FISHER: SEED DEVELOPMENT IN PEPEROMIA 223 
ail 
by Juel](37), and in Smilacina stellata by McAllister (45). The 
case of Smilacina is particularly interesting here since the four 
megaspores show various transitions between the typical axial 
row and the tetrahedral arrangement. Moreover, after the 
megaspores are formed and unmistakably separated by walls, 
the walls disappear. The megaspore nuclei now become the first 
four nuclei of the embryo sac, and at this stage they assume the 
tetrahedral arrangement very similar to that in the four-nucleate 
embryo sac of Peperomia. It is not only in Peperomia that the 
first four nuclei of the embryo sac are arranged tetrahedrally, but 
this is likewise true in all genera in which sixteen free nuclei are 
formed in the eels: Ae see in Gunnera (Schnegg, 64; 
Ernst, 20, 21; Modil ; , 59), in Sarcocolla (Stephens, 
70), in Brachysiphon (Stephens 70), in Penaea (Stephens, 70), and 
in Euphorbia (Modilewski, 49, 51, 52). 
(4) The fact that the tetrad is always complete in number of 
nuclei in Peperomia makes the homology seem rather more 
probable in this genus than it is in plants in which the axial row 
is incomplete, as for example, in one consisting of three cells 
instead of four. 
(5) Eleven species of Peperomia have been shown to have 
sixteen nuclei in the mature embryo sac. In addition to the six 
here investigated are: P. pellucida (Campbell, 6), P. hispidula 
(Johnson, 34), P. Sintenisii, P. arifolia, and P. Ottoniana (Brown, 4). 
No species has been found to have any other number. As Brown 
(4) points out, the presence, in the mature embryo sac of Pepe- 
romia, of the larger number of nuclei than in the sac of ordinary 
Angiosperms, is in harmony with the view that more megaspore 
nuclei are concerned in the formation of the Peperomia type of 
embryo sac. 
(6) The fact that all four nuclei divide, or germinate, can not 
militate against the theory that they are homologous with mega- 
spore nuclei, because there are cases in which megaspores are 
undoubtedly formed and in which they all germinate—that is, 
the nuclei all divide. This is true, for example, in Crucianella 
(Lloyd, 41), in Smilacina stellata (McAllister, 45), and sometimes 
in Epipactis (Brown & Sharp, 5). And then there are interme- 
diate cases in which one or more of the non-functional mega- 
