Banded iron formations could have formed in shallow-water marine conditions from iron transported to the oceans in detrital sediments and silica transported in solution. If atmospheric P o 2 were significantly lower than at present and biological productivity were comparable to present-day values, the entire oceans below the thermocline would be anaerobic ( f o 2 ∼10 −70 ), while the surface mixed zone would be relatively oxidizing. The supply of Fe 2+ to the oceans in sediments is much greater than the supply of dissolved sulfate. Under anaerobic conditions, all the sulfur would be precipitated as pyrite, and Fe 2+ would build up in solution until siderite saturation was achieved (∼10 ppm Fe 2+ ). When the anaerobic deep water welled up onto a shallow platform, the dissolved Fe 2+ would be oxidized to goethite, and slight evaporation would cause precipitation of amorphous silica or magadiite. Siderite would be formed where the supply of organic carbon was sufficient to maintain anaerobic conditions at the sediment-water interface. Silica-secreting organisms are assumed to have been absent, and an oceanic pH value of 7.7 or lower (perhaps caused by a slightly higher atmospheric P co 2 would be required to prevent sepiolite precipitation. Quantitative calculations show that hydrogen ion released by oxidation of Fe 2+ and precipitation of goethite would be sufficient to prevent precipitation of calcite when the water is evaporated sufficiently to deposit a weight of SiO 2 equal to the weight of Fe 2 O 3 precipitated by oxidation of Fe 2+ . The predicted physical environment of deposition is similar to that of present-day shallow-water carbonates. This is in good agreement with sedimentological studies. Apart from the direct consequences of lowered atmospheric Po 2 , slightly increased Pcoa and the absence of silica-secreting organisms, the chemistry and thermal structure of the ocean 2 b.y. ago could have been identical to those of the present-day ocean.