The global ocean landings of all fish and shellfish has held relatively steady in the last few years at about 88 million U.S. tons per year--or 80 million metric tons, the unit we shall use here (Fig. 1). This is somewhat lower than in 1989, when production reached an all-time peak of about 86 million tons, ending a long and steady climb of a factor of four in but fifty years. Any trend in the total catch can be misleading, however, for it obscures larger and sometimes opposite trends in what is caught of different varieties in different parts of the world. Total landings in the Indian ocean continue to increase, while at the same time thirteen of fifteen major marine areas have experienced a recent decline. The catch of cod from the Atlantic Ocean, for example, has fallen by almost a factor of three since a peak in 1970, and that of Atlantic haddock and Pacific bluefin tuna by about a factor of four in the same period.

 Why the total catch has now leveled off is not as simple as sometimes portrayed. Involved are a complex of factors including economics, overfishing, the quality of management, and changes in local and regional environments. Responsible answers to the oft-asked question of whether we are approaching the limit of what we can harvest from the sea are necessarily qualified by present uncertainties in all of these contributing factors.

 The total landings consist of a wide range of species, and include, among others, the cod-like fishes, the herrings, mackerels and tunas, shellfish, and other economically significant varieties such as menhaden that never make it to the table. Some fish--for example, tunas and shellfish--are per pound quite valuable, while others, such as the herrings, are less so. Declining stocks of more highly valued fish are consistently compensated for by increased landings of other varieties.

 Not all fish are eaten fresh: much of what is taken from the sea is frozen into large blocks for further processing or later use. Over the years fish have also been preserved by pickling, drying, smoking, and salting. Today much of the world catch--currently about 40 percent--is not used directly as human food but is converted, in what is a major global industry, into fish meal and fish oil, chiefly for feeding livestock. Commerce in these staples, however, has changed in recent years with advances in the processing of soybeans for the same purposes. The substitution is not a perfect one, for soybean meal lacks two essential amino acids that are naturally present in fish meal. Moreover, the increasing use of soybeans has been counterbalanced by new demands for fish meal as a food source for aquaculture: the "farming" of fish under controlled conditions. The export of fish meal to China, for example, has risen steeply in recent years to satisfy the demands of a very large and growing aquaculture industry.
 
 

Table 1

 ECONOMICALLY-IMPORTANT SPECIES IN TWO MAJOR FISHERY AREAS AND THEIR PRINCIPAL USES

Northwest Atlantic
COD	Fresh consumption; also frozen in various forms
ATLANTIC HERRING or "Maine Sardine"	Canning; fish meal; fish oil
ATLANTIC MENHADEN	Fish meal; fish oil
CLAMS	Canning; fresh consumption; frozen
Northeast Pacific
ALASKA POLLOCK	Frozen filets; fish paste for simulated crab
SALMON	Canning; fresh and frozen whole as filets
PACIFIC COD	Fresh consumption; frozen; now substituting for declining stocks of Atlantic cod
SNOW CRABS	Fresh consumption; frozen
PACIFIC HERRING	Most valuable as herring eggs which in the North Pacific are laid on kelp, and highly prized in the Japanese market

The quantity, or available stock, of any variety of fish is not infinite and varies considerably from time to time and place to place, due to both natural causes and the impacts of fishing and other human activities. These variations can be severely felt in the commercial fishing industry, whose fleet of vessels has doubled in size in the last twenty-five years. Downward trends in the catches of certain varieties have raised valid questions of whether the seas are sufficiently bountiful to meet the needs of an ever-increasing population, and how ocean fisheries can be best conserved and managed and allocated.

 While the fractions of national economies devoted to fishing in the major industrialized nations may not be as large as those in the developing countries, the sometimes large fluctuations in the abundance of fish stocks has had significant impacts on coastal communities in Japan, Indonesia, eastern Canada, New England, and Alaska, to cite but a few examples.

