The Taxoidece. 
99 
two genera. Spiess (12) has found in both cases that when more 
than two ovules are present the extra ones are arranged decussately 
with reference to the normal pair. 
In nearly all the points in which Taxus and Torreya differ from 
one another, Taxus is the less specialised, but in the structure of 
the mature seed the reverse seems to hold good. Professor F. W. 
Oliver (14) says of the seed of Torreya, “ In the apparent retention 
of old features it exceeds either Taxus or Cephalotaxus .” It is 
possible that the comparatively small size of the Taxus seed and 
its coloured aril (said to be attractive to birds) may have something 
to do with its wide distribution, while on the other hand Torreya 
and Cephalotaxus may partly owe their restricted habitats to the 
handicap of their cumbrous seeds. The radially symmetrical 
peltate stamens of the Yew, with their numerous pollen sacs, are 
represented in Torreya by dorsiventral stamens bearing four pollen- 
sacs on the lower side. That this is phylogenetically a case of 
reduction is strongly indicated by Coulter and Land’s interesting 
observation (19) that the dorsal resin cavity in the stamen of 
Torreya taxi folia is formed ontogenetically by abortion and fusion 
of three rudimentary sporangia. The number of archegonia shows 
reduction as we pass from Taxus to Torreya. In the Yew ovule 
there are normally five to eight, hut the number may rise to 
seventeen, whereas in Torreya californica three are commonly found, 
and in T. taxi folia only one. Again, the development of the 
proembryo of Taxus is less specialised than that of Torreya ; Taxus 
produces sixteen to thirty-two free nuclei before forming cell-walls, 
while in Torreya wall formation is initiated at the four-nucleate 
stage. 
The morphology of the female flower of Taxus and Torreya is 
also best explained on the view that Torreya is one degree more 
specialised than Taxus. In the case of Taxus the so-called “ primary 
shoot ” bearing bracts arises in the axil of a foliage leaf, while 
from the axils of the two uppermost bracts on this shoot secondary 
axes arise, each bearing three pairs of bracteoles and terminating 
in an ovule (Fig. 4). One of the secondary axes commonly aborts. 
In Torreya we meet with precisely the same arrangement, except 
that the axes are telescoped down, the number of scale leaves 
is reduced, and two secondary axes are normally, instead of 
occasionally produced (Fig. 5). The erect ovule placed terminally 
at the end of a secondary bracteole-bearing axis, which itself arises 
from a primary bract-bearing axis produced laterally on an axis 
