unshocked and once-shocked fish in each of 

 the two blocks: (1) positive fields moving 

 to the north, and (2) positive fields mov- 

 ing to the south. The assumption that the 

 relative effect of each combination would 

 be the same in each block, even though the 

 tests in one block would be performed later 

 than those in the other, was tested by 

 graphically comparing the trends of the 

 effects between blocks and by analyzing 

 interaction effects. 



If the proportion of fish that entered 

 the north end with power off had changed 

 within a block, then the estimates of rela- 

 tive effects of test conditions run at 

 dJi'ferent times in the block would have been 

 biased. To avoid such a bias the order of 

 testing each combination was random within 

 each block; a new random order was used for 

 each series with unshocked and once-shocked 

 fish. 



While designing the experiment, we 

 surmised that a difference between north 

 and south blocks of tests would represent 

 a possibility of the following effects: 

 (1) a preference by the fish for the north 

 or south end; (2) differences in the pattern 

 of electrical current flow associated with 

 the reversal of direction of the electrical 

 fields; and (3) changes in the experimental 

 environment, such as the water temperature 

 chzmges which did occur between blocks. 

 Furthermore, if differences between blocks 

 due to the pattern of electrical current 

 flow were assumed to be negligible, and if 

 the effect of temperature also were shown 

 to be insignificant, then a pronounced dif- 

 ference between blocks could be ascribed to 

 a preference of the fish for one end of the 

 laboratory tank. During the experiment, we 

 did assume the differences between blocks 

 due to the pattern of electrical current 

 flow to be negligible, and the effect of 

 temperature was insignificant. Although we 

 do not know the cause for the preference, 

 we conclude that the difference between 

 blocks was due to a preference of the fish 

 for one end of the laboratory tank. 



The experiment was, therefore, designed 

 as a 3-^ factorial with the following values 

 of electrical variable factors: potential, 

 60-75-90; pulse frequencies, 2-5-8; and 

 pulse durations, 10-20-30. Eight series of 

 tests (4 on unshocked, and 4 on once-shocked 

 fish) of these 27 combinations were run at 

 random. Four series (2 on unshocked, and 2 

 on once-shocked fish) were tested in each of 



the two experimental blocks: (1) positive 

 fields moving to the north, and (2) posi- 

 tive fields moving to the south. 



Source of Fish 



The adult squawfish ( Ptychocheilus 

 oregonensis ) used in these experiments were 

 transported in an aerated tank from U. S. 

 Fish and Wildlife Service hatcheries at 

 Little White Salmon 2ind Leavenworth, Wash- 

 ington, and from the Bonneville hatchery 

 of the Oregon Fish Commission at Bonneville 

 Dam, Oregon. The fish had been held in 

 outdoor rearing ponds until sufficient num- 

 bers had been obtained for transportation 



NEGATIVE ZONE 



I I I 



bfof' of Sequence 



POSITIVE ZONE 



opproK imofely 17 apart 



(-) 

 (+) 



(-) 



(+) 



Figure 1. — Laboratory experimental area with 

 staggered electrode array. Arrows indicate 

 start of sequence of electrical fields mov- 

 ing toward north end of laboratory tank. 



