64 



BULLETIN OF THE BUEEAU OF FISHERIES 



tions are conducted with their usual intensity, accompanied by the customary 

 feverish raking and scraping to secure the largest pack possible under unfavorable 

 circumstances. If the nature of the season could be predicted with a fair degree of 

 certainty, it would be possible to avoid immediate serious operating losses and at 

 the same time aid in standardizing the runs by securing a more nearly adequate 

 spawning escapement even in a poor year. 



One of the foremost among the objects to be sought in a fisheries conservation 

 program, therefore, is a reliable basis for prediction concerning the magnitude of 

 future fish supplies. This need is appreciated in all fishery investigations, whether 

 of salmon or herring, halibut, cod, or mackerel. The significance of the intensive 

 studies of dominant year groups, of which we hear so much, lies here. In the case 

 of the salmon, in comparison with other marine fisheries, we enjoy certain obvious 

 advantages in developing a technique of prediction, together with certain disad- 

 vantages equally obvious. Among the advantages we find the salmon segregated in 

 relatively small, geographically limited, self-perpetuating races, the entire spawning 

 colonies of which annually may be passed in review, and permit detailed quantitative 

 and qualitative determinations impossible in the case of species that spawn in the 

 sea. Among the disadvantages stands prominently the fact that the salmon are 

 relatively short-lived, and, after the young enter the sea, are for the most part wholly 

 inaccessible until the year in which they reach maturity, seek their spawning beds, 

 and invariably die. We are largely deprived, therefore, of the important aid to 

 prediction that can be obtained by the investigation of dominant or defective year 

 classes, which can be seen to pass through the herring or other fisheries over a term 

 of years. With the salmon these are available only in a very minor way, which 

 will be discussed later in these pages in connection with the possible significance of 

 grilse runs of different magnitude. 



The great advantage to be found in the use of year classes for purposes of pre- 

 diction, whether abnormally large or small, lies in the fact that by the time these are 

 sufficiently defined the major part of the hazards that confront every brood are already 

 past, and we have to do with the survivors of a long series of attacks to which the 

 brood has been subject in every stage from the egg to the mature fish. Every hazard 

 means also an uncertainty, and with the passing of the hazards the uncertainties grow 

 less and less. The most reliable evidence concerning the final condition of the brood 

 is that obtained from the years immediately preceding maturity, while the earlier 

 years give progressively less that is authoritative. Still less significant than the 

 fry is the number of eggs produced during the season responsible for the brood; and 

 of still less value, the number of spawning pairs. We believe, however, that all of 

 these, when quantitatively known, have significance and value, though in varying 

 degree, and it is this that justifies us in undertaking investigations concerning the 

 relation that will be found to exist between the number of spawning parents, at one 

 end of the chain, and the final number of their mature progeny, at the other. We 

 consider it probable that we shall discover a significant correlation. 



We have no doubt, however, that if we could substitute for the number of spawn- 

 ing fish the number of resulting fry emerging from the gravels, or, of still higher 

 value, the number of fingerlings of the brood that accomplish their seaward migra- 

 tion, we would be in possession of data of far greater predictive value. We have 



