504 SMITH: DIOSCOREA VILLOSA 
occurs, and that the branches fuse in pairs to form the phloem 
groups of the root. At the same time, a rearrangement of the 
xylem groups takes place, with the result that the protoxylem, 
which is external in the hypocotyl, becomes internal in the root. 
CONCLUSIONS 
Apparently the early divisions in the embryo of Dioscorea 
villosa have no very definite relation to the formation of organs 
or to the constitution of the embryo, since no important differences 
are found among the ripened embryos which ‘might correspond to 
the observed differences in the plane of the first division or in the 
shape of the proembryo. 
Widely accepted generalizations on the embryology of mono- 
cotyledons and dicotyledons have been based on cases such as 
Sagittaria and Capsella in which a filamentous proembryo is formed 
whose terminal cell gives rise to the cotyledon in monocotyledons 
and to the stem primordium in dicotyledons. However, as Coulter 
and Land (1914) have pointed out in the case of Sagittaria, the 
origin of organs from particular cells of filamentous proembryos 
has been assumed rather than proved. Moreover, a number of 
genera, both of monocotyledons and dicotyledons, have now been 
investigated in which the proembryo is massive. In several mono- 
cotyledonous genera, including Zannichellia (Campbell, 1897), 
Lilaea (Campbell, 1898), Sparganium (Campbell, 1899), Avena 
(Cannon, 1900), and sometimes Limnocharis (Hall, 1902), both 
cotyledon and growing point are found to originate from the 
terminal segment. However, a considerable amount of growth, 
resulting in a mass of tissue, must occur in the embryo before 
actual differentiation of organs takes place. If the positions of 
the primordia of the cotyledon and stem on this mass of tissue 
can be said to be terminal or lateral, this difference might serve 
as a basis for the distinction between monocotyledonous and 
dicotyledonous embryos. Unfortunately, it is at just this period 
of development that gaps in embryological studies are usually 
found. If position of origin be considered the criterion, the 
embryo of Dioscorea villosa is strictly monocotyledonous. 
Lyon (1901) has advanced the view that an area extending 
entirely around the axis of the embryo is potentially cotyledonary. 
