274 
BULLETIN OF THE BUREAU OF FISHERIES 
years will give too low an estimate of the goodness of growth in these years as compared 
with years not involved in this particular comparison. The possible distorting effect 
of examples such as those just given is obviated in large measure, however, by the 
fact that the evaluation of the goodness of growth of a single year involves compari- 
sons with so many other calendar years that the effect of a single distorted comparison 
will tend to be greatly reduced. 
In Silver Lake (table 52), for example, the evaluation of the year 1927 is based 
on comparisons that involve all other years from 1924 to 1930, inclusive. At the 
best, however, the evaluation of growth in different calendar years shows only their 
relative positions, and a general shifting of values might be expected to follow the 
Figure 6. — Relationship in different calendar years between the deviation of the growth of the Trout Lake cisco from the average 
and the deviation of the air temperature from the normal during the cisco’s growing season. Deviation from the average growth, 
broken line; deviation from normal temperature, solid line. 
elimination or addition of another calendar year in the comparisons. The method 
of comparing goodness of growth in Trout, Muskellunge, and Silver Lakes in different 
calendar years has the advantage that all the single comparisons are based on fish 
of the same age, and that any possibility of a distortion through Lee’s phenomenon 
is thereby eliminated. 
The fact that collections of the Clear Lake cisco were made in only 2 consecutive 
calendar years makes necessary a slightly different procedure in the comparison of 
growth in that lake in different calendar years. In Clear Lake the evaluation of 
the goodness of growth by the method employed in the other three populations would 
mean that the estimate of the relative position of a single calendar year would be 
based not on comparisons with several other calendar years but on comparisons with 
only the single years immediately preceding and immediately following the year in 
