STUDIES OF HEREDITY IN RABBITS, RATS, AND MICE. 



BY W. E. CASTLE. 



I. FURTHER EXPERIMENTS UPON THE MODIFIABILITY 

 OF THE HOODED CHARACTER OF RATS. 



In publications Nos. 195 and 241 of the Carnegie Institution of 

 Washington reports have been made on the results of a series of experi- 

 ments designed to show to what extent a mendelizing character, the 

 hooded pattern of piebald rats, may be altered by selection or by 

 crossing. At the last report (Castle and Wright, 1916) the plus- 

 selection series had been carried without out-crossing through 16 

 successive generations, in the course of which the mean grade of the 

 offspring had advanced from +2.05 to +4.13, in terms of the arbi- 

 trary grading scale depicted in both previous publications. Since 

 then the plus-selection series has been carried through four additional 

 generations of selection (17 to 20) and the mean of the selected race 

 has been raised to 4.61. In some respects this part of the series is 

 less satisfactory than that previously reported on, because smaller 

 numbers of animals were available from which to select and the selec- 

 tion has therefore been less rigorous. The race has unmistakably 

 fallen off in vigor and fecundity in later generations. It is uncertain 

 whether this should be ascribed to inbreeding alone, uncorrected by 

 selection for vigor (as in Miss King's experiments), or to increase in 

 the prevalence of disease, or to both causes. Certain it is, however, 

 that notwithstanding increasing care in regard to feeding and sanita- 

 tion a very large proportion of our breeding-pens in the case of the 

 selected races produce no young at all. 



Tables 1 to 4 show in detail the grade distribution of the young 

 produced by plus-selected parents of generations 17 to 20. The 

 numbers of young produced in each of these generations are respec- 

 tively 351, 420, 280, and 92. The mean grade of the young advances 

 from 4.13 (in generation 16) to 4.48 (in generation 17), remains prac- 

 tically stationary in generations 18 and 19 at 4.46 and 4.49 respectively, 

 and then advances again (in generation 20) to 4.61. 



The minus-selection series which at the last report had been car- 

 ried through 17 generations, with an advance in mean grade of the 

 young from grade minus 1.00 to grade minus 2.70 has now been 



l 



