128 
JOHN T. BUCHHOLZ 
is an organ of competition, the structure upon whose merit the selection of 
the surviving embryo depends, while in Angiosperms it is usually a chain 
of comparatively unelongated cells which carries the embryo into the central 
portion of a very soft endosperm. Ginkgo, which has a relatively short 
suspensor, seems to be about the only Gymnosperm which occasionally 
matures more than one of its several embryos, and even here Cook (ii) 
found this occurring in only two percent of the seeds. It is much more 
infrequent than this in Pinus, and probably in all other Gymnosperms. 
It is clear that the erratic character of polyembryony in Angiosperms should 
not influence our judgment in deciding on the phylogenetic value of this 
feature in Gymnosperms. 
Simple polyembryony is found among cycads, since these have constantly 
several archegonia; but none of them are known to possess cleavage 
polyembryony with its accompanying phenomena, so that polyembryony 
is a feature too uniform to be of value in showing the affinities of the genera 
within the Cycadales. However, the ordinary anatomical characters of 
the proembryo have been used very effectively for this purpose by Chamber- 
lain (6). 
Among conifers, much more precise comparisons of embryonic develop- 
ment are possible when we once understand these greater variations of the 
proembryo and of the early embryo brought about by cleavage polyembry- 
ony; stages of development which one would otherwise expect to find 
rather conservative and uniform. The peculiar splitting of the embryo, 
which was introduced somewhere in the ancestry of Pinus, has persisted for 
some time in the evolution of the conifers, and was suppressed or eliminated 
by a number of distinct methods. The rosette embryos, rosette cells that 
abort, and many other early embryo features are only results of cleavage 
polyembryony. Together with the apical cell (doubtless a Pteridophyte 
character which has persisted), these features give us a splendid array of 
consistent characters to serve as an index to the natural classification of the 
groups. 
Abietineae 
In a very recent paper on "Polyembryony among Abietineae" (4), I 
have made a number of comparisons of the embryos of this group, which 
are summarized with greater accuracy in the diagram of figure 9. While 
the proembryos of all Abietineae (with the possible exception of Pseudot- 
suga) appear to be identical (4), the condition of cleavage polyembryony 
(Cl.p) is restricted to Pinus, Cedrus, and Tusga. Abies balsamea displays 
an occasional embryo with cleavage polyembryony, but normally it has 
only simple polyembryony and is similar to Larix, Picea, and Pseudotsuga 
in this respect. 
It will be seen that the apical cell (A) has the same range of distribution 
within this group. Fusion of embryos seems to be the method by which 
