SMITH: DIOSCOREA VILLOSA 555 
The maximum development in this area takes place at the point 
or points that are in the most favorable position to function; in 
monocotyledons maximum development occurs on only one side, 
while in dicotyledons equal growth occurs at two points dia- 
metrically opposite each other. This notion has been further de- 
veloped by Coulter and Land (1914) in connection with their 
work on some of the South African Liliaceae, particularly A ga- 
panthus. Coulter (1915) even extends the generalization to gym- 
nosperms. According to his view, a variable number of primordia 
appear on the cotyledonary ring; one, two, or more of these 
develop, the growth of the others being checked, principally by 
the growth of other organs of the embryo. As I have pointed out, 
no cotyledonary ring was observed at any time in the embryo of 
Dioscorea villosa, although such a structure has been described for 
D. pyrenaica and Tamus communis (Solms-Laubach, 1878). The 
two primordia that do appear in Dioscorea villosa develop into 
the cotyledon and the first secondary leaf respectively. The only 
way in which this case can be made to fit Coulter’s theory is by 
assuming that the vigorous development of the first secondary leaf 
has entirely checked the development of the cotyledonary zone 
except at one point. However, the figures of Solms-Laubach 
(1878) indicate that the development of the first secondary leaf 
is just as vigorous in Tamus. I have no explanation to offer of 
the entire absence in D. villosa of a sheath covering the first 
secondary leaf, which is so prominent a feature in the embryos of 
other Dioscoreaceae. 
There is so much diversity of opinion in regard to the phylo- 
genetic significance of seedling anatomy that it seems hardly worth 
while to discuss the question to any extent in connection with 
Dioscorea villosa. Miss Sargant (1903) considers the tetrarch root 
such as is found in D. villosa primitive; however, in her opinion, 
this form of root structure is associated with the early development 
of the plumule which is characteristic of climbers. On the other 
hand, Hill and De Fraine (1913) consider diarch root structure 
primitive, but think that the root structure in any given case 1s 
largely dependent upon the size of the seedling. The seedling of 
D. villosa is small so that the formation of a tetrarch root cannot 
be accounted for on the basis of the size of the seedling. The 