 A well-known case in the developing world is the collapse of anchovy stocks off the western coast of South America in the early 1970s. (A "collapse" is said to occur when the catch undergoes an abrupt and continued drop, often drastic, from which stocks may or may not recover.) In but a few years the natural availability of these small fish fell markedly, from a biomass of 20 to about 4 million tons, with severe impacts on the economies of Peru and Chile. It has since risen to about 8 million tons. The California sardine industry experienced a collapse of a similar kind in the 1950s, as did the king crab industry in Alaska in the 1960s. The recent collapse of the stock of northern cod off eastern Canada -- dropping from roughly 3 to about 1 million tons -- has generated a substantial setback to the economy of the area. And the Governor of Massachusetts recently declared the collapse of the cod, haddock and yellowtail flounder fisheries off of New England as a natural disaster and the cause of great economic hardships to the region.

 The Case of the Orange Roughy
[The orange roughy, about 12 inches long, is found in waters off New Zealand and was in recent years extensively exported to the U.S. Unlike many fish that appear on menus or market shelves it comes not from surface waters but from darker and colder depths of the ocean--as much as a mile down--and is harvested by technologies that have recently come into greater and greater use. The species is also longer-lived than we: those that survive reach maturity in their 30s and have been found to live 150 years or more. Their slow rate of growth and development make the orange roughy particularly susceptible to overfishing, and within the last year, the catch has fallen to but a fraction of what was harvested earlier in the decade. ]

 WHERE FISH ARE FOUND

 Although the oceans cover about 70 percent of the Earth's surface and are on average more than two miles deep, they are not uniformly populated with fish. In terms of depth, the region of primary production is the near- surface layer through which sufficient light can penetrate, called the photic zone. In productive coastal waters it is typically ten to thirty meters deep (or about thirty to 100 feet.) Here solar energy allows the growth of the minute, floating aquatic plants called phytoplankton that serve as a first link in an ensuing food chain.

 About 90 percent of all commercial landings, moreover, are taken in coastal waters where the greatest density of fish are found--the estuaries and embayments that extend outward from the coastline to the edge of the continental shelf. The rich nutrient content in these waters--seen as enhanced concentrations of chlorophyll in pictures of ocean color taken from the vantage point of space--is the result of run-off from the land, upwelling, nutrient regeneration, and other ocean processes that nourish life in the sea.

 Nor are all coastal, shallow water areas of the world equal in terms of the numbers and type of fish that are present. Shrimp, for example, are often found where river deltas deliver great volumes of freshwater into the sea. Extensions of the continental shelf, called banks, define regions of particularly high productivity for many fish; well-known examples are Georges Bank off Cape Cod, and Browns Bank and the fabled Grand Banks off Newfoundland. Other regions are extensive, relatively shallow areas of the oceans, such as the central and southern portions of the North Sea between the United Kingdom and the European coast. Coastal, upwelling areas, such as the coasts that are influenced by the Benguela Current off southwest Africa and the Humboldt currents off the western coast of South America, are also exceptionally productive.

 The fishery productivity per area in the open oceans is generally thought to be lower than in the coastal regions. At the same time there are great expanses in mid-ocean that yield large quantities of tuna and other highly- migratory species. A well-known example is the equatorial eastern Pacific: a common fishing ground for purse seines and long-line vessels that catch yellowfin and other species of tuna.

 APPORTIONING A NATURAL RESOURCE

 The commercial harvesting of fish in ocean waters has been often marked by controversy and contention, ranging from disagreements among fishermen to major international confrontations. Examples of the latter are the "cod wars" of recent decades in Icelandic waters, the recent dispute between Canada and Spain on the easternmost edge of the Grand Banks and the adjoining Flemish Cap, the contested "doughnut hole" in the north Pacific Ocean near Alaska, and the adoption of the Hague Line that now divides the jurisdiction of the eastern tip of Georges Bank between the U.S. and Canada.

 These and other disputes have been rooted in two concepts or problem settings, called open-access, and extension of jurisdiction. The basis for the concept of open access is that fish stocks are not "owned" by anyone, and are there to be taken on a first-come, first-served basis. Yet the unrestricted application of this long-standing approach leads to problems that are all too-well known: a wasteful duplication of capital investment on the part of the fishing industry, and overfishing of finite fisheries stocks.

