1907] COKER— DEVELOPMENT OF SEED IN PONTEDERIACEAE 297 



r 



much thicker than in the tissue above {jig. i8). As the growth proceeds 

 their nuclei become very irregular, stain deeply, and show a tendency 

 to fuse together. The protoplasm also becomes much denser than 

 that of the upper endosperm {fig. i8). In fig. ig are represented 

 several cells of the antipodal endosperm more highly magnified; 

 they are from a stage slightly later than represented in fig. i8. 



As the maturity of the seed approaches, and the upper endosperm 

 begins to be filled with starch, the cell contents of the lower endosperm 

 show signs of disintegration. The nuclei become deeply staining 

 and irregular masses, and the protoplasm loses its distinctive structure 

 and begins to disappear. No starch was found in these cells at any 

 time, and at the maturity of the seed they appear, like the cells 

 immediately above them, empty and somewhat crushed at the tip of 

 the embryo. 



The number of cells in the basal endosperm may become quite 

 large. I have counted as many as forty-five in a single section, and 

 there must have been in that case at least tw^o hundred. 



The seed coats of Pontederia are very much like those of Heteran- 

 thera, except that the outer integument is three cell-layers thick, with 

 the cells of the two outer layers thin and boxlike, and those of the 

 inner of the same height throughout, giving a smooth, instead of 

 ridged, surface to the seed. 



EICHHO^NIA 



My material of Eichhornia was not sufficient to follow the develop- 

 ment of the basal endosperm, but its presence is shown in my sections 

 at several stages. It seems to be similar in every way to that of Ponte- 

 deria and Heteranthera. The structure of the seed coats is exactly 

 as in Heteranthera, The same figures would do for both. 



CONCLUSION 



In looking over the literature of the subject I find but a single 

 case where a tissue similar to this has been described, and there its 

 origin has not been determined in detail. Billings (i) has found 

 that in Tillandsia usneoides there is formed in the lower part of the 

 embryo sac, Just above the antipodals, a tissue which in structure and 

 position much resembles that of the Pontedcriaceae. He does not 



