142 
JOHN T. BUCHHOLZ 
Ephedra 
I cannot refrain from including a mention of the Gnetales in this dis- 
cussion, for these have polyembryony, much as have the Coniferales. In 
the embryo of Ephedra (22, 40), for example, cleavage polyembryony is 
the regular occurrence. The free nuclei which are concerned in the organi- 
zation of the embryo initials do not all descend to the bottom of the egg, 
as in Pinus, but remain in comparative independence of each other, and 
this condition of cleavage polyembryony is very striking. Judging from 
the published accounts and figures there is not much probability that an 
apical cell stage is to be found, and this would furnish us an instance in 
which the apical cell is eliminated and cleavage polyembryony is retained. 
It seems to me that this feature of cleavage polyembryony very definitely 
marks the Gnetales as having been derived from the conifers rather than 
from the cycads. 
There is perhaps no better way for me to summarize these many com- 
parisons of embryo development made during the course of this discussion, 
than to use the method of a diagram, as was done for the Abietineae. 
Such a scheme of phylogeny represents only the present status of our 
knowledge concerning embryogeny and is not intended to be accurate to 
the last details, for many of these details have been too inadequately 
described. It doubtless represents fairly accurately the affinities of the 
larger groups. 
In this diagram (fig. 89), the ranges of the various embryonic features 
which have been discussed are plotted in light lines, while the (probable) 
phylogenetic connections suggested by embryogeny are shown in the heavy 
lines. The range of the apical cell (A), and cleavage polyembryony (CLp) 
are marked with some uncertainties, but in general it appears that the latter 
is more limited in its range than is the apical cell. The archegonial complex 
is found in all forms surrounded by the circle Cx. Rosette embryos are not 
plotted, but the forms having rosette cells are marked as follows: R, rosette 
cells ; (R) , occasional rosette cells ; R-em, rosette cells that develop embryos ; 
R-{em), rosette cells that occasionally give rise to embryos, etc. The 
rosette embryos are therefore also primitive features that soon vanish, 
followed by the disappearance of the rosette cells, which are the last vestiges 
of cleavage polyembryony. The embryo cap (Emb.cap) found in Araucaria 
and Agathis is an advanced feature found also in the forms somewhat 
involved in the evolution of the Araucarineae. 
It will be found that in general the cotyledon number decreases as we 
proceed upward in every direction away from Pinus. Though the number 
of cotyledons is somewhat variable throughout, this number is reduced 
sooner or later among conifers (3) to the limit of three or two. 
I have included the range of one well-known taxonomic feature in this 
diagram, namely, the spur shoot (Sp.S) . Sciadopitys with its double needles 
