© Journal of The Royal Society of New Zealand Volume 28 Number 3 September 1998 pp 375 403 Paleocene gymnosperms from Mount Somers, New Zealand Mike Pole* Thirteen conifer macrofossils and a gymnospennous leaf of uncertain affinity are documented from the early-mid Paleocene of the Mount Somers coal mine, Canterbury, New Zealand Podocarpaceae are prominent and diverse Two taxa are formally described as new genera of Podocarpaceae, Mumu somerensis gen etsp nov and Twtio imbncatus gen etsp nov A second species of Kakahuta K drinnann, is described There are also two species of Prumnopitys,P hmamae sp nov and/ 5 sp 'Mt Somers' A further five taxa ot Podocarpaceae are informally described Shoots with decussately-arranged scale leaves are identified as Libocedrus cf L bidwilln (Cupressaceae) Paahake papillatus gen et sp nov is compared with Sciadopitvs (Taxodiaceae) and Torreya (Taxaceae) A single small piece of cuticle belongs in the Araucanaceae and is tentatively identified as Araucana An additional new genus and species, Hoiki mcqueenu gen et sp nov , has an unusual epidermal morphology with transversely oriented stomates but is of unknown affinity These species grew in a clastic swamp where Podocarpaceae formed one of the major elements The Libocedrus and Prumnopitys specimens are the oldest macrotossil records of a conifer genus extant in New Zealand New terms are introduced for stomatal distribution on conifer leaves Keywords Araucanaceae, Cupressaceae, New Zealand, Paleocene Podocarpaceae, paleobotany, Taxaceae, Taxodiaceae cuticle terminology INTRODUCTION The Mount Somers Coal Mine lies approximately 100 km west of Chnstchuch, New Zealand (Fig 1) The coal and associated carbonaceous mud (sitting on a basement of weathered Mount Somers Volcamcs of mid Cretaceous age) formed on a broad alluvial plain along the paleo-east coast (the Taratu Formation of Carter 1988) which was subsequently submerged by a widespread marine transgression Van der Lmgen (1988) described the geology of the Mt Somers area He included the coal measures within the Broken River Formation of Field & Browne (1986) However, Carter's broader application of the Taratu Formation is preferred here Raine & Wilson (1988) documented the palynology of the Mt Somers region and, based on pollen, placed all the sediments into Zone PM3 of Rame (1984) This is correlated with the Teunan (covering most of the Paleocene) - early Waipawan (Late Paleocene-earhest Eocene) stages They considered the upper, marine part of the sequence was "close to the upper boundary of Zone PM3, and is likely to be of Waipawan age" A sample from the "upper unit of the Pottery Clay Quarry" (about 1 km from the coal mine), was shown as stratigraphically equivalent to a horizon immediately above the highest coal in the Mt Somers mine Dmoflagellate cysts dated this sample to the Palaeocystodimum austrahnum Zone of Wilson (1984) This is correlated with the early-middle Teunan stage The plant *Department of Plant Science, University of Tasmania, G P O Box 252C, Hobart, 7001, Australia Present address Department of Botany, University of Queensland, St Lucia, Brisbane, QLD 4072, Australia 376 Journal of The Royal Society of New Zealand, Volume 28, 1998 Fossil Locality Mt Somers Coal Mine South Branch Gorge Road 1 km Fig. 1 Locality map of the Mount Somers coal mine. fossils described here are therefore regarded as early-mid Paleocene (based on correlation of the New Zealand scale with the International in Edwards et al. 1988). The Mt Somers Coal Mine was visited in 1992. No plant macro fossils were found in-situ. However, several blocks of fossiliferous carbonaceous mudstone were collected from a dump at the edge of the pit. The mudstone forms part of the overburden and is likely to come within two or three metres of the top of the coal. Both angiospermous and gymnospermous macrofossils were obtained from this mudstone. The gymnospermous elements are described in this paper. METHODS A number of individual shoots and leaves were located by splitting mudstone and detaching fragments of leaf with tweezers. After removal of silica with hydrofluoric acid the leaves could be mounted directly on a scanning electron microscope (SEM) stub or reduced to cuticle using aqueous chromium trioxide. Cuticle fragments were mounted in glycerine jelly for transmitted light microscope (TLM) viewing, or on stubs for SEM viewing. However, most material was obtained by disaggregating blocks of mudstone in hydrogen peroxide and then sieving to obtain dispersed cuticle and leaves. Formal taxonomy on fossils is here restricted to specimens including cuticle where at least the leaf shape can be determined. Even though the cuticular characteristics of very small fragments of cuticle may appear distinct at the generic level, these are considered too fragmentary on which to erect formal new taxa. Pole—Paleocene gymnosperms from Mount Somers 377 All fossils of a taxon are assumed to include the adult foliage. This is justified on the basis of the extreme unlikelihood of encountering the seedling form of a conifer without any of the adult form which produces orders of magnitude more foliage biomass. Terminology of cuticular features mostly follows earlier works, for instance Florin (1931) and Wells & Hill (1989a). 'Subsidiary cell' refers to any cell in contact with the guard cells and which is distinguished from normal epidermal cells, either by shape, thickness, or texture of overlying cuticle. The restricted sense of Stockey & Ko (1988) where subsidiary cells are identified by a difference in overlying cuticle micromorphology is not adopted. In my opinion this leads to problems in homology with the polar subsidiary cells. 'Encircling cell' here refers to a cell which was evidently descended from the same immediate mother cell as a subsidiary cell, and is distinguished from normal epidermal cells, usually by shape (the term was used by Florin 1931). Some new, or modified terms for stomatal distribution are introduced here as Appendix 1. Where a definite family-level identification has been made, this has been formalised as an IF-THEN rule (Pole 1995). Their purpose is to make clear the characters that have led to identification, and in a manner which can be databased for future inclusion in expert-systems. These rules are listed as Appendix 2. The collection locality has been allocated the Geological Society of New Zealand Fossil Record number K36/f79. Fossil hand specimens and slide-mounted cuticle are catalogued with the prefix 'SB', SEM stubs are prefixed with 'S', and cuticle mounts of herbarium material are prefixed with 'AQ' or 'OPH'. All material is stored in the Department of Plant Science, University of Tasmania. TAXONOMY Coniferopsida Araucariaceae Araucaria Leaf (Fig. 2) form unknown (single fragment only, 1.5 x 2.0 mm), amphistomatic, 20-21 rows of stomates on each surface, assumed multiveined. Stomates in discrete rows, short, tight chains, longitudinally oriented, dicyclic, paratetracyclic. Subsidiary cells sunken, inserted lower than level of insertion of encircling cells (no raised rim around stomatal pore), walls smoothly curved. Stomatal aperture elongate. Guard cells deeply sunken, polar extensions prominent, lateral extensions reduced. Epidermal cells straight walled, smooth, outer surface subdued, outline isodiametric to slightly elongate. Glabrous. Family placement follows Araucariaceae leaf placement rule #1.2 (Appendix 2). The specimen is tentatively placed in Araucaria. It is none of the other known genera in the Araucariaceae (Agathis, Araucarioides (Bigwood and Hill 1985), and Wollemia (pers. obs.) have obliquely oriented stomates). However, the virtual absence of elongate epidermal cells, which are characteristic of all extant species of Araucaria (pers. obs.) hints that more complete specimens may suggest a new genus. If it truly belongs in Araucaria, there are not enough characters known to allow a definite sectional placement, although the longitudinally oriented stomates and flattened leaf form rule out section Eutacta (Bigwood & Hill 1985). Araucaria