Beyond this general knowledge of behaviour, a spawning investigation may sometimes be adapted to give quantitative estimates for two of the variables which describe the structure of the fish population First, under certain circumstances the number of eggs in a given area may provide a measure of the potential recruitment of young fish to the stock in later years Secondly, if the average number of eggs produced by one female fish is known, the total number of female fish spawning in that area may be computed. Since the sex ratio is approximately unity, twice this figure gives the total population of spawning fish, which is an estimate of the stock available to commercial exploitation.

To be of value, estimates of this type must not only be made to a sufficient degree of accuracy to detect any changes which occur from year to year, but must also be made in such a way that the magnitude of sampling error can be determined. A perusal of Figs. 4–9 shows that, not only is the distribution of eggs at any one time highly aggregated, but that a very large proportion of the total quota is provided by comparatively infrequent dense patches, the omission of any one of which might contribute a considerable error to the total estimate In some cases these patches may be of considerably smaller width than is indicated by the 4-mile unit sampling lines It is indeed possible that all the difference between a prolific and a meagre spawning season might be made by one very dense patch of eggs, occurring for one day only and in one small area. Contributing to this difficulty is the fact that the snapper egg develops comparatively rapidly and is thus available for estimation for no more than two days as compared with say, the North Atlantic cod which, at temperatures near freezing, remains unhatched for over a month It is thus clear that any truly accurate estimate will require a pattern of samples repeated every day and covering every part of the Gulf It is not possible at present to state the appropriate sampling interval, but it would seem that this should not be greater than five miles. Assuming that the present high-speed plankton sampler gives a completely adequate sample for this purpose, such a programme would require nearly 24 hours towing at 8 knots each day for several months, a working intensity which cannot be approached with present resources. Even then, the time allotted has made no allowance for random sampling which would be necessary to obtain an estimate of sampling error.

It would seem unlikely, therefore, that any useful quantitative estimate of total egg production could be made without employing several vessels. Similarly, some caution is necessary in interpreting the data presented above. Although the relative lateness of the 1950–51 spawning compared with that of the previous year appears to be established with a reasonable degree of likelihood, it is by no means impossible that bursts of spawning which remained undetected could have made a substantial difference to the interpretation of results.