by
J.A. Gulland
Marine Resources Service
FAO Fisheries Department
Rome, Italy
1. INTRODUCTION
The assumption on which the present Consultation is based is that a country (or a group of countries sharing a stock) has determined that in relation to national policies (economic efficiency, greater food production, etc.) the amount of fishing on some given stock is too high and should be reduced.
The simplified rationale behind such a decision is illustrated in Figure 1, which is a typical yield curve as given in numerous volumes and reports on fishery management. This gives the relation between sustained catch and number of standard vessels. The figure also shows the relation between the number of vessels and the abundance (biomass) of the stock. Without any regulation the economic long-term equilibrium is at the point A, where 125 vessels are catching 56,000 tons. A better state (according to the policies of the country) would be B, where, in the long-term, 76 vessels would catch 60,000 tons. Apart from an increased total catch, these 76 vessels should generate much profit, which could go into the pockets of (a limited number of) deserving fishermen, be siphoned off to central government or used in other suitable ways.
In principle there are several ways in which it can be determined that the fishery is being maintained close to the target point (e.g. by checking that the stock abundance is around 90,000 tons, or that only about 70 standard vessels are operating), and several measures by which the necessary controls can be applied (e.g. catch quotas, licence limitations, or various indirect measures). The same measure need not be used at all stages in the process. For example, it may be decided that the best long-term target is to maintain the spawning stock at 90,000 tons; that to do this the catch in 1983 should not exceed 60,000 tons, and this will be achieved by imposing a tax of $50 per ton on all landings, thus discouraging all but a few vessels from going fishing (the level of the tax being set so that just that number of vessels will consider it worth going fishing as is required, given the likely catch rates in 1983, to catch 60,000 tons).
If life were as simple as suggested by Figure 1, then the choice of measure for setting targets or controls would not be critical. Unfortunately, for virtually no fishery does Figure 1 give an adequate representation of biological reality, nor are the commercial and administrative facts of life such that deciding that the quota for 1983 should be 60,000 tons is sufficient to ensure that no more than 60,000 tons are caught. Among the few stocks for which Figure 1 is a reasonable (though not exact) representation of biological reality are whales. The history of whale management does not suggest that biological simplicity alone is a guarantee of successful management.
The choice of the appropriate measures (catches, effort, etc.) for setting targets and controls can be critical. Further, because the nature of the biological, economic, social and administrative complications vary greatly from fishery to fishery, the best measure is not the same in all fisheries but for each fishery has to be carefully chosen after looking at the particular condition of that fishery.
Nevertheless, a superficial examination of fishery management today would suggest that the only proper way of managing fisheries is through catch quotas. The apparently interminable discussions within EEC are concerned very largely with the value of Total Allowable Catches (TACs) and how they should be allocated. Some of the key phrases in the draft convention of the Law of the Sea - “The coastal state shall determine the allowable catch” (Article 61(1)) or “Where the coastal state does not have the capacity to harvest the entire allowable catch, it shall … give other States access to the surplus …” (Article 62(2)) - imply that catch limits are the key to good management. This is far from being the case; later sections in this paper will show that in many situations attempting to control the fishery through catch limits is a second - or third - best method. Also, as shown in another paper, the so-called “surplus” is like a unicorn - very nice in theory and in illustrations of the ideal life but seldom seen in real life. (These comments are, of course, no criticism of the UN Conference on the Law of the Sea, in which fishing was a relatively minor item. Within national delegations the fishery experts were concerned principally in ensuring that their national interests were not hurt. For this purpose meaningless or indeterminate phrases often served well. The only ones among those attending UNCLOS whose prime interest was to see that the text provides as good a framework and guide as possible for rational fishery management and development were some of those from international organizations. These were attending as observers and were relatively powerless to influence the outcome.) To understand why catch quotas have achieved their current prominence, and thereby make it easier to come to a fresh evaluation of the relative strengths and weaknesses of different approaches, it is necessary to give a short historical review.
Figure 1
2. SOME HISTORY
Regulation of important commercial fisheries by means of catch quotas goes back nearly half a century to the North Pacific halibut fishery (Skud, 1973) and, immediately after the second World War, Antarctic whaling (Allen, 1981). However, the key period, when the attention of fishery scientists and administrators became focussed on the control of fishing effort, and when it became commonly accepted that the proper way of doing this was by catch quotas, was in the 1960s, and especially in the two North Atlantic Commissions (NEAFC and ICNAF).
