VARIATION AND HEREDITY IN TWINS 135 



(h) Those in which the mother is normal (without 

 doubhng), but in which doubhng occurs among the 

 offspring; of these there are 41 sets, of which 22 are 

 female and 19 male; in the group it is assumed that 

 the father possessed the factors for doubling. 



{c) Those in which both mother and offspring are 

 entirely without doubling; of these there are 43 sets, 

 of which 22 are female and 21 male; in this group it 

 Hcust be assumed that the father possessed no doubling 

 factors. 



These data would appear to show conclusively that 

 the ''doubling" factor is inherited as a dominant, but 

 that the mode of inheritance is not typically MendeHan ; 

 if '^ doubling" is dominant and ''lack of doubling" is 

 recessive, we should expect a considerable number of 

 individuals to be heterozygous for "doubling" and to 

 produce equal numbers of germ-cells that carry the dou- 

 bling factor and of those that do not. If heterozygosity 

 occurred, we should often get "non-doubling" in sets of 

 offspring from mothers that show doubhng but are hetero- 

 zygous for the character. That this result is not realized 

 is a strange circumstance, and one that can be explained 

 satisfactorily only on the assumption that a segregation 

 of dominant and recessive factors occurs during cleavage 

 so that the blastomeres, which go to produce both soma 

 and germ-cells of the individuals, are pure for the factor 

 in question, and that homozygous offspring are always 

 produced, and never any heterozygous ones. Segrega- 

 tion like that which is supposed to occur during matura- 

 tion divisions, when chromosome reduction accompanies 

 it, would appear to occur here during the process of 

 cleavage in which no reduction of chromosomes takes 



