the Embryos and Seedlings of the Cactaceae. 457 
Very closely related to Cephalocactus is Lophophora , some- 
times made a distinct group. It contains the species E. Wil- 
liamsii. The embryos (Fig. 39) are nearly globular, and 
have very small rounded cotyledons, in this character re- 
sembling those of E. horizontalonius. The epicotyl in 
E. Williamsii bears rounded tubercles like those figured 
(Fig. 38 d) for E. horizontalonius , though less prominent, and 
bears small feathered spines. The spines in the adults, 
though seeming to be absent, are really present in this 
species, as I have elsewhere shown 1 , and the species really 
represents a form like a young horizontalonius (compare 
Engelmann’s figure) without the large spines, and there 
can be no doubt that it is an Echinocactus , closely related 
to the division Cephalocactus . 
Another division often given distinct generic rank is 
Astrophytum , of which I have studied embryos of A. myrio- 
stigma and A. capricornis (Figs. 40, 41). Embryos of these 
two are much alike, and very different from those of any 
others I have studied. The hypocotyl is swollen, but more 
on one side than the other, which is due to the position 
in the seed ; and the root, for the same reason, is somewhat 
on one side. The cotyledons are rather narrow but pointed, 
though less so than in some other divisions. In the epicotyl 
the first clusters at right angles to the cotyledons are very 
prominent, marking the immediate beginning of the ribs, and 
making the plant, as seen from above, almost quadrangular. 
The ribs hence develop earlier in this species than in any 
other I have seen in the family. Two others form over the 
cotyledons, and these four form the starting-point for four 
strong ribs which, by the forking of one, soon become five, 
which distribute themselves over the circumference. In their 
later growth these two species become less alike. The clusters 
of characteristic remarkably pitted hairs occur upon young 
seedlings. The spines in the first-formed clusters are well- 
developed, but later clusters have them very small. 
In the important division Stenocactus I have had no 
1 Flora, op. cit. p. 72. 
I i 2 
