12 



THE VARIATION AND CORRELATIONS OF 



The regression lines are interesting, since they give us the plan, so 

 to speak, upon which the crickets are made. The equations for deter- 

 mining the probable length of the several organs for a given wing- 

 length, together with the probable error of the coefficient of regression, 

 are given in table 6. It is to be noted that in no case is there a marked 

 difference between the two groups in the regression coefficients (the 

 multiplying constants in the equations). However, the slight differ- 

 ence is, in each case, in the same direction, i. e., larger in the short- 

 winged group. What this means is best explained by reference to fig. 

 3, which shows the position of a pair of the regression lines and also 



WINGS. 



Ss 



12 



14 



16 



IS 



10 



12 



1 6 1 8 20 22 24 26 



FIG. 3. Regression lines of tegmina on wings. 



the distribution of the individual crickets studied, each cricket being 

 represented by a dot. The lines, instead of being parallel, as they 

 would be were the regression coefficients equal, approach each other 

 toward the smaller dimensions. In other words, in each group pari 

 pftssu the decrease in wing-length the other organs decrease, but they 

 decrease rather more rapidly in the long-winged group than in the 

 short-winged one. As stated above, the difference is very slight, and 

 we may say that, in a general way, the influences which brought about 

 this marked dimorphism of wing-length have not greatly affected the 

 structural relations between the wings and the other organs studied. 



TABLE 6. Certain regression lines of the Gotha collection. 



Skewness, in a biometric sense, is that condition of the distribution 

 of frequencies in which the average dimension differs significantly from 



