J. ARTHUR HARRIS 



More mature series, r = .9579 .0020 



/ = - - 2.6704 -f- 0.8740 h 

 Combined series, r = .9533 .0018. 



/ = -- 2.5953 + 0.8685 h 



Diagram 5 gives the regression of number of leaves on the number 

 of leaf homologs (cotyledons + leaves). The solid dots representing 

 the empirical mean leaf number lie practically on the theoretical line. 



Correlation between total leaf homologs and the deviation of the 



number of leaves from their probable value : 



Less mature series, ^=.6936 .0226. 



More mature series, r = .7642 .0185. 



Combined series, r= .7378 .0089. 



The correlations between total leaf homologs and the deviations of 

 the cotyledons from their probable value are numerically identical with 

 the foregoing but negative in sign. 



The results show a high degree of consistency of the two series. These 

 final constants show that when the total number of leaf homologs in 

 creases, the variation is due to a far greater extent to the laying down of 

 a greater number of primordial leaves than to the formation of a larger 

 number of cotyledons. The regression equations for both the deviation 

 of the number of leaves from their probable value (^) and the deviation 

 of the number of cotyledons from their probable value (s c ) are given by 



&amp;lt;r z = = .874557 

 = 2 -5954 -- 0.3007 // 

 &quot;i = ~ 2 -5954 + 0.3007 h 



and represented in diagram 6. The results are clearly linear. 



Illustration 7. Change in proportion of parts in developing trout 

 JENKINSON (1912) has given data for total length and length of 

 head for three growth stages in the American rainbow trout. His con 

 stants are : 