 In the setting of extended jurisdiction, nations enjoy exclusive fishing rights in ocean areas that lie off their coastlines--restricted zones which were expanded by most coastal nations in the mid-1970s to reach 200 miles outward from their shores. Today about 90 percent of the total catch comes from these Exclusive Economic Zones, or EEZs. In principle, the concept of extended jurisdiction solved most of the problems of open-access, by providing authority to national states to manage the stocks under their jurisdiction, although it is not always used to best advantage.

 A need to limit access, in either of these settings, is well recognized. A number of schemes have been developed, internationally, to remedy open-access harvesting, although it is fair to say that such methods have yet to be applied in many of the fisheries around the world. Nor has a standard framework been developed on how to approach problems of access in a politically acceptable way. By any formula of ownership, how are resources best harvested to return a maximum benefit to society in terms of food and economic gain?

 The problem comes down to one of "who gets what?" As access is restricted, who is to control it, and thereby own the fish? How should ownership be divided among individual fishermen and among various groups--for example, the trawl fishermen, the hook fishermen, and those who fish for recreation? The problem becomes even more complex when the question is added of how stocks in a given area should be divided among those who are allowed access. Or when arguments are heard from those who oppose fishing itself, based on damage to the environment or harm inflicted on other species such as marine mammals.

 Who gets what?

 The question of who-gets-what from the sea lies at the heart of modern fisheries management. Any decision that allocates wealth in any form is certain to provoke controversy, and commercial fishing is no exception. To many people, moreover, the deeper issue is neither political nor economic, but the prudent management of a global resource: how to increase the production of ocean fisheries in the face of growing human needs, particularly in the developing world; and how to deal with changes in the environment that already affect the yield.

 The challenge of fisheries management can be divided cleanly into two questions. The first requires a knowledge of the carrying capacity of the oceans themselves, or how many fish at any time are in the sea: for sustainable production, how many of any kind, or size, can and should be caught? The second, which depends directly on answers to the first, is that of determining the most beneficial allocation among various user groups.

HOW MANY FISH ARE THERE?

 The number of fish in the sea fluctuates considerably in the course of natural cycles of production; as a result, projections of the quantity of any species available for harvest in any week or month or year carry a varying degree of uncertainty.

 The challenge of predicting the available stocks of fish can be approached most simply by thinking in terms of single populations: the number of a given species, such as Atlantic cod, in a certain geographical area, such as the waters of the Grand Banks. At any time, each such population, as with people, is made up of a collection of age groups or year classes, consisting of all those that were born in a given year--for example, in 1994. For the population of cod on the Grand Banks in 1996, the 1994 year class (or cohort) are all of age group two.

 Such a cohort exists initially in the form of eggs. When hatched, the new-born fish do not at all resemble the adults, and are known as larvae. The mortality rates of both eggs and larvae are high, due to natural predation and other factors, and also quite variable in space and time. As the young fish grow, the mortality rate of the cohort declines, and stabilizes during the juvenile or pre-adult stage. Since they are destined to join the population of harvestable fish of their species, those reaching the pre-adult stage are called recruits.

 As with eggs and larvae, the number of recruits at any time is also a function of mortality. In populations with a high rate of natural mortality (such as anchovies), a cohort will exist for but a few years; in populations with a low rate of natural mortality, such as the orange roughy, a cohort may not become extinct for many decades. Initially, when the fish in the cohort are growing fast, its total weight or biomass increases. In time, the death rate will overtake the gains realized by the growth of individual fish and the biomass of the cohort begins to diminish.

 How estimates are made

 All of the factors just cited influence the quantity or weight of fish in the cohort that can be caught at any time, which is calculated and projected with the help of one or more of several interrelated theories, or models.

 The simplest, production theory, is based on the premise that when the population is either very small or very large, the catch that can be sustained on a continuing basis will be limited: when the population is small, because fewer fish are there to be caught, and when the population is large, because less food is available per individual. At some intermediate level-- with the population neither too small nor too large to constrain production--productivity will be optimized, resulting in what is called maximum sustainable yield.

