the other hand, many of the experiments indicated 

 no mortalities caused by the dam. 



Returns from below the dam by date oj tagging. — 

 The upstream recoveries of tagged sockeye, 

 released below the dam, indicated a fairly uniform 

 pattern of returns for all experiments (figs. 6,7, 

 and 8). As tagging experiments were alternated 

 between left and right hanks on alternate days, 

 these graphs show a fairly consistent pattern of 

 returns for both release areas and for the duration 

 of the tagging season. The percent-recovered 

 curve in 1954 assumed the shape of the numbers- 

 tagged curve, with fewer returns from the tails of 

 the experiments. Percentage returns for 1955 

 and 1956 show a remarkably uniform pattern. 

 The large numbers tagged, shown for July 26 

 and 28 in 1955 and for July 24 in 1956, arc the 

 result of combining left and righi hank experi- 

 ments, as experiments were conducted on both 

 hanks on those dates. Returns of nearly 50 



percent from the August 3, 1955, experiment 

 (fig. 7) are a result of observing a large number 

 of that day's tagged fish passing Tumwater Dam, 

 more fish observed than for any other single ex- 

 periment. It might be that this particular sample 

 contained relatively more Wenatchee fish than the 

 other samples. 



Returns from below dam compared with water 

 How.— In order to determine the. influence of 

 different water levels on fish passage at Rock 

 Island Dam, tag recoveries from releases at each 

 bank below the dam were compared with water 

 (low (figs. 9, 10, and 11). In general, the tag 

 returns from both hanks showed a fairly consis- 

 tent pattern despite large changes in flow during 

 the experiments. 



The returns for 1954 merit attention, since fewer 

 returns are indicated for experiments on August 

 L0-13 during the low river Hows (fig. 9). These 



350 



U.S. FISH AND WILDLIFE SERVICE 



