368 PRECAMBRIAN/Prokaryote Fossils fractal dimensions; continuous laminae of barely varying thickness (typical of crystal growth); lack of wrinkle fabrics typical of biofilms; and lack of fenestrae or other signs of biological processing Such simple stromatolites seem likely to have formed from largely physical processes such as direct chemical precipitation from seawater It is only after about 3000 Ma that fractally complex stromatolites are found in carbonate platform settings, for which the presence of biofilms seems more plausible, if not always demonstrable These tend to have wrinkle mat fabrics, discontinuous laminae, and laminar fenestrae (e.g those from the 2700 Ma old Belingwe Group of Zimbabwe and the Fortescue Group of Western Australia) Microfossils associated with Archaean stromatolites remain questionable, however, until about 2530 Ma and are not really diverse until the 1900 Ma old Gunflint Chert of Canada (Figure 3D) Even here, the biogenicity of the associated stromatolites is open to question because they resemble abiogenic sinters Silicified Microbiotas Precambrian oceans appear to have been supersaturated with silica (SiO2) because it was not being removed from the water column by groups that evolved later, such as the diatoms, radiolaria, and sponges In environments where early diagenetic silica was able to engulf prokaryotic populations, such as in peritidal hypersaline bacterial-mat settings, prokaryotic sheaths and even cell walls were sometimes (but still very rarely) well preserved in three dimensions This prevalence of hypersaline settings may have produced an unfortunate bias towards prokaryotic assemblages in the early fossil record In other words, the inferred dominance of prokaryotic microfossils within Precambrian cherts may be due to their restricted hypersaline setting rather than to evolutionary factors Silicified microbiotas are usually studied by means of standard (30 mm) to thick (300 mm) petrographic thin sections, at magnifications of up to about Â400 This thin-section technique is paramount because of the way in which it provides for contextual analysis, including three-dimensional morphology, mineralogy, rock fabric, and rock history Other techniques, such as maceration (digestion in strong acids), etching of rock chips, scanning electron microscopy, and atomic force microscopy, are also used, but these not provide the requisite information on context and are prone to the inclusion of structures that are later contaminants or ‘artefacts’ of the preparation process A classic example of a silicified Precambrian microbiota is the 1900 Ma old Gunflint Chert, which preserves about 12 taxa of prokaryotes, including forms that superficially resemble coccoid and filamentous cyanobacteria (Figure 3D and insert) but may be more closely allied to extant iron bacteria The putative cyanobacteria Eoentophysalis (Figure 3B) and Archaeoellipsoides are thought to be present in the approximately 2100 Ma old cherts of West Africa and the 2000 Ma old Belcher Group cherts of Canada, respectively The latter is claimed to preserve the specialist heterocyst cells used by cyanobacteria to help fix nitrogen in an otherwise oxidizing atmosphere As one moves back into the Archaean, microfossils become both extremely rare and highly questionable, despite the great abundance of carbonaceous cherts and tufa-like carbonates This may be explained partly by the inference that Archaean cherts were laid down in largely hydrothermal conditions that were often acidic and reducing, and partly by the scarcity of large and resistant cellular materials at that time Bundles of silicified filaments and tiny calcified holes from stromatolites in the 2530 Ma old carbonates of the Transvaal Supergroup of South Africa may be the casts of coccoid and filamentous cyanobacteria but little of diagnostic significance is preserved Intriguingly, such encrusted cyanobacterial filaments are rarely seen before about 1000 Ma, and endolithic microborings are not reported prior to 1500 Ma A single microfossil-like structure from the 2700 Ma old Fortescue Group of Western Australia has been compared with a cyanobacterial filament but its biogenicity and context awaits full documentation Skeins of pyritic filaments found within carbonaceous cherts from 3200 Ma old black smokers of the Sulphur Springs Formation in Western Australia may be the remains of anaerobic hyperthermophile bacteria, though the indigenous and biogenic nature of these intriguing structures has yet to be demonstrated beyond question The oldest cherts containing a supposed diverse prokaryotic microflora (Figure 4A) come from the 3450 Ma old Apex Cherts, which are intimately associated with ultrabasic and basaltic lava flows of the Apex Basalt in the Warrawoona Group of Western Australia At least eleven different kinds of filamentous microfossil have been described from these rocks, some of which have been compared with cyanobacteria (Figure 4A) This has been taken to suggest that photosynthesis had begun to release oxygen into the atmosphere by 3450 Ma and that a substantial amount of evolution had taken place by this time A critical re-examination of the context and fabric of these cherts suggests, however, that all these