AGE AND GROWTH OF THE CISCO 
279 
the growing season should not be taken to show that in these four cisco populations 
the amount of growth in a particular summer does not depend in part on the tempera- 
ture of the lakes’ waters during the growing seasons of the several stocks. It must be 
recognized first that fluctuations in air temperature may not offer a perfect index of 
fluctuations in water temperature. A more probable explanation of the failure of 
annual fluctuations in temperature and in growth to show correspondence is that 
annual differences in growth are so closely dependent on some other factor or factors 
that the effect of annual differences in temperature are almost completely obscured. 
There is reason to believe that in the populations of this investigation the annual 
fluctuations in the goodness of growth may show some dependence on annual fluctua- 
tions in the densities of the different populations. 
It has been shown previously (p. 262) that the growth rates (in weight) of 
these four populations follow the inverse order of the relative densities of their 
populations. In view of this demonstrated relationship between growth rate and 
density of population it might well be expected that within a single population 
changes in the density of the cisco may be accompanied by changes in its growth rate. 
In the study of the age composition of the cisco samples from Trout Lake (table 
45, p. 264) it was pointed out that the years 1926 and 1927 saw the production of 
abundant year classes, and that as the result of the abundance of these two groups 
the 1930 Trout Lake cisco samples were on the average composed of older fish than 
the 1928 samples while the 1931 samples were composed of older fish than either the 
1928 or 1930 collections. This accumulation of old fish together with the contribu- 
tions of young in the years later than 1927 probably caused the density of the popu- 
lation to be rather high in 1930. This increase in the density of the population may 
account in part for the drop in growth from 1929 to 1930. 
The data for the Muskellunge Lake cisco also afford evidence for a dependence 
of growth variations on variations in population density. While the 1928 and 1930 
collections contained only a few individuals more than 2 years old, a large number 
of the 1928 year class individuals were present as the III group in 1931 . (See p. 264.) 
The great abundance of the III group in 1931 suggests that an accumulation of the 
stock may have been occurring in Muskellunge Lake in 1929 and 1930. This accumu- 
lation may possibly account for the drop in the amount of growth from 1929 to 1930. 
The age and year class composition data for the Silver Lake cisco (table 47, 
p. 265) show that here as in Trout Lake there was probably an accumulation of the 
stock in 1930. In the examination of the data of table 47 it should be remembered 
that the large number of 1931 I-group fish causes the representation of the older age 
groups of the 1931 samples to appear relatively too low; the gear used in 1928 and 1930 
captured few I-group individuals. If due consideration is given to the absence of 
the I-group fish in the 1928 and 1930 collections the relative abundance of the older 
age groups in the 1931 Silver Lake collections must be recognized to be higher than 
in 1930, while the 1930 collection in turn shows a greater abundance of old fish than 
the 1928 collection. In Silver Lake as in Trout Lake and Muskellunge Lake the 
accumulation of older fish was accompanied by a decrease in growth from 1929 to 
1930. 
It should be mentioned that the data upon which a suggested explanation of 
the drop in growth from 1929 to 1930 in Trout, Silver, and Muskellunge Lakes was 
based fail to offer any logical explanation for the improvement of growth from 1928 
