Embryo and of that of the Grass in particular . 521 
ment of the sheath of the other cotyledon. No previous case of the kind 
has ever been cited, so that the authors have every reason to make the most 
of it. I say that on the current view their conclusions as to the facts hold 
good. But the view I prefer to maintain here is that in Dicotyledons, 
owing to the complete absence, in the majority of cases, of an epicotyle- 
donary axis, two cotyledons cannot possibly be present, but only a single 
terminal cotyledon which has deeply divided. This argument, therefore, 
also applies to the cases of Agapanthus and Cyrtanthus , although in these 
the appearance of two cotyledons is due 
to quite another cause, which constitutes 
the sole importance of the observation. 
That no phylogenetic significance 
attaches to the phenomenon can be 
deduced from the fact that only a single 
seedling of Agapanthus , according to 
Coulter and Land’s account, possessed 
two cotyledons (Fig. 8). To suppose 
that a unique instance, the only example 
ever known, would be likely to exhibit 
the character of the ancestors of the 
whole group is to my mind next door 
to an impossibility. 
If this had been the ancestral feature, 
instances would certainly be much more 
numerous, not only in this genus, but in 
others as well. 
The phenomenon must therefore re- 
present a new, aberrant departure, of 
progressive, not reversionary, nature. 
The same argument applies to Cyr- 
tanthus ; although the two cotyledons may occur as a normal feature of the 
genus (Fig. 9), yet the features of a single genus cannot be taken as in- 
dicative of the characters, whether modern or ancestral or both, of the 
whole class ; they are much more likely to be progressive and novel. 
That the embryos of one or two members of a modified order like the 
Amaryllidaceae, with its inferior ovary and other idiosyncrasies, would 
exhibit more ancestral characters than those of members of less modified 
orders is in the highest degree improbable. They are more likely, in 
agreement with the other advanced characters of the order, to show a 
progressive type of construction. 
One may securely conclude, therefore, on all these grounds that the 
extra structure, opposed to the cotyledon, in these two genera is without 
doubt a second cotyledon, the result of the very exceptional development of 
Fig. 9. Cyrtanthus sanguineus. A. Ex- 
terior view of young embryo. B. Longi- 
tudinal section through the centre of A. 
C. Exterior view of an older embryo. 
c = lamina of cotyledon ; s — sheath of 
cotyledon. (After Miss Farrell.) 
