OlilKXT.VnuX OV .MltiUATIXG ANADliOMOU.S MSHES 



381 



Figure. 4. — Entrance to experimental trough. 



A. Entrance gate (closed). 



B. Retaining pool. 



C. Retaining-iX)ol gate (open). 



D. Bypass for downstream migrants. 



(The white cloth on the bottom of the trough was for 

 photographic purposes and was not present during the 

 tests. ^ 



its passage tluuuoli tlie i-lianiiel aiul was entirely 

 clear of the troiigli. This system of recordino: 

 decisions permitted the fish to change its choice 

 of channels ut any stage of its progress by turning 

 back and entering the other channel. 



During tlie exi)erinients in which tlie water was 

 being modified by heating, the rehutance of the 

 fish to pass through the heating apparatus in the 

 upstream ends of the channels, resulted in con- 

 siderable delay. To save time a "recording 

 point" (fig. 2) was chosen below the heating ap- 

 paratus. When this arbitrary point which was 

 -20 inches (approxinuitely twice the length of the 

 iish) from the entrance to the channel was 



l'ii.ri;i: .".— Dm r.-imc jcali'. The fish were allowed to 

 enter, imi- al ii liiiH'. lliriiu;;h .-i s<-rei'ii gate (sliown 

 open ) . 

 I Thf whiti' ilcitli was imt present during tile tests.) 



ie;!ched, the lisli was consiih'red to lia\'e made its 

 decision and the result was recordeil. When once 

 the fish had passed the recording point if its for- 

 ward progress was too slow, it was urged to con- 

 linue on out of the trough by taj)s on the trough 

 with a stick immediately behind the fish. AVhen 

 the fish was completely out of the trough, the 

 entrance gate was raised again to allow another 

 to enter. 



Before each series of tests, any necessary ad- 

 justments of experimental conditions were made 

 and water samples and temperature measurements 

 were taken. The fish were then allowed to enter 

 the trough one at a time. It usually was found 

 convenient to run approximately '25 to 30 indi- 

 vidual tests in succession before the measurements 

 were taken again. The average time for such a 

 .series of tests was about 40 minutes, although it 

 fluctuated considerably depending upon the be- 

 luivioi- of the fish. 



CONTROLS 



Kvery I'tl'ort was niadc to kee[i cnnditions, e.g., 

 light, depth of water, rate of How, turbulence, as 

 uniform as possible in both chamiels so that any 

 directional response would be due solely to the 

 factor being tested. Whenever major adjust- 

 ments in the ex])erimeiital conditions had been 

 made, before tests with modified water were be- 

 gun, a series of control tests was run to ensure 



i;iui;i7- 



