156 Sargant . — The Reconstruction of a Race of 
Angiosperm ? If so, the Dicotyledonous pair must have been formed 
either by the fission of a single lateral member, or by the addition of a 
second cotyledon to that already in place. The second suggestion is clearly 
improbable : the first deserves consideration. Such a cotyledon as that 
of Tamus might conceivably divide, and the segments shift their position 
until they faced each other. The embryo of Nelumbium as described 
by Professor Lyon might represent a stage in this process, though — as 
we have seen — this is not his own interpretation of the facts. 
If, on the other hand, the Primitive Angiosperms had two cotyledons, 
and we treat the single cotyledon of Monocotyledons as a lateral member, 
it might be derived from the ancestral pair either by suppression of one 
of them or by the complete union of both. 
The embryology of the Pseudomonocotyledons is of the greatest 
interest in relation to this subject. Scattered here and there within 
Dicotyledonous genera are a few species which have but one cotyledon. 
Hegel maier ( 37 ) followed the development of the embryo in three of these 
species from the fertilized ovum to the ripe seed. They are Ranunculus 
Ficaria , Corydalis cava, Carum Bulbocastanum. 
Unfortunately the embryo of all three species is very little developed 
in the ripe seed. A long period of maturation is necessary before germina- 
tion can take place. Hegelmaier failed to secure the maturation stages, 
and his work was first completed in 1902 by the publication of a post- 
humous paper by Schmid ( 78 ), which carried the history of the embryos on to 
the epoch of germination. About the same time Sterckx ( 83 ) published 
a complete account of the development of the embryo in Ranunculus Ficaria. 
In each of these species the single cotyledon was certainly derived 
from two. They belong to dicotylar genera ; in other words, the ancestral 
Corydalis , Carum , and Ranunculus possessed two cotyledons. The single 
cotyledon in the aberrant species is an adaptation peculiar to themselves. 
Thus the change from a dicotylar to a monocotylar form has occurred in 
comparatively recent times — at any rate since the date of the original 
representative of each genus. Here, if anywhere, we might hope to trace 
the development of so well-marked a character in the young embryo. We 
should perhaps expect to find the rudiments of two cotyledons at first, then 
to trace either the development of the one and the arrest of the other, 
or their fusion into a single member. 
The actual process in all three forms bears a strong resemblance to 
the development of the embryo in Tamus. The cotyledon appears on the 
flattened apex of the pro-embryo as a peripheral ridge. At first it is 
circular, but it soon becomes crescent-shaped by the rapid growth of one 
side. The stem-apex is often late in appearing. It is always formed 
in the central depression outlined by the circular ring. Very soon the 
