INHERITANCE OF CONTINUOUS 



AND DISCONTINUOUS VARIATIONS 



DR. F. B. SUMNER has been mak- 

 ing some studies of inheritance in 

 deer-mice. Writing on the ques- 

 tion of discontinuous variation, 

 his conclusions do not support the views 

 of Bateson and DeVries. 



"The dorsal tail-stripe is entirely lack- 

 ing in a certain strain of my mutants. 

 This stripeless condition is recessive to 

 the striped one. In crosses with normal 

 mice, the stripe appears in its full size 

 and intensity. Nevertheless, the stripe 

 itself was shown in the preceding pages 

 to vary from race to race and from one 

 individual to another. And these va- 

 riations, both racial and individual, were 

 found to be hereditary. 



"The case, of course, is parallel to 

 that of Castle's hooded rats. Since 

 'hoodedness' is recessive to 'self- 

 color' and reappears in one fourth of the 

 Fo generation. Castle argues that it is 

 dependent upon a single unit factor. 

 Nevertheless, this factor itself presents 

 hereditary variations in 'potency,' 

 since it can be modified indefinitely 

 by selection. The Mendelian counter- 

 argument is that 'hoodedness' behaves 

 as a unit character in certain crosses 

 merely because there is some one factor 

 without which it cannot manifest itself 

 at all. The variability in its degree of 

 manifestation is due to the fact that the 

 hooded pattern is modified by the 

 action of a number of independent 

 cumulative factors. The argument seems 

 a bit scholastic, but we must admit that 

 it is logical and consistent." Dr. Sumner 

 states his own conclusions in part as 

 follows. 



"All of these differences, structural 

 and pigmental, are found to be differ- 



ences of degree, revealed through a 

 comparison of mean or modal conditions 

 rather than of individual animals. In 

 comparing the less divergent of these 

 races with one another, the frequency 

 polygons for any given character over- 

 lap broadly. 



"These subspecific differences, and 

 even the minor differences which dis- 

 tinguish one narrowly localized sub-race 

 from the parent form, are found to be 

 hereditary, as evidenced by their per- 

 sistence when environmental condi- 

 tions are interchanged'. 



"The gradations in certain of these 

 characters by which individuals of the 

 same race differ from one another are 

 found to be strongly hereditary. 



"Hybrids between even the most 

 divergent of these four races are pre- 

 dominantly intermediate in character, 

 both in the Fi and the Fo generations. 

 In both of these generations a wide range 

 of variability is exhibited, which, how- 

 ever, is little if any greater in the Fo than 

 in the Fi. 



"In contrast to the sensibly continuous 

 variation and sensibly blended inheri- 

 tance shown in respect to these sub- 

 specific characters, is the behavior of 

 certain 'mutations.' Here we meet 

 with typical discontinuous variation, 

 and inheritance of the strictly alter- 

 native or Mendelian type. It is in- 

 sisted that the burden of proof rests 

 upon those who contend that these two 

 types of variation and inheritance are 

 reducible to a single category — that of 

 discontinuity. Anything like a proof 

 of this contention appears to be thus 

 far lacking."' 



1 Sumner, F. B., Continuous and Discontinuous Variations and Their Inheritance in Pero- 

 myscus — III. In American Naturalist, vol. 52, August-September, 1918. 



191 



