GENETIC STUDIES OF RABBITS AND RATS. 17 



In studying bone-measurements, the length of 3 leg-bones has been 

 investigated, that of the tibia, the femur, and the humerus (tables 9 

 to 11). The bones measured are from the right side of the body. 

 The results are very similar in all three cases. The Polish-Himalayan 

 cross is followed by such increase in vigor that the cross-breds of 

 both the Fi and the Fa generations surpass in bone-dimensions the 

 intermediate between the parent breeds, but the Flemish crosses 

 produce offspring which even in Fi fall slightly below the interme- 

 diate between the parent breeds, and in Fj fall below it still more. 

 In every case Fa is more variable than Fi. 



CORRELATION. 



The question may properly be raised whether the same genetic 

 factors affecting size operate in all parts of the body, or whether 

 there are special factors affecting the size of each part. The latter 

 view is favored by Davenport in his studies of human stature; the 

 former seemed to me to be indicated in the statistics of rabbit meas- 

 urements pubUshed by MacDowell (1914). Wright (1918), from a 

 statistical analysis of the same data, concludes that both general and 

 special factors are indicated. The present investigation should be 

 able to throw further light on the subject. If two parts or dimen- 

 sions of the body are influenced by the same genetic agencies, they 

 should vary in unison, as the genetic agencies are made to vary by 

 means of cross-breeding. If they are influenced by different genetic 

 agencies, they should vary independently of each other. For this 

 reason it is desirable to study the correlation existing between each 

 pair of size-characters studied. 



The results of such a study are contained in the correlation tables 

 (tables 12 to 29). These show strong correlation in every case 

 between size of the body as a whole and size of each of its parts, as 

 well as between different parts of the body. Most of the correlation 

 coefficients (table 30) lie between 0.80 and 0.90, where 1.00 would 

 indicate complete identity of all agencies affecting size, whether 

 genetic or non-genetic. The highest correlations are found in com- 

 paring the lengths of the long bones of the legs, humerus with femur 

 (0.906) and with tibia (0.904), and femur with tibia (0.927). Next 

 in closeness of correlation comes the relation between skull-length 

 and the length of the leg-bones (0.806 to 0.871), but the correlations 

 of weight with bone-measurements are almost as close (0.820 to 

 0.852), except in the case of the tibia, where the coefficient falls to 

 0.758. There is a very similar range in the correlation coefficients 

 between ear-lengths and weight (0.836) and ear-length and the 

 various bone-measurements (0.823 to 0.836), except again in relation 

 to the femur, where occurs the lowest correlation of the entire 15 

 studied, viz, 0.741. 



