250 
BULLETIN OF THE BUREAU OF FISHERIES 
1928 growth from the scale measurements of each of these later collections of the 
year class. The 1930 IV-group samples indicated that the 1926 year class had a 
total growth of 19 millimeters in 1928, while the 1931 Y group gave an average cal- 
culated growth of 17 millimeters for the same year. The combination of the 1930 
and 1931 data gives a calculated total growth of 18 millimeters for the 1926 year 
class in the calendar year 1928. Since the 1928 specimens themselves had grown 
17 millimeters at the time of their capture in late July it is apparent that the 1928 
II group had practically completed its season’s growth at the time of capture of the 
samples. 
It is obvious that comparisons of the sort given above, to be most dependable, 
should be based on different years’ collections of the same year class. By such a 
procedure errors that might arise from unequal growth of the same age fish in differ- 
ent calendar years are completely avoided. The confining of comparisons to differ- 
ent years’ collections of the same year class of course limits this type of analysis to 
those year classes that are well represented in more than 1 year’s collection. How- 
ever, for the question at hand dependability is more important than a diversity of 
material. 
Since data of the above type throw light only on the probable time at which 
growth ends, the presentation of the available data on the amount of the season’s 
growth already made by the different populations at the time of capture of the samples 
should be preceded by a brief consideration of probable time at which the season’s 
growth begins. Juday and Birge (1930) stated that in the lakes of this region “The 
ice disappears about the 1st of May; the shallower lakes usually lose their covering 
of ice a few days before the larger and deeper lakes.” From these facts it may be 
considered valid to assume that in general growth does not occur in any of the four 
populations of this study before early May, and that the time at which growth 
begins does not differ by more than 1 or 2 weeks from one lake to another. 
Table 35 shows a comparison of the amount of calculated growth made by three 
age groups of the Trout Lake cisco between the beginning of the growing season and 
the time of capture of the samples, with the total calculated growth for the season 
as determined from later years, collections of the same year class. The data for the 
1928 II group which were used to illustrate the method of comparison have already 
been discussed. It may be seen at once that the 1930 III and IV groups agree with 
the 1928 II group in indicating that the season’s growth of the Trout Lake cisco is 
practically if not entirely complete by the end of July. The growing season for the 
Trout Lake cisco is not longer than 3 months and, since growth may not start at once 
upon the disappearance of the ice, may actually be somewhat less than 3 months. 
Table 35. — Calculated growth of 3 age groups of the Trout Lake cisco up to the time of capture in the 
summer, compared with the growth for the entire summer calculated from collections of the same 
year class in later years 
[Number of specimens in parentheses] 
Year 
class 
Date of capture 
Age at 
time of 
capture 
Growth in 
summer 
up to time 
of capture 
Growth in entire sum- 
mer based on later col- 
lections of 
1930 
1931 
1926 
July 21, 28, 29, 1928 _ 
II 
17 (102) 
19 (99) 
17 (79) 
1926 
July 29, 30, 31, 1930 
IV 
9 (99) 
9 (79) 
1927 
Do'.— ..1 
III 
10 (347) 
11 (269) 
