The hypothesis is presented that about 3.5 billion years ago oxygen was stored chiefly in carbonate rocks, silicate rocks, and water. Sediments cycled in a reducing atmosphere. Iron cycled as ferrous iron and precipitated chemically as carbonate or silicate under the same general controls as calcium and magnesium. As time went on, and photosynthesis by procaryotic organisms occurred, ferric oxides were formed in the shallow waters of the iron-depositing basins, and sulfates were formed bacterially during weathering, even though little if any free oxygen may have been produced. Organic material in an amount equivalent to the number of moles of oxygen stored in sulfates or iron oxides accumulated, and a similar number of moles of carbonate minerals was converted to other compounds. As sediments cycled, degraded organic matter was re-eroded and redeposited; ferric oxides joined the clastic fraction of stream loads, and sulfates tended to remain as sulfates. Perhaps with the advent, some 2 billion years ago, of eucaryotic photosynthetic organisms, which release free oxygen (Cloud, 1972), and the diminution of reduced mineral reservoirs consuming oxygen or its equivalent, atmospheric oxygen began to rise. Eventually a level was reached that achieved equality between the amount of organic material weathered and oxidized and the amount of new organic material deposited as the residue of photosynthesis minus respiration, decay, and oxidation. At this stage no further accumulation of organic materials was possible, and an essentially stable cycling system was established, with no further growth of oxygen-consuming reservoirs. This condition grossly characterizes Phanerozoic time. It is suggested that a major factor in the change from "Precambrian-type iron-formations" to those typical of the Phanerozoic may have been the simple consequence of a change in the transport of iron from dissolved ferrous iron to ferric iron carried in the suspended load of streams.