222 
BULLETIN OF THE BUREAU OF FISHERIES 
From an examination of the tables for Trout, Muskellunge, and Clear Lakes it 
will be noticed that “Lee’s phenomenon” of “apparent decrease in the calculated 
growth” as it is determined from successively older groups of individuals, wnile pos- 
sibly present, is not a source of any great discrepancies, especially where large numbers 
of individuals are involved. On the whole the calculated growths based on different 
age groups in the same year’s collection or upon samples of the same year class taken 
at different ages, agree satisfactorily for corresponding years of life, with the exception 
of the large calculated growths obtained for the younger age groups (marked with 
asterisks in the tables) in the Trout Lake and Muskellunge Lake collections. These 
large growths can be explained as the result of the selective action of the gill nets used 
for their capture. (See section on gill net selectivity.) The small but fairly consistent 
discrepancies that appear in the corresponding calculated growths of older year groups 
of the Trout Lake fish may be partly the result of slight changes in the body scale 
ratio. It is probable that selective action of gear plays a role here also. At any rate 
the discrepancies are of such small magnitude that they affect but little any conclusions 
that might be drawn concerning the growth of the population as a whole. 
LEE’S PHENOMENON IN THE SILVER LAKE CISCO 
In the Silver Lake data (table 5) there appear much more pronounced disagree- 
ments among the corresponding calculated growths of different age groups than are 
present in the other three populations. Not only are these discrepancies relatively 
large, but the comparison of any two age groups shows that in general they tend to be 
cumulative with increased age. These discrepancies appear in the comparisons of 
calculated growths based on the same year class but taken in different seasons as well 
as in calculated growths based on the different age groups of a single year’s collection. 
To illustrate these two points there may be examined (table 5) first the calculated 
growths of the III, V, and VI groups of the 1925 year class and second the calculated 
growths of the III, IV, V, and VI groups of the 1931 collection. With certain ex- 
ceptions in the first year of life it will be noticed that each successively older age group 
tends to give smaller values for the calculated lengths for corresponding years of life. 
Thus it may be seen that the observed discrepancies in the calculated growths depend 
on the differences in the age of the groups studied. 
It will be noticed further that the apparent change of growth rate as it appears in 
the fish older than the II group differs from Lee’s phenomenon as it ordinarily occurs, 
in that it is the calculated growths of the later rather than the earlier years of life 
that are affected most. 
The calculated lengths of the Silver Lake ciscoes at the end of the first year of life 
require some special consideration. If size of sample is taken into account, the cal- 
culated growths for the first year of life show fairly good agreement for all age groups 
older than II. The calculated growths for the first year of life as based on the II 
group are somewhat less than those based on the I group, but those of both groups are 
considerably greater than the growths based on the older-age groups. The high cal- 
culated growths for the first year of life as based on the I group can be explained in 
part by the selective action of the gill nets used in collecting the samples, for there is 
reason to believe that even the smallest mesh of these nets may have taken only the 
larger individuals of this group. Such an explanation, however, does not seem to hold 
for the fish of the II group, since on the basis of comparisons with samples from other 
lakes the range of size of the Silver Lake II group seems to be such that very little 
