1959) reported little difference in tree growth for seed sources collected between 47°N and 52°N and then planted at 

47°N. Although the general rule for photoperiod is that the further north the origin, the more photoperiod sensitive the seed 

source, Beuker (1994) states that the exact effect of photoperiod cannot be predicted. Although changes in photoperiod are 

undoubtedly important, the significant relationships found here for two relatively tight latitudinal bands, strongly suggest that 

the differences in seed source performance in the present study were greatly influenced by temperature. 

Another possible reason for the poorer growth of cold-adapted seed sources at warmer sites, is that they may have 

higher dark respiration rates than local sources (Kozlowski et al. 1991, Mebrahtu et al. 1991). Consequently, when growing 

in a warmer environment, cold-adapted sources expend more energy and have lower net assimilation rates than local sources. 

Reduced rates of net assimilation can help explain greater susceptibility to borer attack since compromised trees would likely 

produce fewer defensive compounds and have reduced rates of diameter growth (Herms and Mattson 1992). Slow stem 

growth has been shown to increase susceptibility of paper birch to the bronze birch borer (Herms 1991, Herms and Mattson 

1991). 

In Michigan, as was true for much of North America, 1988 was a year of severe heat and drought (Kerr 1989, USGS 

1991). In the Great Lakes region, outbreaks of several cambial-feeding forest insects occurred in 1988, including the bronze 

birch borer (Haack and Mattson 1989, Jones et al. 1993). Part of the rapid increase in birch mortality and borer attack 

between years 12 (1987) and 15 (1990) in the present study (Figs. 3B, 3C) could have been triggered by the 1988 drought. In 

fact, it appears from the slopes of tree survival (Fig. 3B) and borer attack (Fig. 3C) for the four CD treatments, that the CD2 °, 

CD3 °, and CD4 ° trees were much more negatively impacted by the heat and drought than were the CD1 ° trees, suggesting 

that an increase in annual mean temperature of about 1°C will have limited detrimental effects on paper birch, but that 

increases of 2°C and more will be very damaging. In support of this contention are studies in the Upper Peninsula of 

Michigan by Reed et al. (1992a) and Jones et al. (1993) that report slower growth and higher mortality rates for paper birch, 

following the warm and dry years of 1987 and 1988 when mean summer temperatures were about 1°C above normal. In 

controlled studies on yellow birch, Betula alleghaniensis Britton, limited rootlet mortality was observed with a I°C increase 

in soil temperature, but extensive rootlet mortality occurred with increases of 6-7°C (Redmond 1955, 1957). 

241

As global warming occurs, I predict that those tree species with associated cambial-feeding insects that invade the 

trunk will be the first tree species to exhibit widespread mortality. This guild of trunk-infesting, cambial-feeding insects 

poses the greatest lethal threat to their host trees (Haack and Slansky 1987, Mattson et al. 1988, Haack and Byler 1993). 

Trees are typically highly resistant to cambial-feeding insects, but during periods of stress, they become more susceptible to 

such insects (Mattson and Haack 1987a, 1987b; Millers et al. 1989). Some of the boreal tree genera that occur in Michigan 

that can be killed by trunk-infesting, cambial-feeding insects, and thus will likely be more at risk as climatic warming occurs, 

include Abies, Betula, Larix, Picea, Pinus, Populus, and Tsuga (Table 3). Although climatic warming in the northern 

hemisphere is predicted to bring about tree mortality along the southern edge of a species range, the northern range of these 

same tree species may expand (Cannell et al. 1989, Woodward 1992, Sykes and Prentice 1995). In fact, paper birch has been 

documented to be expanding its range northward into the tundra, possibly in response to climatic warming (Woodward 1992). 

Table 3.--Boreal tree genera of the Great Lakes region that will likely experience widespread insect-induced mortality as a 

result of climatic warming and the corresponding trunk-infesting, cambial-feeding insect that will likely be the major 

mortality agent. 

Host tree Insect species 

Genus Common name Species Common name Family 

Abies Fir Pityokteins sparsus (LeConte) Balsam fir bark beetle Scolytidae 

Betula Birch Agrilus anxius Gory Bronze birch borer Buprestidae 

Larix Larch Dendroctonus simplex LeConte Eastern larch beetle Scolytidae 

Picea Spruce Dendroctonus rufipennis (Kirby) Spruce beetle Scolytidae 

Pinus Pine Ips pini (Say) Pine engraver beetle Scolytidae 

Populus Aspen Agrilus liragus Barter and Brown Bronze poplar borer Buprestidae 

Tsuga Hemlock Melanophilafulvoguttata (Harris) Hemlock borer Buprestidae 

The present study shows that data from genetic test plantings of forest tree species can provide insight into how treeinsect 

interactions may change as air temperatures increase in the future. Similar studies should be conducted by others. 

This would be particularly useful in areas where the local geography allows seed to be collected from several temperature 

regimes within a narrow latitudinal band and thereby allow the effects of temperature and photoperiod to be partly disentangled. 


I wish to thank the staff of Michigan State University, Fred Russ Experimental Forest, for field assistance in this 

study; Paul Bloese, Michigan State University, Department of Forestry and Michigan Cooperative Tree Improvement 

Program, for making available early tree performance data from this paper birch plantation; and Paul Bloese, Daniel Herms, 

Dow Gardens, and Robert Lawrence, USDA Forest Service, for reviewing an earlier draft of this manuscript. 


AKERS, R.C. and NIELSEN, D.G. 1990. Reproductive biology of the bronze birch borer (Coleoptera: Buprestidae) on 

selected trees. J. Entomol. Sci: 25: 196-203. 

ANDERSON, R.F. 1944. The relation between host condition and attacks by the bronzed birch borer. J. Econ. Entomol. 37: 

588-596. 

242

AUCLAIR, A.ND., WORREST, R.C., LACHANCE, D., and MARTIN, H.C. 1992. Climatic perturbations as a general 

mechanism of forest dieback, p. 38-58. In Manion, P.D. and Lachance, D., eds. Forest Decline Concepts, American 

Phytopathological Society Press, St. Paul, Minnesota. 

AUCLAtR, A.N.D., LILL, J.T., and REVENGA, C. 1995. The role of climate variability and global warming in the dieback 

of northern hardwoods. Water, Air, and Soil Pollution (In press). 

AYRES, M.P. 1993. Plant defense, herbivory, and climate change, p. 75-94. In Kareiva, RM., Kingsolver, J.G., and Huey, 

R.B., eds. Biotic Interactions and Global Change. Sinauer Associates, Sunderland, Massachusetts. 

BALCH, R.E. and PREBBLE, J.S. 1940. The bronze birch borer and its relation to the dying of birch in New Brunswick 

forests. For. Chron. t6: 179-201. 

BALL, J. and SIMMONS, G. 1980. The relationship between bronze birch borer and birch dieback. J. Arbor. 6:309-314. 

BARTER, G.W. 1957. Studies on the bronze birch borer, Agrilus anxius Gory, in New Brunswick. Can. Entomol. 89: 12- 

36. 

BARTER, G.W. t 965. Survival and development of the bronze poplar borer Agrilus liragus Barter and Brown (Coleoptera: 

Buprestidae). Can. Entomol. 997' 1063-1068. 

BENGTSSON, L. 1994. Climate change: Climate of the 21st century. Agric. For. Meteorol. 72: 3-29. 

BEUKER, E. 1994. Long-term effects of temperature on the wood production of Pinus sylvestris L. and Picea abies (L.) 

Karst. in old provenance experiments. Scand. J. For. Res. 9: 34-45. 

BOTKIN, D.B. and NISBET, R.A. 1992. Projecting the effects of climate change on biological diversity in forests, p. 277- 

293. In Peters, R.L. and Lovejoy, T.E., eds. Global Warming and Biological Diversity. Yale Univ. Press, New Haven, 

CT. 

BRAATHE, R 1995. Birch dieback - caused by prolonged early spring thaws and subsequent frost. Norwegian J. Agric. Sci. 

Supple. No. 20: 1-59. 

CAMMELL, M.G. and KNIGHT, J.D. 1992. Effects of climatic change on the population dynamics of crop pests. Adv. 

Ecol. Res. 22: 117-162. 

CANNELL, M.G.R., GRACE, J., and BOOTH, A. 1989. Possible impacts of climatic warming on trees and forests in the 

United Kingdom: a review. Forestry 62: 337-364. 

CARLSON, R.W. and KNIGHT, F.B. 1969. Biology, taxonomy, and evolution of four sympatric Agrilus beetles borer 

(Coleoptera: Buprestidae). Contrib. Amer. Entom_ol.Inst. 4: 1-105. 

CHITTENDEN, F.H. t898. A destructive borer enemy of birch trees, with notes on related species. U.S. Dep. Agric. Div. 

Entomol. Bull (new series) 18:44-51. 

CLARK, J. and BARTER, G.W. 1958. Growth and climate in relation to dieback of yellow birch. For. Sci. 4: 343-363. 

DEWAR, R.C. and WATT, A.D. 1992. Predicted changes in the synchrony of larval emergence and budburst under climatic 

warming. Oecologia 89: 557-559. 

EICHENLAUB, V.L., HARMAN, J.R., NURNBERGER, F.V.; and STOLLE, H.J. 1990. The Climatic Atlas of Michigan. 

University of Notre Dame Press, Notre Dame, Indiana. 165 p. 

FARMER, R.E., Jr. 1993. Latitudinal variation in height and phenology of balsam poplar. Silvae Genetica 42: 148-153. 

243

FLEMING, R.A. and VOLNEY, W.J.A. 1995. Effects of climate change on insect defoliator population processes in 

Canada's boreal forest: Some plausible scenarios. Water Air Soil Pollut. 82: 445-454. 

FRANKLIN, J.F., SWANSON, F.J., HARMON, M.E., PERRY, D.A., SPIES, T.A., DALE, V.H., MCKEE, A., FERRELL, 

W.K., MEANS, J.E., GREGORY, S.V., LATTIN, J.D., SCHOWALTER, T.D., and LARSEN, D. 1992. Effects of 

global climatic change on forests in northwestern North America, p. 244-257. In Peters, R.L. and Lovejoy, T.E., eds. 

Global Warming and Biological Diversity. Yale Univ. Press, New Haven, CT. 

HAACK, R.A. 1992. Tree-stress-buprestid interactions: Our current understanding and future needs, p. 75-76. In Allen, 

D.C. and Abrahamson, L.E, eds. Proceedings: North American Forest Insect Work Conference. 25-28 March 1991, 

Denver, CO. U.S. Dep. Agric. For. Serv. Gen. Tech. Rep. PNW-294. 

HAACK, R.A. and ACCIAVATTI, R. E. 1992. Twolined chestnut borer. U.S. Dep. Agric. For. Serv. Forest Insect and 

Disease Leaflet I68. 12 p. 

HAACK, R.A. and BENJAMIN, D.M. 1982. The biology and ecology of the twolined chestnut borer, Agrilus bilineatus 

(Coleoptera: Buprestidae), on oaks, Quercus spp., in Wisconsin. Can. Entomol. 114:385-396. 

HAACK, R.A. and BYLER, J.W. 1993. Insects and pathogens: regulators of forest ecosystems. J. For. 91 (9): 32-37. 

HAACK, R.A. and MATTSON, W.J. 1989. The long, hot summer of '88: They nibbled while the forests burned. Natural 

History (January) 56-57 

HAACK, R.A. and SLANSKY, F. 1987. Nutritional ecology of wood-feeding Coleoptera, Lepidoptera, and Hymenoptera, p. 

449-486. In Slansky, F. and Rodriguez, J.G., eds. Nutritional Ecology of Insects, Mites, and Spiders. John Wiley, 


HAWBOLDT, L.S. 1947. Aspects of yellow birch dieback in Nova Scotia. J. For. 45: 414-422. 

HEDDEN, R.L. 1989. Global climate change: Implications for silviculture and pest management, p. 555-562. In Proceedings, 

Fifth Biennial Southern Silvicultural Research Conference. U.S. Dep. Agric. For. Serv. Gen. Tech. Rep. SO-74. 

HERMS, D.A. 1991. Variation in resource allocation patterns of paper birch: Evidence for physiological tradeoffs among 

growth, reproduction and defense. Ph.D. Dissertation, Michigan State University, Department of Entomology, East 

Lansing, MI. 

HERMS, D.A. and MATTSON, W.J. 1991. Does reproduction compromise defense in woody plants? p. 35-46. In 

Baranchikov, Y.N., Mattson, W.J., Hain, F.R, and Payne, T.L., eds. Forest Insect Guilds: Patterns of Interaction with 

Host Trees. U.S. Dep. Agric. For. Serv. Gen. Tech. Rep. NE-153. 

HERMS, D.A. and MATTSON, W.J. 1992. The dilemma of plants: To grow or defend. Quart. Rev. Biol. 67: 283-335. 

IPCC (Intergovernmental Panel on Climate Change). 1992. The Supplementary Report to the IPCC Scientific Assessment. 

Houghton, J.T., Callander, B.A., and Varney, S.K., eds. Cambridge University Press, Cambridge, UK. 198 p. 

JONES, E.A., REED, D.D., MROZ, G.D., LIECHTY, H.O., and CATTELINO, RJ. 1993. Climate stress as a precursor to 

forest decline: paper birch in northern Michigan, 1985-1990. Can. J. For. Res. 23: 229-233. 

JONES, E.A., REED, D.D., and DESANKER, RV. 1994. Ecological implications of projected climate change scenarios in 

forest ecosystems of central North America. Agric. For. Meteorol. 72:31-46. 

KARL, T.R., HEIM, JR., R.H., and QUAYLE, R.G. 1991. The greenhouse effect in central North America: If not now, 

when? Science 251" 1058-1061. 

244 :_: 

i

KELLOMAKI, S., HANNINEN, H., and KOLSTROM, T. 1988. Model computations on the impacts of the climatic change 

on the productivity and silvicultural management of the forest ecosystem. Silva Fennica 22: 293-305. 

KERR, R.A. 1989. 1988 ties for the warmest year. Science 243 891. 

KOZLOWSKI, T.T., KRAMER, RJ., and PALLARDY, S.G. 1991. The Physiological Ecology of Woody Plants. Academic 

Press, San Diego. 657 p. 

LOERCH, C.R. and CAMERON, E.A. 1983. Determination of larval instars of the bronze birch borer, Agrilus anxius 

(Coleoptera: Buprestidae). Ann. Entomol. Soc. Amer. 76: 948-952. 

LOERCH, C.R. and CAMERON, E.A. 1984. Within-tree distributions and seasonality of immature stages of the bronze 

birch borer, Agrilus anxius (Coleoptera: Buprestidae). Can. Entomol. 116: 147-152. 

MACCRACKEN, M.C., BUDYKO, M.I., HECHT, A.D., and IZRAEL, Y.A. 1990. Prospects for Future Climate: A Special 

US/USSR Report on Climate and Climate Change. Lewis Publishers, Chelsea, MI. 270 p. 

MATTSON, W.J. and HAACK, R.A. 1987a The role of drought stress in provoking outbreaks of phytophagous insects, p. 

365-407. In Barbosa, P. and Schultz, J.C., eds. Insect Outbreaks: Ecological and Evolutionary Perspectives. Academic 

Press, Orlando. 

MATTSON, W.J. and HAACK, R.A. 1987b. The role of drought in outbreaks of plant-eating insects. BioScience 37:110- 

118 

MATTSON, W.J., LAWRENCE, R.K., HAACK, R.A., HERMS, D.A., and CHARLES, RJ. 1988. Plant defensive strategies 

for different insect feeding guilds in relation to plant ecological strategies and intimacy of host association, p. 1-38. 

In Mattson, W.J., Levieux, J., and Bernard-Dagan, C. eds. Mechanisms of Plant Resistance to Insects. Springer- 

Verlag, New York. 

MATYAS, C. 1994. Modeling climate change effects with provenance tests data. Tree Physiol. 14: 797-804. 

MDA (Michigan Department of Agriculture). 1971. Climate of Michigan by stations. Michigan Department of Agriculture, 

Climatological Division, Lansing, MI. 

MDA (Michigan Department of Agriculture). 1974. Supplement to the climate of Michigan by stations: mean temperature 

maps for the period 1940-1969. Michigan Department of Agriculture, Climatological Division, Lansing, MI. 14 p. 

MEBRAHTU, T., HANOVER, J.W., LAYNE, D.R., and FLORE, J.A. 1991. Leaf temperature effects on net photosynthesis, 

dark respiration, and photorespiration of seedlings of black locust families with contrasting growth rates. Can. J. For. 

Res. 21" 1616-1621. 

MILLER, R.O., BLOESE, RD., HANOVER, J.W., and HAACK, R.A. 1991. Paper birch and European white birch vary in 

growth and resistance to bronze birch borer. J. Amer. Soc. Hort. Sci. 116: 580-584. 

MILLERS, I., SHRINER, D.S., and RIZZO, D. 1989. History of hardwood decline in the eastern United States. U.S. Dep. 

Agric. For. Serv. Gen. Tech. Rep. NE-126. 

NASH, R.W., DUDA, E.J., and GRAY, N.H. 1951. Studies on extensive dying, regeneration, and management of birch. 

Maine For. Serv. Bull. 15.82 p. 

OVERPECK, J.T., BARTLEIN, RJ., and WEBB, III, T. 1991. Potential magnitude of future vegetation change in eastern 

North America: Comparisons with the past. Science 254: 692-695. 