By the 1960s, regulation of the size of trawl meshes had been in force in several parts of the North Atlantic for some time. It was clear that while such regulations did give benefits, there were limits to how far they were effective in multi-gear, multi-species fisheries, i.e. the conditions in most of the area except some of the most northern parts (Bear Island, Labrador) which were nearly pure trawl fisheries for cod. It was also clear that such regulations could do little (and then only in the short- to medium-term) to tackle the economic problems of excess effort and hence excess coasts. ICNAF in particular was active in promoting basic studies of the problem (e.g. Templeman and Gulland, 1965; ICNAF, 1968). These studies showed that there were two basic choices that need to be made: (a) whether the amount of fishing should be measured and controlled in terms of catch or effort (e.g. number of days at sea), and (b) whether there should be a single global quota or whether there should be allocations to countries of specific shares in the total quota.
The second question was easily answered. Both theoretical studies and practical experience in the fisheries for whales and halibut showed that without allocation there would be a scramble for the quota (e.g. the “Whaling Olympics”) and most, if not all, the potential economic benefits would be dissipated. Though a single unallocated quota still continues to be used in some fisheries, e.g. the Eastern Pacific yellowfin tuna, horse mackerel in the South-East Atlantic, these are exceptions. In practice the system of exemptions used in the tuna fishery meant that there were de facto allocations, with the amount left after the exemptions were dealt with being in effect the US allocation. There are doubts about the effectiveness of the ICSEAF measures.
To answer the first question - measuring the amount of fishing in terms of catch or effort - it was therefore necessary to look at the implications for allocations, as well as the other factors such as the constancy of the relation between catch or effort on one hand, and fishing mortality on the other. In the conditions of the North Atlantic Commissions this tipped the balance very clearly towards using catch.
Catch allocations are easily understood; if one country has a quota of 50,000 tons and another 100,000 tons, the second is, in at least one obvious sense, getting twice the benefits of the first. Whether the real benefits are twice as much is another matter, depending on the ways in which the countries use their quotas and manage their fisheries to reduce costs and increase the gross returns (e.g. by timing the landings so as to get the best prices). Everyone concerned - the ordinary fisherman as well as the administrators round the negotiating table - could see what was happening to their shares, in absolute terms and relative to others. In contrast, with the variety of vessels and gears involved, effort allocation was viewed as complicated and difficult to understand. It is far from clear what combination of numbers of Portuguese dory vessels, English wet-fish trawlers, and Russian factory ships would be needed to achieve a target fishing mortality and a fair allocation between countries.
Further, and this was a point given some emphasis in the North Atlantic Commissions, the efficiency of each of these groups will change with technical improvements, and change by different amounts. Thus the target efforts, in total and for each country, would have to be continuously changed. It would be difficult scientifically to determine the necessary change but even more difficult, politically and administratively, to get the necessary changes agreed to, especially to get a country whose fleets have increased their efficiency a lot to agree to a greater than average reduction in its allocation in terms of nominal effort.
It was recognized that the target catch would have to be altered from year to year to allow for changes in stock abundance (e.g. presence of a good year-class). However, it was felt that these adjustments - that did not have to involve differential adjustment to national allocations - could be dealt with. In the context of the main fisheries being discussed - principally the northern cod fisheries - this was reasonable. In most of the stocks the fish are long-lived; typically fish recruit at three to five years old, but do not make their biggest contribution to the fisheries until perhaps six to eight years old, and some of the fisheries on spawning fish depend on even older fish. Thus the strength of most year-classes in the fishery could be estimated from their occurrence in the commercial catches of previous years. The exceptions - the one or two youngest year-classes - did not contribute significantly to the catches and even these, it was believed, could be estimated if necessary by research-vessel surveys of the pre-recruit fish.
To a large extent these expectations have been fulfilled. Allocation of national shares of some Total Allowable Catch has been established as a workable method of controlling the total amount of fishing in a multi-national fishery. Scientists have found it possible to agree among themselves on the values of the TAC to be set each year and these recommended values of the TAC have been, at least in a good number of cases, approximately correct in terms of the state of the stocks and the objectives of the managers. The fishing industry and administrators have learnt that it is convenient to follow the recommendations of the scientists on the value of the TACs.
