298 BOTANICAL GAZETTE [October 



give the exact origin of this tissue, but says: ''the nuclei resulting from 

 the first divisions of the endosperm nucleus take position at either 

 end of the sac^ leaving, however, a few to form a thin parietal layer 

 between. At the antipodal end, cell formation with walls begins 

 at once, and a number of large cells form a tissue which stands out 

 conspicuously in the cavity of the sac, which othenvise contains only 

 a few free endosperm nuclei." Tillandsia is considered a near rela- 

 tive of the Pontederiaceae, and it does not seem improbable that the 



r 



origin of this tissue is the same in both cases. A difference to be 

 noted is that in Tillandsia the basal endosperm tissue, according to 

 Billings, is cellular from the first. 



The appearance of the cross-walls at the low^er end of the sac at 

 the first division of the endosperm nucleus will at once recall a similar 

 condition in Sagittaria (Schaffner 13), Potamogeton pauciflorus 

 (WiEGAND 16), and Ruppia (Muhbeck 12). In these cases, how- 

 ever, the nucleus in the lower chamber does not divide to form a tissue, 

 as in the Pontederiaceae.' 



joliosus 



same 



pposed 



The 



wall that cuts off the lower end of the sac is derived from the division 

 of one of the first four nuclei of the sac, its sister nucleus forming the 

 three antipodals which occupy the little pocket below. The author 

 thus regards the large nucleus which occupies the lower end of the 

 sac as an antipodal, as it is the sister nucleus of the lower polar. No 

 upper polar nucleus is supposed to be formed and the endosperm 

 develops from the lower polar. This whole procedure is so peculiar 

 that it seems open to some doubt until confirmed. 



em 



general quite normal, but in two or three cases a large nucleus was 

 discovered in the lower end of the sac, cut off above by a wall. He 

 thinks it may have originated in the same way as I have described for 

 Pontederia. 



In Castalia odorata^ C. ampla, C. puhescens^ Nymphaea advena, 

 Brasenia purpurea, and Cahomha piauhiensis, Cook (5, 6) has found 

 that an antipodal haustorTura, cut off from the rest of the sac by a 



I In Sagittaria, according to Schaffner, it may divide once or twice. 



V 



