282 
BULLETIN OF THE BUREAU OF FISHERIES 
This bimodality of the length frequencies of the calculated growth for the first 
year of life may be found in the 1928 year class from Trout Lake and in the 1926 and 
1928 year classes from Muskellunge Lake. All the other year classes from these 
two lakes, and all year classes from Silver Lake and Clear Lake, where the samples 
are large enough to give reliable results show unimodal distributions for the calculated 
growth of the first year of life. The fact that bimodality in the first year’s growth 
occurs only in two populations and in only one or two year classes of these populations 
suggests that the occurrence of the phenomenon depends on the nature of the local 
conditions within each lake and that these conditions vary from year to year. 
The most reasonable explanation for bimodality in the first year’s growth lies 
in the assumption that in certain years there are two hatchings rather than a single 
one. In early spring a period of warm weather with brilliant sunshine and no strong 
winds can warm the waters of the shallow littoral region to a temperature several 
degrees above that of the main body of the lake. At such a time the development 
of the eggs of the cisco would be accelerated and some might hatch. If, however, 
there occurred before the completion of hatching season a period of cold, windy 
weather the temperature of the water of the littoral region would undergo a sudden 
drop of several degrees. 30 The development of the unhatched eggs would be retarded, 
and their hatching might be delayed for several days or even weeks. Such a situation 
would explain the observed cases of bimodality in the amount of growth during the 
first year of life. Eggs that develop in more exposed regions of a lake or in a lake 
more swept by winds would be less affected by fluctuations in weather conditions. 
GROWTH COMPENSATION 
The phenomenon of “growth compensation” — the tendency for individuals that 
grow relatively slowly in the early years of life to grow relatively rapidly during the 
later years — has been observed by numerous investigators and in several species of 
fish. The only study of growth compensation in the cisco was made by Van Oosten 
(1929) on the Saginaw Bay herring of Lake Huron. He concluded that “the large 
fish of an age group were the large fish in each preceding year of life * * * but 
that the differences between the small and large yearlings diminished each year of 
age- — that is, the small yearlings were rapid growers and the large yearlings slow 
growers.” Thus he found that compensation did occur, but that it was not sufficient 
to overcome completely any advantage in length which a large individual might 
hold at the end of the first year of life. 
A comparison of growth compensation in age groups with unimodal and bimodal 
length distributions at the end of the first year of life should yield information as to 
the effect of the dispersion of the length frequency in early life upon the manner of 
growth in later life. In the 1931 collection from Muskellunge Lake (table 58) the 
distribution of the calculated lengths at the end of the first year is unimodal in the 
II group and bimodal in the III group. Both age groups are represented by large 
samples (258 for the II group and 347 for the III group). For these reasons they 
were selected as the basis for a study of growth compensation. 
Table 58 shows the frequency distributions of the calculated lengths for both age 
groups at the end of the first year of life. The calculated length distributions of the 
3° Forel (1892) pointed out that the development of high temperatures in the littoral region depends on days of great calm and 
brilliant sunshine, and that this warming process proceeds slowly. The negative changes in temperature proceed much more 
rapidly. Forel observed that with a strong south wind the temperature of the littoral waters at the port of Geneva at times dropped 
as much as 6° C. to 8° C. or more from one day to another. 