There have been exceptions. The scientific estimates of the appropriate TAC have sometimes been in error, and the countries have sometimes been able to reach agreement only by agreeing on a TAC in excess of the scientific recommendations. On occasion these exceptions have been serious, e.g. in relation to some herring stocks (Saetersdal, 1980). One cause of scientific error is relevant to the present discussion because it arises from control by catch, rather than effort. To calculate TACs it is common practice to use Virtual Population Analysis (VPA). In this the choice of the terminal F (i.e. the fishing mortality assumed to be occurring in the last year for which data are available) is critical (Pope, 1972). If, in say 1982, when the TAC for 1983 is being calculated on the basis of 1981 data, the value of F for 1981 is under-estimated, this means that the stock in 1981 will be less than one thought and the catches during 1981 will take a higher proportion than expected of this smaller stock. Thus the degree of over-estimation of the stock at the beginning of 1983 (and of the TAC for 1983) may be considerably greater than the original degree of under-estimation of ‘terminal F’.
However, even with these exceptions, the events of the past years have confirmed that catch quotas were appropriate to the conditions of the multi-national fisheries. Nevertheless, the extent to which attention of managers and others have been focussed on catch quotas, and specifically on determining the value of the TAC and its allocation, has had two serious effects.
The first effect has been that within the North Atlantic fisheries (or at least in the north-east), the TACs have become ends in themselves. The general impression given by much of what is said about current problems of managing North Sea fisheries is that if only the correct TACs were agreed on, and complied with, then the fisheries would be properly managed. (This also appears, with a nationalistic flavour, in the belief that if all the foreigners would accept a proper, low share of a TAC, then all would be well.) In fact a catch quota is only a means to an end; unless further action is taken, the imposition of a catch quota goes only a little way towards reducing excess costs and realizing the large economic benefits pointed out by the ICNAF studies of nearly two decades ago. This is not the place to attempt an analysis of what is wrong with current management practices in the North-east Atlantic, still less to suggest prescriptions for improvement. However, it does seem that some shift in attention away from catch quotas could be useful.
The second effect has been more widespread. This is the belief that catch quotas are the only correct way to manage a fishery. Nothing in North Atlantic experience, or in the arguments put forward in favour of catch quotas for the North Atlantic cod fisheries of the 1960s, prove that they would be the most suitable way of controlling the amount of fishing in most present-day fisheries. One factor favouring catch quotas (that a large number of countries all had free access to the stocks) now only applies to tuna and a few other species that can be taken on the high seas. Another (that the stocks were of long-lived fish, without large and unpredictable changes in abundance from year to year) is not typical of most other resources, especially in warmer waters. This point is discussed further in the following section.
3. COMPLICATING FACTORS
The assumption that Figure 1 gives an adequate representation of real life and hence that, from the resource point of view, the choice of management measure is unimportant, can break down in various ways. This can lie within the model itself or can lie outside the model.
Within the model it is not necessarily true that fishing mortality, fishing effort and catch have a neat one-to-one relation. In different years the same catch or the same nominal fishing effort will give rise to different values of fishing mortality. If, as is common, the basic strategy of management involves maintaining a certain fishing mortality, this is undesirable and can be dangerous. Similarly, different catches or different levels of fishing effort (or fishing mortality) will be needed to maintain a certain spawning stock, depending on the current abundance. An annual catch quota must therefore be based on a reliable and up-to-date estimate of the abundance of the total stock, and as necessary of the spawning stock.
The assumptions of the economic aspects of the model, that the costs and value of the fishery are proportional to the fishing effort and the weight caught, respectively, are not always true. If they are not, then any changes in the pattern of fishing brought about by management will not have the economic effects predicted. To the extent that the departures from the assumptions are actually caused by the management action, the results can be very significant. This is particularly true when, as is usually the case, the effect of the regulations is either to increase the costs of a given level of fishing effort (or, more correctly, of exerting a given level of fishing mortality) or to reduce the value of a given catch. When managing an over-exploited fishery, with excess catching capacity, it is often easiest to achieve the immediate objective of reducing the fishing mortality by reducing the efficiency of the fleet (shorter open seasons, prohibition of certain types of gear, etc.). This will increase the costs of exerting the target mortality. Any disruption of the normal pattern of fishing will also tend to reduce the value of a tonne of fish, e.g. by concentrating all the landings in a short open season.