sp. (Fig. 2) Referred specimen: SB 1401 (one cuticular surface), S913 (the other cuticular surface) Family Cupressaceae Libocedrus Shoots (Fig. 3) with opposite and decussately inserted leaves. Some shoots with distinct 378 Journal of The Royal Society of New Zealand, Volume 28, 1998 Fig. 2 Araucariaceae gen. et sp. indet: (A) TLM, showing stomatal rows, SB 1401, scale = 200 urn; (B) TLM, detail of stomates, SB1401, scale = 50 um; (C) SEM, detail of outer stomatal surface, note lack of any raised rim, S913, scale =10 urn; (D) SEM, detail of inner cuticle of stomate, S913, scale = 10 (im. facial and lateral leaves, others without clear distinction. Cuticle very delicate. Stomates probably on adaxial surface (no stomates visible when directly mounting a shoot fragment for SEM viewing), in overlapping rows, monocyclic (no obvious differentiation in thickness of subsidiaries from ordinary epidermal cells), paratetracyclic, aligned parallel to long axis of shoot, with distinct ring of fused papillae around aperture. Guard cells deeply sunken, polar extensions prominent, lateral extensions reduced. Glabrous. Family placement follows Cupressaceae placement rule #1 (Appendix 2). The material is identified as Libocedrus based on leaf form and the longitudinally oriented stomates in overlapping rows. Specific identification is not attempted (due to difficulty in preparing adequate cuticle) although the leaf form is comparable with one of the two extant New Zealand species, L. bidwillii Hook, f., which has indistinct facial and lateral leaves on unflattened shoots, but is clearly different from the other, L. plumosa (D. Don) Sargent, which has very distinct facial and lateral leaves on flattened shoots. Libocedrus sp. cf. L. bidwillii (Fig. 3) Referred specimens: Several shoot fragments S446, S455, 456, S458, S886, SB1397, 1398 Family Podocarpaceae Kakahuia Shoot (Fig. 4) with spirally arranged, distichous, spreading leaves. Leaves straight, singleveined, bifacially flattened, lanceolate, length 9-10 mm, width c. 1.2 mm, margin smooth, apex acute, base narrowing to false petiole. Stomatal distribution hypostomatic, not clear if Pole—Paleocene gymnosperms from Mount Somers 379 A Fig. 3 Cupressaceae gen. et sp. indet. A-E, SEMs of shoots: (A) S455; (B) S445; (C) S458; (D) S456; (E) S446; A-C at same scale, D, E at same scale (in mm units); (F) TLM detail of stomates, note overlapping lateral subsidiary cells, SB 1397, scale = 20 um; (G) TLM detail of stomate, SB 1398, scale = 20 u,m; (H) SEM detail of outer stomatal surface with prominent rings of fused papillae, S886, scale = 10 u.m; (I) SEM detail of inner cuticle of stomate, S886, scale = 10 urn. 380 Journal of The Royal Society of New Zealand, Volume 28, 1998 stomates in distinct zones. Stomates in about 10 overlapping rows, widely-spaced within rows (such that the nearest neighbour of a stomate is often in an adjacent row), in loose chains or isolated, separated within rows by typically markedly isodiametric polar encircling (and perhaps some ordinary epidermal) cells, cells forming outer stomatal rows more elongate. Stomatal complex dicyclic, basically paratetracyclic, longitudinally oriented. Subsidiary cells inserted at same level as encircling cells, raised to form Florin ring. Polar subsidiaries mostly short, sometimes abutting adjacent stomate in row, outline flattened. Guard cells deeply sunken, polar extensions pronounced, slightly expanded, lateral extensions reduced. Stomatal aperture elongate. Many, but not all, encircling cells (or other cells in stomatal row except subsidiary cells) display low papillae - one per cell and normally at one end of cell. Papillae always circular in section, broad and low, isolated. Abaxial epidermal cells rectangular, smooth, low, but distinct papillae on some cells, glabrous. Adaxial epidermal cells markedly butressed, not sinuous, rectangular (broader than lower epidermal cells), glabrous. Family placement follows Podocarpaceae placement rule #3 (Appendix 2). The characters of the fossil fall within the diagnosis of Kakahuia, another Paleocene conifer from New Zealand (Pole 1997). Kakahuia also has buttressed upper epidermal cells, and papillate lower ones. There is a large difference in degree however, between the type species, K. campbellii, and the Mt Somers fossil. K. campbellii has very prominent papillae, often two or three, on all lower epidermal and neighbouring cells, and sometimes even on the polar subsidiary cells, which contrasts markedly with the single and more scattered papillae of the Mt Somers fossils. K. campbellii does not have the isodiametric cells in the stomatal rows. The stomatal distribution of Kakahuia is similar to Prumnopitys (see comments below) and in my opinion the two may be related. Stomatal distribution and leaf form is similar to Retrophyllum Page and Willungia Hill & Pole (where the leaves taper gradually from a maximum width near the base and are not falcate). Strong buttressing of the epidermal cells is found in Willungia, though on the stomatal surface. Kakahuia Pole (Pole 1997, 1998) Kakahuia drinnanii sp. nov. (Fig. 4) Holotype: SB990 (portion of shoot on matrix). Diagnosis: A Kakahuia with stomates separated within rows by distinct isodiametric cells, many cells having a single, low papilla. Etymology: Named for Andrew Drinnan, who was responsible for our visit to the Mt Somers locality. Mumu gen. nov. Shoot arrangement unknown, probably spreading. Leaf (SB989, Fig. 5) long, thin, bifacially flattened, falcate, 28 mm long, c. 1.2 mm wide, single veined, margin smooth, apex acute, base narrows gradually to long false-petiole; stomates distributed in two zones on abaxial leaf surface (hypostomatic), 7-10 stomatal rows per zone, rows in contact (not separated by rows of epidermal cells), often overlapping, cuticle over stomatal rows much thicker than over normal epidermal cells, (in some stomates cuticle over lateral subsidiary cells much thicker than over polar subsidiary cells); Stomatal complex dicyclic, paratetracyclic, typically with four subsidiary cells (but variable, both polar and lateral subsidiary cells frequently divided), subsidiary cells from adjacent rows often touching, sometimes separated by one or two, irregular, cuspate encircling cells (staining darker than epidermal cells), orientation longitudinal. Polar subsidiary cells either shared or abutting adjacent stomate in row (rarely with additional cell in between), elongate (with polar subsidiary cells projecting far beyond lateral subsidiary cells) or ovoid (polar subsidiary cells not projecting beyond lateral subsidiary cells). Guard cells deeply sunken, polar extensions pronounced, lateral extensions reduced. Pole—Paleocene gymnosperms from Mount Somers 381 Fig. 4 Kakahuia drinnanii sp. nov., SB990: (A) Shoot on bedding surface, scale = 1 mm; (B) TLM, stomatal zone on lower surface, arrow indicates papilla, scale = 50 |im; (C) TLM, buttressed epidermal cells on upper surface, scale = 50 u.m; (D) SEM detail of outer cuticle surface, note stomate surrounded by raised rim and low, irregular papillae on other cells, S915, scale = 10 (O.m; (E) SEM detail of inner cuticle of stomate, scale =10 urn. Stomatal aperture elongate. Subsidiary cells flush with rest of leaf surface (no raised rim around stomatal pore). Abaxial and adaxial epidermal surface subdued, epidermal cells long, smooth, straight-sided, end-walls flat or oblique, glabrous. Family placement follows Podocarpaceae placement rule #3 (Appendix 2). Densely packed stomatal rows are found in three extant genera of podocarps; Acmopyle, Saxegothea, and some Podocarpus, while the variable subsidiary cell number is particularly comparable with Acmopyle. All Podocarpus species differ from the fossils in having stomates in clear chains and which are very regular in their outline and paratetracyclic construction. Saxegothea differs in having a regular stomatal form with usually isodiametric polar subsidiary cells which are 382 Journal of The Royal Society of New Zealand, Volume 28, 1998 often shared, linking the stomates into very long tight