 Yield-per-recruit theory attempts to establish the appropriate rate of withdrawal of various sizes or ages, based on considerations of the average growth and natural mortality of individuals in the cohort. Because some fish can grow in size by more than an order of magnitude during their exploited lives, there is an optimal age at which total yield will be maximized, and a specific time when the cohort biomass is at a peak. In practice, of course, it is never possible to concentrate the commercial fishing effort so precisely in time. As a remedy, the model is taken a step further to provide additional guidance on how more distributed efforts can be organized over longer periods to maximize what is caught, which is commonly expressed in terms of the yield per recruit.

 The so-called theory of recruitment-and-stock estimates the stock size that will maximize the number of recruits, based on the relationship between the number of parents and the number of resulting recruits. This is no easy matter, for the number of recruits in any population is known to fluctuate, from year to year, by factors of four to ten or more. While this variation is indeed large from the point of view of the fishing industry, it is also surprisingly small when one considers the likely variation in the immense number of eggs--often as many as a million per parent--that are initially produced. The consistent reduction in variability from factors of hundreds of thousands or more in the total number of eggs down to factors of four to ten in the number of eventual recruits gives awesome testimony to the way that natural forces regulate the stability of fish populations.

 Each of the three theories has proven quite useful in practice, in spite of the acknowledged simplifications that they employ. For example, production theory ignores the year-to-year changes in abundance that are known to occur. Similarly, yield-per-recruit theory generally carries the implication that the number of recruits will be constant from year to year, which is also not the case.

 Limitations of the theory

 The lack of precision that is at times a feature of the models lends considerable uncertainty to any site- or species-specific projection of ultimate yield. What is more, any projection will be further complicated by larger problems of a broader context: the effects of other, interacting species and the consequences, discussed below, of natural or human- induced changes in the environment in which fish live and grow. These factors introduce major uncertainties regarding the quantity of fish that can be caught.

 Were we able to estimate fish stock abundance, growth, and mortality rates with the precision with which foresters can project the volume of wood in a developing stand of trees, the problem of allocation would be vastly different. But because the total number of fish that can be caught in any year is itself uncertain, allocations among users must also be uncertain. These uncertainties inevitably raise the level of political uneasiness in fishery management. They also increase the likelihood of either overfishing or underfishing the stocks. Moreover, because environmental impacts are inevitably entwined with natural fluctuations in the stocks themselves, the cause of a given decline or collapse cannot always be clearly established, limiting what can or should be done by way of restoration. What is certainly true is that under heavy fishing, several poor year classes in a row, due to whatever cause, will reduce the stock very rapidly.

 ISSUES OF ALLOCATION AND ACCESS

 At the heart of most fisheries controversies are issues of allocation that pervade, in various forms, almost every aspect of fisheries management. Among these are questions of how to allocate a population of fish with regard to site, or season of the year; among those who employ different types of fishing methods; between established, commercial fishing fleets and newcomers, and in some cases between commercial and recreational fishers; and between our own needs and those of marine mammals and sea birds.

A key element of allocation is that of access: who and how many should be allowed to exploit natural fishing grounds?

 A comparison of forestry and fisheries management can again be instructive. A timber company that owns a thousand acres of forest will generally harvest and steward that resource in ways that will sustain production. The economic incentives in the competition for common ocean fishery resources are inherently different and not as much determined by precepts of conservation or sustainability. Economic factors drive any competitor to harvest all that he can, as fast as he can. If access is unrestricted, as is more traditionally the case, the end result is to drive stocks downward, often to a level where, on average, there is little economic incentive to harvest them.

 A number of schemes have been advanced to alleviate problems of open access, and these have been put in use by many nations, including New Zealand, Canada, and the U.S., although performance is still difficult to measure. In one scheme, fishermen are granted quotas that are transferable among themselves: a consequence is that it can be mutually advantageous for the less efficient fishermen to sell their allocated portion of the catch to those who are more efficient .

