64 PLANT GROWTH SUBSTANCES 



formed, and two of these produced a single good seed only after several 

 hundred pollinations. In some cases, especially when long-styled species 

 are pollinated by short-styled species, the barrier to crossability may be 

 due to inability of pollen tubes to reach the ovary before the style 

 abscises, because of their slow growth or early bursting. The major 

 barrier to crossability in our group, however, is the cessation of develop- 

 ment of the hybrid embryos. What stops the growth of the embryo? 

 For this we have some evidence. 



In selfs and compatible crosses. Dr. Satin has found that the cells of 

 the integument are filled with starch grains at early stages of develop- 

 ment. Apparently the endothelium, a single layer of cells immediately 

 surrounding the embryo sac, functions as nurse cells and through its 

 activities the cells of the endosperm become filled with fat and granules 

 of aleurone which are ultimately digested and passed on to the normally 

 developing embryo. The endothelium itself also becomes digested. In 

 incompatible crosses, however, the picture is quite different. The number 

 of starch cells of the integuments increases and the endothelium, instead 

 of remaining a single layer of cells, proliferates and may invade the 

 embryo sac and form a tumorous tissue. In consequence, apparently, 

 the embryo aborts. This abortion is evidently not due to any lack of 

 food but rather to inability of the embryo for some reason to utilize 

 the abundant food present or to some factor responsible for the inhibi- 

 tion or digestion of the embryo. 



Rappaport has given further information regarding the causes of 

 embryo abortion. He finds that there is a water-soluble, thermostable 

 inhibitor which can be extracted from the disintegrating material which 

 surrounds the aborting or arrested hybrid embryos. This inhibitor stops 

 the growth of normal selfed embryos both in vitro and in vivo. Normal 

 ovules which have been arrested by injection of a solution of the Inhibitor 

 furnish an extract which Inhibits embryo growth when Injected Into 

 another young ovary which again develops an extractable Inhibitor. 



Why wide crosses between certain species cause the production of 

 ovular tumors and embryo inhibitors is not clear. It cannot be due to 

 Incompatibility of different kinds of genes from the two parents since 

 the same type of ovular tumors with embryo abortion had been earlier 

 found in ihc cross /\n X 2;; and 2/2 X ^n when both the cHploid and 

 tetraplold parents were In the same highly inbred line. Moreover a 

 similar production of ovular tumors is obtained, as shown by van Over- 