chains. One of the extant species of Acmopyle {A. sahniana Buchh. et Gray) has stomates in two zones per surface, foliar trichomes and distinctive isodiametric hypoplastic stomates. Fossil leaves with these characters are easily identified as Acmopyle. The other extant species {A. pancheri (Brong. et Gris) Pilger), has two stomatal zones, distinctive hypoplastic stomates, but no trichomes. Fossil leaves with these characters may also be placed in Acmopyle. A single fossil has been placed in Acmopyle (A. tasmanica Hill & Carpenter) even though it has only a single stomatal zone, no trichomes, has no hypoplastic stomates, and differs from extant Acmopyle (amphicyclic: Florin 1931) by having all monocyclic stomates. This broadening of the generic boundary increases the diagnostic importance of other, more subtle characters. In Acmopyle, the cuticle over subsidiary cells is virtually indistinguishable from that of normal epidermal cells based on thickness/staining characteristics using TLM. However, SEM studies have indicated the cuticle over subsidiary cells in at least some species is distinguished by its granularity from that over normal epidermal cells. For A. tasmanica this feature, as well as the deeply sunken guard cells compared with heavily cutinised "epidermal" (presumably subsidiary cells as used here) cells, was used by Hill & Carpenter (1991) for placement in Acmopyle. Undoubted Acmopyle also have a distinctive broad stomatal form where both the polar and lateral subsidiary cells are very wide - about twice as wide as the width of the guard cell pair, and the lateral subsidiary cells tend to be short and not enveloping the polar subsidiary cells. This feature does vary within a leaf though; some subsidiary cells are not as wide, whereas some polar subsidiary cells are even wider. Thus to be at least considered as an Acmopyle, a leaf must have at least one of the following characters: trichomes, isodiametric hypoplastic stomates, or the distinctive stomatal shape. The Mt Somers fossils deviate sufficiently to exclude them from Acmopyle. They do not have trichomes or hypoplastic stomates, do have a generally elongate stomatal outline, and have much thicker cuticle over subsidiary and encircling cells than epidermal cells. Finally, while both extant Acmopyle species, and all fossil species (as presently known), are bilaterally flattened, the Mt Somers fossils are bifacial. They are distinct from all known podocarp genera and are placed in a new genus, Mumu. Mumu gen. nov. Etymology: Mumu is Maori for "a gentle noise" and is employed here to refer to the sound of wind through long, narrow leaves (Tregear 1891). Diagnosis: a conifer which has single-veined, linear, falcate, bifacially flattened leaves; stomates in densely packed rows within which cuticle over all cells is markedly thicker than over normal epidermal cells; stomates essentially paratetracyclic, elongate and longitudinally oriented (but which show much variation in subsidiary cell number, and polar subsidiary cell length) and occur in short, tightly chained rows. Type species: Mumu somerensis Mumu somerensis sp. nov. (Fig. 5) Diagnosis: As for genus. Etymology: From the type locality. Fig. 5 Mumu somerensis gen. et sp. nov., (A-C) SB989, (D-F) S640, (G) S640: (A) Leaf on bedding surface, scale = 1 mm; (B) TLM, stomatal zone, scale = 0.30 mm (this specimen later transferred to S625); (C) TLM, detail of stomates, scale = 50 |im; (D) SEM, inner surface of stomate. Note elongate polar subsidiary cells, scale = 20 urn; (E) SEM, inner surface of stomate. Note moderate-length polar subsidiary cells, scale = 20 urn; (F) SEM, inner surface of stomates. Note very short polar subsidiary cells enclosed by lateral subsidiary cells, scale = 20 um; (G) SEM, outer surface of stomatal zone. Note obscure, plugged stomatal pores, scale = 20 urn. Pole—Paleocene gymnosperms from Mount Somers 383 384 Journal of The Royal Society of New Zealand, Volume 28, 1998 Holotype: SB989 (nearly complete leaf on matrix). Referred specimens: SB1121, 1123, 1135-1141 SEM mounts: S625 (dispersed cuticle is very common) S625, 640 Prumnopitys sp. nov. Shoot (Fig. 6) with spirally arranged, distichous, spreading leaves. Leaves bifacially flattened, falcate, single veined, linear, length 10-12 mm, width c. 1.0-1.2 mm; margin smooth, apex acute, base narrowing to false petiole. Stomates distributed in two zones on each leaf surface (equally amphistomatic), each comprised of 3-5 sometimes overlapping rows, often widelyspaced along each row (nearest neighbour usually in adjacent row), typically isolated or in loose chains, polar subsidiaries sometimes abutting, rarely shared (if at all - preservation poor), projecting slightly. Stomatal complex dicyclic, basically paratetracyclic, narrowovoid, orientation longitudinal; outline flattened. Subsidiary cells inserted at same level as encircling cells, raised to form distinct (but sunken) Florin ring. Guard cells deeply sunken, polar extensions pronounced, lateral extensions reduced. Partially-formed stomates present. Encircling cells often large, as long as or longer than stomatal apparatus. Abaxial and adaxial epidermal cells polygonal, moderately elongate to nearly isodiametric, smooth, straight walled, unbuttressed (or very slightly buttressed) anticlinal walls, outer surface highly sculptured clearly showing outlines of epidermal cells, glabrous. Family placement follows Podocarpaceae placement rule #3 (Appendix 2). Pole (1992a, p. 445) claimed that "Podocarpus has stomates in well defined, well spaced uniseriate rows, whereas in Prumnopitys the stomatal rows are not clearly uniseriate; the nearest neighbour of a stomate often lies in an adjacent row." Using the terminology introduced in this paper (Appendix 1; Fig. 16), the stomatal rows of Podocarpus are discrete, and the stomates are typically in very long, tight chains, while in Prumnopitys both discrete and overlapping stomatal rows occur and the stomates are typically in loose chains, or isolated. Stockey & Frevel (1997, p. 217) have recently remarked that "such rows [well-defined and well-spaced uniseriate] may in fact occur in some [Prumnopitys] species, e.g. P. ferruginoides and P. amara". One of their figures of P. ferruginoides (fig. 32) certainly does show some stomates having a nearest neighbour in the same row and also shows three stomates with polar cells abutting. To test my contention of where the nearest stomatal neighbours lie in P. ferruginoides I sampled nine herbarium specimens of this species (AQ141420, 141424, 141425, 391463, 391478, 391489, 391502,415616,415627) and observed that of 475 stomates over 80% had the nearest neighbour in the adjacent row. Out of this number only two pairs of stomates were tightly linked and the longest of these rare tight chains found anywhere was only three stomates long. As a generalisation my 'rule of thumb' would seem to hold. As regards P. amara, this species was not considered by me as Page (1988) had formally placed it into its own genus, Sundacarpus. Stockey & Frevel's (1997, p. 217) observation that this species has stomata "oriented parallel to the long axis of the leaf overlooks the fact that there is a wide range of variation. In one typical specimen (on 352 stomates of AQ141241, pers. obs.) 18% of the stomates are oriented more than 20° away from the axis, with some at 90°. This range of orientation makes it utterly unique amongst the broad leaved podocarps. While my observation was made to distinguish Prumnopitys from Podocarpus, a more important question is whether that kind of stomatal distribution uniquely identifies Prumnopitys. In my opinion the only other genus with leaves approaching this stomatal distribution is Retrophyllum (Pole pers. obs.). Thus stomatal distribution needs to be used in conjunction with other characters for more certain identification. Other features in the fossils described above suggest that Prumnopitys is indeed the correct identification. These features are not found in all Prumnopitys but combine to sway a value judgement. They include the spirally disposed, falcate