PASTOR, J. and POST, W.M. 1988. Response of northern forests to CO2-induced climate change. Nature 334: 55-58. 

245

PETERS, R.L. 1992. Introduction, p. 3-114. In Peters, R.L. and Lovejoy, T.E., eds. Global Warming and Biological Diversity. 

Yale Univ. Press, New Haven, CT. 

PORTER, J.H., PARRY, M.L., and CARTER, T.R. 1991. The potential effects of climatic change on agricultural insect 

pests. Agric. For. Meteorol. 57: 221-240. 

REDMOND, D.R. 1955. Studies in forest pathology. XV. Rootlets, mycorrhizae, and soil temperatures in relation to birch 

dieback. Can. J. Bot. 33:595-627. 

REDMOND, D.R. 1957. The future of birch from the viewpoint of diseases and insects. For. Chron. 33: 25-30. 

REED, D.D. and DESANKER, RV. 1992. Ecological implications of projected climate change scenarios in forest ecosystems 

in northern Michigan. Int. J. Biometeorol. 36: 99-107. 

REED, D.D., JONES, E.A., HOLMES, M.J., and FULLER, L.G. 1992a. Modeling diameter growth in local populations: A 

case study involving four North American deciduous species. For. Ecol. Mangt. 54: 95-114. 

REED, D.D., JONES, E.A., LIECHTY, H.O., MROZ, G.D. and JURGENSEN, M.E 1992b. Impacts of annual weather 

conditions on forest productivity: A case study involving four North American deciduous forest species. Int. J. 

Biometeorol. 36:51-57. 

SCHLESINGER, M.E. and JIANG, X. 1991. Revised projection of future greenhouse warming. Nature 350: 219-221. 

SCHMIDTLING, R.C. 1994. Use of provenance tests to predict response to climatic change: loblolly pine and Norway 

spruce. Tree Physiol. 14: 805-817. 

SCRIBER, J.M. and GAGE, S.H. 1995. Pollution and global climate change: Plant ecotones, butterfly hybrid zones and 

changes in biodiversity, p. 319-344. In Scriber, J.M., Yoshitaka, T., and Lederhouse, R.C., eds. Swallowtail Ecology 

and Evolution. Scientific Publishers, Gainesville, FL. 

SHANDS, W.E. and HOFFMAN, J.S., eds. 1987. The Greenhouse Effect, Climate Change, and U.S. Forests. The Conservation 

Foundation, Washington, D.C. 304 p. 

SLINGERLAND, M.V. 1906. The bronze birch borer. Cornell Agric. Exp. Sta. Bull. 234: 63-78. 

SOLOMON, A.M. and BARTLEIN, RJ. 1992. Past and future climate change: Response by mixed-deciduous coniferous 

forest ecosystems in northern Michigan. Can. J. For. Res. 22: 1727-1738. 

SOLOMON, J.D. 1995. Guide to Insect Borers in North American BroadleafTrees and Shrubs. U.S. Dep. Agric. For. Serv. 

Agric. Handbk. AH-706. 735 p. 

SPAULDING, R and MACALONEY, H.J. 1931. A study of the organic factors concerned in the decadence of birch on cutover 

lands in northern New Englan d. J. For. 29:1134-1149. 

SYKES, M.T. and PRENTICE, C. 1995. Boreal forest futures: Modelling the controls on tree species range limits and 

transient responses to climate change. Water Air Soil Pollut. 82:415-428. 

SWAINE, J.M. 1918. A new forest insect enemy of white birch. Can. J. For. 14: 1928-1929. 

USGS (United States Geological Survey). 1991. National Water Summary 1988-1989 -- Hydrologic Events and Floods and 

Droughts. U.S. Geological Survey Water-Supply Paper 2375. 591 p. 

VAARTAJA, O. 1959, Evidence ofphotoperiodic ecotypes in trees. Ecol. Monogr. 29:91-111. 

246

WARGO, RM. and HAACK, R.A. 1991. Understanding the physiology of dieback and decline diseases and its management 

implications for oak, p. 147-158. In The Oak Resource in the Upper Midwest: Implications for Management Pro- 

ceedings, 3-6 June 1991, Winona, MN. Minnesota Extension Service, Univ. Minn. NR-BU-5663-S 

WILLIAMS, D.W. and LIEBHOLD, A.M. 1995. Forest defoliators and climate change: Potential changes in spatial distribu- 

tion of outbreaks of western spruce budworm (Lepidoptera: Tortricidae) and gypsy moth (Lepidoptera: 

Lyrnantriidae). Environ. Entomol. 24: 1-9. 

WILSON, L. F. and HAACK, R.A. 1990. The bronze birch borer. U.S. Dept. Agric. For. Serv., North Central For. Exp. Sta. 

Northern Hardwood Note 7.09. 3 p. 

WOODWARD, EI. 1992. A review of the effects of climate on vegetation: Ranges, competition, and composition, p. 105- 

123. In Peters, R.L. and Lovejoy, T.E., eds. Global Warming and Biological Diversity. Yale Univ. Press, New Haven, 

CT. 

WRIGHT, J.W. 1976. Introduction to Forest Genetics. Academic Press, New York. 

247

VARIATIONS IN SPRUCE NEEDLE CHEMISTRY AND IMPLICATIONS FOR 

THE LITTLE SPRUCE SAWFLY, PRISTIPHORA ABIETINA 

C. SCHAFELLNER, R. BERGER, J. MATTANOVICH, and E. FUHRER 

Institute of Forest Entomology, Forest Pathology, and Forest Protection 

Universit_it ftir Bodenkultur, Hasenauerstrasse 38, A 1190 Vienna, Austria 

INTRODUCTION 

The plant-herbivore system of Norway spruce, Picea abies Karst., and the little spruce Sawfly, Pristiphora abietina 

Christ (Hymenoptera: Tenthredinidae), provides a good opportunity to study the synchronization between the host's bud burst 

and the beginning of larval feeding. Adults of the sawfly emerge in spring at about the same time as spruce bud break and 

oviposit in the soft, partially-expanded needles. Every single spruce bud is suitable for oviposition for only a few days. The 

emerging larvae almost always remain on the expanding bud where they hatch and only under starving conditions do they 

switch buds. After 2 or 3 weeks, depending on weather conditions, the larvae are fully grown and drop to the ground to form 

cocoons. They overwinter in the litter until the next spring. 

The larvae of the little spruce sawfly are early-spring feeders and restricted to feeding on expanding spruce needles. 

Their success and survival depend on the exact phenological coincidence of bud break and oviposition, because larvae can 

initiate feeding only in flushing buds. Seasonal variations in the timing of bud break can therefore be of defensive value for 

the trees: trees that are genetically predisposed to flush needles earlier or later than the rest of the stand, thereby preceding or 

following the sawfly swarming, may escape larval feeding. 

During the last 3 decades the sawfly has become an important spruce pest in several parts of Austria. The most 

severe attacks occur in lowland areas where the natural deciduous forests have been replaced by monocultures of Norway 

spruce. As the larvae feed exclusively on current-year needles, the affected trees do not die, but repeated infestations lead to 

deformed, bushy crowns and to reduced height and volume increments. In recent years, an increase in frequency and 

duration of epidemics has been observed, particularly in areas exposed to air pollutants like sulfur and nitrogen (Sierpinski 

1985, Berger 1992). There is good reason to believe that some air pollutants may favor the sawfly's success, either by 

improving food quality due to changes in needle chemistry and/or by weakening the tree's defense system (Schafellner et al. 

1993, Schafellner et al. 1994, Berger und Katzensteiner 1994). 

This paper reports a study on the most important food quality parameters of newly emerging needles during the short 

period of larval feeding, and the changes that occur during the weeks of rapid growth. Additionally, the effect of excess 

nitrogen input (via fertilization) on needle chemistry and the nutritional value of the spruce needles for the feeding larvae is 

demonstrated. 

METHODS 

Three even-aged stands of trees 16 years old, were selected for study to see if food quality parameters show significant 

site-specific variation. Stand A is located at ca. 550 m elevation in Hochstrass, Lower Austria, and stands B and C are 

located at ca. 720 m elevation in the Hausruck, Upper Austria. Moderate sawfly attack was observed at stand A during the 

mid- 1980's. Throughout the 3 years of attack, tree flushing and sawfly attack for every individual tree was recorded (Holzer 

t988). The stand was attacked irregularly so that affected and unaffected individuals often stood close together, although the 

trees did not vary in the timing of bud break and should have been suitable for oviposition. From spring 1988 onwards, the 

sawflies vanished from the area so that during the 1989 sampling year, no frass was collected. Based on the data collected 

during the years of attack, 15 trees that had been affected once and 15 trees that remained unaffected, were selected for 

Mattson, W.J., Niemel_i, P., and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


248

investigation. Sampling was done on 3 occasions at 2-week-intervals during needle expansion in spring: needles were taken 

from twigs of the upper crown region, where the sawflies prefer to oviposit. To record differences between the physiologically 

older needles at the bottom and the physiologically younger needles at the top of a given shoot, on the second and third 

occasion needles were separated according to their location on the expanding bud: top (t), middle (m), and bottom (b) 

needles. 

In contrast to stand A, the trees on stand B were undergoing an actual heavy sawfly attack during the sampling period 

in spring 1989. Additionally, large variation in the timing of bud burst of the trees was found: the early flushing individuals 

flushed 4 weeks before the late flushing ones. Sawfly attack, however, occurred only on the late flushing trees. The early 

flushing trees (n = 10) were harvested 3 times (1,2,3), the late flushing trees (n = 10) only twice (la, 2a). After the second 

sampling of the attacked, late flushing trees (2a), the larvae were fully grown and were no longer feeding. Although the 

sampling on early and late flushing trees differed by more than 4 weeks, needles from sampling date 1 corresponded phenologically 

with 1a and needles from sampling date 2 corresponded with 2a. 

A fertilization experiment was done on stand C with late flushing, attacked trees. The stand was fertilized with 130 

kg nitrogen per ha/year, applied on 4 occasions: in summer 1989, in spring and summer 1990, and in spring 1991. Sampling 

was done during needle expansion in June, 1991, from 15 fertilized and 17 unfertilized (i.e., control) trees. 

Samples for chemical analyses were put into liquid nitrogen, then freeze-dried and milled. Nitrogen (total and 

soluble) was determined by micro-Kjeldahl technique with selenium as catalyst. Soluble carbohydrates (sugars, cyclitols) as 

well as the organic acids (quinic acid, shikimic acid) were determined by GLC, starch content was estimated enzymatically as 

glucose-equivalents. Needle phenolics were defined and determined as the potential of the needle extract (done with boiling 

50% aqueous methanol) to precipitate bovine serum albumin (BSA). Absorption was measured at 590 nm. Fiber (cellulose, 

hemicellulose, lignin) was estimated after a method to determine the fiber content of cattle food: after hydrolyzation with 

strong acids and alkalines and heating up to 550 degrees, the remaining ash residue was weighed. All values are expressed on 

a dry weight basis. 

Larval experiments were performed at stand C with late flushing, fertilized, and unfertilized trees in spring 1991. 

Larvae from naturally attacked buds as well as larvae (lst to 3rd instar) placed on expanding buds and enclosed in cages were 

studied. After cocoon formation, larvae were collected, sexed, and weighed. Parasitized larvae were excluded from the 

analyses. 

RESULTS 

During initial feeding in mid-May (bud length 4 cm) the young sawfly larvae are confronted with both high total and 

soluble nitrogen (2% N tot. and 0.2% N sol.) and organic acid concentrations (12%). On the other hand, carbohydrates 

(sugars and cyclitols) and starch are at rather low levels (8% and 2%). Fiber was approximately 10% of the needle dry 

weight (dw). For all these traits, damaged and undamaged trees did not differ (Fig. 1, upper graph). The crucial difference, 

however, was the phenolics: concentrations were significantly lower in the attacked trees than in the unattacked trees 

(p

40 ........................................ i.......................... T.......................... i........................... [........................... '..........=_ ......":.......................... 

! i LDa,fe 

lii;iiiiiiii i i cte 

30 ............................. i ......................................................................................... ................... 

_' 20 ............................. 

"0 

_ '!'!'!!!!!!!!!_!_!'! 

.............. ...... i 

10 I 

............ 

..... 

i 

]iiiii!i_iiiiii]ii_!_l 

iiEii::iiiiiiiiii!i!i I 

.... 

,_i_,_,_,E _ 

! !ii[::::iiii_ii!i_i!_ ........ 

, ii!iiiiii_i_iliiiiiii_i_l i!i!ii]i]iiii!]!iiii_;i! 

. ::ii_iiii_,i!i!i!i!i!i!i! 

0 i iiiii;iiii_;i!iii::i::iiil I ' _........... ::: :'_'_ 

bud phenolics* total soluble soluble starch organic fiber 

expansion nitrogen nitrogen carbohydr, acids 

cm mg mg 

0 ........................................................................................................................................................................................................................... 

30 ............................................................................................................. ........................... _ ....................................... D affected 

_'20 .......................... _............. _.......................... _.......................... _........................... r.......................... i............................ i.... 

l°una ecte°] 

Io OIl!iil ........ ................ ............. 

bud expansion phenolics** total nitrogen soluble soluble starch organic acids fiber 

cm mg nitrogen mg carbohydrate 

Figure 1.--Spruce needle chemistry during initial (upper graph) and at the end (lower graph) of larval feeding on stand A 

(identical flushing time). Asterisks indicate significant differences between once affected (n = 15) and unaffected 

(n = 15)trees (levels of significance: *p < 0.05, ** p < 0.01). 

250

Figure 2.--Changes in protein precipitating ability ("phenolics") of spruce needle extracts during the period of larval 

feeding on stand A (identical flushing time). Asterisks indicate significant differences between once-affected (n 

= 115)and unaffected (n = 15) trees (levels of significance: * p < 0.05, ** p < 0.0l). Dates of sampling: I) 18 

May 1989, 2) 1 June 1989, and 3) 14 June 1989. 

500 I 

"_ 400 - I ........" 

_) 

= -i 

- _.......... - ____ 

300 .................................................................................. 

I.....1....................................................................... 

O_ 

"5 200-t- 

< 

co 

o ! 

i 

1O0 ............ i 

0 

-i i 

1/la total 2/2a basal 2/2a top 3 basal 3 middle 3 top 

bud needles needles needles needles needles 

Figure 3.---Changes in protein precipitating ability ("phenolics") of"spruce needle extracts of early flushers (i.e., unaffected; 

n = 10), and late flushers (i.e., actually affected; n 10) during the period of needle expansion and larval 

feeding on stand B. Dates of sampling: early flushers: 1) 3 May 1989, 2) 17 May 1989, and 3) 1 June 1989; late 

flushers: la) 1 June 1989, and 2a) 14 June 1989. 

251

the control trees. Within the following 4 weeks, the phenolics decreased in the controls, but did not irathe attacked ones; 

concentrations being almost unchanged throughout larval feeding. 

The effect of N-fertilization on the expanding spruce needles was clear: total and soluble nitrogen were significantly 

higher in the fertilized trees and so were carbohydrates (p

q) 

J 

400 T J ! i i i i o control trees 

I _._1.................................................................................................................................................. [ I ......... 

300 ................... ................. ....... i i i................................................. 

B 

ca i 

J D [] I I 

caE200-)_......................................................................................................... 

N................................................................................................................... 

_. m 

Q3 J I 

I 

} := 

1,5 1,75 2 2,25 2,5 2,75 3 3,25 3,5 

% nitrogen dw 

Figure 5.mProtein precipitating ability Cphenolics") and nitrogen content of expanding spruce needles from individual 

fertilized (n = 15) and unfertilized (n = 17) trees. Date of sampling: 15 June 1991. 

I 

E 

-t [] 

[] 

D control trees 

I 

[] fertilized trees [] 

x:: [] ! 

"_ 14 .... [] 

12 .................................................................................................... 

_ ....................................................................... I 

c:_ [-3 

_- 

(D 

> 

[] 

10 ....................................................... _ ......................... ] ............. m ..................... ................................... 

' i 

0 50 100 150 200 250 300 

mg BSA precipitated/g needle dw 

Figure 6._The relationship between the protein precipitating ability ("phenolics") of expanding spruce needles from 

individual fertilized (n = 9) and unfertilized (n = 8) trees, and the average weights of female larvae (n = 30-40 per 

tree) feeding on these trees. 

253

1992). However, several authors have recorded higher levels of phenolics in immature than in rnature leaves (Coley 1983, 

Puttick 1986, Mauffette and Oechel 1989, Hatcher 1990), as did the present study, but finally a gradual increase in concentration 

occurred with needle maturation (Schafellner et al. 1993). The decline of phenolics during the first month of needle 

expansion may be attributed to a dilution effect resulting from the addition of other leaf constituents and/or as phenolics are 

important precursors to lignin, to an incorporation into the cell wall. 

Flushing needles of Norway spruce have the highest concentrations of key nutrients (total nitrogen, amino acids, 

water), however, they also contain the highest concentrations of secondary compounds (quinic acid, phenolics). Although 

nitrogen is an important factor for growth and survival and positive correlations between needle carbohydrates and insect 

performance have been demonstrated (Schopf 1986, Jensen 1988), we propose that between-tree differences in needle 

phenolics caused the observed pattern of sawfly attack. The amount of phenolics in the newly emerging needles may be 

crucial for the feeding larvae. High phenolic levels suggest a potential defense mechanism of the host tree and seems to be 

determined more by intraspecific genetic differences than by site-dependent or seasonal factors (Lunderstfidt 1980). Additionally, 

we found no hint that needle quality deteriorated for the years after sawfly attack: nitrogen and carbohydrates did 

not decline and phenolics did not increase. 

