McKay et al.

We are aware that some mineral surfaces of ALH84001 display very narrow elongated forms, often subparallel, in the nanometre size range (Fig. 2a, b here and Fig. 1a, b and d of Bradleyet al. ). They are very common on pyroxene surfaces, but also occur on carbonate surfaces. We have also imaged some of these features, which Bradley et al. propose are simply emergent edges of pyroxene lamellae, with a field emission scanning electron microscope (FE-SEM). We have previously reported the presence of small amounts of clay minerals in ALH84001 from TEM images and electron diffraction data11. We propose that the features in Fig. 2a, b here and Fig. 1b of Bradley et al. may be weathering features caused by incipient formation of clay minerals. In places, these figures have a ‘platy’ appearance characteristic of some clay minerals especially in the uncoated ALH84001 sample (Fig. 2a ). Regardless of their exact origin, it is unlikely that these features are biogenic.

Figure 2: FE-SEM images of ALH84001 samples.
figure 2

a, Fine-scale platy texture on a carbonate surface in ALH84001. The sample was imaged at 1 kV without any conductive coating. b, The same surface after coating with a thin (∼10nm) carbon layer. c, TEM image of a shadowed replica of part of a carbonate surface from ALH84001. The replica was made from an uncoated surface so textures in the image represent true textures of the surface. These elongate features are much larger than either the platy structures in (a) or the ‘lamellae’ and whisker magnetite described by Bradley et al.6. d, Texture on the surface of an uncoated carbonate in ALH84001 imaged at 1 kV. Note the fine-grained wispy texture which we interpret as an extremely thin clay mineral alteration of the carbonate. e, The same sample now coated with ∼10 nm of carbon and imaged at 10 kV. Note that all of the large strand features are visible and relatively unchanged although space between them has been partly filled. The fine wispy surface now looks smooth.

Full size image
Figure 1: ‘Elongated forms’ in the meteorite ALH84001.
figure 1

a, Secondary electron image of pyroxene surface. Parallel elongated forms (from upper left to lower right) are emergent lamellae related to substrate cleavage direction. Inset, pyroxene surface oriented such that elongated forms resemble nanofossils. Both surfaces have a 10-nm-thick coating of gold, measured by TEM imaging of ultramicrotomed cross-sections. b, Secondary electron images of highly fractured pyroxene surface. Inset, high-magnification image of lamellae. Coating: 20 nm Au, resulting in more pronounced segmentation. c, Secondary electron image of carbonate crystals (coated with 9 nm of Au/Pd). The average grain size is 9. d, Secondary electron image of carbonate surface within fracture zone. Inset, high-magnification image of elongated forms on carbonate, some of which may be magnetite whiskers6. Coating: 10 nm Au.

Full size image

By contrast, the features that we imaged previously (Fig 6b in ref. 1) are not subparallel. They display intersecting alignment, have significant curvature and are more isolated. In addition, features tentatively interpreted as martian nanofossils are larger than the lamellae or elongated magnetite reported by Bradley et al.6. Curved and S-shaped features from the surfaces of carbonate globules are up to 0.75 μm in length (Fig. 2c ), an order of magnitude larger than the ‘lamellae’ or the elongated magnetite observed by Bradley et al.6. Additionally, some of the features we proposed as possible martian microfossils are ovoid rather than elongated (Fig. 6a in ref. 1).

It is well known that conductive coatings can alter appearance, cover fine details and produce artefacts. We ran a series of controls for coating artefacts (ref. 50 in ref. 1). Carbon coating of ALH84001 samples obscures fine honeycomb-like surface texture and many of the larger features become rounded or softened (Fig. 2d, e ). However, the dimensions of the larger features are not appreciably affected (Fig. 2d, e ).

Previous studies on latex spheres15 show that coatings of Au produce larger artefact grains than Au/Pd; grain size of Au is 2-3 times larger than Au/Pd16. Bradley et al. used Au in producing their high-magnification images, so the purported artefact segmentation would naturally be more prominent than if Au/Pd were used. We used Au/Pd, carbon, or completely uncoated samples in our studies, never gold alone, and have also made control coatings on lunar glass surfaces using 30 and 60 seconds (∼10 and 20 nm thick, respectively) of Au/Pd from our coater. The glass is not exactly analogous to the minerals in ALH84001 (except perhaps for the feld-spathic glass), and artefacts are substrate-dependent, but its smooth surface is ideal for detecting coating artefacts. At 30 s there are no artefacts and at 60 s only a fine uniform cracking or crazing texture is apparent. For our original ALH84001 studies, we generally used 30 s or less (and never more than 60 s) for sample coating. Although some of our images may have a fine crazing resulting from the coating, we have not observed major decoration in any of our control samples. ALH84001 substrates may accentuate and amplify coating artefacts more than our controls, but we have no evidence for such anomalous behaviour.

Bradley et al. suggest that the purported martian nanofossils are too small for fossilized bacteria. In fact, the lower size limit of fossilized bacteria has not been determined. In mammalian blood, living bacteria as small as 70 nm in diameter have been found which contain identifiable DNA and can reproduce17,18. Also, soil bacteria as small as 80 nm have been found19.

Bradley et al. suggested that the purported elongated microfossils may be high-temperature, vapour-formed magnetite whiskers, but other studies have shown that the carbonate globules and their included magnetites could not have formed at temperatures above 100-300 °C20,21,22 whereas elongate magnetite grains can be produced by bacteria at low temperatures23,24. They also assert that TEM imaging is more appropriate for identifying nanofossils, as it is possible to distinguish internal microstructures. SEM and TEM imaging are complementary and the use of both methods on the same samples would provide the best evidence for biogenic activity, particularly when internal structure is destroyed or replaced by mineralization25,26.