26 



C. C. Eii&berp 



The closeness with which these results agree is very remark- 

 able. Tn order that they may be compared more readily, thev 

 are tabulated below. 



CORRELATION OF LARGE AND SMALL HOOKS 



Another question of importance is this : From the mean of the 

 large hooks of any worni can we predict within reasonable limits 

 the mean of the small hooks of that worm? To determine this, 

 correlation tables were made in which the means of the large 

 hooks of each head were entered as subject, and the correspond- 

 ing means of the small hooks as the relative. In this manner the 

 correlation existing between the large and the small hooks of a 

 worm was found to be .3 for T. scrrata, and .38 for T. serialis. 

 The probable errors of these determinations are ±.ll, and ±.14 

 respectively, a trifle over one-third of the correlation in each case, 

 not sufficiently large to destroy it. We thus see that the corre- 

 lation between the large and the small hooks of a worm is a real, 

 though not very large quantity. 



SUMMARY 



The results obtained from these mathematical considerations 

 may be summed up as follows : 



1. There exists a quite high correlation between characters a 

 and b in both the large and the small hooks of both species of 

 tapeworms, it being on the average about .7. 



The correlation between a and c is also fairly large, about .5. 



The correlations of c to b and d are small. 



216 