Fertilization with N increased the nutritional value of spruce needles for P. abietina larvae. This increase in nutritional 

value is based on an increase in the concentration of needle nitrogen and carbohydrates and a reduction in needle 

phenolics. Our results are consistent with several reports that N fertilization increases the concentration of nitrogen and 

reduces the concentrations of carbon-based secondary compounds in leaves of woody plants (Waring et al. 1985, Haukioja et 

al. 1985, Glyphis and Puttick 1989) and that such changes in leaf chemistry are causally related to changes in leaf nutritional 

value for immature insects (Bryant et al. 1987). 

An oversupply of N tends to change the metabolism of carbohydrates and nitrogen. Probably because of a higher 

demand for carbohydrates during N-absorption and N-assimilation, the phenolic content of spruce needles decreases (Tuomi 

et al. 1990), whereas the concentration of total nitrogen and free amino acids increase with N fertilization or N deposition 

(N_isholm and Ericsson 1990). The low concentrations of phenolics in the leaves of fertilized trees may explain the increased 

insect susceptibility of trees exposed to N deposition and excess N (Berger and Katzensteiner 1994). 

In our study N fertilization also caused a significant increase in the needle concentrations of soluble carbohydrates 

(sugars, cyclitols), probably due to a stimulated growth of photosynthetic tissue, and a slight decrease in starch. N fertilization 

did not cause changes in beech foliar concentrations of total sugars, starch or soluble protein, but total phenolic compounds 

decreased by about a third (Balsber-Pahlsson 1992). Changes in total sugar content were found in the current-year 

needles of Scots pine (Ericsson 1979). 

Our results indicate that a shift in the carbon/nutrient balance of spruce by N fertilization increases the nutritional 

value of the needles for P. abietina larvae in two ways: by increasing the concentration of nutritional chemicals (carbohydrates, 

nitrogen), and by reducing the concentrations of toxic and/or digestion-inhibiting chemicals (phenolics) in the needles. 

The experiment demonstrated that nitrogen fertilization enhanced larval growth. Low needle nitrogen concentrations and 

high concentrations of phenolics have been correlated with reduced larval weights of P. abietina feeding upon fertilized and 

unfertilized spruce trees. This supports the idea that phenolics are specific agents for protein deactivation. 

It has often been shown that the effect of air pollution may be favorable to insect herbivores (Holopainen et al. 1991 ). 

Excess nitrogen supply (via atmospheric N-input) may promote the sawfly's success by improving its food quality: stimulated 

tree growth with high carbohydrate and nitrogen levels at the cost of limited allocation of resources for the synthesis of 

defensive compounds (phenolics). 

SUMMARY 

The present study on the plant-herbivore system Norway spruce, Picea abies, and the little spruce sawfly, Pristiphora 

abietina, illustrates variations in spruce needle chemistry as a potential factor for successful insect attacks. While nitrogen, 

carbohydrates, starch, organic acids, fiber, and water content of the new emerging needles showed no significant betweentree 

or site-specific differences during the time of larval feeding, the levels of the needle phenolics were significantly lower in 

254

attacked trees than in control trees. Successful larval feeding seems to be only possible on trees with low needle phenolics. 

High phenolic concentrations during early larval feeding suggest an effective defense mechanism of the host tree and seems 

to be genetically determined. 

A significant negative correlation between needle nitrogen and phenolics in individual trees as well as the positive 

effect of nitrogen and the negative effect of phenolics on larval weight indicate a causal relationship between nitrogen and 

phenolics and larval growth. 

We also demonstrated the effects of excess atmospheric nitrogen input via N fertilization upon different chemical 

properties in spruce needles and the quality of these expanding needles as food for the larvae. Nitrogen fertilization resulted 

in decreased phenolics, starch, organic acids, and fiber, increased nitrogen and carbohydrates, and in a better food value for 

the sawfly larvae. Larval weight was negatively correlated with needle phenolic concentrations. Our findings indicate that 

the quality of spruce needles as food for the P abietina larvae are influenced in two ways: by increasing the concentrations 

of positive factors (e.g., nitrogen, carbohydrates) and decreasing the concentrations of negative factors (e.g., phenol ics, fiber, 

organic acids). 


This study was funded by the Austrian Fonds of Scientific Research, P 7093-CHE: Excess Nitrogen in Forest 

Plant-Herbivore Associations. The authors wish to thank Ing. G. Mottlik, A. Stradner, and Mag. A. Straka for technical 

assistance with the chemical analyses. 


BALSBERG-PAHLSSON, A.M. 1992. Influence of nitrogen fertilization on minerals, carbohydrates, amino acids and 

phenolic compounds in beech (Fagus sylvatica L.) leaves. Tree Physiol. 10: 93-100. 

BERGER, R. 1992. Massenauftreten der Kleinen Fichtenblattwespe Pristiphora abietina (Christ) im Hausruck. Teil I: 

Okologische Rahmenbedingungen in einem "untypischen" Befallsgebiet. Anz. Sch_idlingskde., Pflanzenschutz, 

Umweltschutz 65: 105-114. 

BERGER, R. and KATZENSTEINER, K. 1994. Massenauftreten der Kleinen Fichtenblattwespe Pristiphora abietina 

(Christ) im Hausruck. Teil II: Immissions6kologische Einfltisse. J. appl. Ent. (in press). 

BRYANT, J.E, CHAPIN, F.S., KLEIN, D.R. 1983. Carbon/nutrient balance of boreal plants in relation to vertebrate 


BRYANT, J.E, CLAUSEN, T.E, REICHARDT, EB., McCARTHY, M.C., and WERNER, R.A. 1987. Effect of nitrogen 

fertilization upon the secondary chemistry and nutritional value of quaking aspen (Populus tremuloides Michx.) 

leaves for the large aspen tortrix (Choristoneura conflictana (Walker)). Oecologia 73:513-517. 

COLEY, ED. 1983. Herbivory and defensive characteristics of tree species in a lowland tropical forest. Ecol. Monogr. 53: 

209-233. 

ERICSSON, A. 1979. Effects of fertilization and irrigation on the seasonal changes of carbohydrate reserves in different 

age-classes of needle on 20-year-old Scots pine trees (Pinus sylvestris). Physiol. Plant. 45: 270-280. 

FEENY, EE 1970. Seasonal changes in oak leaf tannins and nutrients as a cause of spring feeding by winter moth caterpillars. 

Ecology 51:565-581. 

FEENY, EE 1976. Plant apparency and chemical defense. Recent Adv. Phytochem. 19: 1-40. 

255

GLYPHIS, J.R and PUTTICK, G.M. 1989. Phenolics, nutrition and insect herbivory in some garrigue and maquis plant 

species. Oecologia 78: 259-263. 

HATCHER, RE. 1990. Seasonal and age-related variation in the needle quality of five conifer species. Oecologia 85: 200- 

212. 

HAUKIOJA, E., NIEMELA, R, and SIREN, S. 1985. Foliage phenols and nitrogen in relation to growth, insect damage, 

and ability to recover after defoliation, in the mountain birch Betula pubescens ssp tortuosa. Oecologia 65:214-222. 

HOLOPAINEN, J.K., KAINULAINEN, E., OKSANEN, J., WULFF, A., and KARENLAMPI, L. 1991. Effect of exposure 

to fluoride, nitrogen compounds and SO 2on the numbers of spruce shoot aphids on Norway spruce seedlings. 

Oecologia 86: 51-56. 

HOLZER, K. and SCHULTZE, U. 1988. Die Bedeutung yon Herkunft und Austrieb bei einem Fichtenblattwespenbefall. 

Cbl. ges. Forstwesen 105:207-216. 

JENSEN, T.S. 1988. Variability of Norway spruce (Picea abies L.) needles; performance of spruce sawflies (Gilpinia 

hercyniae Htg.). Oecologia 77:313-320. 

LUNDERST,ADT, J. 1980. Zur 6kophysiologischen Bedeutung yon Phenolen und Proteinen in Nadeln der Fichte (Picea 

abies Karst.). Z. Pflanzenern_ihr. Bodenkd. 143: 412-421. 

MATTSON, W.J. 1980. Herbivory in relation to plant nitrogen content. Ann. Rev. Ecol. Syst. 11: 119-161. 

MAUFFETTE, Y. and OECHEL, W.C. 1989. Seasonal variation in leaf chemistry of the coast live oak Quercus agrifolia 

and implications for the California oak moth Phryganidia californica. Oecologia 79: 439-445. 

NASHOLM, T. and ERICSSON, A. 1990. Seasonal changes in amino acids, protein and total nitrogen in needles of fertilized 

Scots pine trees. Tree Physiol. 6: 267-281. 

PUTTICK, G.M. 1986. Utilization of evergreen and deciduous oaks by the Californian oak moth, Phryganidia californica. 

Oecologia 68: 589-594. 

SCHAFELLNER, C., BERGER, R., MATTANOVICH, J., and FOHRER, E. 1993. Pristiphora abietina (Hym., 

Tenthredinidae) - ein Bioindikator ftir Luftverschmutzung? Besonderheiten unreifer Fichtennadeln als Larvenfutter. 

Forstw. Cb. 112: 116-128. 

SCHAFELLNER, C., BERGER, R., MATTANOVICH, J., and Ff0HRER, E. 1994. Food quality of spruce needles and the 

performance of the Little Spruce Sawfly, Pristiphora abietina Christ (Hym., Tenthredinidae). The protein precipitating 

ability of the young needles. Acta horticulturae (in press). 

SCHOPE R. 1986. The effect of secondary needle compounds on the development of phytophagous insects. For. Ecol. 

Man. 15: 55-64. 

SCRIBER, J.M. and SLANSKY, E 1981. The nutritional ecology of immature insects. Ann. Rev. Entomol. 26: 183-211. 

SIERPINSKI, Z. 1985. Luftverunreinigungen und Forstsch_idlinge. J. appl. Ent. 99: 1-6. 

TUOMI, J., NIEMEL,_, R, and SIREN, S. 1990. The Panglossian paradigm and delayed inducible accumulation of foliar 

phenolics in mountain birch. Oikos 59: 399-410. 

WARING, R.H., McDONALD, A.J.S., LARSSON, S., ERICSSON, T., WIREN, A., ARWIDSSON, E., ERICSSON, A., and 

LOHAMMAR, T. 1985. Differences in chemical composition of plants grown at constant relative growth rates with 

stable mineral nutrition. Oecologia 65: 157-160. 

256

ACIDIC DEPOSITION, DROUGHT, AND INSECT HERBIVORY IN AN ARID 

ENVIRONMENT: ENCELIA FARINOSA AND TRIRHABDA GEMINATA IN 

SOUTHERN CALIFORNIA 

T.D. PAINE, R.A. REDAK, and J.T. TRUMBLE 

Department of Entomology, University of California, Riverside, California 92521, USA 

INTRODUCTION 

Except at the higher elevations, much of the arid southwestern United States and northern Mexico is not forested. 

The plant communities are dominated by widely spaced shrubs that are 1-3 m in height. Because these areas receive less than 

20 cm of annual precipitation and mean maximum summer temperatures of 38-43 °C, plant growth is often limited by the 

availability of moisture. Many of the species that are components of the Creosote Bush Scrub community are droughtdeciduous 

and remain dormant during the dry periods (Munz and Keck 1968). Even in the slightly more moderate Coastal 

Sage Scrub community that receives up to 40 cm of annual precipitation and less extreme mean maximum summer temperatures, 

the plant species are adapted to extended periods without precipitation. Growth and flowering are often limited to 

those periods with available moisture. 

However, as the population of urban areas adjacent to these biotic communities grows, plants in those communities 

are subject to additional stresses arising from human activities. One of the most important anthropogenic stresses on plants is 

air pollution (Cowling 1982, Lee 1982, Linthurst et al. 1982, Treshow 1984, Shriner 1986). Although there are many 

different atmospheric pollutants capable of affecting plants, including ozone, NO x, SO, and PAN (peroxyacetyl nitrate), the 

impact on plants may be enhanced when the pollutants are combined with moisture. Wet acidic deposition in the form of 

acidic fog is of particular concern in southern California. Acidic fog episodes in southern California last from 2 to 12 hours 

and often have pH's of 2.0-3.0 (Waldman et al. 1982). Physical changes in leaf surface structure, lesions, leaching of foliar 

nutrients, decreased photosynthesis, and reduced growth and yield may result from long-term exposure to acidic fogs 

(Granett and Taylor 1981, Granett and Musselman 1984, Musselman and Sterrett 1988, Musselman and McCool 1989, 

Paoletti et al. 1989, Takemoto et al. 1989, McCool and Musselman 1990, Mengel et al. 1990, Trumble and Walker 1991). 

Pollution-stressed plants also may be more susceptible to insects and diseases (Endress and Post 1985, Trumble et al. 1987, 

Trumble and Hare 1989, Jones and Coleman 1988, Paine et al. 1993). Consequently, if acidic deposition in the form of acidic 

fog occurs when the plants are actively growing and when the herbivores also are active, the pollution stress may have a 

significant impact on the fitness of the plant. 

We have studied the influence of acidic deposition on the interactions between brittle brush, Enceliafarinosa Gray 

(Asteraceae), a dominant shrub in the Creosote Bush Scrub and Coastal Sage Scrub communities of southern California, and 

its primary herbivore, Trirhabda geminata Horn (Coleoptera: Chrysomelidae), a leaf beetle that feeds in both larval and adult 

stages on E. farinosa. The system is important because it is both widespread and an indicator system for damage that may be 

occuring in these fragile communities occupying harsh environments. The impact of pollution stresses on the interactions 

between plant and herbivore has been much less studied in desert than in forest ecosystems (e.g., Johnson and Siccama 1983, 

Jones and Coleman 1988, Mengel et al. 1990); however, there may be much less environmental buffering in these desert 

communities for additional stress imposed by anthropogenic encroachment. 

Acidic Deposition as a Single Stress 

Our previous studies have demonstrated significant effects of acidic deposition on the nutritional quality and palatability 

of E. farinosa to Z geminata in laboratory tests. Plants fogged three times for 3 hours each time at pH 2.75 were 

compared to plants fogged with a control fog at pH 5.80. Acidic-fogged foliage had significantly higher concentrations of 

Mattson, W.J., Niemel_i, E, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


257

total nitrogen (2.33%) and higher soluble protein content (4.96 mg/g leaf tissue) when compared to leaves from controlfogged 

plants (1.94% total nitrogen and 3.94 mg/g leaf tissue soluble protein) (Paine et al. 1993). There were no differences 

between treatments in water content of the leaf tissues (74.8% for acidic-fogged plants and 75% for control-fogged plants). 

Both larvae and adult beetles preferred to feed on leaf tissue from acid-fogged plants (Paine et al. 1993). The studies 

demonstrated that the changes in Encelia foliage that alter insect consumption and growth indices can occur within 7 days of 

treatment with acid fogs. 

Acidic deposition has an indirect effect through E. farinosa upon the growth rate and biomass of T. geminata. Acidic 

fogging (pH 2.75 vs. a control fog of pH 5.8) orE. farinosa resulted in a 34% increase in average larval biomass gain 

(Control = 2.23 mg vs. Acid= 2.98 mg) and a 3 I% increase in larval growth rate (Control= 0.16 mg'mgl'd _vs. 

Acid=0.21 mg, mg_-d_). The effect of fogging on growth rate and biomass gain was strongest for insects initiating feeding 14 

days after treatment applications. The effect of acidic fogging and the duration of the effect (days post treatment) were 

independent of each other (no interaction); the effect of acidic fogging upon growth rate and biomass gain was statistically 

constant through time. For adult beetles, tissue consumption of acidic-fogged foliage was 65% greater than tissue consumption 

of control foliage (Control = 0.0964 cm2 vs. Acid = 0.16 cm2). For larval beetles, tissue consumption of acidic-fogged 

foliage was over 400% greater than control foliage (Control = 0.057 cm2 vs. Acid = 0.259 cm 2) (Redak et al. In Prep.). 

Simultaneous Impact of Acidic Deposition and Drought as Multiple Stresses 

Drought stress has been shown to mitigate the impact of gaseous air pollutants and, to a certain extent, acidic wet 

deposition (acid rain and fog) on plant physiological and growth processes (Tingey and Hogsett 1985, Meier et al. 1990, 

Bender et at. 1991, Showman 199l, Beyers et al. 1992, Temple et al. 1992). Our field data investigating the impact of acidic 

fog upon the E. farinosa-Z geminata system are consistent with the theory that drought stress may ameliorate the impact of 

acidic fog upon the susceptibility of plants to insect herbivory (Warrington and Whittaker 1990, Von Sury and Fluckiger 

1991). Drought stress may make the plant less susceptible to the effects of the air pollution, or alternatively, the two stresses 

may act independently but antagonistically on the plant so that the net effect on herbivore success is neutral. Using an 

undeveloped area (remnant coastal sage-scrub habitat) on the UCR campus, we applied acidic (pH = 2.0) or control (pH = 

5.8) fog treatments (3, 3-hour exposures, every other day for 5 days) to 40 mature E. farinosa plants (20 plants per treatment 

application). Following tog exposure, we caged 50 second instar T.geminata larvae in nylon screen bags on each treatment 

plant. Larvae were allowed to feed and grow on the plants for 16 days at which time they were removed and their fresh and 

dry masses were determined. Additionally, 7 days after fog treatments, we conducted a 72-hour laboratory bioassay determining 

the consumption and growth rates of third instar T. geminata fed foliage taken from the field treated plants. After 

removal of experimental insects, foliage was collected from each plant and analyzed for total N and percent water. 

