EXPERIMENTAL EMBRYOLOGY 393 



would be to know what changes take place inside the embryo sac 

 in known cases of haploid parthenogenesis and then attempt to 

 duplicate them artificially. From the very meager information 

 that we possess in this regard it seems that although the egg can 

 frequently be made to develop parthenogenetically, the real diffi- 

 culty lies in an initiation of endosperm formation. Thus, Katayama 

 (1932) found several-celled embryos in unpollinated ovaries of Tri- 

 ticum fixed on the ninth day after castration, but further develop- 

 ment stopped owing to lack of endosperm. On the other hand, in 

 those ovaries in which pollination is not unduly delayed, triple 

 fusion is presumed to take place normally, resulting in endosperm 

 formation which enables the maturation of the haploid embryos. 

 Similarly, Kihara and Yamashita (1938), who used X-rayed pollen, 

 believe that owing to the greatly reduced rate of growth of the pollen 

 tubes arising from such pollen grains, the egg begins to divide 

 parthenogenetically but the polar nuclei are eventually fertilized 

 in the usual way. 



These ideas need to be verified by making proper developmental 

 studies, for the objection arises as to why, in cases of delayed pol- 

 lination or the use of foreign pollen or X-rayed pollen, the egg 

 should begin dividing first, when in most angiosperms the endo- 

 sperm nucleus is the first to divide and the zygote undergoes its 

 first division only after a few endosperm nuclei have been formed. 



Production of Adventive Embryos. Since embryos arising asex- 

 ually from the cells of the nucellus or the integument carry the 

 full chromosome complement of the maternal parent, they are ge- 

 netically identical with the latter. This phenomenon is of great 

 importance in some of our cultivated plants, especially fruit trees. 

 In Citrus it is used to yield large numbers of uniform rootstocks, 

 since much of the variation in the size and productivity of the 

 trees is due to the variability of the stocks. In Mangifera, those 

 varieties which form adventive embryos can be as safely prop- 

 agated by seeds as by budding or grafting. In others, with only a 

 zygotic embryo, seed propagation does not give a type true to the 

 mother and one must resort to vegetative propagation. Since seed- 

 lings can be raised much more cheaply, a method of inducing the 

 formation of adventive embryos would obviously have a great 

 advantage (see Leroy, 1947). 



Recognizing the importance of adventive embryony in horticul- 



