•76 
Sinnolt . — The Morphology of the 
wide apart, run down through the epimatium on its adaxial face. In P. 
fcrrugmeus these two bundles, after leaving the chalazal end, pass down 
exactly opposite one another, one on either side of the ovule and each just 
at the inner margin of the epimatium. This latter condition, obviously less 
primitive than that in P. spicatus , is precisely what we find in the ovule of 
Cephalotaxus (and of Torreya as well), where two opposite and lateral 
bundles pass upward from the chalaza towards the micropyle just at the 
limit of the two integuments. In Podocarpus , of course, there are the two 
dorsal strands which represent the vascular system of the ovuliferous scale 
between the axis and the attachment of the ovule, but if we imagine that 
this part of the scale, after being greatly shortened, as in Dacrydium , has 
disappeared entirely, as in Phyllocladus , and that the ovule has become 
erect while at the same time retaining its integumentary and vascular 
structures, we obtain a close approximation to conditions in Cephalotaxus . 
A study of the little known New Caledonian genus Acmopyle should prove 
very interesting in this connexion, for here the cone has been reduced to 
a fleshy receptaculum bearing a single erect ovule which has an inner stony 
and, apparently, an outer leathery integument. This seems to provide an 
intermediate condition between Podocarpus and the Taxineae. 
The resemblance between Cephalotaxus and the sub-genus Stachy carpus 
of Podocarpus is further emphasized by the structure of the female gameto- 
phyte and embryo, for in both the reproductive cycle extends through two 
seasons ; spongy tissue is almost completely absent ; the archegonia are only 
two or three in number and are very long and narrow ; they are eventually 
deeply sunken in the endosperm ; the neck is composed of but one tier of 
cells ; the first two sporophytic mitoses occur in the upper portion of the 
archegonium, and wall formation occurs at the sixteen-nuclear stage, and is 
followed by rapid and irregular cell-division which gives rise to a proembryo 
different from that of almost all other Conifers. It is composed of four tiers : 
a single terminal cell, more or less protective in function ; a group of cells 
immediately behind this, which is to give rise to the embryo ; a suspensor 
tier, and a poorly developed rosette. 
In opposition to this evidence from sporophytic anatomy and the 
female gametophyte, it may be urged that the occurrence in Cephalotaxus 
of two ovules to a bract instead of one is a strong objection to the derivation 
of this genus from forms close to Podocarpus . The other two members of 
the Taxineae have but one ovule to a bract, and although the occurrence of 
a single ovule is constant throughout the Podocarpineae (two nucelli were 
found in one epimatium of P. spicatus ), the number varies greatly in other 
coniferous families. It seems very probable that the multiplication of 
sporangia which is so evident in the male cone of the Taxineae may have 
affected the female strobilus as well. The occurrence of two megasporangia 
to each megasporophyll in Ginkgo and the Abietineae makes it also very 
