IQII] SHULL REVERSIBLE SEX-MUTANTS 357 



In the second section of table I the hypothetical gene H for 

 hermaphroditism is given limitations which make it fit all the 

 empirical results of both the first and subsequent .generations; 

 but when the significance of the limitations is taken into account, 

 it becomes evident that there is small advantage gained by the 

 postulation of such a gene. Indeed the only advantage lies in the 

 fact that in case the female is a positive homozygote, it keeps open 

 the question whether or not there is a synaptic mate of F in the 

 normal male; for a newly arisen hermaphrodite gene (H) might 

 conceivably become a synaptic mate of F, even though the latter 

 had had no synaptic mate in the normal male. 



If the female is a neutral homozygote, that is, if the female 

 gene F has a male gene M as its synaptic mate in the male, the 

 hermaphrodite gene (if it exist at all) must be absolutely coupled 

 with this male gene. In like manner, if the female is a negative 

 homozygote FFmm, the H (if present) must be coupled with the 

 male gene M. It is simpler, however, to assume that the hermaph- 

 rodite determiner is a modified form of the sex gene itself, than 

 to suppose that it is a separate gene invariably coupled with the 

 sex gene. This conception that hermaphroditism results from a 

 mutative change in the sex gene, or in its homologue (?), the " Y- 

 element," is made the basis of the last section of table I, but can 

 apply only to those cases in which a male gene is present, or if 

 not a male gene, then its homologue, a sexually indifferent gene 

 which takes the place of M in the male; for if the hermaphrodite 

 character is assumed to be due to a change in the female gene 

 (F), as it must be if the latter has no " synaptic mate," the scheme 

 will not work. 



It appears to me impossible at the present time to determine 

 whether the females of Lychnis are positive, neutral, or negative 

 homozygotes. The facts seem to be equally well met by any of 

 these assumptions; but the definite limitations of the hermaphro- 

 dite character to the males makes inapplicable the extreme form 

 of the "presence and absence" hypothesis (that is, the hypothesis 

 of unpaired genes) unless the female is a negative homozygote 

 with reference to a male sex gene (M ) . While the possibility must 

 be kept open that this is the relationship of the sexes in Lychnis, 



