CENTRAL AND WESTERN ALASKA INVESTIGATION. 155 



The mere statement of the case disproves it. Even the tributaries 

 of one river vary wideh^ year by year in the proportion of the total 

 run which enters each of them. If tnis be true of the tributaries of a 

 single river, how much more probable of four separate rivers, which 

 join only at their mouths. It is certain that they would varv" inde- 



gendently and that the oscillation might be of large dimensions, 

 ►ue to manner of fishing, one of these rivers might experience a 

 progressive reduction of its run that was not felt by the others. 

 Prior to 1908, the Wood River fish ran the same gantlet in the outer 

 bay as did those bound for the other rivers, and in addition were 

 subjected to further reduction b}^ traps and gill nets operating the 

 entire length of Wood River. The main Nushagak and the Snake 

 Rivers, at least, were wholly free from tnis further drain; tneir propor- 

 tion of escape was demonstrably liighcr, and their nins should have 

 fared better. In this complicated case, then, not only the natural 

 oscillations in the nins to the difl'erent rivers, which miglit be at any 

 given time in opposite directions, but also the possibility of pro- 

 gressive changes in the run of any of them, due to its different histoiy 

 must be contended with. If tlie Wood River run should for a term 

 of years diminish relatively to the others, its escapement would 

 diminish relatively to the escapement to the other three streams. 



For these reasons, there has not been an attempt to establish a 

 relation between the size of the Wood River escapements and the 

 size of tlie resulting runs of salmon to the Nushagak ^\ith any high 

 degree of expectation. There are too many unknow n factors entering 

 into the equation. Only on the assumption that the Wood River 

 run so far overshadows the sum of all the others that the latter may 

 be considered negligiiile, is the expectation warranted that close posi- 

 tive results can be achieved. 



It is noted at the outset tliat the recorded escapement from 1908 

 to 1912 showed an alarming progressive decrease both in actual 

 numbers of fish and in percentage of escape, but the five-year period 

 that foUows gives scant evidence of correspondingly decreased runs. 

 It is also noted that the largest escapement l)y far — that of 1908 — 

 was responsil)le for tlie four-year fish of 1912, and the latter was one 

 of the vers' poorest runs within the 10-year period under investiga- 

 tion. The 1908 s})awning escape was recorded as 1,600,000 hsli. 

 The year 1912 had the ver\- low record of 325,000 to reach the spawn- 

 ing beds, yet it produced tlie five-year contingent of the nin of 1917, 

 which furnished the largest nin of any year since 1908. Tliese are 

 glaring failures, and indicate clearly enougli that no such close 

 relation exists between spawning escape on Wood River and the 

 Nushagak run as will warrant predictions regarding the latter. 



But if search is made for correspondonf'cs, which have a high degree 

 of probability in their favor, such can be found. The years 1911 and 

 1912 were jointly res])onsible for 1916, when the total recorded run 

 was the next to the smallest during the period of 10 years. The 

 os<apes both in 1911 and in 1912 were far below the average; in fact, 

 with one exception they wore the two poorest recorded i>scapes in the 

 10-vear period. Whatever mav \hi thous^ht of the higher escapes, it 

 looks as though the 325,000 of 1912 and the 354,000 of 1911 were 

 sufFKiently below an acceptable minimum to make a decided im])res- 

 sion on the total run to Nushagak Bay. The ])ossibilitv of a chance 

 coin' idcnco can not bo eliminated here, and there is no similar case 

 with which to check up. No other instance is recorded in the series 



