280 
BULLETIN OF THE BUREAU OF FISHERIES 
to 1929. Unfortunately there are no 1929 collections upon which to base a com- 
parison of the age composition of that year with the years 1928, 1930, and 1931. 
In the Clear Lake cisco the evidence for a dependence of annual fluctuations in the 
amount of growth on annual fluctuations in the density of the population is somewhat 
stronger than in the Trout, Muskellunge, and Silver Lake populations. The years 1923, 
1 924, and 1925(p.265, table 48, fig. 5) were undoubtedly excellent years for the production 
of young. The presence of three successive good hatches must have led to a consider- 
able accumulation of the cisco stock. Corresponding with this accumulation the 
amount of growth per year decreased from 1923 to 1927 (fig. 8). The production of 
young was poor in 1926 and 1927. The occurrence of 2 poor production years com- 
bined with natural mortality could be expected to lessen the crowding of the popula- 
tion. The growth improved from 1927 to 1929, fell back a little in 1930, and im- 
proved again in 1931. 29 
The relationship indicated in the Trout Lake, Muskellunge Lake, Silver Lake 
ciscoes, and especially in the Clear Lake cisco, between annual fluctuations in popula- 
tion density and annual fluctuations in growth is in agreement with Jarvi’s observa- 
tions (1920, 1924) on the kleine Marane ( Coregonus albula). 
The failure of variations in the amount of growth in different calendar years to 
show any close general dependence on either annual variations in temperature or 
annual variations in population density suggests that possibly these variations in 
growth depend closely on both factors, and that the failure of these factors to operate 
in the same direction in the same year tends to obscure the effect of each of them. 
BIMODALITY IN THE CALCULATED GROWTH FOR THE FIRST YEAR OF LIFE 
The examination of the frequency distributions of the calculated growth for the 
first year of life in the best represented age groups (tables 57 to 60) shows that some 
of these distributions have a distinct bimodality, which appears to be characteristic 
for a year class and present regardless of the age of the fish upon which the calculated 
growths were based. 
Table 57. — Trout Lake cisco — Frequency distribution by 5 -millimeter intervals of the calculated 
growth in length during the first year of life 
Year class 
1926 
1927 
1928 
1929 
Length 
II 
IV 
V 
III 
IV 
II 
in 
II 
1928 
1930 
1931 
1930 
1931 
1930 
1931 
1931 
no 
1 
1 
1 
105 
1 
1 
1 
1 
1 
1 
100 
4 
1 
1 
7 
4 
5 
6 
2 
95 
6 
2 
14 
4 
8 
15 
6 
90 
7 
8 
5 
17 
16 
6 
34 
13 
85 
9 
21 
15 
32 
24 
7 
31 
20 
80 
15 
30 
23 
74 
48 
5 
26 
12 
75 
30 
20 
27 
85 
78 
3 
31 
6 
70 
26 
11 
6 
83 
72 
26 
1 
65 
3 
6 
2 
30 
18 
3 
60 
5 
3 
55 
1 
28 The good growth in 1931 may have been in part due to the climatic conditions of that year as well as to the reduced number 
of ciscoes in the stock. The temperature over the period, May to September, inclusive, was only slightly above normal (table 56) 
but the autumn temperatures were exceptionally high (p. 253). It is probable that the warm autumD of 1931 gave the Clear Lake 
cisco a longer growing season in that year than it usually enjoys. 