 Much involved in the concept of quotas, from a practical standpoint, is the effort required to harvest what is allocated. Because of year-to-year variations in the size of stocks, the same numerical quota may not prove as valuable one year as the next. This can be understood in terms of traditional measures of fish-stock abundance, which is expressed in units of catch per unit of effort. For the commercial cod fisherman it is tons of fish landed per hour trawled, and if in one year the landings per hour are half that of the previous year, the abundance of the cod population is seen as having dropped by a factor of two.

 Fish are generally more abundant earlier in the season than later. This means that the catch-per-unit-of-effort in the beginning of the fishing season will be higher, as will the profit realized. As a consequence, those with quotas compete to fill them as fast as possible: to catch the fish most easily caught before the others do. The result is an inefficient use of capital and labor that constrains the total effectiveness of any allocation scheme.

 EFFECTS OF CHANGES IN THE ENVIRONMENT

 One of the principal shortcomings of models now employed to estimate the available catch is the fact that they do not include, explicitly, the possible impacts of changes in the environment. The conditions in which the larval fish develop are particularly important, and are believed to be the principal contributor to year-to-year variations in recruitment. Subtle changes in the mortality and survival rates of the vast numbers of fish that exist as larvae can have a very large effect on the total number that survive. A small percentage increase in survival rate of the larval fish can result in a several-fold enhancement in the eventual recruitment.

 Predation and nutrition are the major determinants of the number of larvae that will survive. The impact of more or fewer natural predators is easily predicted. Nutrition enters the equation in a more complicated way. A variety of factors dealing with the availability of the smaller organisms on which larvae feed can affect their ability to escape the larger fish who would prey on them. Environmental conditions affect the general physiology of both the larvae and that of their predators and prey, as well as the rate of encounters between them.

 Water temperature is an obvious variable that alters the behavior and abilities of all organisms in a marine ecosystem. Fish, in particular, cannot regulate their own body temperature and must either seek waters of their particular temperature preference or allow their responses to adjust to what is at hand. A subtle warming or cooling, moreover, may affect different populations in different or even opposite ways. Consider a larva and its typical prey. A one degree increase in temperature may (i) improve the performance of both; (ii, or iii) give relative advantage to one at the expense of other; or (iv) diminish the abilities of both of them. Which will apply depends upon the temperature preferences of the two, as well as the history of temperature change and the relative abilities of the larva or its prey to accommodate.

 The life-or-death matter of encounters between fish and their prey can be complex in any ecosystem setting. A key factor is how disturbed or turbulent the water is--not on the scale of waves or ripples but in dimensions of tenths of millimeters. Turbulent flow on so minuscule a scale is an ever- present feature in most moving water. In the contest of the chase, it replaces the level playing field of still water, boosting the velocity of predator and prey unequally, to either increase or decrease encounter rates between the larva and its food. The effect of small-scale turbulence can be subtle, depending not only on the degree of turbulence, but on the distance between the chaser and the chased. For example, as the distance between them shortens, the effect of turbulence on encounter rates is reduced.

 Small-scale turbulence and its impact on the balance between aquatic predators and prey can result from environmental disturbances of a vastly larger scale, including changes in the pattern of winds that blow across the surface of entire ocean basins. Since almost any change in global or regional climate will alter the wind field, the amount and distribution of turbulence beneath the water will also be affected. Through this chain of events, the natural balance among fish of any kind will also be disturbed.

 The example of Pacific salmon

 An interesting example of environmental impacts is that of salmon productivity in the North Pacific, which, while dependent on many factors, seems to be correlated with long-term changes in the intensity of a persistent barometric feature of the area called the "Aleutian low." The reported correlation--that the catch of salmon is highest when the atmospheric pressure is persistently low--has held over the last several decades, suggesting a possible physical linkage between fish stocks and natural changes in the environment. In addition, during the most recent, 1970s intensification of the Aleutian low--when data on other populations were also available-- substantial increases were also found in the abundance of minute, floating aquatic animals called zooplankton, and of squids, and Hawaiian monk seals.

