94 E. CARLETON MACDOWELL 



tions of extras in Fo is connected with the modification of their 

 distribution. A strongly restrictive group of accessory restrict- 

 ing factors, even in the absence of the main restrictor may pro- 

 duce a normal fly, especially if the fly happens to be small. 



Much of the recent genetics work supports the principle of 

 accessory or multiple factors. According to this principle more 

 than one independent factor may influence a single character. 

 Besides the cases which prove unquestionably the existence of 

 such multiple factors by means of definite ratios, as the ligula 

 in oats and the red grain in wheat (Nilsson-Ehle '09), yellow 

 endosperm in maize, (East and Hayes, '11) and the triangular 

 form of capsule in shepherds purse (Shull '14) there are a large 

 number of investigations in which the presence of multiple 

 factors is strongly indicated. Since the author has already 

 discussed such cases (MacDowell '14 a and b) it will be necessary 

 only to add references to the recent work of Wichler ('13), Ikeno 

 ('14), Hayes ('14), Davenport ('13), Lotsy ('13), Phillips ('14) 

 and Punnett ('14), and to call attention to the critical discussion 

 of Shull ('14). Investigations showing transgressing segregation, 

 in which the F2 distributions form continuous gradations but 

 with modes dividing the individuals into groups corresponding 

 to 3 :1, or 1 :2 :1, (Balls '07, Leake '11, Bifl'en '05) have been 

 interpreted by an hypothesis involving one main factor and 

 accessory factors, similar to the one employed above. Still 

 closer resemblance to the work herein described is that of Castle 

 and Phillips ('14) with piebald rats. They dealt with a variable 

 character that proved to be influenced by a Mendelian factor. 

 The distribution of the variations of this character was modified 

 by crosses. No hypothesis other than that of accessory factors 

 could be found to explain all the results. However, Castle 

 still holds that the continued success of selection goes to prove 

 the modifiability of a Mendelian factor. The main difference 

 between the investigations on the piebald rats and the extra 

 bristled flies, aside from the fact that no minus race of flies has 

 been established, lies in the fact that in the case of the flies the 

 progress of selection does not seem to continue after the sixth 

 generation, while in the rats it appears to continue as far as the 