leaves, the sunken Florin ring (with a corresponding crease in the subsidiary cell cuticle - see Stockey & Frevel 1997), the Pole—Paleocene gymnosperms from Mount Somers 385 Fig. 6 Prumnopitys limaniae sp. nov. SB992: (A) Shoot on bedding surface, scale = 1 mm; (B) TLM, stomatal zone, scale = 50 |im; (C) TLM, stomatal zone, scale = 50 u,m; (D) SEM, outer cuticle surface, note distinct but sunken Florin ring and outlines of other epidermal cells, S884, scale = 10 |am; (E) SEM detail of inner cuticle of stomate, S884, scale = 10 urn. outline of the epidermal cells visible on the outer leaf surface (also noted by Stockey & Frevel 1997), and stomatal rows that often overlap. The new fossil differs from the other equally amphistomatic Prumnopitys species, P. portensis Pole, from the Tasmanian Eocene, in having non-sinuous and non buttressed epidermal cell walls, and a narrow stomatal outline. It is therefore regarded as a new species. Prumnopitys limaniae sp. nov. (Fig. 6) Etymology: Named after Maria Limani, for supporting my research. Diagnosis: A Prumnopitys having leaves which are under 10 mm wide, equally amphistomatic with stomates in two zones on each surface, non-sinuous and non buttressed epidermal cell walls, and a narrow stomatal outline. Holotype: SB992 (portion of shoot on matrix). Prumnopitys sp. 'Mt Somers' Complete leaf unknown (Fig. 7). Leaf bifacially flattened, single veined, c. 1 mm wide, linear. Stomates distributed in (probably) two zones, of six overlapping stomatal rows. 386 Journal of The Royal Society of New Zealand, Volume 28, 1998 mpBm B .. »3ferA t, i>r-\ Fig. 7 Prumnopitys sp. 'Mt Somers': (A) TLM, abaxial surface with single stomatal zone and probable gap over midvein at upper edge, scale = 200 urn; (B) TLM, adaxial surface, showing epidermal cells with very angular-sinuous walls, scale = 200 um; (C) TLM, stomates, note in central stomate both lateral subsidiary cells have divided, scale = 50 urn; (D) TLM, detail of stomate. Note typical 'inflated' outline with small polar subsidiary cells, scale = 20 urn. Stomates typically widely-spaced along each row (nearest neighbour usually in adjacent row), isolated or in loose chains. Stomatal complex dicyclic, basically paratetracyclic, outline inflated; polar subsidiary cells typically small, isodiametric, lateral subsidiary cells broad, sometimes divided, orientation longitudinal. Subsidiary cells inserted at same level as encircling cells, raised to form distinct Florin ring. Guard cells deeply sunken. Abaxial epidermal cells elongate, sinuous, buttressed; adaxial epidermal cells elongate, sinuous but with more sharply angled than abaxial cells, glabrous. Family placement follows Podocarpaceae placement rule #3 (Appendix 2). Furthermore, this fragment of cuticle exhibits typical characters of Prumnopitys and is placed in that genus. These include the stomatal distribution, stomatal outline, and epidermal cell shape. The specimen is likely a distinct species but because the complete leaf form is unknown, no formal taxonomy will be attempted. Prumnopitys sp. "Mt Somers" Fig. 7 Referred specimen: SB 1400 Tiotio gen. nov. Shoot (Fig. 8) with spirally arranged, loosely imbricate leaves. Leaves not flattened, broadly keeled, single veined, elongate, curved, smooth margined, length 4—5 mm, width c. 1 mm, Pole—Paleocene gymnosperms from Mount Somers 387 apex acute, base not narrowing, attached by long decurrent adnate base Stomates distributed in four equally spaced zones (two zones on each leaf surface - equally amphistomatic), zones of 2-3 discrete rows and short, tight chains, irregular outer edge of encircling cells Withm each zone, stomatal rows separated by one or two epidermal cells Stomatal complex dicyclic (sometimes partly tncychc), paratetracychc, typically with four subsidiary cells, oriented longitudinally, aperture elongate, shape irregular, from ovoid with polar subsidiary cells projecting to foreshortened Polar subsidiary cells often shared or abutting, enclosed by lateral subsidiaries Subsidiary cells inserted at same level as epidermal cells, flush with rest of leaf surface (no raised rim around stomatal pore) Guard cells deeply sunken, lateral and polar extensions reduced Partially formed stomates noted Epidermal cells rectangular to slightly wavy, smooth, glabrous Family placement follows Podocarpaceae placement rule #1 (Appendix 2) The leaf shape, with four equally spaced zones of stomates which are linked in chains, suggests the fossils have an affinity with extant Dacrycarpus (e g compare with species illustrated in Wells & Hill 1989a, b, Hill & Carpenter 1991, Pole 1992b) However, while extant Dacrycarpus consists of nine species which all have very similar epidermal features, fossil material has been included in this genus on the basis of epidermal features which he well outside the limited variation shown today The New Zealand species, Dacrycarpus dacrydioides, is not distinct as claimed by Wells & Hill (1989a, b), in that it has completely cutmised guard cells and buttressed epidermal cell walls These comments follow partly from Wells & Hill's undermacerated specimen attributed to D dacrydioides Boulter (1970, 1971) pointed out that conifer guard cells are usually lignified and that such material remains if maceration has been incomplete When specimens have been gold plated and are viewed under SEM, the distinction between hgnin and cutm cannot be seen However, under TLM ligmn does not stain along with cutin I can confirm that D dacrydioides does not have cutmised guard cells, although it, and other species of Dacrycarpus, retain a hgnm sheath around the guard cells when they are deliberately under-macerated I have not seen any D dacndioides (including the herbarium specimens observed by Wells and Hill), or any other fully macerated Dacrycarpus with buttressed epidermal cell walls Very rarely there are sinuosities in the walls but they are never associated with buttressing The two illustrations given by Wells and Hill of D dacrydioides are puzzling in other respects Dacrycarpus stomates are typically elongate, though they are often nearly rounded, as shown in their illustrations However, when D dacrydioides stomates become compressed to this form, thev almost invariably share a polar subsidiary cell When the stomates are in such close proximity that the ends of lateral subsidiary cells are in contact, and thus enclose the polar regions, the reduction to a single, shared polar subsidiary cell is absolute (pers obs) In Wrells & Hill's illustrations there are three instances where the stomates are compressed to this degree, but two polar subsidiary cells remain The cuticle in these photographs is more similar to Dacrvdium cupressinum In the fossil record, two specimens placed into Dacrycarpus by Wells and Hill, D acutifohus and D involutus both have monocyclic stomates This does not occur in any extant Dacrvcarpus, in fact amongst extant Podocarpaceae it only occurs in Microcachrys and some stomates of Microstrobos D acutifohus has a very characteristic thickened cuticular rim around the guard cell cuticle This is entirely unknown in Dacrycarpus but is very comparable with that shown by Mesibovia rhomboidea Wells and Hill, later in the same paper (which is also monocyclic) This new genus, which was regarded as clearly a member of the Podocarpaceae at the time, was later (Hill et al 1993) placed in the extant Tasmaman genus Athrotaxis of the Taxodiaceae I suggest that family is where the affinities of D acutifohus and D involutus, lie Jordan (1995) described a new species of Dacrycarpus, D carpenteru, from the Tasmaman Pleistocene Along with prominently buttressed epidermal and subsidiary cell walls, it has 388 Journal of The Royal Society of New Zealand, Volume 28, 1998 monocyclic stomates where lateral subsidiary cells are sometimes shared between stomates of the same row. None of these features occur in Dacrycarpus. The last feature is restricted, being noted in Microcachrys (Podocarpaceae; Wells & Hill 1989) and Neocallitropsis (Cupressaceae; pers. obs.). D. carpenterii is not a Dacrycarpus