(376) 
it seems more reasonable to regard the ovary consisting of two 
carpels as derived by reduction than as representing a more 
primitive type. Inthe development of Svcyos there is distinct 
evidence of such reduction (see text-figure z). The earlier 
stages of the development of the gynoecium are marked by the 
appearance of three equidistant elevations on the torus, which 
by their position and behavior, must be regarded as incipient 
carpels. We have here a reduction in the number of func- 
tional carpels as shown by the formation of a single placenta, 
a case therefore quite similar to what might appear should 
one of the three placental lobes of a form like Bryonopsis or 
Coccinza bear a single ovule and the other placental lobes 
become reduced or obliterated. While the ovules of Bryo- 
nopsis and Coccinza are horizontal, it is quite probable that if 
they were reduced to one as in the case of Svzeyos its position 
would be changed to the vertical. 
Cyclanthera ina comparison with Stcyos would represent a 
condition intermediate between that of Sycyos and a radially 
symmetrical form such as Bryonopsis or Fevillea. If in the 
ovary of Cyclanthera explodens the two rows of ovules should 
be replaced by a single ovule in vertical position, the condi- 
tion would be entirely similar to that of Szcyos. In the incipi- 
ency of the gynoecium three lobes may be recognized, though 
not so distinctly as in Szcyoes, and these soon become obliterated 
in the stigma. It seems quite evident that the ovary of 
Cyclanthera explodens has also been derived from a tricarpel- 
lary form. By the abortion of two of the three placental 
ridges in a form like Bryonops7s a condition would result in 
all essential respects similar to that existing in the ovary of 
Cyclanthera explodens. 
The growth of the placenta and the origin of the ovules 
are matters of interest and are worthy of attention. Evi- 
dently the simplest form of placentation is that in which 
numerous seeds arise on parietal placentae in a unilocular 
ovary (Alsomztra). From this condition it is easy to trace the 
gradual ingrowth of the edges of the carpels and the develop- 
ment of the various features of placentation, already described 
