variations in new production and, hence, atmospheric carbon dioxide drawdown in Antarctic waters.

The rest of this story is well known. John Martin gained considerable media attention with the radical notion that iron addition in the Southern Ocean could be used to "engineer down" atmospheric carbon dioxide. This marked the first time that biological oceanography per se commanded prime-time media attention, and it was no surprise that Martin's proposed iron enrichment method to draw down atmospheric CO₂ met with considerable negative publicity and was unpopular with biological oceanographers, environmentalists, and federal agencies. Martin himself kept his radical notion, which he always mentioned with a playful grin, separate from his serious determination to test the Iron Hypothesis. His critics did not or would not recognize this distinction.

John H. Martin died in June 1993, but his iron hypothesis was tested successfully in an in situ transient iron enrichment experiment in September 1993 (Martin et al., 1994) and again in May 1995 (Coale et al., 1996). It has now become evident that iron is a limiting or regulating nutrient in many marine and freshwater habitats for many organisms, not just primary producers. At a recent American Society of Limnology and Oceanography (ASLO) meeting on aquatic sciences more than 50 papers referred to iron effects. As with the ubiquity of chemosynthetic ecosystems, the question is, How could we have missed the importance of iron for so long?

Martin's proposed research to test the iron hypothesis with an in situ transient iron addition in the equatorial Pacific Ocean was controversial from the start (Chisholm, 1995). There was significant opposition because of worries that confirmation of the hypothesis would lead immediately to reckless climate engineering. Furthermore, no one had ever modified and marked a patch of open-ocean water, and many oceanographers were dead certain that it couldn't be done. Two courageous program managers, Ed Green of ONR and Neil Anderson of NSF, devised a Byzantine funding arrangement to get Martin's experiment done despite their agencies' aversion to controversy. Without heroic efforts by these two individuals, the rapid progress in testing the Iron Hypothesis would not have taken place. It is regrettable that at present there are no in situ iron experiment projects under way by U.S. investigators; fortunately, other countries are forging ahead boldly with work in the Antarctic and North Pacific oceans.

1990 Martin, J.H. 1990. Glacial-interglacial CO₂ change: The iron hypothesis. Paleoceanography 5:1-13.

1994 Martin, J.H., et al. 1994. Testing the iron hypothesis in ecosystems of the equatorial Pacific Ocean. Nature 371:123-129.

1995 Chisholm, S.W. 1995. The iron hypothesis: Basic research meets environmental policy. Reviews of Geophysics 33:1277-1288.

1996 Coale, K.H. et al. 1996. A massive phytoplankton bloom induced by an ecosystem-scale iron fertilization experiment in the equatorial eastern Pacific Ocean. Nature 383:495-501.

Work on the preceding achievements was set in motion and doggedly pursued by individual investigators: deep-sea diversity by Howard Sanders and Bob Hessler; new and regenerated productivity by Dick Dugdale and Dick Eppley; zooplankton milieu by Rudy Strickler; and the iron hypothesis by John Martin. Of course, science in general (and oceanography, in particular) is a team activity, and these individuals had important and essential collaborators, but for the breakthroughs described here, these individual investigators were key to the achievement. In this context, these achievements are quite unlike the first two—the discovery of vents and the gaining of a global perspective through satellite imagery—and the following three, all of which were set in motion by teams.

Over the past 25 years our vision of the pelagic food web structure has changed dramatically. We now view the traditional "diatom-copepod-fish" foodweb as a relatively minor component. The food web consistently present in all oceanic habitats is based on pico-and nanoplankton-sized autotrophs and heterotrophs, which are efficiently grazed by flagellates and ciliates. The pelagic food web is microbe-centric. ("Microbe" in this context means small autotrophs, heterotrophs, and mixotrophs, and refers to both prokaryotes and eukaryotes.) Pioneering work by Malone (1971) introduced these ideas regarding picoplankton productivity and micrograzer regulation, but it was not until the late 1970s that this revolution gathered momentum.

The microbial revolution was the easiest achievement to select. In our informal survey it was by far the first choice for inclusion as a landmark achievement, and it was the accomplishment that one of the authors (RTB) suggested at the OEUVRE meeting as the major advance of the past 20 years. There is wide consensus that the microbial revolution is of paramount importance for biological oceanography. It is a revolution still in progress and it appears to be different things to different people (Azam, 1998; Steele, 1998).

In 1974, Larry Pomeroy's paper titled "The Ocean's Food Web: A Changing Paradigm" foretold the microbial revolution by asking a logical sequence of questions:

• Do small autotrophs carry out a major portion of oceanic primary production?