108 SEVENTH REPORT. 



From this table we get: mean coefficient for homologous joints .9622 =b. 0022, and 

 mean coefficient for non-homologous joints .9583 ±.0022. Here then the mean coeffi- 

 cient for homologous joints — mean coefficient non-homologous joints=.9622-.9583==: 

 .0039 ±.0031. This difference is not significant. 



Putting all the results together it appears that there is not nearly so great a difference 

 in degree of correlation between homologous joints on the one hand, and non-homologous 

 joints on the other hand, as would probably be predicted on a priori grounds. Probably the 

 fairest comparison is that instituted in Table I., which shows that the slight differences 

 in the direction of correlations in homologous joints are insignificant in comparison with 

 their probable errors. In the case of Tables II and IV, for the reasons stated above, the 

 difference obtained cannot be considered to represent the true state of the case. 



The whole question of the relative degree of correlation in homologous 

 and non-homologous pairs of organs will be more fully discussed in the 

 complete report of this study of the crayfish. Until the more complete 

 analysis of the material has been finished we may safely conclude, I think, 

 that, so far as our data indicates, morphologically homologous joints of 

 the legs of the crayfish show approximately the same degree of correla- 

 tion, one with another, as is exhibited by non-homologous joints. While 

 the coeflScients of correlation between homologous joints are in most 

 instances slightly higher than those between non-homologous joints, yet 

 the differences are generally found to be insignificant when compared 

 with their probable errors. In other words, there is no marked or uni- 

 form advantage of homologous over non-homologous joints in regard to 

 degree of correlation. 



Zoological laboratory. University of Michigan, Ann Arbor, Mich. 



