MODES OF INHERITANCE IN HYBRIDS 473 



6. F. majalis 9 X Cyprinodon (f. In two experiments per- 

 formed at different times none of the eggs lived through the 

 cleavage period but began to disintegrate during advanced cleav- 

 age stages. 



It is clear that for some obscure reason all of the Cyprinodon 

 egg hybrids succeed better than any of the three types of Fundu- 

 lus egg hybrids. It may be that this difference in success between 

 reciprocal crosses is due to differences in the sizes of eggs and 

 that it will be found that greater success generally is associated 

 with the smaller egg, less burdened with yolk. . Whatever fac- 

 tors may be found to condition the differences in developmental 

 success between reciprocal crosses we have at least the clear evi- 

 dence of such differences and these facts militate against the 

 idea that the degree of developmental success in teleost hybrids 

 is simply a function of phylogenetic relationship. To make my- 

 self clear on this point let me review the facts concerning recipro- 

 cal crosses between F. diaphanus and Cyprinodon. The Cyrpi- 

 nodon egg hybrid succeeds in passing safely through the period 

 of gastrulation and in making a considerable start at embryonic 

 differentiation, while the F. diaphanus egg hybrid succeeds only 

 in barely beginning gastrulation. There is a marked difference 

 here in developmental success without any difference in phylo- 

 genetic relationship. There must then be some factor other than 

 phylogenetic relationship at the basis of the difference between 

 such reciprocal crosses. 



It should be noted also that all of these six inter-generic crosses 

 under discussion, though the parent species all belong to the 

 same family, Poeciliidae, develop much less successfully than 

 several crosses between species belonging to different orders, as 

 for instance some of those cited by Moenkhaus. A cross between 

 Fundulus heteroclitus of the order Haplomi and Gasterosteus 

 bispinosus (stickleback) of the order Hemibranchiae, gives hy- 

 brids in which, although the germ ring is only one-third around 

 the yolk, the "embryo is short and completely formed." Again 

 a cross between F. heteroclitus of the order Haplomi and Menidia 

 notata of the order Acanthocephali give hybrids in which the 

 embryo is described as being sufficiently differentiated to show 



