No. G27] SITOETKR ARTTCLES AND DISCUSSION 



371 



tions. So we resorted to trapping: immature wild rats from a 

 single locality and using; these as a foundation stock. Crosses 

 with the plus race were then started successfully, but the corre- 

 sponding experiment with the minus race was hard to get going 

 and so has lagged behind the plus crosses. A report on the re- 

 sult of the plus crosses was made in 1916 (Castle and Wright). 

 The crosses with the minus race were not then sufficiently ad- 

 vanced to show what their outcome would be and this was still 

 true when reply was made to the criticism of MacDowell, as it 

 had been previously when reply was made to Muller and to 

 Pearl, and subsequently when I addressed the Washington Acad- 

 emy of Science on the role of selection in evolution (1917). But 

 since then the minus crosses have given what seems to be con- 

 clusive evidence that tlie single gene had not been altered by 

 selection, although the inherited complex responsible for the 

 hooded character had steadily been altered in opposite directions 

 and these alterations were permanent in the sense that they rep- 

 resented racial modes, stable so long as the race was not out- 

 crossed. 



I still have on hand a few representatives of the plus and of 

 the minus races which because of their low fecundity it has been 

 impossible to select furtlier for several generations. The two 

 races are very dillVrtMit in jippt^innicc The plus riice shows no 

 white except on tlic iiiidfr sidr ;iih1 sotnctiinos aloim- the Hank. 

 The minus nirr slmw^ no l.l.n-k .A,vpi .i short hoo.l 1\ in- anterior 

 to the shoulders, and in an occasional individual a small black 

 spot or tM'o in the middle of the back or on the tail. Yet the 

 variability of each race is still considerable; as measured by our 

 "grades" it has not appreciably diminished in recent genera- 

 tions. The somatic ditferences entailed by the selection experi- 

 ments with the hooded character of rats are seemingly greater 

 than those secured by Sturtevant or by MacDowell in regard to 

 bristle number in Drosophila, yet I doubt not they may be ex- 

 plained on similar grounds. 



Crossing with a wild race affects very differently the plus and 

 the minus selected races. See Tables I and II. The plus race 

 was much less affected than the n»inus race. Its mean grade was 

 lowered, by three successive ci-osses with tlie wild race, not over 



doubled by the first cross. Thai is tlie variability of the hooded 

 character, when extracted ui h\, from the iirst wild cross, was 



