LAKE HERRING OF GREEN BAY, LAKE MICHIGAN 



117 



mean percentages of deviations from the average 

 weight for all years: 1949, —2.3 percent; 1950, 

 — 1.8 percent; 1951, 1.3 percent; and 1952, 1.7 

 percent. This period of increasing weight was 

 also one of generally improving growth rate 

 (table 22). 



Hile (1936) found that the length-weight relation 

 and condition coefficient varied from year to year 

 in three of four populations of ciscoes in north- 

 eastern Wisconsin lakes. Annual differences in 

 the length-weight relation were reported by 

 Deason and Hile (1947) for Leucichthys Iciyi and 

 by Jobes for L. reighardi (1943), L. alpenae 

 (1949a), and L. hnyi (1949b). 



INFLUENCE OF METHOD OF CAPTURE ON 

 WEIGHT 



Discussions of seasonal and annual fluctuations, 

 and sex differences in the length-weight relation 

 have been based entirely on fish taken from pound 

 nets. Gill-net samples were omitted from these 

 comparisons because of the bias to length-weight 

 data introduced by gill-net selectivity. Farran 

 (1936) treated this problem in detail and estab- 

 lished limits of selectivity (in terms of length and 

 girth) of different sizes of mesh of gill nets in cap- 

 turing marine herring. Deason and Hile (1947) 

 demonstrated that within a sample of kiyi from 

 Lake Michigan that was homogeneous as to age, 

 sex, and locality and date of capture, the coeffi- 

 cient of condition decreased with increase in 

 length of fish taken by gill nets of the same mesh 

 size but increased in fish of the same length with 

 increase of mesh size. 



Although materials for the study of effects of 

 gear selection on length-weight data in the Green 

 Bay lake herring are scanty, those that are avail- 

 able (table 31) demonstrate conclusively that gill 

 nets tend to take heavier fish than do pound nets 

 operating in the same area and season, but because 

 of the small numbers of fish on which the individual 

 averages are based, a number of exceptions oc- 

 curred. The records for females taken during the 

 spawning season show almost no difference be- 

 tween samples from the two gears. The extent of 

 the bias in the remaining comparisons is suffi- 



ciently great, however, to make exclusion of the 

 gill-net samples desirable in detailed studies of the 

 length-weight relation. 



Table 31. — Weights of lake herring taken in gill nets and in 

 pound nets at different times of the year, 1950 and 1961 



[Weight in ounces) 



' Collected from a 2?8-lnch-mesh gill net at Oconto on November 30, 1950. 

 2 Collected from a pound net at Fox on November 29, 1950. 

 > Collected from a 25s-lnch-mesh gill net at Pensaukee on February 20, 1951. 

 * Collected from a pound net at Schumachers Point on February 22, 1951. 



GENERAL GROWTH IN WEIGHT 



The differences in weight according to sex, 

 season, and year of capture detract from the use- 

 fulness of the general growth curves of the Green 

 Bay lake herring. The best means of depicting 

 growth is to compute weights for the calculated 

 lengths of the best-represented age groups for 

 all pound-net data (table 17). Weights for age 

 groups III, IV, and V calculated from the general 

 length-weight relation (p. 114) are given in table 32. 

 Growth curves for these age groups are given in 

 figure 10. 



