550 SMITH: DIoSCOREA VILLOSA 
that of the cells at the six-celled stage. At the sixteen-celled 
stage, the embryo assumes one of two typical shapes, each of 
which has been observed in several cases: it may be either elon- 
gated (Fic. 7), or nearly globular (Fic. 8). The exact sequence 
of divisions cannot be followed with any certainty after the eight- 
celled stage, nor is it possible to trace the development of the 
organs of the full-grown embryo from particular cells. 
As the embryo develops a region is differentiated at the apex 
in which division is especially active, as indicated by the smaller 
size of the cells (Fic. 9, m). An elongation of the embryo, to- 
gether with the increased rate of growth in its apical region, 
renders it a little later roughly pear-shaped (Fig. 10). The sus- 
pensor is now definitely marked off from the rest of the embryo 
(FIG. 10, s), and reacts differently to stains in this and succeeding 
stages. Now, too, the dermatogen is differentiated except on 
the side from which the first secondary leaf is to develop (FIG. 10, 
1; compare also Fic. 11,2). An embryo of the age of that shown 
in Fic. 10 bears a fairly close resemblance to one of Lilium 
philadelphicum figured by Coulter (1897, pl. 34, f. 31). Solms- 
Laubach (1878) shows no embryo of Tamus communis or of 
Dioscorea pyrenaica at a similar stage of development. Older 
stages than that of my Fic. 11 are represented in his figures only 
by drawings of surface views of whole embryos. According to 
Solms-Laubach’s description the original terminal arch of the 
embryo becomes somewhat flattened and laterally displaced by 
the growth of one side of the embryo. From the flattened portion 
the plumule develops, surrounded by a ring-shaped wall of tissue. 
In my preparations of D. villosa, it appears clear that the lateral 
swelling shown at Fic. 11, 1, which probably corresponds to that 
described by Solms-Laubach, is the primordium of the first 
secondary leaf, whose further development is shown in FIGs. 
15-24. The growing point of the stem lies in the axil 
of the first secondary leaf, and it is obvious that both these 
structures are lateral in origin. The cotyledon (F1G. II, ¢) de- 
velops in a terminal position. In Tamus communis and Dioscorea 
pyrenaica, according to Solms-Laubach, first the cotyledon and 
then the sheath develop from opposite points in the ring-shaped 
primordium, and are hence of lateral origin. 
