AUG 

 DATE OF TAGGING EXPERIMENT 



Figure 12. — Seasonal variation In Intertidal stream life 

 of pink salmon in Olsen Creek, 1961. 



4 e 12 16 20 24 28 I 6 9 13 17 22 26 30 



JULY AUG SEPT 



DATE OF TAGGING EXPERIMENT 



Figure 13. — Regression of seasonal change in length of 

 stream life of pink salmon in Olsen Creek, 1961. 



measurements were treated separately by sex. 

 Measurements of males were more widely 

 dispersed from the mean, as indicated by 

 length frequency polygons, ranges in lengths, 

 and standard deviations from the means (figs. 

 15, 16, and 17). The graphic nnethod of com- 

 paring measurements devised by Dice and 

 Leraas (1936) is used to show differences in 

 size of spawners. Vertical overlap or lack of 

 overlap of the black bars (figs. 16 and 17), 

 which represent two standard errors of the 



mean on either side of the mean, provide a 

 measure of statistical significance. This 

 method is reliable only if the sample sizes 

 are large. For smaller samples sizes the 

 modification of the Dice-Leraas diagram by 

 Simpson, Roe, and Lewontin (1960) gives more 

 precise results. 



In 1960 the mean length of 1,381 males 

 was 42 cm., and the mean length of 1,168 

 females was also 42 cm. In 1961 the mean 

 length of 1,901 males was 44 cm., and the 



15 