and the affinities might lie outside of Podocarpaceae. In the Mount Somers material, there are subtle but consistent differences from extant Dacrycarpus species (Fig. 9). The stomatal complexes in the fossils are usually compressed lengthwise, often so much that the polar subsidiary cells, which are usually isodiametric, are enclosed by the lateral subsidiaries of two stomatal complexes. This happens infrequently in extant Dacrycarpus. The lateral subsidiary cells in the fossils often have an angular, or irregular outline, compared with the usually smooth curve in extant Dacrycarpus. The outer edge of the encircling cells usually form a straight edge in extant Dacrycarpus (Fig. 9) but the fossils lack this regularity. Finally, Florin rings are absent from the outer cuticular surface although they are present in extant Dacrycarpus. This suggests the fossils belong to a genus distinct from Dacrycarpus. Supporting this conclusion, Dacrycarpus pollen is absent from both Mt Somers (Raine & Wilson 1988), and indeed, from pre-Eocene sediments in New Zealand (Mildenhall 1980) although Raine (1984) lists Dacrycarpites sp. in the Late Cretaceous (but not in the Paleocene). For these fossils a new genus is erected, Tiotio. Tiotio gen. nov. Diagnosis: a conifer with single-veined, unflattened, spirally disposed and loosely imbricate leaves which have longitudinally oriented, paratetracyclic stomates with an angular outline, polar subsidiary cells often shared, outer edge of stomatal rows (as defined by neighbour cells) irregular, and no Florin rings. Etymology: Tiotio is Maori for "having sharp points or projections" (Tregar 1891), referring here to the shoot. Type Species: T. imbricatus Tiotio imbricatus sp. nov. (Fig. 8) Diagnosis: As for genus. Etymology: From the imbricate nature of the leaves. Holotype: SB1117 (TLM mount of cuticle from isolated leaf). Referred specimens: whole leaves or shoots: SB991; dispersed cuticle on TLM mounts: SB996, 997, 1116, 1120, 1142-1145; dispersed cuticle on SEM mounts: S448, 452, 454, 623, 624 (a common component of the Mount Somers assemblage). Mt Somers podocarp sp. A. (Fig. 10) Reference specimen: SB 1004. Referred specimens: SB1002, 1003, 1127-1134. Description: Shoot with spirally arranged, imbricate leaves. Leaves not flattened, broadly keeled, single veined, scale-like, length c. 1 mm, width c. 0.5-1 mm; apex acute with marginal frill, possibly epistomatic (uncertain); stomates randomly distributed in single zone. Stomatal complex dicyclic, encyclocytic with typically 4 subsidiary cells, typically circular, or polar subsidiary cells projecting slightly beyond margin of lateral subsidiary cells, oriented essentially parallel to axis of leaf, but sometimes oblique. Subsidiary cells inserted at same level as encircling cells. Guard cells deeply sunken, polar and lateral extensions reduced. Stomatal aperture unclear. Subsidiary cells raised to form Florin ring. Epidermal cells slightly rectangular to irregular, isodiametric, smooth walled, glabrous. Family placement follows Podocarpaceae placement rule #2.2 (Appendix 2). There is no Pole—Paleocene gymnosperms from Mount Somers 389 Fig. 8 Tiotio imbricatus gen. et sp. nov.: (A) SEM of shoot portion, S448; (B) SEM of shoot portion, S452; (C) SEM of shoot portion, S454; (D) Shoot on bedding surface, SB991; (E) TLM, stomatal rows, scale = 50 urn, SB 1120; (F) TLM, detail of stomates, SB 1120, scale = 100 um; (G) SEM, inner surface, stomatal row, S623, scale = 50 um; (H) SEM, outer surface, stomatal apertures, note lack of Florin rings, S624, scale = 20 um. obvious identity with an extant genus. Overall morphology combined with the generally longitudinal orientation of the stomates suggests comparison with Lepidothamnus and Microstrobos. However the fossils do not have the large, inflated style of stomate of the former, or the tightly chained stomates of the latter. Further identification will await better material. 390 Journal of The Royal Society of New Zealand, Volume 28, 1998 Fig. 9 Extant Dacrycarpus spp. (TLM): (A) D. cummingii (Parl.) De Laub., E/1040 (B) D. dacrydioides (A. Rich.) De Laub., OPH2686; (C) D. expansus De Laub., E/0924; (D) D. imbricatus (Blume) De Laub., OPH2109; (E) D. kinabaluensis (Wassch.) De Laub., E/1505; (F) D. steupii (Wassch.) De Laub., E/1446; (G) D. viellardii (Parl.) De Laub., E/1445; (H) D. cummingii, OPH2687. A-F scale = 100 urn, G-H scale = 50 um. Pole—Paleocene gymnosperms from Mount Somers 391 Fig. 10 Mt Somers podocarp sp. A.: scale -leaf: (A) TLM of single scale-leaf, SB1003, apex is to left, scale = 10 urn; (B) TLM of stomatal zone, SB 1004. Long axis of leaf is approximately horizontal - note some obliquely oriented stomates, scale = 100 u.m; (C) TLM of leaf apex, SB 1004, scale = 10 u.m; (D) TLM detail of single stomate, SB 1004, scale = 20 um. Mt Somers podocarp sp. B (Fig. 11 A, B) Reference specimen: SB994. Description: Shoot form unknown. Leaf flattened, size unknown. Stomatal distribution amphistomatic; number of zones unknown. Stomates in discrete rows, paratetracyclic; in short, mostly tight chains, longitudinally oriented. Polar subsidiary cells mostly shared, projecting strongly from overall stomatal outline (not enclosed). Lateral subsidiaries strongly arched, so stomates appear fore-shortened. Encircling cells forming straight-edges to rows. Epidermal cells rectangular to slightly wavy, smooth, glabrous. Family placement follows Podocarpaceae placement rule #3 (Appendix 2). Mt Somers podocarp sp. C (Fig. 11C, D) Reference Specimen: SB995 (dispersed cuticle only). Description: Shoot form unknown. Leaf probably flattened; stomatal distribution over leaf unknown, probably in two zones either side of clear region over midrib. Stomates basically paratetracyclic; longitudinally oriented, in widely spaced (separated by 7-13 rows of epidermal cells) rows. Within each row stomates may be widely separated or joined by abutting polar subsidiaries into short, tight chains 2-3 stomates long. Encircling cells irregular. Epidermal cells moderately buttressed, glabrous. 392 Journal of The Royal Society of New Zealand, Volume 28, 1998 Fig. 11 (A,B), Mt Somers podocarp sp. B: (A) TLM, stomatal rows, SB994, scale = 50 um; (B) TLM, stomatal rows, SB994, scale = 50 urn; (C,D), Mt Somers podocarp sp. C: (C) TLM, short stomatal row, SB995, scale = 50 urn; (D) TLM, short stomatal row, SB995, scale = 50 um; (E, F), Mt Somers podocarp sp. D: (E) TLM, stomatal zone, SB 1119, scale = 10 um; (F) TLM, detail of stomates, SB 1119, scale = 50 um. Family placement follows Podocarpaceae placement rule #3 (Appendix 2). Mt Somers podocarp sp. D (Fig. HE, F) Reference specimen: SB 1119. Description: Shoot form unknown. Stomatal distribution unknown. Stomates paratetracyclic; longitudinally aligned; in short, loose rows (but polar subsidiaries sometimes abutting), generally well spaced. Outline of stomates nearly smoothly ovoid, polar subsidiaries either quite small or tightly curved on their outer margin, not projecting from overall outline. Epidermals long, thin, straight walls, glabrous. Pole—Paleocene gymnosperms from Mount Somers 393 Fig. 12 Mt Somers podocarp sp. E, SB 1124: (A) TLM, stomatal zone, scale = 50 urn; (B) TLM, normal epidermal cells, scale = 50 |im; (C) TLM, single stomate, scale = 20 um; (D) TLM, single stomate, scale = 20 um. Family placement follows Podocarpaceae placement rule #3 (Appendix 2). Mt Somers podocarp sp. E (Fig. 12) Reference Specimen: SB 1124 (dispersed cuticle only). Description: Shoot form unknown. Leaf probably flattened; stomatal distribution hypostomatic; number of zones unknown. Stomates basically paratetracyclic; longitudinally oriented; occurring in short, tight chains, 2-5 stomates long, where polar subsidiary cells share or abut adjacent stomate in row. Lateral subsidiary cells strongly semi-circular, polar subsidiary cells projecting. Epidermal cells long, rectangular, buttressed, glabrous. Family placement follows Podocarpaceae placement rule #3 (Appendix 2). Taxaceae or Taxodiaceae Paahake gen. nov. Shoot arrangement unknown (Fig. 13), probably spreading. Leaves bifacially flattened, single veined, lanceolate (interpreted from fragments), apex acute, margin smooth. Stomates distributed in two zones on abaxial leaf surface (hypostomatic), 7-11 overlapping rows per zone, rows densely packed, not separated by normal epidermal cell files, all (subsidiary and neighbouring) cells in zone papillate. Stomatal orientation longitudinal (sometimes oblique). Stomatal complex monocyclic, of distinct type with 7-8 subsidiary cells common (often two polar subsidiary cells at each pole, and two lateral subsidiary cells on either side), circular. Guard cells deeply sunken, polar extensions pronounced, lateral extensions pronounced. 