To illustrate the problem using the approach of Ohmoto et al., I use their lowest oxygen partial pressure (pO2). Solar ultraviolet and lightning would dynamically maintain pO2 at 10−13 atm (ref. 2). Ohmoto et al. correctly state that this is far above the critical pO2 for siderite stability, namely 10−65 ± 5 atm. The partial pressure at siderite stability has only mathematical significance; there is no actual molecule of oxygen in any realizable volume of water at the pO2. Rather, the dissolved oxygen at Archaean pO2 in water, which is about 10−16 mol kg−1, is too small to affect the chemistry of a soil observably, as illustrated by a simple mass balance. Basalt contains about 10% by mass of ferrous oxide, FeO, so it would take 2 × 1015 kg of oxygen-saturated water to oxidize the ferrous oxide in a kilogram of rock to magnetite, Fe3O4. With a typical rainfall of 1 m yr−1, it would require 1015 years of rain to oxidize a metre of section. This conclusion applies as long as oxygen is a trivial component. Even at a partial pressure of 10−6 atm, it would take 108 years to oxidize a metre of section.

Rather, the trivial oxygen dissolved in soil water at pO2 = 10−13 atm reacts with the rock, producing an undetectable amount of ferric iron. This leaves a solution that is quantitatively depleted in oxygen and buffered by ferrous silicates, where siderite is stable if there is enough pCO2. At a higher pO2 of 10−6 atm, trace oxygen in soil water would produce observable ferric iron before it was exhausted, again leaving a buffer with ferrous silicates. Moreover, if the oxidation reaction is kinetically inhibited at trivial concentrations, the presence of oxygen in the air is irrelevant to the presence of siderite.