GENETIC STUDIES OF RABBITS AND RATS. 13 



lower in class 24, the mean being 2,472 or about three and a half 

 classes lower than Fi. The variability of F2 is about 40 per cent 

 greater than that of Fi, being 230 grams instead of 162. 



In neither of the Flemish crosses does the Fa variation extend up- 

 ward into the range of pure Flemish, but is separated from it by from 

 four to seven vacant classes. In other words, the larger parental 

 type is not recovered in Fa, doubtless because the number of inde- 

 pendent genetic factors involved is too great. If the larger parental 

 type is not recovered, it is not to be expected that the smaller type 

 would be recovered, since this would involve a chance recombination 

 of factors equally improbable of realization in a limited number of 

 offspring; but in reality the F2 range does in every case extend down- 

 ward into the range of the smaller parent type; therefore either the 

 smaller type is more readily recovered, or what appears to be the 

 smaller type recovered in F2 is really not such, but is a new genetic 

 combination or combinations which resemble in gross weight the 

 original parental type. The latter alternative seems more probable. 

 This view is supported by the observations on ear-length, which 

 character is closely correlated with weight, and yet in regard to 

 which neither the small. nor the large parental type is recovered in 

 Fj, all Fa variates occurring in the intermediate region. 



EAR-LENGTH. 



Ear-length is a character easier to study than body-weight, because 

 the adult condition is attained earlier and the races studied are more 

 sharply distinguished as to ear-length than as to weight. The ears, 

 as a rule, have attained their full growth at the age 150 days, or even 

 earlier in the case of the Polish race, whereas the weight continues 

 to increase slowly after that age. 



The ear-length was recorded at the same time that the rabbits 

 were weighed, though less frequently, as little change was noted in 

 the ear-length after the age 150 days. After a rabbit had attained 

 that age his ears were measured merely often enough to make sure 

 that no further change had occurred and that the earlier measure- 

 ments had been accurate. A variation of 1 or 2 millimeters was 

 frequently noted in the measurements, and the final rating of each 

 rabbit was accordingly based on the approximate mean of the 

 recorded measurements. The measurement taken was read from a 

 ruler placed between the base of the right ear and the head, the ear 

 being then held vertically against the ruler and slightly stretched 

 and the reading made at the ear-tip. Small races of rabbits have 

 short ears; large races have both longer ears and longer skulls, as we 

 shall see. 



The variation in ear-length of the groups of rabbits studied is 

 shown graphically in figure 7. The Polish rabbits studied range 



