A number of investigators have used chemical profiles of paleosols to reconstruct the evolution of atmospheric oxygen levels during the course of Earth history (Holland, 1984, 1994; Kirkham and Roscoe, 1993; Ohmoto, 1996). Over the past decade Holland and his co-workers have examined reported paleosols from six localities that formed between 2.75 and 0.45 Ga. They have found that the chemical profiles of these paleosols are consistent with a dramatic change in atmospheric PO2 between 2.2 and 2.0 Ga from < or = 0.002 to > or = 0.03 atm (Holland, 1994). Ohmoto (1996) examined chemical data from twelve reported paleosols ranging in age from 2.9 to 1.8 Ga. He concluded that these chemical profiles indicate that atmospheric PO2 has not changed significantly during the past 3.0 Ga. We seek to resolve the conflict between these reconstructions through a broader examination of the paleosol literature, both to determine which reported paleosols can be definitively identified as such and to determine what these definite paleosols tell us about atmospheric evolution. We here review reports describing over 50 proposed paleosols, all but two are older than 1.7 Ga. Our review indicates that 15 of these reported paleosols can be definitively identified as ancient soils. The behavior of iron uring the formation of these 15 paleosols provides both qualitative and semiquantitative information about the evolution of the redox state of the atmosphere. Every definitely identified pre-2.44 Ga paleosol suffered significant Fe loss during weathering. This loss indicates that atmospheric PO2 was always less than about 5 x l0(-4) atm prior to 2.44 Ga. Analysis of the Hokkalampi paleosol (2.44-2.2 Ga) (Marmo, 1992) and the Ville Marie paleosol (2.38-2.215 Ga) (Rainbird, Nesbitt, and Donaldson, 1990) yield ambiguous results regarding atmospheric PO2. Loss of Fe during the weathering of the 2.245 to 2.203 Ga Hekpoort paleosol (Button, 1979) indicates that atmospheric PO2 was less than 8 x 10(-4) atm shortly before 2.2 Ga. The presence of red beds immediately overlying the Hokkalampi, Ville Marie, and Hekpoort paleosols suggests that by about 2.2 Ga there was an unquantified but substantial amount of oxygen in the atmosphere. Iron loss was negligible during formation of the 2.2 to 2.0 Ga Wolhaarkop (Holland and Beukes, 1990) and Drakenstein (Wiggering and Beukes, 1990) paleosols and during formation of all the later paleosols we previewed. Thus, atmospheric PO2 probably has been > or = 0.03 atm since sometime between 2.2 and 2.0 Ga.

中文翻译：

---

古土壤和大气氧的演化；批判性审查

许多研究人员使用古土壤的化学特征来重建地球历史进程中大气氧含量的演变（Holland，1984，1994；Kirkham 和 Roscoe，1993；Ohmoto，1996）。在过去的十年中，Holland 和他的同事们研究了来自六个地区的报道的形成于 2.75 至 0.45 Ga 之间的古土壤。他们发现这些古土壤的化学特征与大气中 PO2 在 2.2 至 2.0 Ga 之间的剧烈变化一致。 < 或 = 0.002 至 > 或 = 0.03 atm（荷兰，1994 年）。Ohmoto (1996) 检查了 12 个已报道的古土壤的化学数据，年龄范围从 2.9 到 1.8 Ga。他得出的结论是，这些化学特征表明，在过去的 3.0 Ga 中，大气中的 PO2 没有显着变化。我们试图通过对古土壤文献进行更广泛的检查来解决这些重建之间的冲突，以确定哪些报道的古土壤可以被明确地识别为这样，并确定这些明确的古土壤告诉我们关于大气演化的什么。我们在此审查了描述 50 多种拟议古土壤的报告，除两个外，所有古土壤的年代都超过 1.7 Ga。我们的审查表明，这些报告的古土壤中有 15 种可以明确确定为古代土壤。铁在这 15 种古土壤形成过程中的行为提供了关于大气氧化还原状态演变的定性和半定量信息。每个明确确定的 2.44 Ga 之前的古土壤在风化过程中都遭受了显着的 Fe 损失。这种损失表明大气中的 PO2 在 2.44 Ga 之前总是小于大约 5 x l0(-4) atm。 Hokkalampi 古土壤 (2.44-2.2 Ga) (Marmo, 1992) 和 Ville Marie 古土壤 (2.38-2.215 Ga) 的分析)（Rainbird、Nesbitt 和 Donaldson，1990 年）产生了关于大气 PO2 的模棱两可的结果。2.245 至 2.203 Ga Hekpoort 古土壤风化过程中 Fe 的损失（Button, 1979）表明大气中的 PO2 在 2.2 Ga 之前不久小于 8 x 10(-4) atm。红色层的存在直接覆盖在 Ville 的 Hokkalampi Marie 和 Hekpoort 古土壤表明，到大约 2.2 Ga，大气中存在未量化但数量可观的氧气。在 2.2 到 2.0 Ga Wolhaarkop（Holland 和 Beukes，1990）和 Drakenstein（Wiggering 和 Beukes，1990) 古土壤以及我们预览的所有后期古土壤的形成过程。因此，从 2.2 到 2.0 Ga 之间的某个时间，大气中的 PO2 可能已经 > 或 = 0.03 atm。