of forms from a common parent by the loss 

 of dominancy in its several character-deter- 

 minants accounts for the general presence of 

 a recessive unit, corresponding with each 

 dominant unit, in all the nearly related forms. 



No suggestion has been made as to the na- 

 ture of the change by which a dominant allelo- 

 morph becomes recessive, but if this change 

 be looked upon as a degenerative one which 

 may be followed later by complete disappear- 

 ance of the unit it would account for the fact 

 that hybrids between nearly related forms are 

 usually Mendelian, while those between more 

 distant ones are not. 



I may summarize briefly as follows: 



(a) What appear to be unit characters may 

 be, and probably usually are, compound char- 

 acters. 



(fc) New characters appear by the change of 

 one or more character determinants from the 

 dominant to the recessive condition. 



(c) Some of the partial products resulting 

 from this process of analysis have no ex- 

 ternally apparent distinguishing characteristic, 

 and these supply instances of so-called ' latent ' 

 characters. 



(d) Mendelian hybridization results in an 

 F t which is a partial or complete synthesis of 

 an ancestral condition. 



(e) This conception gives an explanation of 

 the general presence of recessive units corre- 

 sponding to the dominant units in each closely 

 related form. 



(f ) If the change from the dominant to the 

 recessive condition is a degenerative process 

 which may be followed by complete disappear- 

 ance of a unit, an explanation is found for 

 the fact that Mendelian behavior is a function 

 of nearly related forms but not of more dis- 

 tantly related ones. 



