liminary data for the distribution of the rare earth elements (REEs) in paleosols suggest that Eu in paleosols has been present in the +3 valence state since at least 2750 Ma, and that the valence state of Ce in paleosols changed from +3 to +4 between 2750 and 1800 Ma. These results are consistent with those for Fe and indicate that the REEs may join iron as useful indicators of oxygen evolution in the Precambrian atmosphere.
The data for paleosols developed on igneous rocks have been supplemented recently by information on a paleoweathering profile developed on carbonate facies Kuruman Iron Formation in Griqualand West, South Africa (Holland and Beukes, 1990). The profile was probably developed ca. 1900 Ma. It is highly oxidized, and the high degree of iron retention during weathering can be used to show that PO2 was probably in excess of 15% PAL (Figure 1.3). This is a higher minimum for PO2 than that set by the behavior of iron in paleosols developed in igneous rocks, and indicates that the O2 content of the atmosphere rose from about 1 to >15% PAL between 2200 and 1900 Ma (Figure 1.1).
The transition inferred from paleosol data is consistent with that inferred from the time distribution of iron formations and postdates the last known occurrence of detrital uraninite ores by several hundred million years (Knoll, 1979; Walker et al., 1983). Isotopically very light organic matter in late Archean and earliest Proterozoic sedimentary rocks has also been interpreted in terms of an early appearance of environments capable of sustaining aerobic metabolism, at least locally (Hayes, 1983). These data suggest that PO2 may have increased in at least two steps: an initial rise from extremely low oxygen tensions to levels about 1 to 2% PAL, and a later increase to levels >15% PAL approximately 2100 Ma (see also Walker et al., 1983).
Why oxygen levels should have increased in this manner is not clear. The origin of oxygenic cyanobacteria is poorly constrained in time, but it certainly occurred before 2100 Ma. Fossils morphologically diagnostic for the group are known only from about 2000 Ma (Golubic and Hofmann, 1976), but plausibly cyanobacterial remains have been found in early Archean cherts (e.g., Schopf and Packer, 1987). Buick (1992) has argued on sedimentological and geochemical grounds that stromatolites in lacustrine carbonates of the 2800 Ma Fortescue Group, Australia, must have been built by oxygenic photoautotrophs. If Hayes' interpretation of the carbon isotope record is correct, cyanobacteria radiated 2800 Ma or earlier. The pre-2800 Ma sedimentary record has been sampled too poorly to establish whether anomalously light carbon was widespread in early Archean lacustrine environments.
Increases in atmospheric oxygen were probably occasioned by increases in primary productivity and/or decreased rates of oxygen consumption. The increase from very low O2 levels to 1 to 2% PAL may have been related to productivity increases associated with rapid continental growth and stabilization during the late Archean/earliest Proterozoic (Knoll, 1979, 1984; Cameron, 1983). In contrast, the later increase to >15% PAL does not seem to be related to a major tectonic event. The high oxygen level in today's atmosphere must be related to the role of PO2 in maintaining the redox balance of the atmosphere-biosphere-ocean-lithosphere system. However, the nature of the connection is still in dispute. Atmospheric PO2 determines the concentration of O2 in surface ocean water, but the influence of the O2 concentration in seawater on the burial efficiency of organic matter within marine sediments seems to be slight (see, for instance, Betts and Holland, 1991). Nutrients are a more likely link between PO2 and the burial rate of organic matter, and hence between PO2 and rates of long-term O2 generation. A plausible argument can be made that links the marine geochemistry of PO43- to that of iron and hence to the O2 content of the atmosphere today. If this argument turns out to be valid, then the history of atmospheric O2 may have been controlled by a complicated feedback system involving the marine geochemistry of iron and phosphorus. The rapid increase in PO2 ca. 2100 Ma may have marked the passage of the system across a threshold from one steady state to another.
At first glance, the fossil record appears to provide strong support for the linkage of environmental and biological evolution. The oldest known fossils of probable eukaryotic origin are spirally coiled, megascopic remains