394 Journal of The Royal Society of New Zealand, Volume 28, 1998 Subsidiary cells inserted at same level as encircling cells, once, sometimes twice papillate. Papillae partially fused to form very irregular ring around (not overarching) stomatal aperture. Papillae on other cells in stomatal zone round, essentially unconnected with other papillae. Abaxial and adaxial epidermal cells typically rectangular with wavy-sinuous, unbuttressed (or slightly buttressed) anticlinal walls, outer epidermal surface subdued, glabrous. The closest comparison of the Mt Somers fossils appears to be with two extant genera in two other families; Sciadopitys (Taxodiaceae s.l.) and Torreya (Taxaceae). Sciadopitys verticillata (Thunb.) S. et Z. (Fig. 14A-D) has bifacially-flattened, lanceolate leaves, but has only a single stomatal band. Within this band all the cells have distinctly thin, elongate papillae (see Florin 1931, fig. 26 for sectional views) which often overarch and interdigitate above the stomatal pore. The stomates are monocyclic, typically with 8-12 subsidiary cells. Torreya spp. (Fig. 14E-H) have very delicate cuticle with a stomatal form very similar to Sciadopitys (8-12 subsidiary cells) but with stomates in two zones. Papillae in Torreya are quite different from Sciadopitys, and the fossils. They sit on (and are interconnected by) a series of longitudinally oriented ridges where clusters of papillae radiate out in an almost dendritic fashion (see Florin 1931, figs 4,5 for sectional views). The non-stomatal epidermal cells of the fossils (rectangular and wavy-sinuous) differ markedly from Sciadopitys (elongate and fusiform) and Torreya (rectangular and straight-walled). Taxus and Austrotaxus (other genera in the Taxaceae with papillae) differ by having distinctly fused papillate rims around the stomatal pore and they have amphicyclic stomates. Placement in the Cupressaceae (characterised by papillate epidermal cells) is rejected on the basis of leaf shape, stomatal form, and form of the papillae around the stomatal pore. Flattened, lanceolate leaves do not occur in adult, extant Cupressaceae (but do occur in some seedling foliage; pers. obs.). The stomates, although typically also monocyclic, have subsidiary cell numbers of 4-6 (up to 2 lateral subsidiary cells on each side). Numbers less than 6 are absent in the fossils, while the 7-8 found in the fossils are absent or exceedingly rare in Cupressaceae. In the Cupressaceae papillae (when they are present at all; see Oladele 1983) on the subsidiary cells are always elongated around the stomatal pore and form a regularly shaped ring, while in the fossils the papillae are not distinctly elongate and the shape of the ring is very irregular. The combination of flattened, lanceolate leaf form having two stomatal zones where subsidiary and encircling cells all have broad, round simple and essentially unconnected papillae is unique and it is regarded as indicating an extinct conifer genus. Present data do not warrant placing the Mt Somers specimens in either the Taxodiaceae s.l., or the Taxaceae. They are described as a new genus, Paahake. Family Indet. (cf. Taxodiaceae and Taxaceae) Paahake gen. nov. (Fig. 13) Etymology: a maori word for 'ancient times' (Biggs 1981). Diagnosis: a conifer with flattened, lanceolate leaves, two stomatal zones, within which all surficial cells have simple, low, rounded, and mostly unconnected papillae. Type species: P. papillatus sp. nov. Paahake papillatus gen. et sp. nov. Fig. 13 Etymology: after the papillate leaf surface. Diagnosis: As for the genus. Reference specimen: SB 1122 Referred specimens: SB998-1001, SB 1122, S781, 883, 784 Pole—Paleocene gymnosperms from Mount Somers 395 Fig. 13 Paahake papillatus : (A) SEM of mid-apical portion of leaf, adaxial surface uppermost, note two stomatal zones, S781, scale = 1 mm; (B) TLM of stomatal zone, SB 1122, scale = 10 um; (C) TLM detail of stomatal zone, SB1122, scale = 50 um; (D) SEM, outer stomatal surface, S927, scale = 10 um; (E) SEM, inner stomatal surface, S783, scale = 50 um; (F) SEM, detail of inner stomate, S783, scale = 20 um; (G) SEM, detail of inner stomate, S784, scale = 20 um. 396 Journal of The Royal Society of New Zealand, Volume 28, 1998 H% 1 "^ -- \ Fig. 14 A-D, Sciadopitys verticillata (herbarium material of extant species): (A) SEM, outer stomatal zone, OPH2204, scale = 100 um; (B) SEM, inner stomatal zone, OPH2203, scale = 50 urn; (C) TLM, stomatal zone, normal epidermal cells above, OPH2203, scale = 10 u.m; (D) TLM, detail of single stomate, OPH2204, scale = 50 u,m; E-F Torreya taxifolia (herbarium material of extant species); (E) SEM of outer cuticle surface, note papillae on longitudinal ridges, S878, scale = 10 |0.m; (F) SEM detail of inner cuticle of stomate, subsidiary cells not clear, possibly obscured by mesophyll material, S878, scale = 10 (im; (G) TLM of stomatal zone, the numerous subsidiary cells are clear, AQ141823, scale = 200 |im; (H) TLM detail of stomates, AG141823, scale = 50 urn. Pole—Paleocene gymnosperms from Mount Somers 397 A Tl f f ., < i , .1;' /* J ZM Fig. 15 Hoiki mcqueenii gen. et sp. nov.: (A) TLM of mid portion - apex of leaf, SB 1146, scale = 1 mm; (B) TLM of stomatal zone, SB 1118, scale = 100 urn; (C) TLM of single stomate, SB 1118, scale = 50 um; (D) SEM, inner stomate, S642, scale = 20 um; (E) SEM, inner stomate, S642, scale = 50 um; (F) SEM, outer surface of stomate, S642, scale = 20 um. Incertae Sedis Hoiki gen. nov. Shoot unknown. Leaves awl-like (Fig. 15), flattened (?bifacially), apex with up-turned tip (see SB 1146), hypostomatic, stomates in two zones, oriented transversely to long-axis of leaf, well-separated, not in rows, two distinct lateral subsidiary cells, no polar subsidiary cells (paracytic). Guard cells surficial (or very nearly so), no raised rim around stomatal pore, stomatal pore slit-like, no clear polar or lateral extensions, but slight polar thickenings. Epidermal cells isodiametric, straight-walled, smooth, outer surface subdued, glabrous. Transversely oriented stomates occur throughout a large number of plant taxa, including 398 Journal of The Royal Society of New Zealand, Volume 28, 1998 mosses, gymnosperms and angiosperms (Butterfass 1987). There is little about these enigmatic fossils which could clearly place them into even larger groups like these. They are included here partly for completeness although the shape of the fossil leaves (particularly of SB 1146), the relative thickness of the cuticle (monocots and Casuarinaceae rarely survive the dispersedcuticle process), suggests they may be gymnosperms. However, within this group, the paracytic structure of the stomates, the non-sunken guard cells, and the transverse orientation of the stomates differs from all extant taxa including the conifers and cycads. The extinct Bennettitales had transversely oriented stomates but they differed markedly in other cuticular features (see Sincock & Watson 1988 for a review of Bennettitalean cuticle), from the Mt Somers fossils. They typically had very sinuous and buttressed epidermal cell walls and prominent lateral extensions of the guard cell cuticle. Harris (1935) illustrated similar transversely oriented paracytic stomates from the broad, multiveined Podozamites punctatus Harris (Jurassic) and Doludenko (1967) illustrated transverse stomates in the similar leaves of Podozamites aff. eichwaldii Schimper (Late Jurassic-Early