Campbell: The embryo-sac of Pandanus 217 



believes that the single sporogenous cell in Pepero^iiia and Gunucra 

 is less primitive than the four cells found in so many angiosperms, 

 but he does not believe that this suppression of the divisions Is In 

 any way associated with the peculiar characters of the embryo-sac 

 itself. He thinks that the sixteen-nucleate embryo-sac should be 

 recognized as an independent type not in any way derived from the 

 ordinary eight-nucleate sac, and this we believe will really prove 

 to be the case. Whether the ordinary eight-nucleate type has been 

 derived from this sixteen-nucleate form may be questioned ; but 

 the condition in Giutiiera macrophylla^ which is almost intermedi- 

 ate in character between Peperomia and the ordinary angiosper- 

 mous type, makes this not at all improbable. I believe we are 

 justified in assuming that the condition in Gunnera inacropJiylla 

 really does represent a condition nearly intermediate between the 

 very primitive condition found in Peperomia and the familar eight- 

 nucleate sac of the typical angiosperms. The recent announce- 

 ment of still another sixteen-nucleate sac in the Penaeaceae may 

 be referred to again in this connection. 



The peculiar embryo-sac of Pandanus cannot be explained on 

 the theoiy of reduction from the typical eight-nucleate form. The 

 embryo-sac in this case is one of three sporogenous cells, and even 

 if we admit that it is composed of two megaspores, in one of these 

 the nuclei are twelve in number instead of eight, while the other 

 has only two. 0( the forms hitherto described, Peperomia hispi- 

 dula most nearly resembles Pandanns in the arrangement of the 

 nuclei in the embryo-sac. 



Summary of results 



1. The pistillate flower of Pandajius may consist of a single car- 

 pel, or of several loosely united carpels. Each carpel bears a single 



anatropous ovule. 



2. The primary sporogenous cell is separated from the epider- 

 mis of the nucellus by several layers of parietal or tapetal cells, 

 which are presumably derived from the division of a single tapetal 

 cell. 



r 



3. The primary sporogenous cell divides transversely into two 

 cells, of which the lower, and larger one, gives rise at once to the 

 embryo-sac; the upper cell divides by an anticlinal wall again into 



two equal cells. 



