44 2 Ganong . — The Comparative Morphology of 
develop by removing the epicotyl in O. bernardina and one 
or two other species. A remarkable feature occurred in one 
seedling of O. echinocarpa (Fig. 3 c). One of the cotyledons, 
possibly in reality two congenitally united, was forked near 
its tip, and in the fork bore a bud producing hairs and spines 
of an altogether normal sort, which persisted long after the 
epicotyl developed. Such a case is difficult to explain on the 
basis of a formal morphology, for leaves, which the coty- 
ledons undoubtedly are, do not produce buds normally in this 
family. 
The development of the epicotyl is simple, and is as 
described earlier for the family in general. From the first, 
the general mode of formation of leaves and axillary buds 
agrees with what is found in the adults, except that the 
characteristic bristles (Borsten), the remarkable hair-sheath 
of the spines in Cylindropuntiae, and the nectar-glands 
formed of metamorphosed spines are all absent from the 
earliest clusters and appear later, though I have not de- 
termined exactly when nor how. 
The epicotyl in all species of this genus is at first cylindrical. 
In O. vulgaris and other Platopuntiae it is only when the 
epicotyl is at least a centimetre long that it begins to flatten, 
and then it is parallel to the faces of the cotyledons. It would 
be of much interest to determine whether in the seedlings, as 
in the young shoots described by Goebel 1 , the flattening is 
dependent upon the presence of light, and hence is a phe- 
nomenon of irritable response in which the growth-effect has 
not yet become hereditary. 
The epicotyl in its development forces apart the cotyledons, 
which however retain a connexion with one another at their 
bases, so that the epicotyl seems to rise from a sort of sheath. 
When the epicotyl and the axillary buds of the cotyledons 
are removed, the hypocotyl grows far above the normal size, 
but the cotyledons wither as they do under normal con- 
ditions (Fig. 2 e). 
1 Flora, lxxx, 98. 