Acidic fogging showed no effect on larval growth or consumption values or any plant foliage quality values (Table 

1). For the field experiment described here, we feel the drought status of the experimental plants may account for the lack of 

treatment effects (acidic fog effects) upon insect performance. At the time of the experiment, southern California was in the 

sixth to seventh year of a sustained drought. For the year of this study, 1990, the UCR area had received less than 15% 

(approx. 40-50 ram) of normal precipitation by the end of the growing season (CIMIS weather data). Indeed approximately 1 

week after treatment applications, several experimental plants (of both treatments) began to show characteristic signs of 

drought-induced leaf senescence (leaf curling, browning, dropping older more mature leaves), and by the second week 

following removal of experimental insects, all of the plants were undergoing severe leaf-loss. Enceliafarinosa normally is a 

drought deciduous plant (Ehleringer and Clark 1988); however, in this case, the period of leaf-drop began approximately 6 

weeks to 2 months earlier than normal. Given the sustained 6 years of drought and the very early period of leaf abscission, 

we are confident that all experimental plants were experiencing relatively severe drought stress. 

The response of E. farinosa to drought stress includes (among other processes) an increase in leaf pubescence 

(Ehleringer and Bjorkman 1978, Ehleringer 1982). We suspect that (1) the experimental plants were so severely drought 

stressed that additional stress, in the form of acidic fog deposition, may have no further influence on the plants' suitability as 

a host for 7: geminata and (2) the morphological changes that occur in E. farinosa with drought stress are such that the plant 

may be physically protected to some extent from the impact of acidic fog (i.e., drought is functioning as a "filter" to the 

impact of acidic fog). The dense covering of pubescence (as a result of drought stress) could have physically shielded and/or 

buffered the plant from acid tbg deposition (Caporn and Hutchinson 1987, Musselman 1988). 

258

Table 1.--Impact of acidic fog upon the foliage quality of Enceliafarinosa and performance of Trirhabda gerninata _._ 

A. Field Bioassay. 

Final Insect Total Nitrogen Total Water 

Biomass (rag) Content (%) Content (%)3 

Fog pH 2.00 0.()049+0.0001 2.83+0.06 60.82+0.54 

Fog pH 5.80 0.0049_+0.0001 2.80+0.07 60.08+0.54 

B. Laboratory Bioassay. 

Insect Biomass Insect Growth Insect Consumption 

Gain (mg) Rate (rag.tugt.d: _) Rate (mg.mg_*d 1) 

Fog pH 2.00 0.0027+0.0002 0.9008+0.0816 27.729+4.171 

Fog pH 5.80 0.0026-+0.0003 0.8252+0.0826 26.179-+4.259 

There were no significant differences in any of the measured parameters between acid- and control-fogged treatments, 

ANOVA. 

2Values shown are means + 1 standard error. 

3Analyses carried out on arcsine transformed data. Values shown are back transformed. 

CONCLUSIONS 

The implications of altered host plant quality for insect herbivore development are considerable (Rhoades 1983, 

Strong et al. 1984, White 1984). Researchers have clearly demonstrated that nitrogen availability and plant secondary 

chemistry affect such basic insect life processes as growth rates, survival, and reproductive capacity (Onuf 1978, Rosenthal 

and Janzen 1979, Prestidge 11982,Prestidge and McNeill 1983). Since acidic fog incidents typically occur over broad 

geographical areas (Perkins 1974, Hoffman et al. 1985), the cumulative effects on insect herbivores at the population level 

may be quite significant. Thus, potential changes :in plant physiology due to stress or direct injury resulting from acidic fog 

may well have more serious consequences than previously believed. The dynamics of herbivorous insect populations 

associated with areas subjected to acidic fogs may be dramatically affected depending on the type and magnitude of the 

chemical changes that can occur in plants subjected to this type of pollution. 

By using E. farinosa and its most common natural herbivore, Z geminata, we are able to draw from and build upon a 

relatively thorough understanding of the relationships between the effects of acidic fog on E. farinosa foliage quality and T. 

geminata success. Previous studies have shown that the major defensive compounds in E. farinosa are the sesquiterpene 

farinosin and the two chromenes euparin and encecalin (Wisdom and Rodriguez 1982, 1983; Wisdom et al. 1983; Wisdom 

1985, 1988). These compounds reduce growth rates, survivorship, and quantity of food consumed for early instar larvae of T. 

geminata. Additionally the foliar concentrations of these compounds are highest at that time of year when early instars are 

present and actively feeding. We are currently investigating whether or not the concentrations of these compounds are 

affected by exposure to acidic fogs. Changes in the concentrations of the secondary chemistry of the plants may partially 

explain the results presented here. However, Wisdom (1985) showed that T. geminata larvae apparently feed on tissues high 

in total nitrogen content. We have shown that acidic fog exposure may change primary plant chemistry, resulting in increased 

plant total nitrogen and soluble protein. These increases are associated with increased larval preference (Paine et al. 1993) 

and may also be associated with better larval performance. If the herbivore consumes more plant tissue as a direct consequence 

of an anthropogenic stress and decreases plant fitness, there may be a significant impact on the plant community if the 

stress continues for a significant period of time. Although drought may ameliorate the impact of the anthropogenic stress, the 

long-term consequences of the pollution stress on the ecosystem are not clear. 

259


BENDER, J., TINGEY, D.T., JAGER, H.J., RODECAE K.D., and CLARK, C.S. 1991. Physiological and biochemical 

responses of bush bean (Phaseolus vulgaris) to ozone and drought stress. J. Plant Physiol. 137: 565-570. 

BEYERS, J.L., RIECHERS, G.H., and TEMPLE, P.J. 1992. Effects of long-term ozone exposure and drought on the 

photosynthetic capacity of ponderosa pine (Pinus ponderosa Laws.). New Phytol. 122:81-90. 

CAPORN, S.J.M. and HUTCHINSON, T.C. 1987. The influence of temperature, water and nutrient conditions during 

growth on the response of Brassica oleracea L. to a single, short treatment with simulated acid rain. New Phytol. 

106:251-259. 

COWLING, E.B. 1982. A status report on acid precipitation and its biological consequences as of April 1981. In D'itri, F., 

ed. Acid Precipitation-Effects on Ecological Systems. Ann Arbor Press, Ann Arbor, MI. 

EHLERINGER, J.R. 1982. The influence of water stress and temperature on leaf pubescence development in Encelia 

farinosa. Am. J. Bot. 69: 670-675. 

EHLERINGER, J.R. and BJORKMAN, O. 1978. A comparison of photosynthetic characteristics of Encelia species possessing 

glabrous and pubescent leaves. Plant Physiol. 62: 185-190. 

EHLERINGER, J.R. and CLARK, C. 1988. Evolution and adaptation in Encelia (Asteraceae), p. 221-248. In Gottlieb, L. 

and Jain, S., eds. Plant Evolutionary Biology. Chapman and Hall, New York. 

ENDRESS, A.G. and POST, S.L. 1985. Altered feeding preference of Mexican bean beetle Epilachna varivestis for 

ozonated soybean foliage. Environ. Pollut. 39: 9-16. 

GRANETT, A.L. and MUSSELMAN, R.C. 1984. Simulated acidic fog injureslettuce. Atmosph. Environ. 18: 887-890. 

GRANETT, A.L. and TAYLOR, O.C. I981. The effect of designated pollutants on plants. Fifth annual report, AMRL-TR- 

80-112, USAF Aerospace Med. Res. Lab., Wright-Patterson AFB, Ohio. 

HOFFMAN, M.R., JACOB, D.J., WALDMAN, J. M., MUNGER, J.W., and FLAGAN, R.C. 1985. Characterization of 

reactants, reaction mechanisms, and reacting products in atmospheric water droplets: fog, cloud, dew and rain water 

chemistry. Final report for contractARB No. AZ-048-32. California Air Resources Board. 

JOHNSON, A.H. and SICCAMA, T.G. 1983. Acid deposition and forest decline. Environ. Sci. Technol. 17: 299A-304A. 

JONES, C.G. and COLEMAN, J.S. 1988. Plant stress and insect behavior: Cottonwood, ozone and the feeding and oviposition 

preference of a beetle. Oecologia 76:51-56. 

LEE, J.J. 1982. The effects of acid precipitation on crops. In D'itri, E, ed. Acid precipitation-effects on ecological systems. 

Ann Arbor Press, Ann Arbor, MI. 

LINTHURST, R.A., BAKER, J.P., and BARTUSKA, A.M. 1982. Effects of acidic deposition: a brief review. In Frederick, 

E, ed. A Specialty Conference on Atmospheric Deposition. Air Pollution Control Association, Pennsylvania. 

MCCOOL, P.M. and MUSSELMAN, R.C. 1990. Injury of three ornamental flower crops from simulated acidic fog. Plant 

Dis. 74: 310-312. 

MEIER, S., GRAND, L.E, SCHOENEBERGER, M.M., REINERT, R.A., and BRUCK, R.I. 1990. Growth, ectomycorrhizae 

and nonstructural carbohydrates of loblolly pine seedlings exposed to ozone and soil water deficit. Environ. Poll. 64: 

11-27. 

260

MENGEL, K., BREININGER, T., and LUTZ, H.J. 1990. Effect of simulated acidic fog on carbohydrate leaching, CO 2 

assimilation and development of damage symptoms in young spruce trees (Picea abies L. Karst). Environ. and Exp. 

Bot. 30: 165-173. 

MUNZ, P.A. and KECK, D.D. 1968. A California flora and supplement. University of California Press, Los Angeles. 

MUSSELMAN, R.C. 1988. Acid neutralizing capacity of leave exposed to acidic fog. Environ. and Exp. Bot. 28: 27-32. 

MUSSELMAN, R.C. and MCCOOL, EM. 1989. Effects of acidic fog on productivity of celery and lettuce and impact on 

incidence and severity of diseases. Ann. Appl. Biol. 114: 559-565. 

MUSSELMAN, R.C. and J. L. STERRETT, J.L. 1988. Sensitivity of plants to acidic fog. J. Environ. Qual. I7: 329-333. 

ONUF, C.R 1978. Nutritive value as a factor in plant-insect interactions with an emphasis on field studies. In Montgomery, 

G., ed. The Ecology of Arboreal Folivores. Smithsonian Institution Press, Washington, DC. 

PAINE, T.D., REDAK, R.A., and TRUMBLE, J.T. 1993. Impact of acidic deposition on Enceliafarinosa Gray 

(Compositae: Asteraceae) and feeding preferences of Trirhabda geminata Horn (Coleoptera: Chrysomelidae). J. 

Chem. Ecol. 19:97-105 

PAOLETTI, E., GELLINI, R., and BARBOLANI, E. 1989. Effects of acid fog and detergents on foliar leaching of cations. 

Water, Air, and Soil Poll. 45: 49-61. 

PERKINS, H.C. 1974. Air Pollution. McGraw-Hill, New York. 

PRESTIDGE, R.A. 1982. Instar duration, adult consumption, oviposition and nitrogen utilization efficiencies of leafhoppers 

feeding on different quality food (Auchenorrhyncha: Homoptera). Ecol. Entomol. 7: 91-101. 

PRESTIDGE, R.A. and MCNEILL, S. 1983. The importance of nitrogen in the ecology of grassland Auchenorrhyncha. In 

Lee, J.A., McNeili, S., and Rorison, I.H., eds. Nitrogen as an Ecological Factor. 22nd Symp. Brit. Ecol. Soc. 

Biackwell Scientific Publ., Oxford. 

REDAK, R.A., PAINE, T.D., and TRUMBLE, J.T. In prep. A determination of the direct impact of acidic fogging upon the 

success (mortality and growth) of Trirhabda geminata (Coleoptera: Chrysomelidae). for Oecologia. 

RHOADES, D.F. 1983. Herbivore population dynamics and plant chemistry. In Denno, R.F. and McClure, M.S., eds. 

Variable Plants and Herbivores in Natural and Managed Systems. Academic Press, London. 

ROSENTHAL, G.A. and JANZEN, D.H. 1979. Herbivores: Their Interaction with Secondary Plant Metabolites. Academic 

Press, London. 

SHOWMAN, R.E. 1991. A comparison of ozone injury to vegetation during moist and drought years. J. Air Waste Manage. 

Assoc. 41" 63-64. 

SHRINER, D.S. 1986. Terrestrial ecosystems: wet deposition. In Legge, A.H. and Rupa, S.V., eds. Air pollutants and their 

effects on the terrestrial ecosystem. Advances in Environmental Science and Technology, v 18. John Wiley and 

Sons, New York. 

STRONG, D.R., LAWTON, J.H., and SOUTHWOOD, R. 1984. Insects on Plants. Harvard University Press, Cambridge, 

MA. 

TAKEMOTO, B.K., JOHNSON, A.G., PARADA, C.R., and OLSZYK, D.M. 1989. Physiology and yield of field-grown 

Brassica oleracea L. exposed to acidic fog. New Phytol. 112: 369-375. 

261

TEMPLE, RJ., RIECHERS, G.H., and MILLER, RR. 1992. Foliar injury responses of ponderosa pine seedlings to ozone, 

we and dry acidic deposition, and drought. Environ. and Exp. Bot. 32:101-113. '_ 

TINGEY, D.T. and HOGSETT, W.E. 1985. Water stress reduces ozone injury via a stomatal mechanism. Plant Physiol. 77" 

944-947. 

TRESHOW, M. 1984. Air Pollution and Plant Life. John Wiley and Sons, New York. 

TRUMBLE, J.T. and HARE, J.D. 1989. Acidic fog-induced changes in host-plant suitability. J. Chem. Ecol. 15: 2379-2390. 

TRUMBLE, J.T. and WALKER, G.E 1991. Acute effects of acidic fog on photosynthetic activity and morphology of _ 

Phaseotus lunatus. HortScience 26:1531-1534. 

TRUMBLE, J.T., HARE, J.D., MUSSELMAN, R.C., and MCCOOL, E 1987. Ozone-induced changes in host-plant 

suitability: interactions of Keiferia lycopersicella and Lycopersicon esculentum. J. Chem. Ecol. 13:203-218. 

VON SURY, R. and FLUCKIGER, W. 1991. Effects of air pollution and water stress on leaf blight and twig cankers of 

London planes (Platanus X acerifolia (Ait.) Willd.) caused by Apiognomonia veneta (Sacc. and Speg.) Hohn. New 

Phytol. 118: 397-405. 

WALDMAN, J.M., MUNGER, J.W., JACOB, D.J., FLAGEN, R.C., MORGAN, J.J., and HOFFMAN, M.R. 1982. Chemical 

composition of acid fog. Science 218: 677-680. 

WARRINGTON, S. and WHITTAKER, J.B. 1990. Interactions between sitka spruce, the green spruce aphid, sulphur 

dioxide pollution and drought. Environ. Poll. 65: 363-370. 

WHITE, T.C.R. 1984. The abundance of invertebrate herbivores in relation to the availability of nitrogen in stressed food 

plants. Oecologia 63" 90-105. 

WISDOM, C. 1985. Use of chemical variation and predation as plant defenses by Enceliafarinosa against a specialist 

herbivore. J. Chem. Ecol. 11" 1553-1565. 

WISDOM, C. 1988. Comparisons of insect use and chemical defense patterns of two Sonoran desert shrubs, p. 36-49. In 

Zahary, R.G., ed. Desert Ecology 1986: A Research Symposium. Southern California Academy of Sciences and the 

Southern California Desert Studies Consortium. 

WISDOM, C. and RODRIGUEZ, E. 1982. Quantitative variation of the sesquiterpene lactones and chromenes of Encelia 

farinosa. Biochem. Syst. and Ecol. 10: 43-48. 

WISDOM, C. and RODRIGUEZ, E. 1983. Seasonal age-specific measurements of the sesquiterpene lactones and 

chromenes of Enceliafarinosa. Biochem. Syst. Ecol. 11"43-48. 

WISDOM, C.S., SMILEY, J.T., and RODRIGUEZ, E. 1983. Toxicity and deterrency of sesquiterpene lactones and 

chromenes to the corn earworm (Lepidoptera: Noctuidae). J. Econ. Entomol. 76: 993-998. 

262

THE RESISTANCE OF SCOTCH PINE TO DEFOLIATORS 

V.I. GRIMALSKY 

Research Institute of Forestry, Gemol, Belorussia 

SUMMARY 

Scotch pine, Pinus sylvestris, is the most widely distributed tree species in Eurasia. It is often damaged by several 

defoliating insects: Dendrolimus pini, Panolisflammea, Bupalus piniarius, and Diprion pini. The foci of mass outbreaks of 

these insects typically occur in pure pine stands on poor, dry sandy soils, and on richer soils (loamy sands and sandy loams) 

depleted by long agricultural use. 

Chemical analyses of needles showed that N contents are always higher in pines growing outside (i.e., on richer, 

moister soils) than inside the foci of mass outbreaks. With respect to sugars, no differences were found. Larvae on trees 

enclosed in gauze bags, showed a high mortality of early instars (I-III) in stands on rich, moist soils. Young larvae cause 

oleoresin droplets to fbrm where they chew into needles, the mechanism Ofpine resistance to them. Such exudations are 

weak from trees on poor, dry soils. Therefore, larvae can feed on them without impediment, and their mortality is low. 

