298 BULLETIN OF THE BUREAU OF FISHERIES 
SELECTIVE ACTION OF GILL NETS USED IN COLLECTING SAMPLES OF CISCOES FROM 
TROUT, MUSKELLUNGE, SILVER, AND CLEAR LAKES 
In this investigation, although other phases of selectivity will be considered 
briefly, the study of the selective action of gill nets will be concerned primarily with 
the effect of selection upon the determination of growth in the various populations. 
The analysis will involve particularly the comparison of average lengths as deter- 
mined for a single age group from samples taken in nets of different sizes of mesh. 
The analyses will show the basis for the elimination of certain selected groups in the 
computation of general growth curves (see footnote, p. 226) and will furnish a better 
idea of the validity of the use of gill-net samples in the study of growth. Such analysis 
is of especial importance here since different ranges of mesh sizes were employed in 
taking the different years’ collections. 
Attention should be called to certain criteria that are of general value in deter- 
mining the reliability of gill-net samples for growth study. It may be pointed out 
that in general a sparse representation in a sample of a young age group whose average 
length is near the lower limit of effectiveness of the nets used, is a source of suspicion 
as to the reliability of the sample of that particular group. If this same sparsely 
represented group gives calculated growths that are in serious disagreement with 
those of the older age groups it should be eliminated from the data used for the study 
of growth in the population as a whole. A further check on the reliability of the 
material that represents a particular age group lies in the comparison of the average 
lengths for the group as determined from samples in different sizes of mesh and as 
determined from the different combinations of such samples. The examination of 
the length distributions of the catch of nets of different size mesh may also give an 
idea as to the adequacy of the sampling with respect to particular age groups. 
As was mentioned in the introduction the 1928 data do not include records of 
gear for individual fish. For this reason the 1928 data cannot be used in the study 
of gill-net selectivity. 
For purposes of convenience the gear used in 1928, 1930, and the early part of 
the 1931 season will be designated throughout this section as "old”, 36 and the gear 
used during the latter part of the 1931 season and through the 1932 season will be 
known as “new.” Descriptions of the two groups of nets were included in the 
section on methods. 
Tables 71, 72, 74, 75, 76, and 77 show the effect of size of mesh on the deter- 
mination of average size of the different age groups in the populations of Trout, 
Muskellunge, Silver, and Clear Lakes. Because of the generally small number of 
specimens upon which the average of lengths of the various age groups depend, 
attention must be given to the general trend of the differences that result from 
differences in gear rather than to the actual amount of the difference in any specific 
case. Since each population represents a distinct and separate problem in the study 
of selectivity, the data for each lake will be discussed separately. 
TROUT LAKE 
With the exception of a single specimen all the ciscoes collected in Trout Lake 
in 1930 were obtained from 1 %-inch mesh nets, while the 1928 samples were probably 
taken in nets of 1%-inch (trammel), 1%-inch, l^-inch, and perhaps a few in 1 %- and 
m Although the “old” gear as fished in 1930 and in the early part of 1931 included 5 different sizes of mesh, only the lH-inch 
and 2-inch nets caught ciscoes. All collections from Clear Lake were made with the new gear. The date of change in gear was 
July 22, 1931. 