 Further investigation by marine scientists has shown that during the 1970s the depth of what is called the mixed layer of the ocean--the regime in which the warm water of the surface circulates and mixes with the much colder water below--went through significant changes. At this time the mixed layer in the central Pacific Ocean increased in depth by as much as 70 percent; in the North Pacific it decreased by 10 to 20 percent. The implication in the first case is that a deeper mixing layer stirred more nutrients downward, increasing the productivity of fisheries. In the case of the North Pacific, where a change of the opposite kind was also associated with an increased catch, the explanation offered is that the calmer, surface water increased the exposure of phytoplankton to light, thus increasing the food supply. The rate of production of phytoplankton, as with any plant, is determined in part by the amount of sunlight they receive.

 The significance of these findings, if indeed valid, could reach well beyond the sampled instances in which they were observed. If the productivity of fisheries follows interannual, decadal, or other long-term climatic changes, then changes in the catch--such as the notable decline of certain fisheries in recent years--need to interpreted in the context of these natural cycles, as do ensuing decisions of the management and allocation of stocks.

 It can also be assumed--if the environment is their driver--that the large decadal changes in salmon abundance are but a sample of more basic and larger-scale changes in ocean productivity. In this case we should expect multidecadal changes not only in salmon abundance, but perhaps in the overall yield of the entire North Pacific ocean. Knowledge of natural cycles of production would obviously aid the development of management programs for wild stocks. Natural cycles of production, in the case of salmon, could also significantly affect the returns expected from the several hundred million dollars that are invested each year in raising salmon in hatcheries on the west coast of the Americas. In years when the natural production is high, the return for each young fish released into the ocean will be very much less than in years when it is low.

 Probable impacts of global climate change

 A working group on fisheries of the Intergovernmental Panel on Climate Change (IPCC) has recently concluded a study of the probable impacts of enhanced greenhouse warming. The study is admittedly limited by the often case-specific state of what is known of effects of the environment on fish stocks, and by uncertainties in the parameters of climate change on which the study is based, including the impacts of global warming on the structure and strength of ocean currents. A general picture emerges, nonetheless.

 Projected global warming would have but a small effect on fish stocks, if we consider only the direct effects of changes in sea surface temperature, and ignore the local economic consequences of shifts in stocks of fish from one region to another. With higher temperature, global fisheries production should remain about the same, because of the thermal inertia of the ocean and the relative ease with which most fish can migrate to seek suitable conditions. The total catch, however, will probably be distributed differently from today, and there are likely changes in the dominance or commercial importance of certain species. Overall, the changes we should expect may not be negligible.

 A slight change in ocean temperature, due to the way that off-shore fisheries have now been nationalized, could result in systematic shifts in stocks of certain species from one nation's EEZ to that of an adjoining country. On a more local scale, a geographic shift in stocks within the boundaries of an EEZ would be felt economically by local or regional industries, and might most severely impact the subsistence fishers in the developing countries.

 Potentially far more important are what might be called the side effects of global warming, and specifically the interactions between the oceans and the atmosphere that affect the movement of water in the sea. Changes in surface winds could have a large impact on fisheries in that they would alter both the delivery of nutrients into the photic zone and the strength and distribution of ocean currents. Significant changes in ocean currents will affect the transport of larval stages of fish and alter upwelling and exchange of water over highly productive banks.

 In addition, the rise in sea level of the amount projected by the IPCC would have an undeniable effect on the commerce in fish, since about 70 percent of global fish resources spend critical parts of their lives near the shore or near the mouths of rivers. Any change in marshes or wetlands, or in the extent to which seawater intrudes into rivers, would be felt by many ocean fish. For shellfish--that live almost exclusively near the shore--a sufficiently rapid rise in sea level could tax their ability to relocate: a shift in habitat that would also be constrained by the roads and other structures that now mark so much of the shoreline. A sea-level change would also have significant impacts on the existing facilities of what is now a large aquaculture industry.

 As with agriculture, there would be winners and losers: global warming would likely be implicated in the downturn of some regional fisheries and in the expansion of others. As with agriculture, the negative effects of global warming on ocean fisheries could fall disproportionately on those who could least afford to make needed adjustments.

