INBREEDING AND CROSSBREEDING IN SEED DEVELOPMENT 95 



sterile, they considerably exceed both the parents in capacity for vegetative 

 growth. The inference is clear that the genie combination resulting from this 

 cross yields markedly different results in the endosperm and the sister 

 sporophy te. The difference in part may be a consequence of the 2 : 1 balance 

 of L. pimpinellifolium and L. peruvianium genes in the embryo as compared 

 with the 4: 1 ratio in the endosperm. The important point, however, is that 

 the mechanism of seed formation in the flowering plants is such that the two 

 products of a given double fertilization may be quite differently endowed in 

 terms of the genes necessary to perform their respective functions. 



The second example to be discussed in this connection will enable us to 

 visualize the limits which may be reached in endosperm disfunction with 

 retention of embryo viability. 



Fertilization freely occurs when squirrel-tail barley, Hordeum jubatum is 

 pollinated by cultivated rye, Secale cereale. The resulting seeds all die, how- 

 ever, within less than two weeks. Space does not permit me to recount here 

 the steps leading to the breakdown. They have been described in detail else- 

 where (Cooper & Brink, 1944; Brink & Cooper, 1944). The endosperm early 

 becomes completely disorganized. Some of the embryos formed, however, 

 reach a stage previous to collapse at which time they may be dissected from 

 the seed and successfully reared on an artificial nutrient medium. A single 

 plant was grown to maturity from an embryo treated in this way. The 

 plant was thrifty, although sterile. Representatives of the parent species 

 grown under comparable conditions were not available, so that a valid com- 

 parison of relative vigor could not be made. The hybrid, however, appeared 

 to be intermediate in stature and number of tillers. 



The extreme character of the endosperm disturbances in the H. jubatum X 

 S. cereale seed indicates that this hybrid could not arise under field condi- 

 tions. Although the embryo is demonstrably capable of continued develop- 

 ment its growth is terminated in the seed due to failure of the associated 

 endosperm. Death of the embryo, as an indirect result of endosperm disfunc- 

 tion following wide crosses, appears to be commoner than was thought before 

 the physiological implications of the secondary fertilization in flowering 

 plants were recognized. Realization of this fact has stimulated additional 

 interest in circumventing the phenomenon by excising such embryos from 

 the seed and rearing them artificially. 



Artificial methods of cultivating embryos removed from abortive seeds 

 often have been used to extend the area within which gene transfers may be 

 effected. Numerous interspecific hybrids have thus been grown which other- 

 wise are not realizable. The nature of the general problem involved may now 

 be seen in somewhat broader perspective. Two points of particular interest 

 may be noted. 



The first, briefly adverted to above, is that the frequency with which em- 

 bryos are formed following matings between distantly related plants is much 