To monitor this resistance, the ends of living needles (about 1/3 - 1/4 their length) on live trees, were cut. After 5 

minutes the clipped needles were scored as follows: 0 - oleoresin does not exudate from the needle; 1 - a thin film of 

oleoresin appears on the cut, or separate tiny droplets which do not converge to a single lens; 2 - a small oleoresin lens 

appears (no more than 0.5-1.0 mm thick); 3 - a larger oleoresin lens appears. The following indices of July oleoresin exudation 

intensity (I) and efficiency (E) were calculated: 

I = (nl+2*n2+3*n3) / N E = (n2+n3) * 100 / N 

where I is a mean index of oleoresin exudation; nl, n2, n3 are number of needles with scores of 1, 2, or 3; N = total number 

of investigated needles (including those scored as 0); and E = oleoresin exudation efficiency. 

The relationship between mortality (M) of larvae (I-III instars) and (I) was tested using regression: M = a*I - b, 

where a and b are parameters specific for each insect species. Thus, for Dendrolimus pini, M = 65.0 1 - 34.4, for Acantholyda 

stellata, M = 37.3 1- 8.3, for Diprion pini, M = 66.9 1- 51.2, and for Neodiprion sertifer, M = 42.4 1 - 29.9. The r2 values for 

these regressions vary from 0.53 - 0.79. 

One can calculate that Scotch pine is resistant to young larvae of Neodiprion sertifer beginning at I _>1.7 and E > 70. 

For other defoliators, Scotch pine is resistant at I > 1.4 and E > 40. It should be noted that July oleoresin exudation intensity 

must be determined no later than 3-4 days after heavy rains when the humus layer moisture is > 6-8%, and the moisture of the 

lower layers (down to 50 cm depth) is _>3-4%, and air temperature, _>10 *C. 

Oieoresin exudation intensity from the needles of several pine species show that they can be divided into three 

groups: (1) low oleoresin exudation intensity (Pinus banksiana, P mugo, P maritima), (2) intermediate intensity (P 

sylvestris, P. pallasiana, R nigra, P strobus), or (3) high intensity (P. sibirica, P. cembra, P. pithyusa). 

The connection between the oleoresin exudation intensity of various pine species and their resistance to various 

defoliators is not perfect. The resistance of pine species depends also upon terpene composition of oleoresin. 

Mattson, W.J., NiemeRi, R, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


263

COMPANION PLANTING OF THE NITROGEN-FIXING GLIRICIDIA SEPIUM 

WITH THE TROPICAL TIMBER SPECIES MILICIA EXCELSA AND ITS IM- 

PACT ON THE GALL FORMING INSECT PHYTOLYMA LATA 

MICHAEL R. WAGNER 1, JOSEPH R, COBBINAH 2, and DANIEL A. OFORI z 

_School of Forestry, Northern Arizona University, P.O. Box 15018, Flagstaff, Arizona 86011, USA 

2Forestry Research Institute of Ghana, UST P.O. 63, Kumasi, Ghana 

INTRODUCTION 

In Ghana, West Africa, there is increasing concern over the loss of valuable commercial timber species that contribute 

to the economic development of this country (Wagner and Cobbinah 1993). Part of the diminishing supply of timber relates 

to the high rates of deforestation (Zerbe et al. 1980, Mergen 1983, Repetto 1988). But an equally important factor that is 

often overlooked is the failure of native species to regenerate following disturbance in the natural forest or when planted in 

plantations (Fontaine 1985). Forest insects can significantly contribute to the poor regeneration rates of tropical timber 

species. An example of this is the complete failure to artificially regenerate Milicia excelsa, an extremely valuable tropical 

timber species in Africa, because of heavy damage by the gall-forming psyllid, Phytolyma lata (White 1968, Wagner et al. 

1991). 

Milicia excelsa and M. regia are important timber species in Africa, known in the timber trade as lroko, Mvule, and 

locally in Ghana as odum. The wood is durable, uniformly grained, and extremely resistant to termites. While the species is 

a common component of the natural overstory in much of the remaining tropical forests of West Africa, it is not currently 

regenerating and has a predicted resource life of less than 10 years in Ghana (Alder 1989). Reforestation efforts have focused 

on establishing plantations of Imko which have failed in part because of massive damage by Phytolyma lata, the odum 

gallfly. 

The odum gallfly oviposits on new foliage of Milicia spp., and the first instar nymphs damage leaf tissue and induce 

galls (White 1968; Orr and Osei-Nkrumah 1978; Cobbinah 1986, 1988). The nymphs feed within the gall tissue. Numerous 

attack sites result in large gall masses. After about 2-3 weeks, the galls burst open releasing the adult psyllid. The gall tissue 

that has burst open exudes gall and leaf fluids that are heavily colonized by saprophytic fungi. These saprophytic fungi result 

in decay of leaf and stem tissue and dieback of young shoots (Wagner et al. 1991). Multiple attacks within the same season 

result in heavy damage and frequent death to young seedlings. 

Considering that conventional plantations had largely failed, we decided to examine the influence of companion 

planting of Gliricidia sepium, a well-known nitrogen-fixing agroforestry species, and Milicia excelsa on odum gallfly attack. 

We hypothesized that Gliricidia sepium would provide nitrogen and overstory shade more similar to conditions that occur in 

the natural tropical forests. In addition, we examined the influence of shading and nitrogen fertilization independently in an 

effort to determine which factor contributed more to the observed effects. In this manuscript we report that companion 

planting, shading, and fertilization all reduce attack and damage by the odum gallfly on Iroko. 

Mattson, W.J., Niemela, E, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


264

EXPERIMENTAL METHODS 

Experiments were carried out at the Forestry Research Institute's experimental nursery about 5 km south of the 

University of Science and Technology, Kumasi, Ghana. The area is within the moist semi-deciduous forest zone (Hall and 

Swaine 1983), a common tropical forest type in West Africa that generally supports populations of Milicia excelsa. Seedlings 

used in the companion planting and shading experiments were collected from a single parent tree (AA 17) from the village of 

Abofour in the Ashanti Region of Ghana. Other studies have indicated that this particular provenance is of intermediate 

susceptibility to Phytolyma lata (Cobbinah and Wagner 1994). Seedlings used in the fertilizer trial were of mixed origin. 

Seeds were collected in 1991, sown on raised beds, and transplanted into polyethylene bags. Trees were kept under shade 

and watered regularly until used in the experiment. The spacing on all experiments was 1m x 1m. 

Companion Planting Trial 

The companion planting trial was established in September 1991. The trial consisted of four rows of 10 (40 total) 

Milicia excelsa planted in alternating rows with Gliricidia sepium cuttings and four rows of 10 Milicia excelsa planted 

without G. sepium. At periodic intervals the following data were collected: seedling height, number of leaves, number of 

galls/plant. Regression analysis was used to estimate the number of galls in large gall masses as described by Cobbinah and 

Wagner (1994). For this experiment, the data were averaged for each row of 10 trees to reduce variability, consequently n = 4 

for both treatments. 

Shading Trial 

The shading experiment was established in October 1992. The shade was created by supporting commercial shade 

cloth (Hummerts International) over wooden structures. Four levels of shade were established: full sun = no shade, light 

shade = 33% shade cloth, medium shade = 57% shade cloth, and deep shade = 82% shade cloth. Under each shade level, 50 

seedlings were planted in five rows of 10 seedlings each. Data collection was as indicated above. For this experiment, n = 

50 per treatment. 

Fertilizer Trial 

The fertilizer experiment was established in December 1992 and ran for 6 months. The effect of fertilizer was tested 

using a randomized complete block design with four replicates (25 seedlings/plot and 100 seedlings/block) and four treatments: 

control (no fertilizer), NPK ® (15-15-15), Phostrogen ® (10-4-22), Sampi® (8-3-3). Fertilizer was applied in the 

irrigation water at the rate recommended by the manufacturer. Seedlings were kept under shade. Data were collected as 

above and the average number of diebacks was determined. 

Statistical Analysis 

Data for the companion planting trial were analyzed using a paired t-test on STATGRAPHICS software. Each row of 

10 trees was treated as the sample unit. Data for the shading trial were transformed [Logl0(x+l)] to meet assumptions of 

homogeneity of variance and were analyzed using an ANOVA on STATGRAPHICS software. Data for fertilizer trials were 

analyzed using ANOVA and LSD multiple range test using STATGRAPHICS software. The P value to reject the null 

hypothesis was set at 0.05. 

RESULTS 

Companion Trial 

Milicia excelsa seedlings growing together with the nitrogen fixing Gliricidia sepium were not significantly taller 

(Fig. 1) and they did not produce more leaves (Fig. 2) than Milicia excelsa seedlings growing alone during any sample 

period. However, there were significantly fewer galls during two sample periods on Milicia grown with Gliricidia (Fig. 3). 

There was also a tendency for fewer galls in the presence of Gliricidia for most of the 1992 sample periods (Fig. 3). 

265

250 

With Gliricidia 

[] Without Gliricidia P=0,83 

200 .................................. 

P=0,81 

o 150 ............................... P--:o.28 '_.--0.4_,. 

.E P=0.21 

03 

© 

C" 

c 100 

aJ 

50 

0 

P=0.31 

P=0.29 

P=0.12 

P=0.32 

P=0.48 

1/92 2/92 3/92 4/92 9/92 11/92 1/93 3/93 5/93 7/93 

Sample Date 

Figure 1.--Influence of companion planting Gliricidia sepium with Milicia excelsa on M. excelsa growth at nine sample dates 

over 14 months. P-value based on paired t-test. 

C 

800 ................................................ 

_- 600 

O 

With Gtiricidia 

[] Without Gliricidia 

(f} P=,).55 P=0.28 

> _n 

//A 

Z 400 ............................................. //A 

o ///1 

I,- P=0.82 //A 

//A 

//A 

E P=O.S6 //A 

L 200 P=0:66 ....................... 

I'0-" .lo ///I 

//A 

P=0,19 

11/91 12/91 1/92 2/92 3/92 4/92 9/92 11/92 1/93 

Sample Date 

Figure 2.--Influence of companion planting Gliricidia sepium with Milicia excelsa on the number of" leaves of M. excelsa. 

P-value based on paired t-test. 

266

C 

w 

Q. 

1,200 

1,000 

o 800 

C_ 

--- 600 

0 

[] With Gliricidia 

[] Without Gliricidia 

P=0.02 

C'} P=O.07 

IE 400 ............... P=o.s4 /// 

-1 /// ...... P=0.40 ' 

"_ //I 

"+'" I1i 

1-0- 200 ............................ /// 

/// 

/// 

/// 

P=0.96 

0 /// 

1/92 2/92 3/92 4/92 9/9211/92 1/93 3/93 5/93 7/93 

Sample Date 

Figure 3.--Influence of companion planting Gliricidia sepium with Milicia excelsa on the total number of Phytolyma lata 

galls on M. excelsa. P-value based on paired t-test. 

Shading Trial 

Shading had a significant positive impact on height (Fig. 4). After 1 year, trees grown at the highest level of shade 

(82%) were nearly 50% taller than trees at 57% shade and 100% taller than trees grown in full sun. In general, there were 

more leaves produced under shade than in full sunlight, but this pattern was not consistent across sample dates (Fig. 5). 

Despite the fact that more leaves were available to Phytolyma lata under shade, there were fewer galls per plant under shade 

(Fig. 6). While this pattern is statistically significant at only two of the sample dates, the pattern is very consistent across all 

sample dates. 

Fertilizer Trial 

Two of the fertilizer treatments, 15-15-15 and 10-4-22, significantly affected Milicia height when compared to the 

control (Fig. 7). In general the higher the nitrogen the greater the impact. The highest nitrogen content fertilizer also resulted 

in significantly fewer branch diebacks. Diebacks are positively related to the number of galls (Cobbinah and Wagner 1994). 

DISCUSSION 

We are unaware of other studies that have clearly linked companion planting or agroforestry techniques and modifications 

in the incidence and impact of damaging insects in a tropical ecosystem. While the patterns we observed are not 

absolutely consistent for all the sample periods, there is sufficient evidence to warrant further study. 

There is, however, a considerable body of evidence on the relationship between fertilization and insect attack (see 

Stark 1965, and Schowalter et al. 1986 for reviews). Strauss (1987) observed that fertilization tended to have a positive 

impact on sap feeding insects and no effect on chewing insects. Several studies have examined the relationship between 

fertilization and population parameters of the gall-forming adelgid Adelges coolyei (Mitchell and Paul 1974, Johnson et aL 

1977, Mitchell and Miller 1976). In each case they found a general trendof fertilization increasing population parameters of 

Adelges coolyei. While these results are opposite to our own findings, the environmental conditions and plant species are 

267

7O 

Full sun P:O.00 

60 [] Light shade 

Medium shade 

50 []Deep shade 

E 

o 

.-, 40 

x: 30 

E -- P=0,29 -- P=0,06 P=0.0 

P=O.O0 

2O . _ - 

10 .. ...._ 

0 

10/92 11/92 1/93 3/93 9/93 

Sample Date 

Figure 4.--Height (cm) of Milicia excelsa grown under four levels of shade. Light shade = 33%, Medium shade = 57%, 

Deep shade = 82%. 

00 ................................................................ 

_-_Full sun P=O,O3 

[] Light shade ,_,_ 

t_ [] Medium shade _\'_ 

T__150 ................................ _\\, . . . 

_--° [] Deep shade :::.:_ 

(1) P=O.06 : i """" 

__ 100 ....................................................... 

..... \\\. 

0 

E ..... \\\\ t_.tt 

: .... 

c- 50 :::: ,-.-,-, ...... :ii 

:_\'h \\\\ :::: ;ii;_ 

F_. : \\\\ i:ii i N_ 

\\\\ ::::: 

\\\\ ::: 

:.: : \\',.\ ::,: 

,,; 

\\\\ :::; .... 

10/92 11/92 1/93 3/93 

Sample Date 

Figure 5.--Number of Milicia excelsa leaves produced under four levels of shade. Light shade = 33%, Medium shade -- 

57%, Deep shade = 82%. 

268

0 .............................................................. 

Light shade P=0,05 

8 " _ Medium shade ............................... _ ..... 

[] Deep shade 

l _._Control 

E 6 ...................................................... ........ 

(_ 

.. P=0.08 P=0,19 

--...... __ __ 

"-_ P=0.03 

± 

0 -- 

10/92 11/92 1/93 3/93 9/93 

Sample Date 

Figure 6.ENumber of Phytolyma tata galls per Milicia excelsa plant grown under four levels of shade. Light shade = 33%, 

Medium shade = 57%, Deep shade = 82%. 

60 

.................. _1_Height 15 

50 ..................... 

b l_ Diebacks ............ b 'I 

09 

&"4o ..... -_ 

o - 

v _O 03 

O3 "O 

•_ 30 cc- 

O {..- 

E _ 

520 5 _x 

10 

0 0 

Control 15-15-15 10-4-22 8-3-3 

Fertilizer Treatment 

Figure 7.mMilicia excelsa height (cm) and average number of Phytolyma lata induced diebacks as the result of four fertilizer 

treatments. 

i 

269

very different from the Milicia/Phytolyma system we examined. There is very little consistency in the response of herbivores 

to fertilization of their hosts (Stark 1965, Schowalter et al. 1986). Insufficient evidence exists to postulate any general 

pattern in this regard. 

Hall and Swaine (1983) reported that Iroko is not abundant in the evergreen tropical forest because it requires high 

light for germination and establishment. However, as previously noted, plantations of this species planted in full sun have 

been heavily attacked by the odum gaUfly and have largely failed. In contrast to the report of Hall and Swaine (1983), we 

have been told by a tropical tree physiologist that the optimal photosynthesis for Iroko seedlings occurs at 8-10% of full 

sunlight, suggesting this is a shade tolerant species (E.M. Veenendaal, pers. comm.). The results we report in this paper 

indicate that the abundance and subsequent impact of this insect is greater in full sun than under shade. 

We recognize that there would be considerable utility in examining the relationship between Iroko and the odum 

gallfly under natural field conditions. However, naturally regenerating seedlings of this species are difficult to find. In a 

recent field excursion to the Tain II Forest Reserve, which is noted for its high density of Iroko in the overstory, we were able 

to locate only a single seedling in 12 person/hours of searching. The low incidence of Iroko seedlings has been confirmed 

anecdotally by several botanists and forestry field surveys. Growing stock as of 1987 was immature stock (1-7ft girth) 

834,000m 3and mature stock (> 7ft girth) 5,269,000m 3 (Alder 1989). 

The results we report suggest that standard agroforestry practices such as mixed planting with nitrogen-fixing trees 

may have utility as part of a pest management program to control the odum gallfly. Other approaches including selection for 

genetic resistances have resulted in considerable success (Cobbinah and Wagner 1994). Cooperative effort between silviculturists, 

physiologists, and genetists is required to fully understand how to reduce the impact of this insect and provide for a 

continuing supply of one of Africa's most valuable timber species. 

