444 Haploids and Autopolyploids 



that diffuse into the surrounding tissues and regulate the way 

 they develop. One of these developmental changes is the for- 

 mation of a channel of conducting tissues in the integument and 

 megasporangium (nucellus), through which nutrients pass into 

 the growing embryo sac. In the hybrids, endosperm development 

 is slower, and this secretion is probably reduced in amount 

 although it may also differ qualitatively. Whatever may be the 

 cause, these conducting elements fail to develop in the hybrids. 



At the same time, the megasporangium begins to grow. This 

 structure always remains one-celled in N. rustica, but in the hy- 

 brids it becomes several cells thick. The slow development of 

 the endosperm, accompanied by a hyperplasia of the megaspo- 

 rangium, results in a markedly lower endosperm/megasporan- 

 gium ratio in the hybrids. In some hybrids the growth of the 

 megasporangium is so pronounced that this tissue completely 

 surrounds the endosperm; in others a gap remains in the mega- 

 sporangium leaving an opening from the endosperm to the in- 

 tegument as in A^. rustica. When the embryos were in the eight- 

 to sixteen-cell stage, all seeds of N. rustica had this gap, as did 

 75 per cent of the seeds of N. rustica X ^- tahacuyn and only 26 

 per cent in A^. rustica X N. glutinosa. Apparently the nutrients 

 are cut off from the endosperm in practically all the glutinosa 

 hybrids, with the result that practically all the developing ovules 

 collapse early. In some of the tabacum hybrids, however, enough 

 nutrient material seems to reach the endosperm so that these 

 ovules attain a more advanced stage of development and even 

 develop occasionally into shrunken but germinable seeds. This 

 study reveals that the endosperm is an important structure and 

 sometimes may act as a barrier to hybridization between species. 

 The endosperm in this hybrid appears to divide normally, but its 

 whole development is slow. 



Sterility in hybrids between barley and rye also results from 

 the failure of the endosperm to develop in normal fashion. 

 Thompson and Johnston found that such a cross was incom- 

 patible because of abnormal development of the endosperm in 

 the hybrid seed and particularly of the endosperm nuclei. Cooper 

 and Brink made a similar cross in which the seeds aborted com- 

 pletely from the fourth to the thirteenth day. In this inter- 

 generic hybrid, the primary endosperm nucleus often divides 

 somewhat later than normally, and subsequent behavior of the 



