238 WALTER S. SUTTON. 



chiefly upon hybrids, while the cytological data are the result of 

 study of a pure-bred form ; but the correlation of the two is 

 justified by the observation of Cannon ' that the maturation mi- 

 toses of fertile hybrids are normal. This being the case it is 

 necessary to conclude, as Cannon has already pointed out, that 

 the course of variations in hybrids either is a result of normal 

 maturation processes or is entirely independent of the nature or 

 those divisions. If we conclude from the evidence already given 

 that the double basis of hybrid characters is to be found in the 

 pairs of homologous chromosomes of the presynaptic germ- 

 cells, then we must also conclude that in pure-bred forms like- 

 wise, the paired arrangement of the chromosomes indicates a 

 dual basis for each character. In a hypothetical species breed- 

 ing absolutely true, therefore, all the chromosomes or subdi- 

 visions of chromosomes representing any given character would 

 have to be exactly alike, since the combination of any two 

 of them would produce a uniform result. As a matter of fact, 

 however, specific characters are not found to be constant quan- 

 tities but vary within certain limits ; and many of the variations 

 are known to be inheritable. Hence it seems highly probable 

 that homologous chromatin -entities are not usually of strictly uni- 

 form constitution, but present minor variations corresponding to 

 the various expressions of the character they represent. In other 

 words, it is probable that specific differences and individual 

 variations are alike traceable to a common source, which is a 

 difference in the constitution of homologous chromatin-entities. 

 Slight differences in homologues would mean corresponding, 

 slight variations in the character concerned - - a correspondence 

 which is actually seen in cases of inbreeding, where variation is well 

 known to be minimized and where obviously in the case of many 

 of the chromosome pairs both members must be derived from the 

 same chromosome of a recent common ancestor and hence be 

 practically identical. 



In the various forms of parthenogenesis we meet the closest 

 kind of inbreeding and a brief consideration of the variability to 

 be expected in each, from the standpoint of the chromosome 

 theory, may serve as a guide to such research as will test the 



1 Cannon, W. A., loc. cit. 



