130 
JOHN T. BUCHHOLZ 
In Pseudotsuga, the uppermost walled cells of the proembryo elongate 
to form the suspensor, and therefore no aborting rosette cells are found. 
It is interesting to note how these embryo characters relate themselves 
to the results of the anatomists. The distribution of the resin canal char- 
acters described by Jeffrey (19, 20) are included in the diagram (Res.c), 
cutting Pinus off from Cedrus, although these two genera are very similar 
on the basis of embryongeny and also in the spur shoot characters used by 
Engler and Gilg (16) and others as a basis for subdividing the Abietineae. 
On the other hand, the distribution of the apical cell, cleavage polyembry- 
ony, and rosette embryos cuts Pinus off from the rest of the Pinae of Jeffrey. 
The intermediate position of Cedrus between Pinus and Abies was pointed 
out by Jeffrey (19) and by Chrysler (7) on the basis of medullary rays. 
Pinus has such an array of primitive features, both embryological and 
anatomical, that there can no longer be much doubt of its primitive position 
in any natural phylogeny of Abietineae. 
Other Coniferales. 
In the light of the foregoing, let us now consider the affinities of the other 
Coniferales as revealed by their embryo development, where this is sufff- 
ciently known. The interpretations which I shall give, while not generally 
those given by the investigators to whom the particular work is credited, 
do no violence to the facts as they have been described, and for any errors 
due to these interpretations I assume full responsibility. 
I will first" present the line of evolution between Pinus and Araucaria, 
based on embryogeny. There are not many known types that fit in 
between these two very different methods of embryo development, but we 
have steps enough to give us a definite clue to the possible lines of advance, 
and to suggest the kind of embryogeny from which that of the araucarians 
was derived. 
The number of free nuclear divisions is increased and a larger number of 
embryo initials is formed, resulting for a time in greater cleavage poly- 
embryony. Advancing a little further, we find that only a small portion 
of these potential embryo initials function. Cleavage polyembryony is then 
eliminated by the modification of the terminal portion, which organizes 
a cap and prevents the embryos from splitting apart. The apical cell also 
persists for a time in this line of evolution, and it appears that the formation 
of this cap over the advancing end of the embryo is the thing which does 
away with the apical cell stage. Here again we have probably had the 
apical cell stage and cleavage polyembryony eliminated by the same device. 
Even if we take the position that Pinus did not directly give rise to 
araucarians, it is apparent that they had a common origin and were derived 
from a condition of cleavage polyembryony. Pinus has remained in this 
condition with very little modification, while the araucarian embryogeny 
has become specialized, and has completely eliminated cleavage polyembry- 