 THE ROLE OF AQUACULTURE

 Raising fish under controlled conditions is often suggested as a way of compensating for the shortage or decline of natural stocks, and even as a technique for replacing conventional commercial methods of fishing. Certainly not all species can be reared in this way. Still, the practice of aquaculture--or when applied to ocean species, mariculture-- has expanded in recent decades, and today adds about 16 million tons, or 8 percent, to the catch of fish of various kinds that are each year taken from the sea. The average growth of the industry, globally, is about 5 percent per year.

 Fish for human consumption are raised in inland pools or coastal waters by a wide variety of techniques. In France, for example, a very large yield of oysters is sustained by a system of "ranching" in large embayments. The tiny oysters, known as "spat," settle on rocks and other hard surfaces and are relocated to elevated areas of the bottom, called "tables," which are exposed to the air when the tide is low. At these times, and only then, the oyster farmers can tend their crop of oysters with tractor-drawn equipment. The highest quality oysters are finished in special ponds called "clares," in which the oysters are carefully spaced so that each has an ample supply of planktonic algae as a food and a contributor to flavor.

 Common freshwater practices include the culture of trout in North America and of carp in Asia. Most of the effort in saltwater aquaculture has been devoted to shrimp and shellfish such as oysters, although high quality, specialized fish such as sea bream and yellowtail are commonly raised in Japan. Recent advances include the rearing of salmon in pens in Norway and the production of shrimp in Ecuadorian waters, where young shrimp are captured in the wild and then raised in coastal ponds. These techniques have contributed substantially to the global production of these two, commercially-valuable fish.

 We need keep in mind, however, that aquaculture as practiced today does not appreciably reduce our demands on the sea, for many of the fish raised in this way must be fed protein, usually in the form of fish meal, which is derived from other varieties caught commercially in the wild, and in large quantities. The problem is that only a fraction of what is consumed by any fish is converted into growth; the remainder is burned by respiration or excreted. Thus, to produce a pound of fish by "farming" can require several times that weight of other fish, which also constrains the economic feasibility of the process. Moreover, space suitable for mariculture is limited: all but one of the sea lochs in western Scotland, for example, now support salmon farms.

 Aquaculture still holds many promises for an expanded world population, but the advantages--so beguiling at first look--need to be examined in the cold light of costs and technical possibilities. While freshwater aquaculture can make a difference in global food supply, mariculture, as practiced today, appears to be financially successful only in the case of more expensive varieties of seafood, such as shrimp, oysters, and salmon.

 IMPACTS OF FISHING METHODS ON THE SUSTAINABLE CATCH

 As in almost any industry, the economic pressures in commercial fishing drive investors ever in the direction of reducing the labor required, and the number of days that vessels are kept at sea. Examples of techniques now in use include gill nets that entrap fish more effectively, the use of ever larger purse seines, and drift nets that stretch out tens of miles from the large ships that deploy and reel them in. These and similar developments in fishing gear inevitably increase the number of other fish that are too small or less valuable species and are thrown overboard: a figure that is now put at 27 million tons of fish per year, or about a third of the amount that is kept. For some kinds of commercial fishing, the fraction can exceed by far that which is kept. In addition, marine mammals and sea birds are sometimes destroyed by fishing operations.

 The importance of controlling "bycatch" cannot be overemphasized. Little is known of the possible disruptions that may ensue in an ocean ecosystem when so great a mass of dead fish is discarded. In many instances what is returned to the sea is clear waste, as when those fish too small for processing are allowed to expire and are thrown back into the ocean, at the expense of future yields of the same species. In other cases, as in shrimp fisheries, it is not clear whether the substantial bycatch may not contribute to production, in the sense of thinning an overly-abundant stock.

 Factors that impact the balance of marine ecosystems are a valid concern in terms of the sustainable use of the oceans. But while much is said about the effects of modern fishing gear on the conservation of natural resources, the situation can be all too easily overstated. Gill nets and drift nets, for example, pose a threat to fish stocks only to the extent that the number of these nets in use may take quantities of fish greater than what is prescribed as optimal by fishery managers. If we put aside more personal concerns as to whether other species should be subject to the effects of fishing, each fish caught--no matter what the method--depletes the stock by one.

