1918] BUCHHOLZ—PINUS 219 
In his work on Cephalotaxus STRASBURGER (38) shows in pl. 19, 
fig. 53, what is possibly a rosette embryo group, although he does 
not refer to it in the text. It is of interest in this connection to 
note also that the Cephalotaxus embryo has a cap which associates 
it with Araucaria (4, 38) and Agathis (12). All of this suggests 
another possible line of advance from a Pinus-like ancestor, through 
intermediate forms like Cephalotaxus, and a culmination in the 
embryo of the araucarian type. Thus it looks as though nearly 
all the embryos of Coniferales may be derived from an ancestor 
with cleavage polyembryony and an apical cell like Pinus, differ- 
entiating into the several more or less distinct lines of specializa- 
tion. This is a strong argument in support of the theory that 
Pinus is a very primitive and ancient genus. 
PoLycoTyLEeDoNy.—If polycotyledonous gymnosperms have 
been derived from dicotyledonous ancestors, one would expect that 
in the ontogeny of the cotyledons 2 primordial zones would 
first appear, and these 2 zones divide up and give rise to the 
primordia of the separate cotyledons. On the other hand, this 
investigation goes to prove the opposite; namely, that the poly- 
cotyledonous condition is the more primitive, and the tricotyle- 
donous or dicotyledonous condition derived. 
Most of the work which has been done on polycotyledony has 
been based upon the vascular anatomy of the seedling (16). The 
arguments that favor the derivation of polycotyledonous embryos 
by a splitting of cotyledons are based on anatomy and are well 
summarized by CouLTER and CHAMBERLAIN (10), who state that 
“it must be remembered that these same facts may be used also as 
evidence that the dicotyledonous condition has arisen from the 
fusion of more numerous cotyledons.” 
SAXTON (33) also doubts the origin of polycotyledons from 
dicotyledons, and concludes from a study of cross-sections of 
P. pinaster that “the primordia are exactly equal and equivalent 
in origin.” However, he produced no direct evidence to indicate 
that fusions of the many cotyledons may have occurred. 
The study of the ontogeny of the cotyledons brings out facts 
not hitherto considered in connection with this problem. In 
speaking of cotyledonary fusions, it must be understood that full 