SUMMARY 

Planting the nitrogen-fixing Gliricidia sepium in conjunction with the tropical timber species Milicia excelsa reduces 

the abundance and damage caused by the gall-forming pysUid, Phytolyma lata. This effect appears to be related to both the 

increased shading and enhanced nitrogen environment. Shading independently has a strong impact on gall abundance with 

the lowest gall abundance occurring under the highest level of shade (82%). Gall abundance is reduced under shade even 

though more leaves are available to the insect. Fertilization at the highest levels of nitrogen also reduces the abundance of 

galls. 

ACKNOWLEDGEMENT 

These studies were supported by the International Tropical Timber Organization through research grant No. PD75/90 

to J.R. Cobbinah and M.R. Wagner. 


ALDER, D. 1989. Naturalforest increment, growth and yield, p. 47-52. In Ghana Forest Inventory Proceedings. Overseas 

Dev. Agency/Ghana For. Dep., Accra. 

COBBINAH, J.R. t986. Factors affecting the distribution and abundance of Phytolyma lata (Homoptera:PsyUidae). Insect 

Sci. Appl. 7: 1111-115. 

COBBINAH, J.R. 1988. The biology, seasonal activity and control of Phytolyma lata. IUFRO Regional Workshop on Pests 

and Diseases of Forest Plantations, June 5-10, 1988, Bangkok, Thailand. 

COBBINAH, J.R. and WAGNER, W.R. 1994. Phenotypic variation in Milicia excelsa toattack by Phytolyma lata. 

Biotropica. (in review). 

270

FONTAtNE, R.G. 1985. Forty years of forestry at FAO: some personal reflections. Unasylva. 37(148): 5-14. 

HALL, J.B. and SWAINE, M.D. 1981. Distribution and Ecology of Vascular Plants in a Tropical Rain Forest: Forest 

Vegetation of Ghana. W. Junk Publishers, The Hague. 383 p. 

JOHNSON, J.M., LITTLE, M.R, and GARA, R.I. 1977. Effects of different N-fertilizers and concentrations on Adelges 

cooleyi populations. J. Econ. Entomol. 70(4): 527-528. 

MERGEN, F. 1983. Tropical Forestry: A Challenge to the Profession of Forestry. Yale University Printing Service, New 

Haven, CT. 23 p. 

MITCHELL, R.G. and PAUL, H.G. 1974. Field fertilization of Douglas-fir and its effects on Adelges cooleyi populations. 

Environ. Entomol. 3: 501-504. 

MITCHELL, R.G. and MILLER, R.E. 1976. Effects of Douglas.fir fertilization on the Cooley spruce gall aphid. Am. 

Christ. Tree J. 20(1): 29. 

ORR, R.L. and OSEI-NKRUMAH, A. 1978. Progress Report on Phytolyma lata (Hemiptera:Psyllidae). Forest Products 

Research Institute. Kumasi, Ghana. 

REPETTO, R. 1988. The Forest for the Trees? Government Policies and the Misuse of Forest Resources. World Resources 

Institute. Washington, DC. 105 p. 

SCHOWALTER, T.D., HARGROVE, W.W., and CROSSLEY, D.A. 1986. Herbivory in forested ecosystems. Ann. Rev. 

Entomol. 31: 177-196. 

STARK, R.W. 1965. Recent trends in forest entomology. Ann. Rev. Entomol. 10: 303-324. 

STRAUSS, S.Y. 1987. Direct and indirect effects of host-plant fertilization on an insect community. Ecology 68: 1670- 

1678. 

WAGNER, M.R., ATUAHENE, S.K.N., and COBBINAH, J.R. 1991. Forest Entomology in West Tropical Africa: Forest 

Insects of Ghana. Kluwer Academic Publishers, Dordrecht. 210 p. 

WAGNER, M.R. and COBBINAH, J.R. 1993. Deforestation and sustainability in Ghana: the role of tropical forests. J. For. 

91(6): 35-39. 

WHITE, M.G. 1968. Research in Nigeria on the Iroko Gall Bug (Phytolyma spp.). Nigeria For. Inf. Bull. 18. 72 p. 

ZERBE, J.J., WHITMORE, J.L., WAHLGREN, H.E., LAUNDRIE, J.F., and CHRISTOPHERSEN, K.A. 1980. Forestry 

activities and deforestation problems in developing countries. Report to Office of Science and Technology. U.S. 

Department of Agriculture, Forest Service, Forest Products Laboratory. 115 p. 

271

FUNCTIONAL HETEROGENEITY OF FOREST LANDSCAPES: 

HOW HOST DEFENSES INFLUENCE EPIDEMIOLOGY 

OF THE SOUTHERN PINE BEETLE 

ROBERT N. COULSON, JEFFREY W. FITZGERALD, BRYAN A. MCFADDEN, 

PAUL E. PULLEY, CLARK N. LOVELADY, and JOHN R. GIARDINO 

Knowledge Engineering Laboratory, Department of Entomology and Department of Geography, Texas A&M University 

College Station, Texas 77843, USA 

INTRODUCTION 

Insect outbreaks are autogenic disturbances that are normally observed at the landscape scale as levels of herbivory 

above an average or expected amount. When favorable environmental conditions coincide with optimal resource availability, 

populations increase in size and outbreaks occur (Rykiel et al. 1988). Epidemiology is the study of causes for change in the 

distribution and abundance of an insect species at tile landscape level of ecological integration. Typically, studies of epidemiology 

have considered herbivore populations in the context of their impact on food and habitat resources. In this paper we 

focus on the association of forest landscape structure and the process of herbivory, i.e., the effect of landscape pattern on the 

process of herbivory (Turner 1989). Our goal is to examine epidemiology of the southern pine beetle, Dendroctonusfi'ontalis 

Zimmermann (Coleoptera: Scolytidae), in the context of functional heterogeneity of forest landscapes. Our objectives are: 

(i) to review briefly the current state of knowledge on epidemiology of D. frontalis, (ii) to consider the concept of functional 

heterogeneity of forest landscapes and how it relates to epidemiology of D. j?ontalis, (iii) to consider ways and means of 

measuring functional heterogeneity of forest landscapes, and (iv) to illustrate the relation of functional heterogeneity and 

herbivory by the insect. 

Although emphasis is directed to D. frontalis, it is noteworthy that this insect is a member of a guild consisting of 

four other bark beetle species: D. terebrans (Olivier), Ips calligraphus (Germ_), L grandicollis (Eichhofl), and I. avulsus 

(Eichhoff) (Coleoptera: Scolytidae). The population systems of these insects are interrelated. How the members contribute 

to epidemiology of the guild is unknown. 

Background 

Epidemioiogy of the Southern Pine Beetle 

[=lammet al. (1988) provide a review of literature on the natural history of D. frontalis. A general conceptual model 

of epidemiology of the insect was proposed by Coulson et aL (1983, 1985a, b, 1986a) and Rykiel et al. (1988). This model 

considers the process of herbivory in the context of variables thought to influence epidemiology, i.e., meteorological conditions, 

the lighming disturbance regime, forest stand conditions, background levels of the insect, and landscape structure. 

Studies by Flamm and Coulson (1988), Flamm et al. (1992), Lovelady et al. (199 I), and Lovelady (1994) have employed 

empirical methods to address specific aspects of the conceptual model. 

Outbreaks of D. frontalis are the result of population fluctuations observed at the landscape scale. There are three 

separate processes associated with D. fi'omalis epidemiology. The first, which operates at the population and community 

levels of integration, involves factors associated with initiation of infestations. The second, which operates at the ecosystem 

level of integration, deals with factors that influence growth of established infestations. The third, which operates at the 

landscape level of integration, deals with the network of habitat units and population centers occurring in a heterogeneous 

land area consisting of multiple interacting ecosystems. An insect outbreak is an epidemiological condition that represents a 

composite of the three processes operating across different scales of time and space. The actual process of herbivory by D. 

fromalis occurs at the population and community levels of ecological integration and results in mortality to host pines. 

However, the effects are propagated through component ecosystems to the landscape level (Coulson et al. 1986b, Rykiel et 

al. 1988, Lovelady et al. 1991). 

Mattson, W.J., Niemel_, R, and Rousi, M., eds. 1996. Dynamics of forest herbivory: quest for pattern and principle. USDA 


272

Host Defenses 

An important component of epidemiology of bark beetles centers on the interaction of the insect with its host. 

Considerable effort has been directed to research on host defense mechanisms, because of their critical role in regulating 

colonization by bark beetles. In pines (Pinus spp.) both preformed and induced defense systems have been identified. 

During colonization, bark beetles directly confront the defense mechanisms of the host. The colonization process is greatly 

influenced by variation in defense capacity associated with different pines, (e.g., Iongleaf pine is more resistant to D. frontalis 

colonization than loblolly or shortleaf), and with different individuals of the same host species. Within a forest landscape 

there is also a seasonal variation in host defense capability. Lorio (I988) suggested that normal physiological changes 

associated with tree phenological development and response to variable environmental conditions cause regular and predictable 

fluctuations in resistance/susceptibility to D. frontalis. Lovelady (1994) provides a comprehensive review and interpretation 

of seasonal variation in host defense in relation to population dynamics ofD. frontalis. 

Lightning-struck hosts, which have greatly reduced capacity for defense against colonization, serve as refuges for 

bark beetles. For many years foresters have recognized that infestations olD. frontalis are often associated with a lightningstruck 

host. The initial conceptual model of epiderniology of D. frontalis included lightning-struck hosts as a prominent 

component, although only circumstantial evidence supported the contention (Coulson et al. 1983). Subsequent research has 

suggested that lightning-struck hosts are an essential feature of the natural history of the insect. This research considered the 

interaction of D. frontalis with disturbed hosts (Coulson et al. 1986a, Flamm and Coulson 1988, Flamm et al. 1992) as well 

as an analysis of lightning as a disturbance regime (Lovelady et al. 1991, Lovelady 1994). 

Of particular concern in the study of epidemiology is how host defenses are distributed at the landscape scale. 

Herbivory by D. frontalis produces disturbance patches (infestations) in the forest matrix that can be observed relative to 

other features of the landscape, i.e., the crowns of infested trees become discolored and are visually detectable. The disturbance 

patches are typically associated with old-growth forest stands occurring on poor sites. Hazard rating systems have 

been developed to grade forest stands with respect to vulnerability to D. frontalis infestation (Lorio 1980, Branham and 

Thatcher 1985, Mason et al. 1985). Several different methods have been devised. Each rating system has novel features, but 

all involve integration of a subset of variables: tree species, radial growth, height, and DBH; stand basal area (pine, hardwood, 

total), species composition, site index, and degree of crown closure; and landform classification. When applied to a 

forest landscape, the hazard rating systems provide a general view of the distribution and abundance of host defenses against 

the insect. It is noteworthy that the hazard rating systems integrate a substantial knowledge base on the interaction of D. 

frontalis, host plants, and site conditions. 

To summarize, we have described the nature of the interaction of D. frontalis with host defenses and have identified 

several sources of variation important to population dynamics of the insect. Although the hazard rating systems have proven 

useful in defining vulnerability of forest landscapes to herbivory by D. frontalis, they ignore the seasonal dynamics of host 

defenses associated with different tree species, with individuals of the species, and with lightning-struck hosts. Below we 

will consider how the spatial and temporal sources of variation in host defense can be integrated in the context of the natural 

history of the insect. In effect, we view host defense capacity as a variable in landscape structure that can be characterized 

and defined. The distribution of this variable across the forest landscape influences epidemiology of D. frontalis. 

Outbreaks and Forest Landscapes 

Herbivory at the landscape scale is significant because it is at this level of integration where the ecological and forest 

management consequences of insect outbreaks are interpreted (Pickett and White 1985, Forman and Godron 1986, Barbosa 

and Schultz 1987, Turner 1987, Platt and Strong 1989, Turner and Gardner 1991, Holling 1992a, Wiens 1992, Coulson et al. 

1993). Advances in the geographic information system (GIS) technologies and the development of statistics for analysis and 

description of spatially referenced data have greatly expedited studies of insect outbreaks in forest landscapes (Liebold and 

Barrett 1993). 

Historically, interest in epidemiology of D. frontalis has been primarily associated with the economic impact of the 

insect on forest resources. Emphasis has focused on investigations of herbivory in the context of impact on forest stand 

structure, i.e., the research has considered the effect of process (herbivory) on pattern (the configuration of elements that 

constitute the landscape mosaic). For example, Fitzgerald et al. (1994) examined the relation of various suppression tactics, 

employed to modify D. frontalis populations, on the subsequent development of new infestations within the surrounding 

forest landscape. 273

However, in pre-colonization forests of the southern US, herbivory by D. frontalis (along with fire) was likely 

involved in structuring the landscape (Schowalter et al. 198 l, Rykiel et al. 1988). As D. fror_talis infestations are r_ormally 

suppressed in an effort to reduce economic loss by the insect, it has not been possible to evaluate this hypothesis un,it 

recently. Outbreaks of D. frontalis in Rare II Wilderness Areas on National Forests in Texas have been allowed to follow a 

natural course, with only modest intervention to protect endangered species. The outbreaks, which occurred over a period of 

1983 to t994, have radically changed the pattern of the forest landscape mosaic of these wilderness areas (Fig. t). Under this 

epizootic condition, individual infestations coalesced and the pattern and extent of tree mortality substantially modified the 

structure of the forest landscape. 

Figure I .---The impact of herbivory by D. frontalis in pine fores_ landscapes in east Texas. Under this epizootic condition, 

individual inl_stations coalesced and the pattern and extent of tree mortality substantially modified the structure of 

the tbrest landscape. 

The alternate way of viewing insect outbreaks is to consider how landscape pattern affects the process of herbivory. 

Again, because D. frontalis infestations are normally suppressed, it was not possible in the past to examine this relation. 

Figure I clearly illustrates how the process of herbivory is influenced by the pattern of patches that were created by previous 

in|izstations of D. j}ontalis. There are few examples of empirical studies of the eflEcts of landscape pattern on the process of 

herbivory (Turner 1989). 

By definition, epidemiological studies are dynamic and therefore the process of herbivory and the pattern of the tk_rest 

landscape mosaic change through mutual interaction, i.e., the pattern influences the process and the process influences the 

pattern. Although not investigated in a rigorous manner, the scenario observed on the National Forest wilderness areas of 

Texas sugges{s that bark beetle herbivory drives the process of ecosystem succession (from the release to reorganization 

phase in Holling's scheme 1992b) (see also Brown 1991). 

274

Functional Heterogeneity of Forest Landscapes: How Bark Beetles Perceive and Respond to Their Environment 

Studies of epidemiology begin with a consideration of issues associated with landscape structure. The term landscape 

structure refers to the spatial relationships between distinctive ecosystems (ecotopes, see Coulson et al. 1993), i.e., the 

distribution of"energy, materials, and species in relation to the sizes, shapes, numbers, kinds, and configurations of components 

(Turner 1989). Characteristics of landscape structure (porosity of the forest matrix, boundary configurations, patch size 

and shape, etc.) influence epidemiology of bark beetles. Structural characteristics of landscapes are often summarized and 

represented as heterogeneity. Although defined in several ways, the term heterogeneity is generally taken to mean composition 

of'parts of"different kinds. In considering the interaction of D. frontalis with the forest landscape it is necessary to 

distinguish between measured and functional heterogeneity. Measured heterogeneity deals with physical features or components 

of a landscape. Functional heterogeneity deals with how an organism perceives and responds to its environment 

(Kolasa and Rollo 1991). There is obviously a direct link between patches, boundaries, and heterogeneity as boundaries 

define patches, and patchiness is what produces heterogeneity (Wiens 1992, Hansen and di Castri 1992). 

Tied to concepts associated with landscape heterogeneity is the subject of habitat structure. In general, habitat 

structure refers to the physical arrangement of objects in space (Bell et al. 1991). In studies of epidemiology, an essential 

task is to define the distribution, abundance, and location of habitat units. In effect, habitat structure is an element of measured 

heterogeneity, i.e., habitat can be viewed as a mappable element of a landscape mosaic. Southwood (1977) suggested 

that habitat serves as the template for ecological strategies. 

Examination of how bark beetles perceive and respond to their environment at the landscape scale, i.e., evaluating 

functional heterogeneity, was difficult before the availability of (i) GIS technologies, (it) spatially referenced databases, and 

(iii) spatial statistics. Also, until the text by Forman and Godron (1986), there was no standardized nomenclature for describing 

landscapes from an ecological perspective (McIntosh 1991). With these problems partially solved, it is now possible to 

consider the substantial knowledge base on natural history of bark beetles (and other organisms) in the context of the forest 

landscapes where they occur (Dunning et al. 1992). 

In reference to D. frontalis, we have indicated that a great deal is known about host selection relative to the initiation 

and subsequent growth of infestations. In the case of initiation of infestations, lightning-struck hosts play a prominent role. 

For infestation growth, variables associated with forest stand hazard become significant. With knowledge of how these 

variables are spatially and temporally distributed, we can consider scenarios, based on the natural history of the insect, that 

explain causes for changes in the distribution and abundance of D. frontalis at the landscape scale. 

Ways and Means of Measuring Functional Heterogeneity 

Ways and means for measuring landscape heterogeneity have been reviewed by Kolasa and Pickett (I991), Turner 

(1987, 1989), S. Turner et al. (1991), and M. Turner et al. (1991). Although there are several methodologies available, no 

single index is suitable for characterization of functional heterogeneity for a given organism. Each organism perceives and 

responds to its environment in unique ways. The elements of landscape structure that are important in defining herbivory by 

bark beetles, for example, may be substantially different from those needed to characterize the process for deer populations. 