Cretaceous). Both are gymnosperms of uncertain affinities. The affinities of these leaves remain open. Their distinctness warrants a new genus even though higher affinities are unknown. Gymnosperm Incertae Sedis. Hoiki gen. nov. Diagnosis: Leaves flattened, probably bifacially, hypostomatic. Stomates in two zones, either side of single midvein, well-spaced. Stomatal complexes transversely oriented, syndetochelic (or paracytic). Epidermal cells with straight walls. Glabrous. Etymology: Hoiki is Maori for "tapering upwards", referring to the narrowing and turned-up apex of the leaf. Type species: Hoiki mcqueenii. Hoiki mcqueenii gen. et sp. nov. (Fig. 15) Holotype: SB1146 (fragment of leaf). Diagnosis: as for the genus. Etymology: Named for D.R. McQueen, for his work on New Zealand paleobotany. Referred specimens: SB1118, 1147, S642 DISCUSSION These conifers grew in a general 'coal swamp' environment, but their preservation in mud suggests they are most likely derived from a clastic swamp environment adjacent to the peat. In terms of biomass the conifers form roughly the same proportion of the assemblage as angiosperms (pers. obs.). Clearly conifers, and particularly the Podocarpaceae, were important biomass components of the vegetation which the macrofossil assemblage sampled. Palynological research by Raine & Wilson (1988) found conifer pollen forming 45-70 % of the total count (the most numerous conifer grain was Phyllocladidites mawsonii) at the Mount Somers mine, both in the coal and surrounding mud. This suggests that conifers were important in both clastic and peat vegetation. The most striking difference of the Mt Somers assemblage from Cretaceous assemblages in New Zealand is the almost complete absence of Araucariaceae macrofossils at Mt Somers. Araucariaceae are present in, and often dominate, all Cretaceous assemblages known (Pole 1995) contrasting markedly with the single fragment of cuticle found at Mt Somers, forming far less than 1% of the total cuticle biomass. As the Cretaceous assemblages are also from general coal-forming environments, this difference probably does not reflect substrate, but may indicate a climate change. Raine & Wilson Pole—Paleocene gymnosperms from Mount Somers 399 (1988) recorded the araucarian pollen Dilwynites granulatus Harris from sediments overlying the coal at Mt Somers, but not Araucariacidites australis Cookson. The angiosperms which accompanied the conifers included Proteaceae (Pole in press) and Lauraceae (Pole unpublished data). The overall assemblage may be contrasted with two Paleocene assemblages from Kakahu, also in Canterbury, which are angiosperm-dominated (Pole 1993, 1997). With the possible exception of the Libocedrus, all of the species documented here are now extinct. Even the informally described fragments cannot be matched with extant New Zealand conifer cuticle. Most of them probably belong to extinct genera. Araucaria sp. continues the history of the genus in New Zealand since the Cretaceous, while the Libocedrus and Prumnopitys now have the oldest macrofossil records of any extant New Zealand conifer genera. Modern affinities can be suggested for the new genera and the fossils may therefore represent ancestral or sister-taxa. For instance, Mumu shows affinities with Acmopyle, Kakahuia with Prumnopitys and the paired-leaved podocarps, and Tiotio with Dacrycarpus. The Mount Somers assemblage provides a valuable insight into the evolution of coal swamp vegetation in New Zealand as well as contributing to the knowledge of diversity of the Podocarpaceae. ACKNOWLEDGMENTS This material was collected on a field trip to New Zealand financed by A. Drinnan, University of Melbourne via an ARC grant. The research was started in the Department of Plant Science, University of Tasmania with funding from an ARC grant to R.S. Hill, and was completed in the Department of Botany, University of Queensland with funding from an ARC grant to G. Stewart and M.E. Dettmann. Many thanks to J. Douglas (Melbourne), R. Spencer (Royal Botanic Gardens Melbourne), and C. Gee (Institut fur Palaeontologie, Bonn) for their help in supplying herbarium material of Sciadopitys; Queensland Herbarium for supplying herbarium material of Prumnopitys, Sundacarpus and Torreya, and Mike Elliot (Massey University) for supplying Dacrycarpus dacrydioides. Assistance with electron microscopy from staff in the Center for Microscopy and Microanalysis, University of Queensland, and from John Bertram, Dept of Botany with computer graphics, was greatly appreciated. The management of Mt Somers Coal Mine are thanked for allowing access. I thank the two referees, J. Lovis and R.S Hill, whose comments improved the clarity of the manuscript. REFERENCES Biggs, B. 1981. The Complete English-Maori Dictionary. Auckland University Press, Auckland. 227 pp. Bigwood, A. J.; Hill, RS. 1985. Tertiary araucarian macrofossils from Tasmania. Australian journal of botany 33: 645-656. Butterfass, T. 1987: The transverse orientation of stomata. The botanical review 53: 415—441. Carter, R. M. 1988: Post-breakup stratigraphy of the Kaikoura Synthem (Cretaceous-Cenozoic), continental margin, southeastern New Zealand. New Zealandjournal ofgeology and geophysics 31: 405^29. Doludenko, M. P. 1967: Epidermal structure of Podozamites leaves. International Geology review 9: 214-217. Edwards, A. R.; Homibrook, N. de B.; Raine, J. I.; Scott, G. H.; Strong, C. P.; Wilson, G. J. 1988: A New Zealand Cretaceous-Cenozoic geological time scale. New Zealand Geological Survey record 31: 135-149. Field, B. D.; Browne, G. H. 1986: Lithostratigraphy of Cretaceous and Tertiary rocks, southern Canterbury, New Zealand New Zealand Geological Survey record 14. 55 pp. Florin, R. 1931: Untersuchungen zur Stammesgeschichte der Coniferales und Cordaitales. Kunglia Svenska Vetenskapakademiens handlinigan V. 10: 1-588. Harris, T.M. 1935. The fossil flora of Scoresby Sound, East Greenland, 4. Meddelelser om Gronland 112: 1-176. 400 Journal of The Royal Society of New Zealand, Volume 28, 1998 Hill, R. S.; Carpenter, R. J. 1991: Evolution ofAcmopyle and Dacrycarpus (Podocarpaceae) foliage as inferred from macrofossils in south-eastern Australia. Australian systematic botany 4: 449-479. Lingen, G. J. vander 1988: Lower Tertiary transgressive sediments of the Broken River Formation, Mt Somers area, Canterbury, New Zealand. New Zealand journal of geology and geophysics 31: 287304. Mildenhall, D. C. 1980: New Zealand late Cretaceous and Cainozoic plant biogeography: a contribution. Palaeogeography, palaeoclimatology, palaeoecology 31: 197-233. Oladele, F. A. 1983: Scanning electron microscope study of stomatal-complex configuration in Cupressaceae. Canadian journal of botany 61: 1232-1240. Page, C. N. 1988. New and maintained genera in the conifer families Podocarpaceae and Pinaceae. Notes Royal Botanic Garden Edinburgh 45: 377-395. Pole, M. S. 1992a: Eocene vegetation from Hasties, north-eastern Tasmania. Australian systematic botany 5: 431—475. Pole, M. S. 1992b: Early Miocene flora of the Manuherikia Group, New Zealand. 2. Conifers. Journal of the Royal Society of New Zealand 22: 287-302. Pole, M. S. 1993: Keeping in touch: vegetation prehistory on both sides of the Tasman. Australian systematic botany 6: 387-397. Pole, M. S. 1995: Late Cretaceous macrofloras of eastern Otago, New Zealand: Gymnosperms. Australian systematic botany 8: 1067-1106. Pole, M.S. 1997. Paleocene plant macrofossils from Kakahu, South Canterbury, New Zealand. Journal of the Royal Society of New Zealand 27: 371^00. Pole, M.S. 1998. The Proteaceae record in New Zealand. A us tralian systematic botany 11(3) (in press). Raine, J. I. 1984: Outline of a palynological zonation of Cretaceous to Paleogene terrestrial sediments in the West Coast region, South Island, New Zealand. New Zealand Geological Survey report 109. 