1913] Goodspeed: Nicotiana Hyhrids 183 



or genes must be assumed as the basis of the character appearing 

 in modified condition in each stage. Considerations of this sort 

 simply serve to show that certain characters are sufficiently stable 

 within themselves so that the fundamental interacting units upon 

 which they depend for their outward, somatic expression, do not 

 appear while other characters, not so stable, under the stress of 

 hybridization do manifest the fundamental units through the 

 interaction of which they are able to appear in substantial form. 

 The fact that they are manifested in one ease implies that funda- 

 mentally the3% though not apparent, are present in all other cases 

 also. So soon as it can be demonstrated that for any one char- 

 acter the influences upon which it depends, the factors beneath 

 the surface, are actually apparent, in that hypothetical terms to 

 stand for them in mathematical expressions can actually be pro- 

 posed, at once one postulates for all characters the existence of 

 similar influences whether they are similarly apparent or not. 

 In other words, all "unit-characters" are "molecular" in struc- 

 ture. Aggregates of similar "units or genes" evidently are 

 responsible for the outward evidence of "unit-characters" and 

 the unity of this "unit-character molecule" may depend upon 

 the degree of affinity of the units or genes — a true chemical 

 affinity perhaps (see East and Hayes, 1912, p. 35 ) . The more 

 quantitative our investigation becomes, the more clearly is 

 brought out the looseness, rarely the firmness, of the bond which 

 holds the innumerable units or genes to the ' ' unit-character ' ' con- 

 ception. From the Mendelian standpoint it seems clear that we 

 are getting below the surface with reference to the real sig- 

 nificance of the "unit-character" conception. The presence of 

 more fundamental, basic heritable influences within the seemingly 

 indestructible "unit-character" is assumed in order to make the 

 mode of inheritance of both "qualitative" and "quantitative" 

 characters susceptible of interpretation according to the Mende- 

 lian notation. All "unit-characters" mu.st then be assumed to be 

 alike in the possession of these fundamental, "subepidermal" 

 influences whether their presence needs to be assumed or not. 

 Otherwise we have to recognize two categories (1) "unit-charac- 

 ters" — distinct, atomic, indestructible and (2) potential charac- 

 ters — molecular in structure and dependent upon the interaction 