Figure 2 illustrates the relation of functional and measured heterogeneity. In Figure 2 functional heterogeneity is represented 

as a gradient ranging from homogeneity on one end to heterogeneity on the other. Interpretation of the significance of index 

values in the mid-range of the gradient is difficult. We emphasize, also, that intermediate heterogeneity conditions are most 

difficult to interpret from the behavioral perspective of the organism, i.e., it is difficult to predict how an organism will react 

to conditions of intermediate heterogeneity. For this reason measurement of functional heterogeneity for a specific organism 

will likely require use of several indices, each of which is sensitive to different aspects of landscape structure, e.g., connectivity, 

porosity, fragmentation, boundary configuration, etc. To address this issue, we developed (or adapted) three different 

procedures to characterize functional heterogeneity of forest landscapes relative to herbivory by D. frontalis. These indices 

are described briefly below. 

Indices of Heterogeneity 

Our initial goal in defining functional heterogeneity for D. frontalis was to combine knowledge of natural history of 

the insect with information on landscape structure. We developed (or modified) three indices to characterize functional 

heterogeneity. All use a weighted pattern-recognition computational approach, and each is sensitive to a different aspect of 

275

FUNCTIONAL VS. MEASURED HETEROGENEITY 

FunctionalHeterogeneity 

Measured Heterogeneity 

Figure 2.--Relationship of measured heterogeneity (composition of parts of different kinds) to functional heterogeneity (how 

an organism perceives and reacts to its environment). Note that measured heterogeneity for the left and right matrices 

is the same, but the functional heterogeneity is different. 

Interpretation of the significance of index values in the mid-range of the gradient is difficult. The intermediate 

heterogeneity conditions are also the most difficult to interpret from the behavioral perspective of the organism, i.e., 

it is difficult to predict how an organism will react to conditions of intermediate heterogeneity. 

landscape pattern. Application of the procedures involves (i) classification of land elements based on ecological function, (ii) 

tiling of the landscape data (usually maps or images), and (iii) quantification of spatial pattern. Following is a brief description 

of each index along with an explanation of the calculation procedure. These methods apply for any NxN array of 

weighted landscape elements. We are currently investigating the sensitivity of these indices. 

The Angular Moment of Inertia Index (AMI). The AMI is an index that is sensitive to variation in the dispersion 

of landscape elements. It is the normalized product of a second moment calculation. The index is a measure of the dispersion 

of weighted matrix elements about the centroid of the landscape area. The AMI is calculated using the following 

procedures. First, the centroid of the weighted surface (matrix) is calculated. Because each element, a_j,within the matrix has 

a value and locational indices attached, the horizontal (Xo)and the vertical (Yo) centroids can be calculated by 

Xo = (_j_i(aij),j) / _i_j aij 

Yo= (£_Zj(aij).i) / £,Ej aij 

Second, Xoand Yo are used to derive the second moment (M) about the centroid 

M = Z_jc(o. Xo)2 + (i - Yo)2) • au) 

However, because each element in the matrix represents a cell that has a real extent, the second moment for each cell : 

must be calculated. The equation becomes 

276 

M = Z_i

The Connectivity Index (CI). The CI index is based on a run-length encoding strategy that is commonly used to 

encode raster (matrix) data. Run-length encoding is a good compression technique for rasters that have a high degree of 

spatial autocorrelation. Run lengths are measured geometrically rather than arithmetically, giving more than linear value to 

longer runs. The computational procedures for the CI index are given below. 

First, count the runs of matrix elements with the same value (i). For example, 1l 11 would be a run of length L(1) = 

4. A row, column, or diagonal may have several runs. For example, vector 111222233333 has three runs. There can be 

multiple runs for a given i. 

For each row, column, or diagonal, we sum the runs for each distinct element value. For example, vector 

1112222211111 has three runs and two values. Converted to run values, the vector is 123, 12345, 12345. The sum of runs, 

by element value, for this vector is ES(1) = 21 and ZS(2) = 15. For a given run i of length L, the run value is S(i) = L * ( L + 

1 ) / 2. For a given element value i, all possible S(i) are calculated by row, column, and diagonal, with the diagonal S(i) 

divided by 2.2 to compensate for separation distance. The total run value for a matrix and a given i is simply the ZS(i) for 

that matrix. 

For the general NXN matrix, the smallest possible ZS(i), i = 1 to n where n is the number of different values, is 

ZSmin(i) = 2 ( N2+ N2/2 -2). 

ZSmin(i) exists where all run lengths are 1. The largest possible value, ZSmax(i), must be calculated using NXN 

matrix of constant i. 

Thus, the range is 

The unweighted CI is calculated as 

ESmax(i) = NZ(N+I) + 2(Z K(K+I))/2 z - N(N+I)/2 2 

R = ZSmax(i) - ZSmin(i). 

H = (ES(i) - ESmin(i)) / R. 

The CI for a matrix where i values are relative to a function (i.e., hazard rating or probable use) can also be computed. 

The run values are calculated as weighted sums (WCI). 

WCI = 1; i(_:S(i)), i = 1,2 ..... 

The Spatial Interaction Index. The third index of heterogeneity is based upon the gravity model and is sensitive to 

fragmentation of the highest valued landscape elements. When applied to geographic analysis, the gravity model is called 

spatial interaction. Therefore, we have called this index spatial interaction (SI). 

SI is based on Newton's law of attraction between masses. It states that attraction is proportional to the product of 

two masses (assigned weight) and inversely proportional to the distance between them. For a matrix of landscape elements 

represented in matrix form: 

, k = 1, where 

mkand ml are distinct elements of the matrix representing the attraction measure, r is the distance separating mk and m,, and t is 

the maximum value in the matrix. 

Functional Heterogeneity of Forest Landscapes and Host Defenses 

In this section we examine functional heterogeneity of forest landscapes in relation to epidemiology of D. frontalis. 

Our approach is to provide (i) a general overview of the procedure used to characterize functional heterogeneity, (ii) illustrate 

how the methodology facilitates integration of intbrmation on landscape structure and natural history of D. frontalis, and (iii) 

conclude with an interpretation of how host defenses influence epidemiology of the insect. 277

Approach for Application of the Indices to Define Functional Heterogeneity of Forest Landscapes 

Use of the indices of heterogeneity is predicated on having a spatially referenced database that embraces the features 

of landscape structure essential to D. frontalis. Typically, development of GIS databases requires a substantial commitment 

to map and image processing in order to identify and reference pertinent landscape variables. In the case ofD. frontalis, we 

are particularly interested in the habitat of the insect, i.e., patches of pines (graded by vulnerability) and lightning-struck 

hosts. 

In addition to the spatially referenced database, it is also necessary to define how the organism interacts with its 

environment, i.e., to relate the natural history of the insect to habitat structure. We indicated earlier that a great deal is known 

about the interaction of D. frontalis and its hosts. Literature on this subject was reviewed in our treatme:nt of host defenses. 

An essential feature of D. frontalis/host interaction centers on the process of dispersal. The parameters for this process 

determine how the network of population centers (infestations) can be connected to habitat patches and lightning-struck 

hosts. Although the dispersal process as a whole is poorly understood, certain elements of it have been studied in great detail, 

e.g., response of the insect to semiochemicals (Smith et al. 1993). We define boundaries for the dispersal process, based on 

estimates available from the published literature (Coulson et al. 1979, Pope et al. 1980, Wagner et al. 1984). 

With the GIS database and information on how the insect interacts with its environment, we can use the indices to 

examine •functional heterogeneity of the forest landscape. The general approach involves development of three types of 

maps: hazard, modified hazard, and heterogeneity. The hazard map is an abstract of the variables associated with different 

degrees of vulnerability of forest stands to infestation by D. frontalis. The modified hazard map includes additional information 

on location of lighming-struck hosts, infestation centers, and insect behavior. The heterogeneity map integrates these 

variables, i.e., information on habitat units, population centers, and insect behavior. This map represents a view of how D. 

frontalis might interact with its environment. The map is produced by applying the indices of heterogeneity to the modified 

hazard map. A moving-window calculation (Fig. 3) is used. The moving-window approach is a common technique in 

landscape ecological studies (Turner et al. 1991). It is particularly useful in defining functional heterogeneity, as window 

size and shape can be changed to accommodate different spatial scales (Weins 1989, Levin 1992) (Fig. 4). In the particular 

application, we are interested in the spatial scale used by D. frontalis, which is largely set by the distribution and abundance 

of habitat units and population centers and the dispersal capability of the insect. 

MOVING-WINDOW ANALYSIS 

J - . n 

i I J' +1 n-1 

-{ii{ i i 

: , 

........ 1 ....... !........... 

---Inaut Array .... Output Array-- 

.... _ 

L, ..... 

Figure 3.--The moving-window analysis applied to an array. The arrays can be raster renditions of landscape maps. Each cell 

(i,j) represents a 2-d sample of the landscape. In this example the window is 3X3 cells. In our methodology functional 

heterogeneity is calculated for land attributes in the windows of the input array and placed in the output array. 

278 

n-1

FUNCTIONAL HETEROGENEITYANALYSIS : 

A TECHNIQUE THAT USES BOTH PATTERNAND SCALE 

MEASURED LANDSCAFE FUNCTIONAL LANDSCAPES 

Moving-Window Application of 

the Weighted Connectivity Index 

1 2 

* grayshadesindicatemapped • sizeof movingwindowcorrespondsto scaleof energy, 

landscapezones material,ororganismflows 

, grayshadesindicateindexvalues 

.... ;IFT] 

1, closelyresemblesmeasured :3 ° vaguelyresemblesmeasured 

landscape landscape 

°onlyboundariesare ° indexvaluesspatiallyvaried 

heterogeneous ° landscapeis functionally 

heterogeneous 

Figure 4.--Functional heterogeneity analysis using pattern recognition algorithms and moving-window calculations. 

Relating Functional Heterogeneity to Epidemiology of D. frontalis 

Although D. frontalis is among the most thoroughly studied pest species, there is no mechanistic explanation for how 

the natural history of the insect operates at the landscape scale to produce observed patterns of epidemiology (Dunning et al. 

1992). We suggest that bark beetle epidemiology in a pine forest landscape involves a network of [i] lightning-struck hosts 

(Lovelady 1994), which serve as sinks during dispersal (Pullian 1988, Pullian and Danielson 1991); [ii] previous existing 

infestations, which serve both as sources and sinks; and [iii] high hazard stands, which also serve as sinks and are needed for 

development of infestations. With only modest requirements for dispersal distance (1 kin), the bark beetles can reduce high 

measured heterogeneity in pine forest landscapes into high functional homogeneity, i.e., the insects can link dispersed food 

and habitat resources. Specific attributes of landscape structure will influence how this scenario is played out in different 

forest environ ments. 

In the following example, we examine how D. frontalis interacts within a meso-scale (100 to 1,000,000 ha) forest 

landscape. The GIS-based study (McFadden 1994) was conducted on the Sam Houston National Forest in southeast Texas. 

This landscape is vegetated primarily with loblolly, Pinus taeda L., and shortleaf, P echinata Mill, pines and mixed hardwood 

species. Populations of D. frontalis in the forest have cycled from enzootic to epizootic levels for more than two 

decades. 

279

Data for the study were acquired from thematic maps, aerial photographs, lightning-strike records, and forest management 

records. The thematic map information included: ownership boundaries, roads, utility lines, hydrology, and forest 

compartment boundaries. Locations of D. frontalis infestations were obtained by georeferencing and interpreting aerial 

photographs, using a computer-assisted mapping system called MIPS (Map and Image Processing System). Aerial photographs 

were provided by the USDA Forest Service. Once located on aerial photographs, the UTM coordinates for infestation 

centers were collected and entered into the GIS (ARC/INFO) as point locations. Lightning strike data for the region were 

purchased fi'om R-Scan Corporation. Again, UTM coordinates for lightning strikes were entered as point locations. Information 

on tk_reststands was taken from a database (CISC-Continuous Inventory of Stand Conditions), which is maintained by 

the Forest Service. These data sources were used to develop a sequence of three maps: the hazard, the modified hazard, and 

the heterogeneity. Each map is described below. 

The Hazard Map. First in the sequence is a general D. frontalis hazard map (Fig. 5). This map was developed by 

hazard rating lorest stands in the study area. In this specific application, we used tree species and age as the principal 

variables of hazard. We mentioned previously that several hazard rating systems have been developed (Lorio 1980). The 

utility of this map is that it defines patches in the forest matrix that are suitable for the development of infestations, given the 

presence of the insect (and acceptable weather conditions). The map also reflects habitat patches graded by host defense 

capability, i.e., the high hazard stands have the lowest capacity for host defense. This rendition of the landscape will have 

seasonal trend reflecting variation in host defenses associated with the normal physiological changes associated with tree 

phenological development described by Lorio (1988). The general hazard map will change slowly, i.e., it will require 

periodic updating, perhaps every 2 or 3 years, to remain an accurate portrayal of hazard. 

5 ecies Ty e Hazard Rating 

m_hortleaf _White [] No Hazarct 

Pine /HiGkory 

Loblolly Pine/ [] Low Hazar_i 

ne mHardwo, [] Medium Hazar_ 

Mix 

DSweetguml [] High Hazar_ 

Willow BWater 

Figure 5._Maps of a :forestlandscape classified by species types and hazard to herbivory by D. fi'ontalis. Several systems 

for hazard rating have been developed. All involve integration of a subset of variables: tree species, radial growth, 

height, and DBH; stand basal area (pine, hardwood, total), species composition, site index, and degree of crown 

closure; and landform classification. When applied to a forest landscape, the hazard rating systems provide a general 

view of the distribution and abundance of host defenses against the insect. 

280

The Modified Hazard Map. The next level of complexity in the map sequence involves adding information on the 

location of lightning-struck hosts and existing infestations (Fig. 6). The lightning-struck hosts represent sinks for dispersing 

beetles. We indicated above that these hosts have greatly reduced defense capacity, and they can be located, presumably 

through olfactory cues, by D. frontalis. The role of lightning-struck hosts has been examined in detail in Coulson et al. 

(1986a), Flamm et al. (1992), Lovelady et al. (1991), and Lovelady (1994). The existing infestations can serve both as sinks 

and sources for dispersing beetles. The product of adding the lightning-struck hosts and infestation centers to the hazard is 

illustrated in Figure 7 (left panel). The modified hazard map also incorporates information on dispersal distance of the insect. 

We assumed that the insect could disperse a distance of ca 1 krn. This conservative estimate was based on reports from the 

literature cited above. On the modified hazard map (Fig. 7 - left panel), the dispersal distance was represented as an area (a 

circle with a 1 kin radius) around the lightning-struck hosts and infestation centers. The lightning-struck hosts are point 

locations in the forest matrix. However, these hosts have a substantial impact on epidemiology when their locations are 

considered within the context of the dispersal range ofD. frontalis. The modified hazard map contains the elements of 

landscape structure involved in epidemiology, i.e., hazard rated forest stands, lightning centers, and population centers. The 

addition of dispersal distance provides the means for connecting the elements. This rendition of the landscape has a significant 

weather-related annual pattern. Thunderstorm (and thus lightning) activity follows seasonal cycles (Lovelady 1994). 

Insect development and dispersal behavior also have seasonal trends. The information content of the modified hazard map 

will change during the course of a year, i.e., it will require frequent updating, perhaps weekly or monthly, to remain an 

accurate rendition of hazard, 

Lightning Infestations 

June 1989 June 1989 

Lightning & SPB 

A Strikes Infestations 

S_and 

Stan_ _ 13oundarie5 

A Boundaries 

Figure 6. Maps of a forest landscape with point locations for lightning-struck hosts and existing infestation centers. Lightning-struck 

hosts have greatly reduced defense capacity and are used as refuges by D. frontalis, i.e., they are sinks for 

dispersing populations of the insect. The infestation centers serve as both sinks and sources for dispersing populations 

of the insect. 

281

Modified Stand Hazard Ma Here ene t 

Figure 7. Modified hazard map (left panel) incorporates in:formation on lightning-struck hosts and existing infestation 

centers. This map also contains information on the dispersal behavior of D. ¢?o_talis. The dispersal distance was 

represented as an area (a circle with a 1 km radius) around the lightning-struck hosts and infestation centers. Dis- 

persal and host selection behavior of the insect link the network of high hazard stands, infestation centers, and 

lightning-struck host. The heterogeneity map (right panel) was developed using the AMI index applied to the 

modified hazard map, using the moving-window calculation. This map illustrates how the network of high hazard 

:stands, liglhtning-struck hosts, and existing population centers is connected through dispersal behavior of the insect. 

The Heterogeneity Map. Third in the sequence is the functional heterogeneity map (Fig. 7 - right panel). In this 

particular example, tile AMI index was applied to the modified hazard map, using the moving-window calculation described 

above. Recall that the AMI index is sensitive to dispersion of landscape elements. The functional heterogeneity index allows 

for integration of the information on landscape structure and insect behavior known to influence epidemiology of D. frontatis. 

The heterogeneity map (Fig. 7-right panel) is a visualization of the results. This map illustrates how the network of high 

hazard stands_ lightni_g.-struck hosts, and existing population centers is connected through dispersal behavior of the insect. 

Knowledge of this interaction will certainly increase with new research on epidemiotogy, and this added complexity can be 

accommodated by the approach, Figure 7 (right panel) is a characterization of only one point in time. During the course of 

the year, functional heterogeneity of the landscape will change for the reasons outlined above. Functional heterogeneity will 

also change during successive years of an insect outbreak, as herbivory creates additional fragmentation of the forest matrix. 

