440 Ganong . — The Comparative Morphology of 
extreme desert situations than the Cylindropuntiae, some of 
them growing even in woods. Many lines of evidence show 
that the Cylindropuntiae are the stem-group from which the 
Platopuntiae are an offshoot. 
Turning to the embryos, we find that all have a slender, 
non-succulent hypocotyl, and well-developed, though fleshy, 
cotyledons approximately the length of the hypocotyl. With 
respect to the exact form of the cotyledons, however, two 
types are distinguishable — some, such as O. bernardina, 
echinocarpa , serpentina , and tetracantha , have equal tapering 
cotyledons, triangular-cylindrical in section (Fig. 2 c), be- 
coming flatter with age, shorter than the hypocotyl ; while 
others, O. vulgaris , missouriensis , Ficus-indica , tortispina , 
Rafinesquii , Engelmanni occidentalism phaeacantha, and other 
Platopuntiae, have them flatter and more leaf-like (Fig. 5 c) y 
unequal, and usually longer than the hypocotyl 1 . This dis- 
tinction of the thick leaves for the Cylindropuntiae, and the 
thinner for the Platopuntiae, holds in all species that I know 
of, with the following exceptions : — as figured in Engelmann 
O. Whipplei seems to have flat cotyledons, but as the plumule 
is shown developed, and since even in Cylindropuntiae they 
flatten with age, this may not be a real exception ; Opuntia 
basilaris is, however, a real and marked exception, for in 
Lubbock’s figures and my own, the cotyledons are distinctly 
of the Cylindropuntiae type, though this species belongs to 
the Platopuntiae 2 . As however O. basilaris is an extreme 
desert-species, it suggests that the form of the cotyledons is 
determined chiefly by the direct environment, the more 
condensed form occurring in the more extreme deserts, and 
the flatter and more leaf-like in moister climates, where the 
natural tendency of the plant to spread as much leaf-surface 
as possible is allowed more freedom to produce larger and 
flatter leaves. This can be tested by observing on the one 
1 The extreme of this flattening is probably reached in 0. brasiliensis, as shown 
by Schumann’s figure on PL 61, Flora brasiliensis. 
2 O. basilaris is in several respects a remarkable form ; particularly noteworthy 
is its profuse branching from near the base. My seedlings of 2 cm. length (in the 
var. ramosd) showed no trace of it, though a larger one showed one branch. 
