44 BOTANICAL GAZETTE [jaxuar 



other hand, the early presence of a number of sporogenous cells in 

 the nucellus appears to ally it with the two genera mentioned, both 

 of which exhibit a similar peculiarity. 



The development of endosperm in the aroids is usually character- 

 ized from the begmning by the formation of prominent cell walls. 

 This characteristic has been noted by Campbell (/. c.) in Anthurium, 

 Nephthytis liberica^ and Lysichiton, and by myself in Nephthytis ^ 

 Gravenreuthii. Arisaema is an exception in that the first formation 

 of endosperm is by free nuclear division, although cell walls are soon 

 developed. In this respect it resembles Diefjenhachia daraquinianaj 

 whose embr}'0 sac becomes quite filled with endosperm nuclei before 

 the cell walls begin to make their appearance. The resemblance of 

 the suppressed axis to that of Dieffenbachia has already been men- 

 tioned. The anatropous character of the o^'ule in the latter genus, 

 the fact that each of the two o\ailes is inclosed in a separate carpel, 

 and the fact that there is only one archesporial cell, and that the pri- 

 mary parietal cell originates at its inner instead of its outer end, serve 

 to draw a broad distinction between it and Arisaema, in spite of the 

 resemblance in endosperm formation. The peculiarities mentioned 

 would probably indicate that Dieffenbachia is a somewhat less primi- 

 tive form than Arisaema, and one whose development has proceeded 

 along different lines. The presence of staminodia among the pistillate 

 flowers also appears to bear out this view. 



Arisaema, therefore, appears to be a more highly developed form 

 than Spathicarpa, Aglaonema, and Nephthytis, but somewhat more 

 primitive than Dieffenbachia, and much more primitive than Anthu- 

 rium, Spathycma, Orontium, and Acorus, since all the latter have 

 perfect flowers. 



\Mien the foregoing was written, the wxiter was not aware of the 

 existence of a paper by Mottieil^ on the development of the embryo 

 sac in Arisaema. A comparison of the two papers, however, shows 

 that they do not cover the same ground, since Mottier Ihnited his 

 mvestigation to the earlier phases, and did not discuss the development 

 of the endosperm, the growth of the embr}'o, or the possible affinities 

 of the species. 



Geeenkelo, Iowa 



» Mottier, D. M., On tKe development of the embryo sac of Arisaema iriphyJhmu 



Box. G.\ZETTE 17:258-260. iiL 18, iSQ2. 



