STUDIES OF THE DEVELOPMENT OF PEPEROMIA HISPIDULA 383 
the latter is composed of three carpels (Johnson, 1902, 1910). Sauru- 
rus also (Johnson, 1900^?), which is still regarded by Engler as the most 
primitive of the Piperales, shows a similar compound ovary. In other 
words, the carpel also seems simplified through specialization by re- 
duction from the more primitive type characteristic of the Piperaceae 
in general. The mature carpel or fruit is thin and of few layers of 
tissue as compared with the Pipers studied. Not the slightest evi- 
dence has been found that any floral envelopes were ever present in 
this plant. The above-mentioned evidence points to this and other 
Peperomias as being forms with flowers more specialized than those of 
the related Pipers, 
In the single ovule of Peperomia, we have, as has been pointed out 
by Fisher (1914, p. 148), another character which cannot be regarded 
as primitive by those accepting any of the current views of the re- 
lationship of the Piperaceae (see Engler and Prantl, p. 189; Hallier, 
1905; Lotsy, 191 1, pp. 487 ff. ; Engler and Gilg, 1912, pp. 157, 158). 
It must of course be kept in mind that while the number of ovules per 
carpel is rather constant in some families there are others like the 
Rosaceae where the number may vary from one to several. 
3. The development of arches porium, megaspores and cell plate. — 
The hypodermal primary archesporial cell is usually single and by its 
division gives rise to a definitive archesporial cell that becomes the 
mother-cell of the embryo sac, and above this a tapetal cell that forms 
several layers of tapetal tissue in the mature seed. The definitive 
archesporial cell and its nucleus enlarge to double their original size. 
The nucleus after showing a typical synapsis undergoes 2 mitoses by 
spindles perpendicular to each other giving rise thus to four, probably 
haploid, nuclei arranged at the points of a tetrahedron within the 
nearly spherical embryo sac. Delicate cell walls then arise from cell 
plates of the usual type, forming thus 4 tetrahedral megaspores, com- 
parable in appearance and origin with the microspore tetrad of this 
species, or, except for the delicate character of the walls, with the 
megaspore tetrad of Selaginella. The whole development of these 
4 cells, including the chromosome history of the nuclei as far as fol- 
lowed, shows them to be the exact equivalents morphologically of the 
megaspores, which in most angiosperms are arranged in an axial row 
(see Coulter, 1908, p. 364). The unusual arrangement of megaspores 
in a tetrad in this Peperomia is probably to be associated with the 
somewhat rounded form of the spore mother cell, and this depends, 
