118 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



u 



z 



«04 



I 

 O 



^3 



I- 

 O 



u 



I 2 3 4 5 6 



AGE 



Figure 10. — Calculated growth in weight of age groups 

 III, IV, and V of Green Bay lake herring as determined 

 from calculated lengths for all fish taken from pound 

 nets, 1948-52, and the length-weight relation for all fish 

 of these age groups. 



Table 32. — Calculated growth in weight for the Green Bay 

 lake herring of age groups III, IV, and V 



[Calculated from the general length-weight relation, p. 114, and lengths of fish 

 taken in pound nets In 1948-52, table 17, p. 104] 



REPRODUCTION AND EARLY GROWTH 

 SEX COMPOSITION 



As is common among fish, data on sex composi- 

 tion in lake herring populations are highly variable. 

 Some factors that may contribute to variability 

 of sex composition in samples from a population 

 are: 



/. Segregation of the sexes through various 

 periods of the year including segregation resulting 

 from sex differences in age and size at maturity. 



2. Differences in mortality (natural or fishing) 

 between the sexes. 



S. Gear selectivity in relation to sex differ- 

 ences in activity and morphology. 



To evaluate any of tliese factors would be diffi- 

 cult, particularly since they are interrelated and 



some or all of them may affect the sex composition 

 of a sample. 



Reports of various authors on different popula- 

 tions of lake herring (table 3.3) show sex composi- 

 tion, expressed as percentage of females, ranging 

 from 29 percent in Blind Lake (Cooper 1937) to 

 73 percent in Trout Lake (Hile 1936). The Blind 

 Lake collection was made during the spawning 

 period but the paucity of females is not character- 

 istic of spawning fish as may be seen by the sex 

 composition of other samples collected during the 

 spawning period — in Swains Lake (67 percent) 

 and Saginaw Bay (51 percent). 



Six out of 1 1 lake herring populations for which 

 data have been published on the change of sex 

 composition in relation to age (table 33) show a 

 rise in the proportion of females with increase in 

 age (Clear I^ake, Gidl Ijake, Lake of the Woods, 

 Muskelliinge Lake, Swains Lake, and Trout Lake), 

 2 populations show a downward trend (Blind Lake 

 and Saginaw Bay), and 3 exhibit no clear trend 

 (Irondequoit Bay, Lake Nipissing, and Silver 

 Lake). The 6 populations exhibiting an increase 

 in the percentage of females with age were col- 

 lected with gill nets and 1 (Swains Lake) was 

 sampled exclusively during the spawning period. 

 Of the 2 populations with a downward trend, 1 

 was sampled with pound nets (Saginaw Bay) and 

 the other with gill nets (Blind Lake), and both 

 sets of data were based on spawning-run collec- 

 tions. One of the 3 populations showing no trend 

 was sampled with pound nets (Irondequoit Bay) 

 and the other 2 were sampled with gill nets, and 

 all represent samples from more than 1 month and 

 year. It is obvious from these comparisons that 

 the relation of sex composition to age as reported 

 for different stocks is not clearly influenced by 

 collecting gear or sexual activity at time of 

 collection. 



Some information on possible sources of bias 

 in determining the sex composition of a population 

 is brought out in the Green Bay data on fluctua- 

 tions in the sex ratio according to age, gear of 

 collection, and depth, season, and year of capture. 



In pound-net samples, which made up the bulk 

 of the Green Bay collections, the percentage of 

 females was consistently higher in February than 

 during other months of the year, and since no 

 trend was shown in sex composition during the 

 other months, the data for all but the February 

 samples are combined in table 34. This seasonal 



