SALMON ANlD TROUT IK ALASKA. 21 



would thus have been a part of the salt-water approaches. Under 

 the circumstances, however, it was necessary to operate the trap at 

 the rapids below the lagoon, where it was subject to the rise and fall 

 of tide. This necessity may to a certain extent have altered results. 



METHOD USED TO DETERMINE MIGRATORY MOVEMENTS. 



The apparatus used for taking the migrating young was a sort of 

 fyke trap made from a regulation collecting seine used by the Bureau, 

 and was similar in principle of operation to the gear used by Rutter 

 in the California investigation and by Babcock in the Eraser (p. 30). 

 The conversion of the seine into a trap was accomplished by sewing on 

 a bag of coarse bobbinet to surround and extend beyond the netting 

 bag of the seine. This latter was then opened and a small funnel of 

 bobbinet laced to the opening. The large bobbinet bag was closed 

 at the hinder end by tying mth a cord, so that it could be readily 

 opened and the contents shaken into a bucket. With a short seine 

 pole and bridle at either end the contrivance was then ready to hang 

 across the current for operation. For successful results the current 

 must be sufficiently rapid to keep the bag distended — otherwise the 

 fish will not enter it — yet not strong enough to burst the fabric. In 

 water too swift the smaller fish are pressed against the web and dro\Mied. 

 The rise and fall of the tide at station 2 in the Naha vitiated the sta- 

 tistics to a degree. With such a contrivance careful weight must be 

 given to the state of the water. Muddy, swift water as at flood time 

 always shows an increased catch, due doubtless in part to the favor- 

 able influence on the operation of the apparatus as well as to increased 

 movement of the fish. 



This apparatus as ordinarily operated is very defective. Estimates 

 of numbers of migrating fish based upon its results can be only very 

 distantly approximate. In no case can the entire current of a large 

 stream be occupied, and only in exceptional cases will the net at all 

 times reach from top to bottom. In the Naha and Sacramento tests, 

 the lead line was frequently some distance from the bottom ; in the 

 Fraser River inquiry, however, it appears that the net was set on the 

 bottom. Only by repeated tests can it be determined on which side 

 of the stream the net will be most effective. On the Sacramento I 

 made a few tests as to its efficacy at the side and center of the chan- 

 nel, respectively, but results were not definite, since other factors 

 necessarily entered into the question. A complete experiment would 

 require several of these traps to be operated simultaneously in dift'er- 

 ent parts of the stream. In any current where fine-meshed fabric 

 can hold, moreover, the stronger fish can breast the stream and escape. 

 This defect of the device used is called to attention by Mr. Babcock 

 in his report for British Columbia. The objections stand only against 

 estimates of total number of miirrants based on the catch made. 



