Compound Characters by Cross-breeding. 53 



though in differing capacities. From the records of the results 

 thus far there seems no good reason for introducing these diffi- 

 culties, or for regarding the case as one of di- hybridisation. 



In attempting to apply either hypothesis to de Vries' other 

 cases or to Tschermak's results in Phaseolus, various complications 

 are encountered, and it is clear that in those phenomena other 

 processes are concerned. Until there is a fuller experimental 

 basis discussion of these cases is scarcely possible. To this subject 

 I hope to return when the experiments on resolution of characters 

 undertaken by Miss Saunders and myself are further advanced. 

 Nevertheless while our conceptions of allelomorphic characters, 

 simple and compound, are in process of formation, it seemed 

 desirable to point out how an unnecessary complication might be 

 avoided. 



From all that we know of these cases of resolution it seems 

 fairly certain that the whole of a compound character need not be 

 resolved on the first cross, but that the cross-bred may bear some 

 gametes which transmit it in its entirety. 



The visible phenomena of cell-division would suggest that the 

 number of gametes bearing the simple character of one parent — 

 here the white — should equal the sum of those representing the 

 other parent and bearing the compound character and its various 

 components. On the view here suggested, the number of gametes 

 bearing the simple character is regarded as equal to the number 

 of only one of these classes ; and in the case under discussion the 

 latter so far seems the more acceptable account. 



We must, however, remember that in this case the simple 

 character is recessive, the compound being dominant; and we 

 naturally inquire if when a dominant simple character is crossed 

 with a recessive compound character, the result would be the same. 

 Whether the fact has a relation to this matter of dominance or not 

 cannot be said, but in the only case I can adduce of a cross of the 

 latter kind, namely that of a white and a coloured breed of poultry, 

 there seems to be reason for the belief that the white gametes are 

 as numerous as those of the various coloured ones collectively, for 

 in the latter case the ratio 3 light to 1 dark (in generation F 2 ) seems 

 distinctly indicated as the result of breeding the hybrids inter se. 



In conclusion, since the resolution of a compound character 

 may be spoken of as an analysis leading to a distribution of the 

 components among the gametes, the term synthesis should surely 

 be reserved for a recombination that has taken place in such a way 

 that the gametes become bearers of the compound character again, 

 as they were in the pure compound form. In the place referred to, 

 de Vries, however, using " analysis " as above, speaks of the union 

 of F and D gametes to form a red zygote as a " synthesis." But 

 until we know in some given case that the zygote so produced 



