CENTRAL, AND WESTERN ALASKA INVESTIGATION. 153 



average for the fish captured in the outer bay. Where thus selected 

 for their small size there will be more four than five year fish and more 

 females than males. The latter may present an added source of danger, 

 for if in the escape females largely predominate they may be unable to 

 find males with which to pair, and their eggs may remain sterile. 

 There are no observations on the spawning grounds which cover this 

 point. It is possible that the male remains active for a longer period 

 and will mate repeatedly, but it is not known that such is the case. 



The escape through nets must present a fairly constant factor, but 

 the amount of escape attendant on unfavorable fishing weather varies 

 widely with the season. Some years are much more conducive to 

 close gleaning than others. In 1915 the commercial capture was 

 above the average, amounting to more than 5,500,000 fish; but the 

 escape of less than 260,000 fish was the smallest then reported. The 

 cannery pack of 1917 was practically the same as in 1915, but the 

 escapement in 1917 was four times as great; in 1916 the escapement 

 was nearly twice as great, although the commercial yield in 1916 was 

 only about three-fifths that of 1915. 



One reason is here apparent why such indifi"erent success is met 

 with in predicting future runs on the basis of commercial packs of 

 salmon. It is the number that escape to the spawning beds that 

 have their influence on a subsequent generation, not the number 

 sealed in cans, and the latter would often give a ver}- erroneous basis 

 for estimating the former. 



It would seem, however, that where the escapement itself has been re- 

 liably ascertained for a number of years, a basis should be at hand for 

 successful prophecy, with a fair degree oif accuracy. The belief is gen- 

 erally held that increase in number of spawning fish up to the lull capac- 

 ity of the spawning beds will result in an increased run. The s])awning 

 beds of the Wood Uiver chain of lakes to which there must have resort- 

 ed in the old days 5,000,000 fish or more, can not be held fully populated 

 by any number that have escaped the nets during recent years. A 

 million fish at least should certainly find room to spawn in this great 

 watershed without serious intci-ference with one another's nests. 



If they can do so, then the larger the number that spawn within 

 those limits, or any more extensive areas of profital^le spawning, the 

 greater should be the resulting schools of fingerlings that pass out to 

 sea in the spring and summer and the larger the returns in adult 

 salmon wlien these come back to spawn. 



If it should be assumed, however, that only half a million or a (piarter 

 of a million could profitably spawn, and that any larger number merely 

 increased the wastage on the spawning beds and contributed nothing 

 to the crop of fingerlings, on this basis and this basis only could a 

 complete failure to establish a relation between increased spawning 

 escapements and the larger size of the resulting runs be understood. 



The Wood River experiment was inaugurated for the purpose of 

 throwing light on this and on other problems, among them some of 

 the most important that confront the commercial fisheries and fish 

 propagation. If the experiment could establish the percentage of 

 returns which could be expected under natural j)r()pagation from a 

 given number of spawning fish; or, stated difiercntly, the lowest 

 percentage of escape that could be relied on to maintain the nin 

 mtact, a sound basis would be laid lor scientific handling of the 

 fisheries. The results, both scientific and j)ractical, woulcl be far- 

 reaching, and the conduct of the experiment all the more should bo 



