Primitive A ngio sperms . 1 49 
soon becomes bi-lobed, and then gives rise to a pair of very long 
cotyledons (56). 
Here there are three variations on the Capsellct scheme : the absence of 
a suspensor, the size of the undifferentiated embryo, and the first appear- 
ance of the cotyledons as a single crescent rather than as two distinct 
rudiments. Of these anomalies, the last is the most important. Many 
instances are recorded of embryos without a suspensor. It is not very 
uncommon to find an embryo which attains considerable size before the 
cotyledons are indicated. Hegelmaier figures an undifferentiated embryo 
of unusual size in Helleborus foetidus (37, PI. II, Fig. 19), and Johnson 
describes a similar case in Heckeria (48, pp. 328-9 and Fig. 39). 
Professor Lyon considers that in Nelumbium we can trace the develop- 
ment of two cotyledons from a single terminal member, which is derived 
from a similar member in the embryo of the Primitive Angiosperms and is 
homologous with the single cotyledon of Monocotyledons. He faces all 
the consequences of this hypothesis. The ancestral member, being terminal, 
is not a phyllome but an organ sui generis — perhaps derived from the foot 
of Vascular Cryptogams, and homologous with the sucker in seedlings 
of Gnetum and Welwitschia (Bower, 13, 14). The derivatives of such a 
structure have, of course, no claim to be considered as true leaves. Hence 
the leaf-like structure of the cotyledons in most Dicotyledons and some 
Monocotyledons must on this view be attributed to their adaptation to the 
functions of leaves. 
Comparison of the anatomical structure of green cotyledons with that 
of true leaves shows a remarkable resemblance between them. Professor 
Ramaley’s descriptions, for example (66, 67), show identical structure in 
epidermis, stomata, palisade and spongy tissue, and in the details of the 
vascular bundle. The differences on which he insists appear to me of 
minor importance. There are differences in the distribution of stomata, 
in the number of rows of palisade cells, in the distribution of the 
bundles. 
Schlickum (77), after careful comparison of the cotyledon with the first 
leaf in a number of Monocotyledons, came to the conclusion that they were 
homologous with each other. I am confirmed in the same opinion by 
examination of a series of preparations most kindly made for me by my 
friend Miss Berridge. She has mounted sections from the cotyledon and 
first leaf of various Dicotyledons, side by side, and the resemblance in 
structure is very clear. 
In short, to deny that cotyledons are homologous with leaves introduces 
new difficulties into a subject already sufficiently obscure. Such difficulties 
are the logical consequences of the three assumptions on which Professor 
Lyon's theory is based : that the single cotyledon of Monocotyledons is 
terminal, that a terminal member cannot be a leaf, and that both cotyledons 
