334 BOTANICAL GAZETTE [November 



as possible that it may pass to the outer layer and thus keep this 

 in condition to continue the absorption of starch from the peri- 

 sperm. Only a careful chemical study of the cell-contents of 

 all these tissues during the progress of germination can decide 

 these questions definitely. 



The restriction of the functions of the endosperm to the pass- 

 ing on, during germination, of the food material without to the 

 embryo within has, so far as I have been able to learn, been 

 pointed out in Sauriirus cermius only (Johnson, 1900^ p. Zl^^^ 

 Humphrey (1896, p. 16) suggested that the single, aleurone- 

 containing layer of endosperm in Canna might have some such 

 function, but did not follow this out in germination. In the 

 germinating seeds of Canna I find that this layer of endosperm 

 persists till a late stage of this process as a sac about the haus- 

 torium, the relation being such that material from the perisperm 

 must pass through the endosperm to get to the embryo, and it 

 seems probable that it may play the same part as that suggested 

 in Peperomia and Heckeria. 



In many other families it has been shown, by Hartz (1885) 

 and others, that the storage tissue in the seed outside the embryo 

 is perisperm, and not endosperm, as stated in the older (and 

 many of the newer) books. This is true, for example, of the 

 Polygonaceae, Chenopodiaceae,Phytolaccaceae, Caryophyllaceae, 

 and others, and, contrary to the accepted authorities, there is 

 present in a considerable number of these forms that I have 

 studied a small amount of endosperm also. In all cases of these 

 perisperm-containing seeds w^hich were sprouted, the endosperm 

 was found to persist for some time during germination, and thus 

 to be capable of taking part in the transference of food material 

 to the embryo. 



Observations thus far made lead me to believe that in the 

 perisperm-containing seeds mentioned the embryo sporophyteof 

 the second generation is never nourished by the parent sporo- 

 phyte directly, but always through the intermediate gametophyte. 

 In general, then, we find that the food substance supplied to the 

 embryo by the nucellus may pass through the endosperm and be 

 stored in the embryo during the ripening of the seed, as m 



