How low will we go in fishing for dinner? Though news accounts over the past decade have documented the crash of one major fishery after another, many consumers have witnessed no shortage of affordable fish. In large measure, that's because different fish are being marketed. Indeed, species once viewed as "trash" can now command $7 per pound or more. Many resource economists have interpreted this trend to mean that while the most popular fish stocks are in jeopardy, a host of attractive alternatives stands ready to fill in. A new study by fisheries scientist Daniel Pauly of the University of British Columbia in Vancouver and his colleagues now comes to a dramatically different -- and more dire -- conclusion. Those substitutes, they find, have been coming from progressively lower niches in the marine food web. With each successive drop, dramatically more fish become available. Yet despite having made these shifts, and working harder, fishing fleets have not increased their tonnage of palatable catch. Moreover, the new data suggest that the food web's structure -- the proportion of organisms at each level -- is shifting. "The ecological price we're paying for maintaining catch is getting higher and higher," Pauly says. Indeed, the findings argue that current world fishing rates are not sustainable, his team concludes in the Feb. 6 Science. Ecologists measure an organism's niche in terms of its trophic level. In the sea, the base level contains mainly seaweeds and phytoplankton. These serve as food for level two organisms, whose predators, in turn, make up level three. And so it goes up the marine food web to its apex, killer whales at trophic level five. Owing to taste preferences, humans have traditionally fished primarily from levels three and four, Pauly says. However, because such fish may derive their diet from a range of trophic levels, most commercial fish don't fall squarely into a single level. Rather, they have an intermediate designation, such as 4.6 for snapper, 3.5 for cod, 3.1 for herring, and 2.5 for sardines. Pauly and his coworkers have now computed the annual average trophic level of the world's fishing catch. They did this by tracking down the trophic level of 220 fish and invertebrates and considering each species' share of the tonnage of a given year's fishing haul, as compiled by the United Nations, for 1950 through 1994. Their calculations show about a 0.1 decrease in trophic level per decade -- to a current global average of about 3.1. These data show that by overfishing the top predators, "we've eliminated the marine equivalent of lions and wolves and are moving towards the taking of rats, cockroaches, and dandelions," worries Elliott A. Norse of the Marine Conservation Biology Institute in Redmond, Wash. Moreover, he says, "by now moving to eliminate the top predators' prey and the prey of their prey, we may be further impeding [the top predators'] recovery." Both concerns are "implicit in our findings," Pauly believes. "If we have fallen half a trophic level in 40 years or so, then we have already hammered the useful part of the food web." The new study "is clever and meaningful . . . and I think that its conclusions are robust," says marine ecologist Paul K. Dayton of the Scripps Institution of Oceanography in La Jolla, Calif. Gary Matlock, director of the National Marine Fisheries Service Office of Sustainable Fisheries in Silver Spring, Md., also thinks the new study's findings have a lot of merit. Clearly, he says, "the overfishing that has occurred on the upper trophic levels needs to be brought under control." According to the 1996 Sustainable Fisheries Act, he notes, his agency must develop a strategy by September to end such overfishing and to begin rebuilding affected U.S. stocks.