THE CHROMOSOMES IN HEREDITY. 



passage quoted, there can be little doubt that irregular division 

 of chromosomes would be likely to produce marked variation, 

 but as Guyer himself observes, these irregularities increase with 

 the degree of infertility. It seems natural to conclude, therefore, 

 that they are not only pathological but perhaps in part the cause 

 of the infertile condition. Furthermore, on the hypothesis of in- 

 dividuality of chromosomes, which Guyer accepts, the loss of a 

 portion of a chromosome by irregular division would be perma- 

 nent and the effect of repetitions of the operation upon the de- 

 scendants of a single chromosome group (which he regards as 

 transmitted as a whole) would be so marked a depletion of 

 chromatic substance as must lead soon to malfunction and ulti- 

 mately to sterility. 



As already noted (p. 216) the first of these two explanations 

 of the causes of variation would allow only four possible combi- 

 nations of chromosomes in the offspring of a single pair. But 

 we know that except in the case of identical twins, duplicates 

 practically never appear in the offspring of a pair however nu- 

 merous the progeny. Therefore, whatever the number of the 

 offspring, the variations of all except the few provided for by the 

 four normal chromosome combinations must be accounted for by 

 obviously pathological division processes, which tend strongly in 

 the direction of sterility. But in the report of Bateson and Saun- 

 ders to the Evolution Committee we find the statement: "We 

 know no Mendelian case in which fertility is impaired" (p. 148). 

 When we reflect that the vast majority of cases studied by these 

 observers were Mendelian and connect this piece of evidence with 

 the testimony of Cannon : that the maturation processes of variable 

 cotton-hybrids are either normal or so distinctly abnormal as to 

 entail sterility and with Guyer' s own admission that the abnor- 

 malities in mitosis increase with the degree of sterility, the bal- 

 ance is strongly against the efficacy of pathological mitoses as 

 factors in normal hybrid variation. 



I take pleasure in acknowledging my indebtedness to Pro- 

 fessor E. B. Wilson for invaluable counsel in the presentation of 

 a subject offering many difficulties. 



DEPARTMENT OK ZOOLOGY, COLUMBIA UNIVERSITY, 



January 25, 1903. 

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



