‘ Flowers ’ of Cephalotaxus. 645 
lowest pair of bracts and one of the second pair have been 
separated from the rest by a long internode, accompanied by 
a displacement of some of the bracts from their paired posi- 
tion, which possibly may represent a transitional step from 
the whorled arrangement of the bracts to the spiral insertion 
of the leaves of the vegetative shoot. 
The apical region of an unproliferated primary axis of the 
inflorescence is usually considerably swollen, forming a con- 
spicuous whitish object in the midst of the whole agglomera- 
tion of bracts, &c. (Fig. 20). The swollen character is evidently 
due in part to the rudimentary bracts which build up this 
apical region, and sometimes project considerably as isolated 
protuberances in a lateral position and immediately above an 
axillary flower, so that I have often been led to imagine that 
this amorphous, rudimentary bract of the primary axis was an 
enlarged median posterior foliar organ of the axillary flower, 
and am still not perfectly clear as to how matters stand in all 
cases. This feature of the primary axis is exactly homologous 
with the still more swollen, fleshy character of the primary 
axis of Phyllocladus , in which the apex is equally undefined, 
and where several fleshy, sterile bracts frequently occur inti- 
mately fused with the axis bearing them. In fact, the 
inflorescence of Phyllocladus more nearly resembles that of 
Cephalotaxus than is the case with any other genus, the only 
difference being that in the former plant the bracts alternate 
and the floral axis is reduced to a single ovule (this being the 
lowest point reached in reduction of all the five genera), while 
in the latter genus the bracts are arranged in pairs, and the 
floral axis reduced to two ovules. 
We next come to the consideration of the individual axillary 
flowers, which afford several points of considerable interest. 
Proliferated primary and secondary axes may occur on the 
same shoot, but by no means always accompany each other ; 
indications of the latter may be present when the former are 
quite absent. The normal condition of things is that in which, 
on the excessively short axillary axis, either two lateral trans- 
versely-placed ovules are alone present, or sometimes between 
U u 2 
