182 University of California Puhlicatio US in Botany [Vol.5 



words, a great assumption was made in the early years of ]\[en- 

 delian interpretations in that the hypothetical term unit-char- 

 acter was assumed to represent a simple, indivisible, more or 

 less independent and dependable entity and it was this unit- 

 character that was expected to reappear unaffected in F.,. Does 

 not the perennial character of the question relating to the sta- 

 bility of the unit-character depend simply upon the unwillingness 

 or inability to appreciate the fact that blackness is as hypo- 

 thetical a something as the term "unit-character" which is ap- 

 plied to it ? The fact that blackness behaves as an orthodox Men- 

 delian dominant makes the heredity short-hand simple and also 

 seems thoroughly to obscure the fact that three factors may 

 fundamentally be responsible for the outward expression of the 

 black tendency just as truly as that three factors are necessary 

 for the production of the purple color in the aleurone cells in 

 certain varieties of maze. Equallj^ truly it is conceivable that 

 30 factors may be concerned in this purple maize color tendency 

 if 30 factors, units or genes are found to be responsible for the 

 outward expression of flower size in a large flowered form of 

 N. acuminata. Black, purple and large flower size are equally 

 involved in this situation. In other words whereas blackness has 

 apparently been shown to be an entity, sufficiently a unity to 

 remain unchanged in hybridization, the true significance of more 

 fundamental units underlying purple aleurone color, height of 

 plant, number of rows, length of ear and size of seed in maize, 

 and responsible for various fruit sizes, number of leaves in 

 tobacco, etc., has been made apparent by the same means. In other 

 experimental material, however, each of these characters may in 

 time be shown to behave as a simple Mendelian dominant also. 

 It seems probable that in time the majority of characters, both 

 "qualitative" and "quantitative," may be found to be modifiable 

 under selection. Each individual addition or subtraction stage 

 will " Mendelize " — using this term to signify experimental results 

 in accord with the Mendelian notation in its present expanded con- 

 dition — with the same pure line individual in each instance. It 

 has been shown that each may behave as a simple oMendelian 

 "recessive unit" (cf. Castle, 1912, p. 356) or each may so 

 behave that a multiplicity of more fundamental interacting units 