Furthermore, there will be a flux in habitat available to the insect in successive years, as high hazard stands are depleted 

through herbivory and new stands grow into the vulnerable age classes. The index values are a quantitative measure of 

ftmctional heterogeneity and therefore can be used for tile analysis and description of sequential map data. Recall, also, that 

the tt_ree indices are sensitive to dilYerent aspects of landscaPe pattern. Therefk)re, each index will provide alternative insights 

into how O_ ._f?ontatis perceives and reacts to its environment. 

Host Defenses and Epidemiology of Dendroctonus frontalis 

"Ik_summarize, at the onset of this study, we indicated that epidemiology of D. frontalis was related to the distribution 

of host defenses in forest landscapes. Previous research had clearly established that infestations typically occur in high 

hazard stands consisting of hosts with reduced capacity for defense against colonization by bark beetles. Hazard rating 

282

systems facilitate mapping of forests according to their vulnerability to bark beetle infestation. These hazard maps define a 

landscape mosaic comprised of forest stands graded by defense capacity (Fig. 5). Typically, the high hazard stands are also 

the preferred habitat of the insect, i.e., populations of bark beetles are clustered in high hazard stands. As long as population 

density remains sufficient to overcome host defenses, infestations grow in size through colonization of neighboring hosts. 

When population density falls below the level necessary for successful host colonization, the insects must disperse to find 

suitable refuges. Two strategies are likely. The insects can disperse to supplement populations in existing infestations at 

other locations or they can colonize lightning-struck hosts, which have greatly reduced defense capacity. These two elements, 

existing infestations and lightning-struck hosts, add to the picture of host defense for the forest landscape (Fig. 7-1eft 

panel). However, forest landscapes in the southeastern US are often highly fragmented, and the vulnerable stands, existing 

infestations, and lightning-struck hosts are dispersed. Behavioral strategies associated with dispersal and host selection link 

populations of the insect to suitable habitat centers (Fig. 7-1eft panel) (Southwood 11977). Finally, the indices of functional 

heterogeneity serve to integrate the information on the distribution of host defenses with knowledge on how the insect 

perceives and responds to its environment (Fig. 7-right panel). Further insight into mechanisms of epidemiology of D. 

frontalis can be gained through analysis of changes in functional heterogeneity of sequential landscape scenes. 

SUMMARY 

This paper focused on functional heterogeneity of forest landscapes and epidemiology of D. frontalis. Specific conclusions 

from the study include the following: 

1. The general premise of this study was that epidemiology of D. frontalis involves a network of high hazard stands, 

lightning-struck hosts, and existing population centers. This network is connected at the landscape scale through dispersal 

behavior of the insect. 

2. The concept of functional heterogeneity (how an organism perceives and responds to its environment) provides a 

useful way to organize knowledge about landscape structure and insect behavior. Although we have focused on functional 

heterogeneity of forest landscapes and epidemiology of D. frontalis, the approach can be used for other insects and taxa. 

3. The :indices of functional heterogeneity were found to be useful in integrating the landscape variables known to 

influence epidem iology of D. frontalis. 

4. Measurements of functional heterogeneity of forest landscapes incorporate two types of knowledge about host defenses, 

The first deals with knowledge embedded in the various types of hazard rating systems used to grade forest stands by 

vulnerability to D. J?oHtalis. The second deals with knowledge on the interaction of lighting-struck hosts and D. frontalis. The 

functional heterogeneity map provides a visualization of the distribution of host defenses within a meso-scale forest landscape. 

5. The functional heterogeneity indices are a tool that can be used to investigate several fundamental issues in epidemiology. 

For example we hypothesize that (i) functional heterogeneity of forest landscapes influences the rate and extent of 

herbivory by bark beetles and that (ii) initiation, growth, and decline phases of forest insect outbreaks can be predicted by 

definition of the change points in functional heterogeneity of landscapes. 


BARBOSA, R and SCHULTZ, J.C., eds. 1987. Insect Outbreaks. Academic Press, San Diego, CA. 

BRANHAM, S J., and THATCHER, R.C., eds. 1985. Proc. Integrated Pest Management Symposium. Gen. Tech. Rep. SO- 

56. New Orleans, LA: U.S. Department of Agriculture, Forest Service. 

BELl.,, S,S._ McCOY, ED., and RUSHINSKY, H.R., eds. 1991. Habitat Structure. Chapman and Hall, NY. 

BROWN, V.K. 1991. The effects of changes in habit structure during seccessin in terresterial communities. In Bell, S.S., 

McCoy, E.D., and Rushinsky, H.R., eds. 1991. Habitat Structure. Chapman and Hall, NY. 

283

COULSON, R.N., PULLEY, RE., POPE, D.N., FARGO, W.S., GAGNE, J.A., and KELLEY, C.L. 1979. Estimation of 

survival and allocation of adult Dendroctonusfrontalis between trees during the development of an infestation. In 

Berryman, A.A. and Safranyik, L., eds. Dispersal of Forest Insects: Evaluation, Theory, and Management Implications. 

Proc. IUFRO Syrup., Sandpoint, ID. 

COULSON, R.N., HENNIER, RB., FLAMM, R.O., RYKIEL, E.J., HU, L.C., and PAYNE, T.L. 1983. The role of lightning 

in the epklemiology of the southern pine beetle. Ziet. Ang. Entomol. 96: 182-93. 

COULSON, R.N., RYKIEL, E.J., SAUNDERS, M.C., PAYNE, T.L., FLAMM, R.O., WAGNER, T.L., and HENNIER, RB. 

1985a. A conceptual model of the role of lightning in the epidemiology of the southern pine beetle. In Safranyik, L., 

ed. The Role of the Host in Population Dynamics of Forest Insects. Proc. IUFRO Syrup., Banff, Canada. 

COULSON, R.N., FLAMM, R.O., WAGNER, T.L., RYKIEL, EJ., SHARPE, P.J.H., PAYNE, T.L., and LIN, S.K. t985b. 

Population dynamics of initiation and growth of southern pine beetle infestations. In Branham, S.J. and Thatcher, 

R.C., eds. Proc. Integrated Pest Management Symposium. Gen. Tech. Rep. SO-56. New Orleans, LA: U.S. Department 


COUt,SON, R.N., FLAMM, R.O., PULLEY, RE., PAYNE, T.L., RYKIEL, E.J., and WAGNER, T.L. 1986a. Response of 

the southern pine bark beetle guild to host disturbance. Environ. Entomol. 15:859-868. 

COULSON, R.N., RYKIEL, E.J., and CROSSLEY, D.A., Jr. 1986b. Activities of insects in forests: implications for wilderness 

area management. In Kulhavy, D.L. and Conner, R.N., eds. Wilderness and Natural Areas in the East: A Management 

Challenge. Stephen E Austin State Univ., Nacogdoches, TX. 

COULSON, R.N., FITZGERALD, J.W., SAUNDERS, M.C., and OLIVERIA, F.L. 1993. Spatial analysis and integrated 

pest management in a landscape ecological context. In Liebold, A.M. and Barrett, H.R., eds. Proceedings: Spatial 

Analysis and Forest Pest management. Gen. Tech. Rep. NE-175. Radnor, PA: U.S. Department of Agriculture, 

Forest Service. 

DUNNING, J.B., DANIELSON, B.J., and PULLIAM, H.R. 1992. Ecological processes that affect populations in complex 

landscapes. Oikos 65. 169-75. 

FITZGERALD, J.W., COULSON, R.N., PULLEY, RE., FLAMM, R.O., OLIVERIA, F.L., SWAIN, K.M., and 

DRUMMOND, D.B. 1994. Suppression tactics _brDendroctonusfrontalis Zimmermann (Coleoptera: Scolytidae): 

an examination of the development of infestations adjacent to treatment sites. J. Econ. Entomol. (In press). 

FLAMM, R.O. and COULSON, R.N. 1988. Population dynamics of the southern pine bark beetle guild in traumatized hosts. 

In Mattson, W. J., Levieux, J., and Bernard-Dagan, C., eds. Plant Resistance Mechanisms to Insects and Pathogens. 

Spri nger-Verlag, NY. 

FLAMM, R.O., COULSON, R.N., and PAYNE, T.L. 1988. The southern pine beetle. In Berryman, A.A., ed. Dynamics of 

Forest Insect Populations. Plenum Press, NY. 

FLAMM, R.O., PUI,LEY, RE., and COULSON, R.N. 1992. Colonization of disturbed trees by the southern pine bark beetle 

guild (Coleoptera: Scolytidae). Environ. Entomol. 22: 62-70. 

FORMAN, R.T.T. and GODRON, M. 1986. Landscape Ecology. John Wiley and Sons, NY. 

HANSEN, A.J. and di CASTRI, F., eds. 1992. Landscape Boundaries: consequences for biotic diversity and ecological 

flows. Ecol. Studies 92. Springer-Verlag, NY. 

HOLMNG, C.S. 1992a. The role of forest insects in structuring the boreal landscape. In Shugart, H.H., Leemans, R., and 

Bonan, G.B., eds. A Systems Analysis of the Global Boreal Forest. Cambridge Univ. Press, Cambridge. 

HOLLING, C. S. 1992b. Cross-scale morphology, geometry, and dynamics of ecosystems. Ecol. Monographs 62: 447-502. 

284 

) ¸7¸¸7 :!; 

:

KOLASA, J. and ROLLO, C.D. 1991. Introduction: The heterogeneity of heterogeneity: a glossary. In Kolasa, J. and 

Pickett, S.T.A., eds. Ecological Heterogeneity. Ecological Studies 86. Springer-Verlag, NY. 

KOLASA, J. and PICKETT, S.T.A., eds. 1991. Ecological Heterogeneity. Ecol. Studies 86. Springer-Verlag, NY. 

LEVIN, S.A. 1992. The problem of pattern and scale in ecology. Ecology 73: 1943-67. 

LIEBOLD, A.M. and BARRETT, H.R., eds. 1993. Proceedings: Spatial Analysis and Forest Pest management. Gen. Tech. 

Rep. NE-175. Radnor, PA: U.S. Department of Agriculture, Forest Service. 

LORIO, RL., Jr. 1980. Rating stands for susceptibility to SPB. In Thatcher, R C., Searcy, J.L., Coster, J.E., and Hertel, 

G.D., eds. The Southern Pine Beetle. Tech. Bull. 1631. Washington, DC: U.S. Department of Agriculture, Forest 

Service and Science and Education Administration. 

LORIO, RL., Jr. 1988. Growth and differentiation balance relationships in pines affect their resistance to bark beetles 

(Coleoptera: Scolytidae). In Mattson, W.J., Levieux, J., and Bernard-Dagan, C., eds. Mechanisms of Woody Plant 

Defenses Against Insects: Search for Pattern. Springer Verlag, NY. 

LOVELADY, C.N., PULLEY, RE., COULSON, R,N., and FLAMM, R.O. 1991. Relation of lightning to herbivory by the 

southern pine bark beetle guild (Coleoptera: Scolytidae). Environ. Entomol. 20: 1279-1284. 

LOVELADY, C.N. 1994. Landscape Study of Bark Beetle Herbivory and the Lightning Disturbance Regime. Ph.D. 

Dissertation, Texas A&M University. 127 p. 

MASON, G.N., LORIO, RL., Jr., BELANGER, R.R, and NETTLETON, W.A. 1985. Rating the susceptibility of stands to 

southern pine beetle attack. Agric. Handb. 645. Washington, DC: U.S. Department of Agriculture, Forest Service. 

McFADDEN, B.A. 1994. Predicting Forest Insect Outbreaks: The Role of Lightning in the Epidemiology of the Southern 

Pine Beetle. MS Thesis, Texas A&M University (In prep.). 

McINTOSH, R.R 1991. Concept and terminology of homogeneity and heterogeneity in ecology, Chapter 2 (p. 23-46). In 

Kolasa, J. and Pickett, S.T.A., eds. Ecological Heterogeneity. Ecol. Studies 86. Springer-Verlag, NY. 

PICKETT, S.T.A. and R S. WHITE, RS., eds. 1985. The Ecology of Natural Disturbance and Patch Dynamics. Academic 

Press, NY. 472 p. 

PLATT, W.J. and STRONG, D.R., eds. 1989. Gaps in forest ecology (special feature). Ecology 70: 535-76. 

POPE, D.N., COULSON, R.N., FARGO, W.S., GAGNE, J.A., and KELLEY, C.W. 1980. The allocation process and 

between-tree survival probabilities in Dendroctonusfrontalis infestations. Res. Pop. Ecol. 22: 197-210. 

PULLIAN, H.R. 1988. Sources, sinks, and population regulation. Am. Nat. 132: 652-661. 

PULLIAN, H.R and DANIELSON, B.J. 1991. Sources, sinks, and habitat selection: a landscape perspective on population 

dynamics. Am. Nat. 137: 550-566. 

RYKIEL, E.J., Jr., COULSON, R.N., SHARPE, RJ.H., ALLEN, T.F.H., and FLAMM, R.O. 1988. Disturbance propagation 

by bark beetles as an episodic landscape phenomenon. J. Land. Ecol. 1: 129-39. 

SCHOWALTER, T.D., COULSON, R.N., and CROSSLEY, D.A., Jr. 1981. Role of southern pine beetle and fire in maintenance 

of structure and function of the southeastern coniferous forest. Environ. Entomol. 10:821-825. 

SMITH, M.T., SALOM, S.M., and PAYNE, T.L. 1993. The Southern Pine Bark Beetle Guild: An Historical Review of 

Research on the Semiochemical-Based Communication System of the Five Principal Species. Virginia Agric. 

Experiment Stn. Bulletin 93-4. 

285

SOUTHWOOD, T.R.E. 1977. Habitat, the templet for ecological strategies. J. Anita. Ecol. 46: 337-65. 

TURNER, M.G., ed. 1987. Landscape Heterogeneity and Disturbance. Ecol. Studies 65. Springer-Verlag, NY .... 

TURNER, M.G. 1989. Landscape ecology: the effect of pattern on process. Annu. Rev. Ecology and Systematics. 20: 11711- 

97. 

TURNER, M.G. and GARDNER, R.H., eds. 1991. Quantitative Methods in Landscape Ecology. Ecological Studies 82. 

Springer-Verlag, NY. t 

TURNER, S.J., O'NEIL, R.V., CONLEY, W., CONLEY, M.R., and GROVER, H.D. 1991. Pattern and scale: statistics for l 

landscape ecology. In Turner, M.G. and Garner, R.H., eds. Quantitative Methods in Landscape Ecology. Ecological 

Studies 82. Springer-Verlag, NY. 536 p. 

WAGNER, T.L., GAGNE, J.A., SHARPE, EJ.H., and COULSON, R.N. 1984. Effects of constant temperature on longevity 

of adult southern pine beetles (Coleoptera: Scolytidae). Environ. Entomol. 13:1125-1130. 

WIENS, J.A. 1989. Spatial scaling in ecology. Functional Ecology 3: 385-97. _ 

WIENS, J.A. 1992. Ecological flows across landscape boundaries: a conceptual overview, Chapter 10 (p. 217-35). In _ 

Hansen, A.J. and di Castri, E, eds. Landscape Boundaries: consequences for biotic diversity and ecological flows. _ 

Ecological Studies 92. Springer-Verlag, NY. 


We acknowledge and thank Audrey M. Bunting for assistance in the preparation of the figures and text of this paper. 

We are grateful to the USDA Forest Service, National Forests in Texas for providing access to GIS databases and forest 

management records; and to the Texas Forest Service for providing access to their operational records on D. frontalis. The 

work reported herein was supported by the USDA Competitive Grants Program [TEX06961 (Landscape Study of Bark Beetle 

Herbivory and the Lightning Disturbance Regime) and TEX08185 (Predicting Forest Insect Outbreaks)], and by the Texas 

Agricultural Experiment Station. This paper is TA No. 31482 of the Texas Agricultural Experiment Station. 

286 

_U.S. _ PRINTING OFFICE: 1996-757--939

) i,i_i)i_)i_i_ 

U.S. Departme_]t of Agriculture, Forest Service. 

] 996. M at tsot_, \Vi ]1iat-1_J,; N ie :_ni1_i,P ekka; Ross i, M at ti, eds. Dy ha,hies o f fores t 

i!f_sll herbi_,ory: Quest for patter,, and principle. (3c,,. I>ch. Rep. NC-183. SI. Paul, 

MN: LLS, Department of AgriculRare, Forest Service, North Central Forcsi Experi- 

meat Station. 286 p. 

Herbivo U on woody phmts is highly var'ktble in both space and time. This 

: proceedings addresses one of its root causes, the highly intricate and dynamic 

relationships that exist between most herbivores and their hosts plants. It emphasizes 

t:ha_ the conseque_-_ces of herbivory both to the consumer and to the producer plant 

oi_e_ balance on a razor's edge .......... depending o_?the exact timing of herbivore am_cks, 

... ;.l_,.t _he specific plant tissues being injured. Herbivory also varies substantially 

amo_'_g ii*_divMual pkmts in relationship to the inherent resistance/susceptibility of 

_/_ it_dividual plants ........ which itself heavily depends on the patticutar physical and biotic 

(c

View or print this publication - Northern Research Station - USDA ...

You also want an ePaper? Increase the reach of your titles

Delete template?

Save as template?