 A concern of a different kind is the use of gear such as trawls and dredges that by design disturb the ocean bottom. Some, such as the oyster patent tong, combined with heavy siltation from agricultural run-off, have done much to destroy oyster habitat. While bottom trawls are used in many fisheries and affect an extensive area of the bottom, we are only beginning to learn the effects of these gear on fish habitat, and whether they are harmful or beneficial.

 CONCLUDING THOUGHTS

 The quantity of fish taken wild from the sea has risen steadily through most of the present century, roughly keeping pace, until recent years, with the growth in human population. It could be, as some have suggested, that we are nearing the maximum sustainable yield of the oceans, or more accurately, what is possible with current methods of fisheries technology and management. What is more certain is that the levels of today are maintained by supplanting the often declining catch of more desired varieties with those that are less so, and are further aided by augmenting what is caught with an increasing reliance on aquaculture. Declines in recent years in the catch of fish such as cod and haddock give clear warning of what can happen when a fixed and naturally-variable resource is subjected to increased fishing pressures, particularly with the awesome capability of modern technologies.

 Significant climatic change would unquestionably affect ocean ecosystems, in ways that go far beyond a simple poleward shift of any species in response to warmer surface waters. Any significant environmental change will also alter the values of parameters that are customarily used in fishery management models to estimate optimal production, yield, and levels of stock. But until more specific changes in ocean circulation and in off- and on-shore conditions can be better predicted and factored in, a high degree of uncertainty will prevail in the management of fisheries, contributing additional questions to present, imperfect systems of allocation. The result is greater uncertainty in long-term capital investment strategies, added tension in regulatory decision-making, and a more wasteful use of the ocean resources that Nature provides.

 The ultimate carrying capacity of the oceans is not a known quantity, and certainly not a constant, although a potential sustainable yield of about 100 million metric tons per year is often cited, amounting to an increase of about 20 percent above what is now taken. This estimate, if valid, falls far short of the needs of a doubled world population, based on today's patterns of consumption.

 Whether the oceans will prove sufficiently bountiful to provide an adequate supply of fish as food for future generations will depend in part on how well we do in managing the resource. Certainly needed is the development of more enlightened policies on how many fish should be taken where and when, as well as the more effective application of such policies, both nationally and internationally. These will demand, in turn, a deeper understanding of the causes of natural and induced variations in stocks.

 At present we have little capability to forecast how environmental changes affect recruitment, and limited knowledge of how species interact with one another in the ecosystem or in the fishery. The most important needs are extensions of existing theories of natural production, and the development of enhanced data systems to support them. International research activities will contribute considerably to a broader understanding of ocean ecosystems. We can hope that what is found can be applied to a more enlightened management of what is taken from the sea.

FOR FURTHER READING

 Global Fisheries, by B. J. Rothschild. Springer-Verlag, New York, 1983.

 "More Food from the Sea?" by B. J. Rothschild. Bioscience, vol. 31, pp 216-222, 1981.

 Population Production and Regulation in the Sea, by D. H. Cushing. Cambridge Univ. Press, Cambridge, U.K., 1995.

 The Provident Sea, by D. H. Cushing. Cambridge Univ. Press, Cambridge, U.K., 1988.

Reviewers

 Dr. John T. Everett is Acting Director for Research and Environmental Information at the U.S. National Marine Fisheries Service, an arm of NOAA, in Silver Spring, Maryland. He is an active member of the Working Group on Impacts and Adaptation of the Intergovernmental Panel on Climate Change and serves as Chair of the Fisheries Section.

 Prof. James J. McCarthy, a biological oceanographer, is the Director of the Museum of Comparative Zoology of Harvard University, where he is a faculty member in the Department of Biology and the Department of Earth and Planetary Sciences. He was the first chairman of the Scientific Committee of the International Geosphere-Biosphere Programme and the founding editor of Global Biogeochemical Cycles, an interdisciplinary journal of the American Geophysical Union.

Figure 1 Landings of marine fish in millions of metric tons, 1975-1992.
Source: United Nations' Food and Agriculture Organization (FAO), 1993.