The worm turned, and the ocean followed.

U nless you hang drywall or travel in certain biogeochemical circles, the calcium sulfate mineral gypsum is probably not on your radar—nor is the amount of sulfate (SO4 2 ) dissolved in seawater, which determines how much gypsum is left behind when that seawater evaporates. All the same, sulfate is the second most abundant negatively charged ion dissolved in seawater today, and tracking its changing concentration in the ocean over the earth’s 4.5-billion-year history is one of a geochemist’s best windows to the chemical and biological evolution of the early ocean and atmosphere. Now, in a novel slant on the sulfate tracer, Canfield and Farquhar in this issue of PNAS (1) link dramatic increases in seawater sulfate in the early Paleozoic Era, 540–460 million years ago, to a major biological innovation: the invasion of marine sediments by burrowing, mud-churning organisms (Fig. 1). And the net result, the authors argue, was the world’s first massive deposits of gypsum. The premise of Canfield and Farquhar’s isotope-driven numerical model for sulfate and gypsum abundance through time hangs on the simple observation that gypsum first became plentiful at roughly the same time as the first appearance of burrowing organisms. Gypsum forms in arid, restricted settings where evaporation drives up the concentrations of dissolved sulfate and calcium to the levels required for its precipitation, and the higher the initial concentration of sulfate, the more likely this process will occur. It is certain that sulfate delivery to the ocean increased over time, and consequently its concentration, but this overall rise was dampened by the formation in the ocean of another mineral, pyrite. At the other end of the sulfur cycle, the origin of most seawater sulfate is tied to the chemical breakdown of pyrite exposed on the continents under an atmosphere containing at least a modest amount of oxygen, followed by transport to the ocean by rivers. This delivery began in earnest about 2.4 billion years ago as oxygen first accumulated beyond the trace amounts present in the earliest atmosphere. Evidence for this ‘‘Great Oxidation Event’’ (GOE) and the corresponding increase in sulfate delivery to the ocean comes in many forms, including distributions of the isotopes of sulfur preserved in pyrite found in ancient sedimentary rocks. Pyrite formation begins when bacteria convert sulfate to hydrogen sulfide (H2S) as they metabolize organic matter. In doing so, the microbes typically show a strong preference for the light isotope (32S) relative to heavier and less abundant 34S. This discrimination, or fractionation, becomes muted during severe sulfate shortages—that is, when its concentration is less than about 1% of that in the modern ocean. The lack of isotopic fractionation in pyrite that formed more than 2.4 billion years ago reinforces the assumption that sulfate concentrations below this low threshold prevailed in the absence of abundant oxidative weathering on the continents (2). Not only does pyrite formation provide an estimate of the sulfate concentration of early seawater, but that process and gypsum precipitation are the two principal pathways by which sulfate is removed from the ocean. When the H2S reacts with iron to form pyrite, sulfate is lost from seawater, as long as the pyrite is buried and not reoxidized back to sulfate. Pyrite forms under oxygen-free local conditions, and for a variety of reasons its formation, preservation, and burial are favored in oxygen-poor settings such as the modern Black Sea, where oxygen is absent in all but the shallowest parts of the water column. There is reason to believe that oxygen deficiencies may have persisted in the deep ocean for 1–2 billion years after the GOE (3, 4). There is also ample evidence for comparatively low sulfate concentrations in the ocean, perhaps only about 10% of today’s, over much of this interval (5)—just as we would expect from a high rate of pyrite burial. Among the principal evidence for low seawater sulfate is the sparse record of gypsum deposition during this interval. Gypsum is a soluble salt, and so its preservation at the earth’s surface over long periods is favored only under the driest conditions. But the paucity of early gypsum is certain to reflect more than preservational bias. Instead, we can imagine how difficult it was in a low-sulfate ocean to reach the elevated concentrations of sulfate needed to permit gypsum precipitation during evaporation, as Canfield and Farquhar argue (1). Gypsum did precipitate in the ocean during the Proterozoic spanning from about 2.5 billion to 540 million years ago (6, 7), but Canfield and Farquhar assert that abundant gypsum was rare before the early parts of the Paleozoic. Their new model mitigates the impact of preservational skewing by predicting gypsum precipitation over time from isotope mass balance relationships. If we imagine more than a billion years of pyrite burial under the low-oxygen,