408 



FISHERY BULLETIN OF THE FISH AND WILDLIFE SERVICE 



Table 53. — Coefficient in regression equation and R 2 in 

 the study of the relation between environmental factors and 

 growth in length in the second and later years of life of 

 Saginaw Bay yellow perch 



[When regression coefficient is not given that variable was not considered in 

 the derivation of the equation. Values significant at the 5-percent level 

 are indicated by an asterisk] 



Number of variables 

 and months 



4 variables: 



May-October... 

 June-September 



July-August 



Various 1 _. 



Various 2 __ 



3 variables: 



Various 3 



Various * _, 



May-October... 

 June-September 



July-August 



June-August 



Various s 



2 variables: 



Various s __. 



June-September 



July-August 



Various 7 



0.766 

 .668 

 .785 



•.891 

 .782 



.619 

 .679 

 '.707 

 •.671 

 .511 

 .603 

 .550 



.611 

 '.554 



.511 

 •.535 



' Temperature 

 Aug.; turbidity, 



- Temperature 

 turbidity. May. 



3 Temperature 



* Temperature 



* Temperature 

 fl Temperature 

 T Temperature 



June-Aug.; 



April-May. 



July-Aug.; 



July-Aug. 

 July-Aug. 

 July-Aug. 

 July-Aug. 

 July-Aug. 



precipitation, July-Aug.; water level, June- 

 precipitation, July; water level, June-Sept.; 



water level, June-Sept.; turbidity, May. 

 water level, July-Aug.; turbidity, April-May. 

 precipitation, July; water level, June-Sept, 

 turbidity, April-May. 

 water level, June-Sept. 



one combination was significant (temperature, 

 June-Aug. ; precipitation, July-Aug. ; water 

 level, June-Aug.; turbidity, April-May). In 

 this combination, regression coefficients were not 

 significant for water level and turbidity, but 

 were significant for both temperature and pre- 

 cipitation. When three variables were used in 

 the multiple correlation, only two combinations, 

 neither of which included turbidity, gave signifi- 

 cant correlation (temperature, precipitation, and 

 water level for May-Oct., and for June-Sept.). 

 In these two regressions, the only significant re- 

 gression coefficients were those of water level. 

 With 2 variables, 2 of the 3 correlations that 

 included water level were significant (tempera- 

 ture and water level for June-Sept. ; temperature 

 for July-Aug., and water level for June-Sept.), 

 and all three regression coefficients for water 

 level also were significant. In the combination, 

 temperature and water level for July and August, 

 the regression coefficient for water level was sig- 

 nificant but the multiple correlation was not. 



These correlations offer strong evidence that 

 the growth of Saginaw Bay yellow perch in the 

 second and later years of life has fluctuated in- 

 versely with the water level. Despite this appar- 

 ent relation, it is not possible to offer an ecologi- 



cal explanation as to how water level might affect 

 growth. Possibly the adverse effects of deeper 

 water over the usual shallows is greater than 

 beneficial effects from the creation of new shal- 

 lows along the shore. Limnological studies are 

 needed on the food production of inshore areas. 

 The regressions also provide some but much 

 less convincing evidence that growth has been 

 correlated negatively with temperature in June 

 through August and with precipitation in July 

 and August. Jobes (1952) showed a similar nega- 

 tive but insignificant correlation between fluctua- 

 tion in growth and the combined temperature of 

 June to August for Lake Erie yellow perch. 



SEX AND MATURITY 

 Size at Maturity 



The spawning-run samples are considered by 

 many investigators as almost entirely mature fish. 

 Van Oosten (1929) showed that immature fish 

 were not represented in the spawning-run collec- 

 tion of Saginaw Bay lake herring and hence 

 concluded that these samples were not suitable 

 for the estimation of size at first sexual maturity. 

 A similar conclusion was expressed in studies on 

 Great Lakes yellow perch (Hile and Jobes, 1941, 

 1942; Jobes 1952). 



In the present investigation, the comparison 

 of the spawning-run samples for the period 

 1943-55 witli the 1955 collections made out- 

 side the spawning season (summer and fall) re- 

 vealed no significant difference in the percentage 

 of mature yellow perch at corresponding lengths. 

 Consequently, the data on the percentage of 

 mature fish in the collections from all seasons 

 were combined in the preparation of table 54 

 (in the collections of 1950 and June 7, 1955, 

 records of maturity were not available). 



Table 54. — Relation between length and sexual maturity of 

 Saginaw Bay yellow perch in 1948-55 



[All perch longer than those recorded in the table were mature] 



