tests; and the elusive potential of 32Si. Did they all have to be measured? What were the potential gains? What sampling pattern and density were required? How might the results be incorporated into physical models? These questions, first posed within the GEOSECS context (Craig, 1972), have only been answered with any rigor in the last few years. They are fundamentally hard topics, and it is a tribute to the NSF of those days that it ventured into (literally) such uncharted waters.
Moreover each of these tracers had chemical reactivity as well as radioisotopic decay. This had to be constrained too, and for 14C it meant attacking the full ocean CO2 system, while for the radium isotopes the chemical analogue of barium was selected. Each of these efforts had its advocates, and once the science case had been made within the GEOSECS steering committee, it would petition the NSFIDOE for funds.
It was plain at the outset that the GEOSECS program would be fundamentally different in style and scale than anything before. It was also very confusing. At least three major institutions were big players: Scripps, where Harmon Craig had persuaded Arnold Bainbridge to set up the GEOSECS Operations Group that was to craft the advanced instrumentation and staff the technical support activity; Lamont, where Wally Broecker had pioneered many of the radiochemical tracer techniques and gas exchange rate concepts; and Woods Hole, which was to provide the RV Knorr for the first, Atlantic, expedition and where Derek Spencer created the coordinating center. Karl Turekian at Yale and Gote Ostlund at Miami provided wisdom and refereed the sometimes amazing disputes that arose. Incidents involving fire extinguishers, epoxy, and roller derby are best not mentioned here. At each institute there were young scientists eager to be involved, but all had different views on what would be needed and on how to make a personal scientific effort within this large enterprise.
The problems were typical. Big programs need to be staffed with first-class scientists who will remain with the program for years. First-class scientists cannot be cogs in a big wheel, but are very inventive people of rapidly evolving interests who need to create their own identity and establish their own careers. How to balance these conflicting needs often lies at the heart of ambivalent feelings about large programs. Although the decade was dubbed International, it was not clear what this meant for a particular program, but in the case of GEOSECS several individuals in other countries (Yoshio Horibe, Devendra Lal, Wolfgang Roether, Brian Clarke, and Roger Chesselet) made extraordinary personal efforts.
So far as I can tell, the IDOE programs were not successful in making use of SCOR (the Scientific Committee for Oceanic Research) or the IOC (Intergovernmental Oceanographic Commission), the formal ICSU (International Council for Science) and UN-affiliated international bodies, respectively, to carry out their planning or execution, in spite of a strong effort to do so. Lou Brown was brought into NSF to serve as the internationalist, and he remains at NSF today. In practice the dominant new factor was the forcing by NSF of domestic interinstitutional, not international, expeditions and programs. This radical intrusion into the sovereignty over their ships enjoyed by the major institutions was a source of great discomfort to traditionalists, but it opened the door wide to young and ambitious scientists.
The Atlantic Ocean—The GEOSECS Atlantic Expedition in 1972, the Pacific Ocean Expedition in 1974, and the Indian Ocean Expedition in 1978, all presented unique challenges. The Atlantic cruise was preceded by at least two test stations, or cruises, that showed somewhat alarming results (Craig and Weiss, 1970). Measurements of the CO2 system properties made by different techniques gave discordant results, calling into question the basis for using the 14C tracer. It was shown that the precision of the ? 14C measurement, achieved by Gote Ostlund and Minze Stuiver, would be ±4 per mille, giving an age resolution ? 30 years. The product of total CO2 and 14C was required, and the supposedly easier total CO2 measurement should not degrade the signal. But it did, and the confusion was to last for several years. It is a tribute to the drive of the leaders, and the courage of NSF, that the expedition went forward.
This courage was soon tested. The Knorr left Woods Hole for a nine-month cruise on July 18, 1972—and lost the entire, horribly expensive, conductivity-temperature-depth probe/profiler (CTD)-rosette sampling package on the very first station. Apparently a locking pin had not been set in place. In spite of heroic efforts by Arnold Bainbridge, not all the advanced analytical systems worked, arid it would be months before the CO2 system was fully operational. Not everyone trusted the new systems to work at all; Joe Reid had insisted that a separate conventional Natnsen bottle cast be done as backup at every station, and this " as laboriously carded out. The cruise tracks are shown in Figure 1.
Tensions soon arose. The work was long and hard, the cost of supporting the expedition was high, and in contrast to small-or medium-scale science, there was no natural break to stop and write papers. It began, to some hostile critics, to look like a large, expensive, general data gathering hydrographic exercise, possibly similar to the International Indian Ocean Expedition a decade before. A review was held at NSF, and the case for going ahead with the Pacific cruise was made. "What," Admiral Owens asked, "would be the consequence of not funding the Pacific cruise?" Amid the uproar he had made his point: this was NSF research, managed by the Foundation which did have the last word. And there were expectations for individual scientific accomplishment.