The simple model ignores many biological complications even within a unit stock. All fish in the stock are not the same. A catch of a ton of fish will have different effects on the stock and fetch a different value on the market depending on the size or condition of the fish. Very small fish may be worth so little that they are discarded at sea. Management measures may cause the sizes or conditions of fish caught to vary and may change the discarding practices of the fishermen.
No fish stocks exist in isolation, and many fisheries are concerned with only one stock. Biological interactions, e.g. between predators and prey (whales/krill, cod/herring, seals/capelin), have seldom been explicitly taken into account in setting management measures in the past. In the future, though, it is clear that they will have to be more carefully considered. This raises difficult biological problems, e.g. in designing suitable models and still more in fitting the models to data. They also raise difficult social and political problems, especially when marine mammals are concerned, in determining the most suitable management policy, e.g. in balancing the interests of those wishing to harvest krill against those wanting the most rapid rebuilding of whale stocks. These problems concern the strategy of management, e.g. determining the appropriate levels of fishing mortality for the various interacting species. They however do not affect the question at issue here, of the correct tactics and the measures for achieving, e.g. the target fishing mortality.
The tactical questions are the extent to which one or other choice of measure to manage one stock can affect the harvesting of other stocks. Preferably, the controls on an over-exploited stock should encourage the greater use of under-exploited stocks. The practical effect is however often the opposite. Restrictions on the type of the gear that can be used are particularly likely to inhibit the greater exploitation of under-used stocks. Difficult problems are also posed by the incidental catches of the “protected” species by vessels fishing for other species - should they be allowed to keep them and if so how many. A low limit can mean loss by discarding and a high limit a degree of fishing that is really directed at the target species.
As more management schemes are implemented, the incidental costs of different patterns of management and some of the unexpected effects, not covered by simple theoretical models, are becoming apparent. The most obvious costs are of the additional research, administrative and enforcement activities that are needed to determine the regulations for each year in the necessary detail and ensure they are actually followed. Enforcement, particularly when it requires close checking on the activities at sea, can be very expensive. These costs, and also the costs of research, can be greatly affected by the type of measure used.
The most serious indirect effect of regulation occurs when they lead to mis-reporting of basic data. Where catch quotas are in operation but not fully enforced, the actual landings may be considerably in excess of the figure reported. This has, in some cases, reached the stage where the scientists concerned have had to report that it is no longer possible for them to make meaningful assessments of the state of the stock.
The following sections will review the degree to which management by catch quotas can deal with these complications, as judged by the four criteria of maintaining the resource, economic efficiency, equity, and transaction costs. First, though, the actual mechanics of applying a catch limit will be examined.
4. THE MECHANICS OF CATCH LIMITS
The implementation of a limit on catches involves two or three inter-dependent stages - determination of the level of the total allowable catch (TAC), usually though not always the allocation of shares in the TAC to different groups of fishermen, and some control system to ensure that the catches of each group are no more than their allotted share.
Normally, the common practice is for the TAC to be set according to the scientific recommendations. These will be based on some interpretation of the basic management objective (to exert a fishing mortality equal to F0.1, to maintain the stock at the level giving the MSY, etc.), together with the available information on the current stock abundance, etc. In practice the TAC is often modified from the scientific conclusions in order to resolve some immediate problem.
The preparation of the scientific recommendations, which for the ICES area has been described by Hoydal in a paper to this Consultation, meets two major obstacles. First, the value of the appropriate quota is not uniquely determined by the biological conditions but depends on the specific objectives (MSY, F0.1, etc.). These may not be explicitly defined and the scientists will then have to make some assumptions. Second, the data on the current conditions of the stock will never be absolutely precise. Even if pre-recruit surveys are carried out, the strength of the youngest year-class in the stock in the coming season will rarely be accurately known. There will therefore always be some doubts and imprecisions in the estimated quota. Only rarely is there a good understanding between the scientists and the managers on how to deal with these uncertainties. If the scientists take the natural course and express their conclusions as a range - “to exert a fishing mortality equal to F0.1, the quota in 1983 should be between 95,000 and 120,000 tons”, then too often the administrators have taken the easy course, and used the upper limit (i.e. 120,000 tons). The response of the scientists has then often been to play down the uncertainties, and to give a single figure, usually towards the more conservative end of the range. In doing this they are tending to act as though the scientists had the ultimate responsibility for the resource. In the short run this is probably a good thing, with better levels of TAC being set. In the long run, however, it would seem more appropriate for the uncertainties in the scientific advice (which include doubts about the models being used, as well as the estimates of year-class strength, etc.) and in the models to be dealt with in the final decision-making process together with all the other uncertainties.
