It has been noted in Section 3.2 that the data on catch per unit of effort, both for Angola (both Sardinella species combined) and Congo (for each of the species separately) do not show any specific trend. In neither of the cases, however, had allowance been made for changes in fishing power of the fishing vessels.
Although no information is available on the introduction of powerblocks and other improvements of the efficiency of the vessels in the period under consideration, data are available on the increase in size of the seiners. It would seem unlikely that the fishing power of a seiner is proportional to its tonnage. Campos Rosado (1972) shows that the fishing efficiency of seiners of 90 GRT fishing in south Angola, mainly for Trachurus (horse mackerel), is less than twice that of seiners of 80 GRT.
As far as the catch rates of purse seiners give a measure of the fish abundance, it can therefore be taken that the effort uncorrected for increase in boat size will lead to an underestimate of the effort in standard units, and hence to an overestimate of the fish abundance in later years, whereas fishing effort corrected with a factor proportional to the boat size will lead to an underestimate of that abundance. Taking this into account, Figure 5 shows the catch rates in the Angolan and Congolese purse seine fishery, per boat/GRT/year and boat/hold capacity/day respectively. In this graph, the data for S. aurita still do not show any definite trend, but those for S. maderensis now show a substantial decline in Congolese waters between 1964 and 1978.
Thus, it would seem that there has been a decline in the density of S. maderensis in Congolese waters, although neither of the two graphs will have given a proper picture of this decline.
In considering these results, it should be taken into account that the catch rates of purse seiners may not give a good representation of the real changes in the fish stocks, for various reasons discussed in many studies. The most important among these are the possibility of saturation of the vessels at high fish densities, and factors related with the schooling behaviour of the fish and the behaviour of the fishing fleet.
It is, therefore, also possible that the picture given by the graphs of relative low catch rates of S. maderensis in several years when the catch rates of S. aurita are high, and vice versa, is to a certain extent due to the peculiarities of a fishery on schooling fish, e.g., by concentrations of the densest schools, and may not mean that there really is clear opposite fluctuations in the abundance of the two species. However, the fact of the very large fluctuations in the catch rates, especially of S. aurita (over a range from 1.1 to 17 in the uncorrected data) and less pronounced in S. maderensis (from 4.2 to 10) remains and can clearly be seen in both the graphs of catch per effort, whether uncorrected or corrected for changes in boat size. Apart from a possible artificial effect of an underestimate of one species in years of high catch rates of the other species, these fluctuations may be related with fluctuations in the distribution or behaviour of the fish and therefore in its vulnerability to fishing, or with real fluctuations in fish abundance following poor or good year classes. In this regard, it may be noted that the Angolan fishery had very high catch rates in Division 1.2 in the years 1969 and 1972 (though not in Division 1.1), and that the catch rates of S. aurita in Congolese waters were very high, nearly simultaneously in 1969/1970 and 1972/1973. An earlier peak in the Angolan catches in Division 1.2 is observed in 1964, and may be due to the good yearclass 1963 observed in Ghéno and Poinsard age-determination (1969), although the Congolese fishery did not show a high catch of S. aurita in that year. An argument for the suggestion that the high catch rates in Angola and Congo in the two later periods have also been due to very good recruitment may be found in the results of the investigations by Fontana and Pianet (1973) on the gonadosomatic index, referred to in Section 4.3. This index was found to have been particularly high in the years 1967 and 1968, as well as in 1970/1971, and may have given rise to exceptional recruitment in the following years.
Summarizing the available information, it would appear that the fleet fishing for Sardinella spp. in ICSEAF Division 1.2 has remained practically constant in most of the period 1956-1972, with a slight decline in number of boats and a slight increase in their average size. In ICSEAF Division 1.1 no data on sardinella catches are given for the years before 1969, although there was a substantial fleet fishing. However, the sardinella catches in that division after 1969 were only a small fraction of the catches in Division 1.2. Also, in the Congo area there has been some increase in fishing, but the catch in that area is minor compared to that obtained in Division 1.2. Altogether it is likely that in the period 1956-1972 there has been some overall increase in fishing effort for sardinella in the total area, but this increase has probably been rather small. This in itself makes it difficult to apply the traditional techniques of stock assessment to these data.
If it is further taken into account that the data on the total landings may not be very precise as they had to be based on assumptions regarding the species composition (Section 3.1), and the effort for the Angolan fleet is only available in the rough measure of number of boats per year, it is clear that the available observations do not allow to make any estimate of the size of the sardinella resources and their potential yield. Although the probable decline in the catch rates of S. maderensis in Congolese waters may be a local effect of an increasing fishery in these waters, none of the information available so far indicates a serious effect of the fishery on the resources as a whole. It is, therefore, possible that fishing mortality rate is still low and that hence the catch of sardinella can still be substantially increased.
Consideration of the biological characteristics of S. aurita and S. maderensis given above, such as a very high growth rate in the first few years and very slow growth thereafter (i.e., a high value of K), together with a relatively high length at average recruitment (lc perhaps about 50 percent of L¥ according to the Congolese data), indicates that the yield per recruit will continue to increase with fishing until high levels of fishing.
One aspect, however, should be taken into account. It has been experienced in various parts of the world that pelagic resources may suddenly collapse after fairly heavy fishing, even in cases when the catch rates did not show a serious decline. This was, at least in some fisheries, due to a failure in recruitment. It may therefore be important to carry out fishing in such a way that the size of the spawning stock remains above a certain minimum. If the large fluctuations in S. aurita indeed represent fluctuations in stock size rather than in vulnerability to fishing, it might well be harmful to exercise heavy fishing in years of very low abundance, which might bring the spawning stock to too low a level. In particular, in view of the present tendency of increasing fishing both by national and foreign fleets, it is important that the developments in the fishery and in the stocks of each of the two species separately are carefully monitored and that all data, together with whatever data are still available in offices and files but unpublished, are thoroughly and regularly analysed. This will allow for better judgement if and when the need for management arises and preparation for management action if necessary.
One final point should be mentioned. Both the studies and the eventual management of the fisheries are greatly affected by the question of whether the resources of each species in the area all belong to the same stocks, or are separated into different, more or less independent, populations. In the first case, all studies and management should be directed at the stock of each species as a whole, and be undertaken jointly by everyone concerned, whereas in the latter case each population can be handled separately.
The limited knowledge of the distribution and migration of the species in the Congo-Angola area suggests that the picture here may well present a mirror image of the situation in northern West Africa from Guinea Bissau to Cap Blanc, an area of similar oceanographic conditions. It would then appear that S. aurita may belong to one stock migrating throughout the whole area, whereas it is possible that S. maderensis is separated into a number of more or less independent populations along the coast.
In this respect, it should be noted that the reduced fishing effort in Angola in 1975/1976 does not appear to have had any effect on the catch rates in the Congo area, which could indicate a certain stock separation, or a low fishing mortality rate even before 1975.