82 pp. Raine, J. I.; Wilson, G. J. 1988: Palynology of the Mt Somers (South Island, New Zealand) early Cenozoic sequence (Note). New Zealand journal of geology and geophysics 31: 385-390. Sincock, C. A.; Watson, J. 1988: Terminology used in the description of bennettitalean cuticle characters. Botanical journal of the Linnean Society 97: 179-187. Stockey, R. A.; Frevel, B. J. 1997: Cuticlemicromorphology of Prumnopitys Philippi (Podocarpaceae). International journal of plant science 158: 198-221. Stockey, R. A.; Ko, H. 1988: Cuticle micromorphology of some New Caledonian podocarps. Botanical gazette 149: 240-252. Tomlinson, P. B. 1974: Development of the stomatal complex as a taxonomic character in the Monocotyledons. Taxon23: 109-128. Tregear, E. 1891: The Maori-Polynesian Comparative Dictionary. Lyon and Blair, Wellington. 675 pp. Wells, P. M.; Hill, R. S. 1989a: Leaf morphology of the imbricate-leaved Podocarpaceae. Australian systematic botany 2: 369-386. Wells, P. M.; Hill, R. S. 1989b: Fossil imbricate-leaved Podocarpaceae from Tertiary sediments in Tasmania. Australian systematic botany 2: 387—423. Wilson, G. J. 1984: New Zealand Late Jurassic to Eocene dinoflagellate biostratigraphy - a summary. Newsletters on stratigraphy 13: 104—117. Received 19 September 1997; accepted 12 January 1998 Fig. 16 New, or modified terminology for conifer stomatal distribution. See Appendix 1 for details. Subsidiary cells are shaded and surround a clear rectangle representing the guard cells: (A) Stomates in discrete rows. Two rows of dicyclic stomates are shown separated by a row of epidermal cells; (B) Stomates in tight chains. In the top row dicyclic stomates are linked by abutting polar subsidiary cells, in the middle row dicyclic stomates are linked by shared polar subsidiary cells, in the bottom row monocyclic stomates are linked by shared polar subsidiary cells; (C) Stomates in loose chains. Stomates are linked by a mixture of abutting and shared polar encircling cells; (D) Stomates in overlapping rows. The two upper rows are dicyclic stomates, the three lower rows are of monocyclic stomates; (E) Stomates in rows and networked. Stomates are all monocyclic; (F) Stomates random and networked. The white polygons are normal epidermal cells; (G) Stomates random. Stomates are dicyclic. Pole—Paleocene gymnosperms from Mount Somers A. Stomates in discrete rows B. Stomates in tight chains ?==? [ t C S t o m a t e s in \ JBSSJJSSJA I&&,ij^\i i/g^i loose chains l D. Stomates in overlapping rows E. Stomates in rows and networked F. Stomates random and networked G. Stomates random 401 r 1 l '^ ^ ' ' ^ gF| r v ' ° '• ° i/^ 402 Journal of The Royal Society of New Zealand, Volume 28, 1998 APPENDIX 1. NEW OR MODIFIED TERMS FOR STOMATAL DISTRIBUTION IN CONIFER LEAVES Florin (1931, fig 41) observed that many different stomatal distributions of gymnosperms can be derived from three rows of protodermal cells The middle row differentiates to form guard cells and polar subsidiary cells The outer cells may differentiate directly to form monocychc stomates, divide once (parallel to the guard cells) to form dicychc stomates, or twice to form tncyclic stomates Tomhnson (1974) developed a similar idea of three initial rows for monocots It is a matter of observation that the derivatives of the three rows remain distinct in many taxa, while in others, there is typically a degree of overlap These patterns seem to have developmental and phylogenetic significance The terminology presented below and illustrated in Fig 16 recognises these degrees of overlap as well as the apparent lack of any initial rows Additionally I have introduced terms for how close stomates are within a row A) 'stomates in discrete rows.' The guard cells and polar cells (and only those cells) are derived from the same row of initial cells over an extended distance although they may be widely separated along the row The lateral subsidiary cells and encircling cells are derived from adjacent rows of initial cells over an extended distance and are not shared with an adjacent stomatal row This is probably the most widespread condition in conifers Among others it is found in the Araucanaceae and Podocarpaceae B) 'stomates in rows and tight chains.' This is a special case of stomates in rows where stomates are also 'linked' end to end, either by abutting or sharing polar subsidiary cells for extended distances In the Podocarpaceae this is common in Dacrycaipus, Podocarpus, and Saxegothea In the figure the top chain is formed by dicychc stomates abutting polar subsidiary cells, the middle chain is formed by dicyclic stomates sharing polar subsidiary cells, and the lower chain is formed of monocychc stomates sharing polar subsidiary cells C) 'stomates in rows and loose chains.' This is a special case of stomates in rows where stomates are also 'linked' end to end, either by abutting or sharing polar encircling cells for extended distances This is typical of Prumnopitys and Retrophyllum (Podocarpaceae) D) 'stomates in overlapping rows.' This is where guard cells and polar cells (and only those cells) are derived from the same row of initial cells over an extended distance but the adjacent rows of initials are shared between two stomatal rows The one row of initials differentiates lateral subsidiary cells to rows on both sides In the monocychc condition this is found in the Pinaceae and the Cupressaceae In the dicychc condition it is found in the Cephalotaxaceae and in Podocarpaceae where it is typical of Prumnopitys, and occurs sometimes in Saxegothea E) 'stomates in rows and networked.' This involves monocychc stomates where lateral subsidiary cells are shared between adjacent rows and are simultaneously the polar subsidiary cells of other rows It is characteristic of the Cupressaceae and Taxodiaceae F) 'stomates random and networked.' This involves monocychc stomates where subsidiary cells are shared but there is no pattern of orientation or of clear polar and lateral subsidiary cells It is found in Taxodiaceae (e g Athrotaxis) G) 'stomates random.' There is no obvious pattern of derivation from three rows of initials Generally there is no obvious distinction of polar and lateral subsidiary cells It is found in Taxodiaceae (e g Sequoia) and in some small Podocarpaceae leaves (e g the imbricate leaves of Monoao and Halocarpus) APPENDIX 2. TAXONOMIC 'PLACEMENT RULES' USED IN THIS PAPER Araucanaceae leaf placement rule #1 2 (modified from Pole 1995) If leaves are flattened, and have multiple-veins, and stomates occur in rows, and stomatal pore is not surrounded by a Florin ring, then family is Araucanaceae Cupressaceae placement rule #1 If leaf arrangement is opposite-decussate, and stomatal pore is surrounded by a ring of fused papillae, and stomates are monocychc, then family is Cupressaceae Podocarpaceae placement rule # 1 If stomates are predominantly paratetracychc, Pole—Paleocene gymnosperms from Mount Somers 403 and stomates are not surrounded by a raised ran, and stomatal orientation is predominantly parallel to the long axis of the leaf, then family is Podocarpaceae Podocarpaceae placement rule # 2 2 ( m o d i f i e d from P o l e 1995) If leaves are imbricate and scale-like, and subsidiary cells show no clear distinction of polar and lateral, and stomatal outline is circular, then family is Podocarpaceae Podocarpaceae placement rule # 3 If stomates are predominantly paratetracychc, and stomates are surrounded by a Florin ring, and stomatal orientation is predominantly parallel to the long axis of the leaf, then family is Podocarpaceae APPENDIX 3. KEY SEPARATING MT SOMERS GYMNOSPERM MACROFOSSILS, IDENTIFIED TO GENERIC LEVEL. 1 Leaves imbricate, scale like, opposite and decussate, stomates monocychc with a ring of fused papillae around stomatal pore Libocedrus cf L bidwrfhi 1 Leaves spreading, linear 2 2 Stomates transversely oriented Hoiki mcqueenn 3 2 Stomates longitudinally oriented 3 Papillae present 4 3 Papillae not present 5 4 Stomatal zone densely papillate, epidermal cells not buttressed Paahake papillatus 4 Stomatal zone with scattered, small papillae, adaxial epidermal cells buttressed Kakahuia drmnanu 6 5 Stomatal distribution amphistomatic 5 Stomatal distribution hypostomatic Mumu somerensis 6 Leaves linear, not flattened Tiotio imbncatus 6 Leaves flattened Prumnopitys hmaniae