The allocation of the TAC between various groups is the key stage so far as economic efficiency and equity is concerned. Without some allocation, implementation of a catch quota involves an ever-shortening open season, during which there is an open competition between all the fishermen to get as much as they can while the season lasts. The chances of much economic benefit is small.
In international fisheries the importance of allocation to countries is well recognized. As noted in the first section this can be easily done only in terms of catch, which is why catch quotas have dominated the thinking about international fishery management.
Allocation is equally important within national fisheries, until the sub-division of the fishery is fine enough for there to be little competition within a group. How much division is needed will vary. Antarctic whaling was carried out by large expeditions, so that even at its peak each national quota had to be shared by at most two or three big companies. Near-shore fisheries, dominated by small family-owned boats, may require several layers of allocation - perhaps first by port, type of gear, or size of vessel, etc. and then sub-divisions within these categories.
Allocation within a country need not be expressed solely in terms of catch, even when the overall limit is so expressed. Limits, for example, on the total number of vessels that can be operated, or on the numbers operated by a particular company, will prevent much of the uncontrolled competition that dissipates the benefits of an un-allocated total catch quota. Such forms of allocation can avoid some of the problems that arise when all allocations are made in terms of catch. Then there can be a lack of distinction between allocation of a proportional share (in some unspecified total) and a share expressed as so many tons of fish. The latter runs into trouble if the stock should decline, e.g. due to poor recruitment; then the sum of the allocated shares may exceed what can be taken from the stock.
The problems of enforcement of catch quotas have, until recently, been neglected, at least in the published literature. When there was little detailed allocation, it might not unreasonably be assumed that catches were correctly reported. The problems were those of having a sufficiently quick administrative system, so that fishing could be stopped rapidly once the target catch was approached. Once each small group is allocated a share in the total, then there will be an incentive for the fishermen in that group to under-report their catches. This mis-reporting, which if extensive, can undermine the whole statistical system on which the assessment and management of the fishery is based, has apparently reached the stage in some areas where the scientists have to state that they can no longer make reliable assessments of the state of the stock.
5. EVALUATION
5.1 Effect on resources
In principle, given adequate knowledge of the abundance of the stock in the coming season, a catch quota enables the impact of the fishery to be maintained at the desired level, whatever that may be. In practice, knowledge of the resource is less than perfect and the catch quota set will not be the correct value to achieve the desired input on the stock. The extent to which its effect will be closer to the desired one than those obtained through other types of control (e.g. on effort) depends on the year-to-year variability of the stock, and the degree to which these can be measured and predicted. At one extreme, whales live a long time and the population varies little. Catch quotas are highly suitable for whales. Stocks of long-lived fish (halibut, cod in northern waters, etc.) live a long time and, though year-classes vary, it is possible to monitor these changes reasonably well in advance of their full impact on the commercial catches. For these too, control by catch is suitable, though probably no better than effort limits. For the more short-lived stocks, e.g. those in the North Sea, and especially in tropical areas, it is difficult to make accurate estimates of the abundance in the following season. A significant part of the catch comes from the newly recruited year-class which can only be estimated by pre-recruit surveys. These tend to have both high costs and high variance. For these stocks other types of control, e.g. on effort, may provide better control of the resource. Exceptions are some of the shoaling pelagic species. Because for these the catchability coefficient varies inversely with abundance, limits on the nominal fishing effort gives little control of the actual fishing mortality. On the other hand, it may be possible to use acoustic surveys at the beginning of the season to get a direct estimate of the current abundance, and thus use catch limits successfully.
Catch limits are useful only if enforced. This point is discussed further in Section 5.4. Enforcement is however almost inevitably of landings, not catches. Some forms of “catch” control can lead to increased discards. The landing of “regulated” species (e.g. Pacific halibut) by fisheries directed to other species may be prohibited. Alternatively a certain percentage of by-catch may be permitted. This can lead to deliberate fishing for the “regulated” species early in a trip to make sure of the allowed percentage, followed by the discarding of any further catch. To the extent that such discards are poorly recorded and not taken account of in the statistical data, the impact on the resource is greater than expected.
