372 



THE AMERICAN NATURALIST [Vol. LIII 



about twice as great as the variability of the hooded character 

 in the uncrossed plus selected race. In the second and third 

 crosses the variability declined somewhat, but was still consid- 

 erably greater than that of the uncrossed race. It was indeed 

 very similar to that of the plus race in the first seven generations 

 of the plus selection experiment. (See Castle and Wright, p. 

 186.) 



TABLE I 



Eace 





Race wit 













Once extracted hooded F2 young 



. +3.73 

 . +3.17 



.36 







I +3.34 









.1 +3.04 







TABLE 11 



Results of Repeatedly Crossing the Min 

 Race 



[js Selected Race w it 













Control, uncrossed minus race, generation If. . . 

 Once extracted hooded Vz younj? 





.27 





Thricr extracted hooded iVyourin' 









The crosses of the minus race wnv s,,,rt,,l mn rai m.,. 

 later in the course of the selection experiiiieiits. with iiiiiinals of 

 generation 16, minus selection series. Tliey show effects much 

 more striking than those of the plus crosses. See Table II. The 

 minus selected race had now attained a mean of — ^2.63. A 

 single cross, with the same wild race used in the crosses of the 

 plus series, lower the grade to — .38, extinguisliing all the 

 changes in mean grade made by sixteen generations of selection, 

 and leaving the extracted hooded character in a higlily variahlf 

 state (standard deviation 1.25, nearly five times what it had been 

 before). A second cross with the same wild race converted the 

 extracted hooded individuals for the most part into a plus group, 

 mean + 1.01, but with variability somewhat decreased, standard 

 witli the wild race lias given ex- 



