Bot.— Vol. I.] CAMPBELL— SPARGANIUM. 317 



fusion, but in reality it is more nearly a case of fragmenta- 

 tion. Hegelmaier (1885) figures a number of similar cases, 

 and they have been often observed, but so far as I know 

 the origin of the secondary nucleoli has not been studied. 



As the seed ripens, numerous very large crystalloids are 

 formed, at least in S. simplex, in the endosperm-cells sur- 

 rounding the embryo (figs. 63, 64). These are beauti- 

 fully fixed by the action of alcoholic corrosive-sublimate, 

 and stain readily with anilin-safranine. They vary in shape, 

 the most perfect ones being rhombic in outline. Often they 

 were aggregated in large masses, but usually it was evident 

 that these masses were composed of separate crystals. 

 Sfarganium simplex may be recommended as an admi- 

 rable subject for demonstrating these bodies. The material 

 collected of the other species was either too young or had 

 not been treated with the proper reagents, so that I cannot 

 speak as to the occurrence of these crystalloids in the endo- 

 sperm of the other species. They have, however, been 

 noted in S. ramosum Sachs (1887). 



No hard testa is formed about the seed of Sparganium, 

 although this outer part of the nucellus and the integuments 

 persist until a late stage, and probably permanently. Eng- 

 ler's statement (1889) that the endosperm is only separated 

 from the pericarp by the integuments needs confirmation. 

 In the oldest stages examined by me, there were several 

 layers of nucellus cells still evident. The development of 

 the hard inner pericarp is readily followed. The cells early 

 have thickening layers deposited upon their walls, and these 

 later become very thick and have deep pits developed in 

 them. This sclerenchyma is rather better developed in 

 S. Greenii than in S. simplex, and sometimes in the former 

 the thickening is so great as to almost obliterate the lumen 

 of the cell. 



Hegelmaier (1874) has studied in S. ramosum the devel- 

 opment of the curious structure to which he applies the 

 name "Samen-deckel." The structure of this in S. sim- 

 plex agrees closely with his account. It is formed from an 

 enlargement of the micropylar part of the integuments, 



