310 
BULLETIN OF THE BUREAU OF FISHERIES 
33. In each cisco population the amount of growth made by fish in the same year 
of fife but in different calendar years was found to vary considerably. There is no 
apparent correlation between these variations in the amount of growth in different 
calendar years and annual variations in the average air temperature during the 
growing season. There is some indication that annual fluctuations in the density of 
each individual population may affect the amount of growth made in different calendar 
years, but with the possible exception of Clear Lake the data in support of such a rela- 
tionship are not convincing. 
34. The 1928 year class of Trout Lake and the 1926 and 1928 year classes of 
Muskellunge Lake show a bimodal distribution of the calculated lengths at the end 
of the first year of life. It was suggested that the reason for these bimodal distribu- 
tions lies in exceptional weather conditions that bring about two hatchings in the 
spring. However, the effectiveness of weather conditions in bringing about two 
hatchings appears to depend on local conditions within each lake. 
35. The 1928 year class in Muskellunge Lake which has a bimodal distribution 
of the calculated lengths at the end of the first year of life, and the 1929 Muskellunge 
Lake year class, which has an unimodal distribution of the calculated lengths at the 
end of the first year of life, were used as the basis of a stud}^ of the effect of the dis- 
tribution of lengths at an early stage in the life history on the phenomenon of growth 
compensation in later years. It was found that growth compensation is much more 
intense in the group with the greater dispersion of the length distribution at the end 
of the first growing season. In both groups the larger fish at the end of the first year 
of life tend to be the larger fish at the end of the second year of life and at the time of 
capture in the third or fourth summer. Growth compensation in later years reduces 
the advantage of the large first year fish but fails to eliminate it completely. 
36. The fish associations of which the ciscoes are part vary from lake to lake. In 
midsummer the cisco is taken in Trout Lake only with the typical deep water forms, 
the whitefish, the lake trout, and the burbot; in Muskellunge Lake large numbers of 
perch and other shallow-water fish are taken with the cisco ; in Silver Lake the cisco 
spends the summer in practical isolation; and in Clear Lake the pike-perch is relatively 
abundant in the strata inhabited by the cisco. 
37. Because of the scattered nature of the available published data, a review 
was made of earlier investigations concerning the selective action of gill nets. 
38. The examination of the size and age composition of the gill-net samples 
showed that the action of a net or group of nets in taking samples of a cisco population 
cannot be predicted on the basis of experiences with other populations of the same 
species, but that each population presents its own problem of gill-net selectivity. 
Although gill nets are very selective in their operation, the judicious elimination of 
selected age groups can make gill-net samples highly reliable for the determination of 
growth curves. Extreme caution must be employed, however, in the use of gill- 
net samples for the study of age and year class composition. 
BIBLIOGRAPHY 
Birge, Edward A., and Chancey Juday. 1911. The inland lakes of Wisconsin. The dissolved 
gases of the water and their biological significance. Wisconsin Geol. and Nat. Hist. Sur., 
Bull. XXII, Scientific Series No. 7, 1911. xx+259 pp. Madison. 
Bjerkan, P. 1917. Age, maturity, and quality of North Sea herrings during the years 1910-13. 
Report, Norwegian Fish, and Mar. Inves., vol. Ill, no. 1, 1917, 119 pp. Bergen. 
Blegvad, H. 1933. Plaice transplantation. Jour, du Cons., Cons. Perm. Internat. Explor. Mer, 
vol. VIII, no. 2, 1933, pp. 161-180. Copenhague. 