5.2 Economic efficiency
This depends on the extent to which the total catch is allocated to individual groups. At one extreme, with a single global quota, management leads to a scramble for the quota, a short season, and great economic inefficiency (excess capital employed, high catching costs, and often reduced value of the catch). At the other extreme, when each individual enterprise is allocated its own share, economic efficiency is high. If a fisherman knows that he can take say, 100 tons at some time during the year, he can arrange to take this at the most profitable time taking account of seasonal variations in fish conditions and market demand, the costs of catching, and the opportunities to engage in other fisheries.
Compared with control of effort, catch limits - provided that there is some allocation - have the economic advantage that they do not lead to distortion of fishing practices as fishermen try to out-manoeuvre the manager, e.g. by building larger boats, or more powerful boats within a tonnage limit. These distortions must inevitably lead to some degree of inefficiency.
5.3 Equity
This too depends on the nature of the allocation rather than catch limitation per se. Without allocation, not only will much of the economic benefits be dissipated, but such benefits as are gained will go to those who can maximize their share of the catch. These will be those fortunately located with respect to fishing at the beginning of the season, or who can best build up their catching capacity. These will seldom be those who the managers would actually wish to give particular preference to.
When there is allocation, the pattern of allocation can obviously be arranged so as to favour any desired group. With detailed allocation the social impact of control by catch is likely to be similar to that of control by effort. Similar problems will be raised - who should get licences/quotas, for how much, can they be transferred. An apparent advantage of catch limits is that they are clearly comparable - a man with a quota for 100 tons has twice the “benefit” of someone with 50 tons. This is certainly an advantage in multi-national fisheries. It may be more apparent than real at the national level, where the question is whether or not an individual can continue to run his boat, or whether a company has to reduce its fleet, and by how much.
5.4 Transaction costs
The year-to-year adjustments in catch limits, which are essential in all but the least variable of fisheries, are costly. At the least, they involve research staff in time-consuming re-calculations, often under severe pressure of time (dealing with this year's data in order to set next year's catches). The costs in terms of diversion of scarce scientific man-power from other tasks, including the adequate consideration of the broader aspects of resource assessment and management advice, may be large. Often there will also be considerable financial costs to carry out, for example, pre-recruit surveys. These scientific costs are likely to be more than those involved in other forms of control.
In contrast, to the extent that controls on the amount landed (but not the amount caught) can be checked in part, costs of enforcing catch regulations are at least no higher than those of enforcing other measures. However, if several stocks of the same species are being managed independently, and the fish from different stocks cannot be distinguished on the market, then control must be carried out at sea. Also, particularly for high-priced species (lobster, salmon, sole, etc.) it is possible for much of the catch to be sold without passing through recognized markets (or to pass through those markets identified as something else). Control of catch limits may therefore not be as easy as might be supposed from a superficial examination, and may well involve costly control at sea. In addition, it may be less obvious to other fishermen that catch limits are being observed than limits of other kinds - it will be fairly clear what vessels (and of what size and type) are fishing in an area, but less clear how much they are catching. It may be harder in the case of catch limits than for effort limits to build up a mutual trust between fishermen that the regulations are being observed by all those in the fishery, although this trust may be the key to reliable implementation of management measures. Finally, it should again be noted that poorly enforced catch quotas can undermine the supply of reliable catch statistics, which is the basis of the scientific side of fishery management.
REFERENCES
Allen, K.R., 1981 Conservation and management of whales. Seattle, University of Washington, Press, 107 p.
ICNAF, 1968 Report of the Working group on joint biological and economic assessment of conservation actions. Annu.ICNAF, (17)
Pope, J.G., 1972 An investigation of the accuracy of virtual population analysis using cohort analysis. Res.Bull.CIEM, (9):65–74
Saetersdal, G., 1980 A review of past management of some pelagic stocks and its effectiveness. Rapp.P.-V.Réun.CIEM, (177):505–12
Skud, B.E., 1972 A reassessment of effort in the halibut fishery. Sci.Rep.Int.Pac.Halibut Comm., (54):11 p.
Templeman, W. and J.A. Gulland, 1965 Review of possible conservation actions for the ICNAF area. Annu.Proc.ICNAF, (15):47–56
