316 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



no visible evidence of excessive mortality was 

 observed by me. Frequently some fish were 

 weakened by handling and swam near the surface 

 in the live box. When released, these individuals 

 remained visible from the surface and became 

 targets for osprey after the boat left the area for 

 the next station. Such fish, visibly weakened, 

 were never marked. In 2 years, only one dead 

 marked fish was found. 



A third assumption is that recruitment is negli- 

 gible during the time recoveries are being made. 

 Recruitment contributed both trout and grayling 

 to the 1953 population, biasing estimates upward. 

 Although all grayling which were sampled during 

 the first year were at least in their second summer 

 of life, the smallest 2-year-olds (less than 4 inches) 

 undoubtedly passed through the finest mesh 

 (1 ^-inches), at least for the first part of the recap- 

 ture period. The 1953 nettings captured no 

 grayling between 7.4 inches (largest 2-year-olds) 

 and 8.8 inches (smallest individual more than 2 

 years old) . To minimize the effects of recruitment 

 in 1954 only grayling more than 8.4 inches were 

 marked. 



Trout presented a different problem. In 1953, 

 trout as small as 3.3 inches were marked, however, 

 only 7 fish less than 5 inches were measured and 

 these few trout captures were all that recruitment 

 might have contributed to the marked population. 

 This error was negligible. In 1954, although 

 hybrids as small as 5 inches were to be marked, 

 no trout less than 5.7 inches in total length were 

 captured by the seine, or less than 6.1 inches by 

 the nets. Since the 1954 gear would retain 5-inch 

 hybrids, recruitment only slightly biased the 1 954 

 estimates. 



A fourth assumption is that no mortality, either 

 natural or by fishing, acts to reduce the population 

 during the recapture period. Grebe Lake was 

 subject to light fishing pressure both years during 

 the population study. Approximately 301 hybrids 

 and 343 grayling were caught by anglers in 1953 

 and 275 trout and 472 grayling in 1954 during 

 the marking period. Only four fin-clipped fish 

 were reported by anglers in the 2 years. That 

 some marks escaped the notice of fishermen, is 

 probable, since of 290 fish caught by angling and 

 marked by me in 1953 during the study, 10 of 

 these were recaptures. Thus more than 3 percent 

 of my captures had been marked, but only 0.3 

 percent were reported by anglers as fin-clipped. 



The result of mortality during this experiment 

 undoubtedly increased estimates. 



The fifth assumption is that there is no differ- 

 ential availability of marked and unmarked fish 

 (that marked fish did not become "trap-shy" as 

 a result of handling). A means of checking the 

 number of recaptures by the net would be to 

 sample the population by a means other than trap 

 netting and compare the percentage of recaptures 

 by the new method to results obtained from trap 

 netting. Such samples were available in 1953 as 

 personal angling, and during 1954 as seining. I 

 was unable to demonstrate any significant differ- 

 ence between 1953 recaptures by nets and recap- 

 tures by fishing of either hybrid trout or grayling. 

 In 1954, the ratio of marked to unmarked fish 

 was greater in the seine than in the trap nets. 

 A chi-square test indicated a significant difference 

 at the 95 percent level between grayling recap- 

 tured in nets and recaptures by seining. Such 

 evidence suggests that marked grayling avoid 

 trap nets more frequently than unmarked indi- 

 viduals. The result of such differential avail 

 ability in grayling was to increase the population 

 estimate. Trout recaptures by the seine and by 

 nets were not significant at the 95-percent level. 



Of the estimated populations, 16 percent of the 

 trout were captured in 1953 but only 4 percent 

 of the grayling. Thus there is evidence that the 

 trout are more susceptible to trap netting than 

 are the grayling. 



An additional error appeared in the 1953 

 estimate when the mean length was different for 

 fish captured by the different types of gear (two 

 small-mesh hoop fykes, two trap nets, and one 

 large-mesh hoop fyke) (fig. 7). An analysis of 

 variance (table 8) of these 1953 data indicated 

 that this difference in mean lengths was significant 

 at the 95 percent confidence limits. The various 

 nets therefore sampled different parts of the 

 population, and each did so with its own effec- 

 tiveness. It was for this reason that four trap 

 nets of a single design were used in 1954, but 

 unfortunately all bias was not overcome because 

 the two newer nets were preserved with a light- 

 colored copper naphthanate and had a different 

 mesh size in the wings. 



An analysis of variance (table 9) showed that 

 in 1954 the mean sizes of fish from the different 

 trap nets were significantly different from one 

 another and also from the seined sample. This 



