MODES OP INHERITANCE IN HYBRIDS 485 



It is still an open question whether all of these types could 

 successfully be reduced to an orthodox Mendehan basis, were 

 the unit factors involved all analyzed out. Personally I would 

 not enjoy the prospect of untangling these ultimate inheritance 

 units from the conglomeration that exists in the hybrids of the 

 Fi generation. If such an analysis were possible at all it could 

 be made only from data derived from F2 hybrids. Any attempt 

 to proceed very far with such an analysis upon the Fi hybrids 

 would seem to be pushing the factor hypothesis too hard, with- 

 out being of any value to the subject of genetics, for the prin- 

 ciple of dominance is at best of minor significance as compared 

 with the principle of segregation. In the Fi generation of hybrids 

 we may readily have pure dominance, blends, intermediates, 

 mosaics or particulate mixtures without doing violence to the 

 essentials of Mendelian inheritance, for the original unit factors, 

 whether pure dominants, blends or mixtures, segregate out in 

 the pure state in the next generation of hybrids, thus satisfying 

 all of the demands of the theory. 



This then is my conclusion: that the Fi generation of hybrids 

 affords unfavorable material for the study of Mendelian inher- 

 itance and that unless we can interbreed the Fi hybrids we can 

 make little progress in the analysis of the inheritance factors. 

 Yet certain facts of much importance from the standpoint of 

 developmental physiology may be determined by means of teleost 

 hybrid material, some of which we have discussed and shall more 

 fully discuss in this paper. I still feel as I did when I wrote my 

 earlier papers that for the purpose of Mendelian breeding work 

 teleosts are unfruitful material, but that they throw a great deal 

 of light on the process of heredity. In this respect teleost mate- 

 rial is no better and no worse than echinoderm material, which 

 has been so extensively exploited. 



