boccata , with Remarks on the Antiquity of the Taxineae . 13 1 
integument which, as already suggested, is a comparatively recent for- 
mation. 
In attempting to assign a place to Ginkgo in relation to the Cordaitales 
the fact must be taken into account that there is apparently nothing in the 
seed of this plant to represent the 4 outer ’ vascular system. There may, 
however, be some significance in the fact, observed by Sprecher (1907, 
p. 137), that the superficial cells of the young sclerotesta of Ginkgo are 
histologically indistinguishable from tracheides (see Sprecher’s Fig. 150, 
p. 137). May the outer vascular system of Ginkgo be regarded as having 
fused up with the outer surface of the stone ? The following remark by 
Sprecher is of some interest in this connexion : 4 Le noyau sclereux me 
parait etre en rapport avec le tissu de transfusion de l’arille ; il est sa con- 
tinuation dans PariHe.* If this view is justified Ginkgo would appear to 
make a remarkably near approach to the Palaeozoic seed Cardiocarpus , in 
which the 4 outer ’ system of strands comes off from the main supply before 
the latter enters the stone (cf. Nos. 1 and 2 in Fig. 7). This would corro- 
borate the low position of Ginkgo indicated by the motile sperms and the 
well-developed pollen-chamber and 4 internal * vascular system, as well as 
by the presence of a well-developed 4 tent-pole ’. Another fact pointing in 
the same direction is the proximity of the two supply bundles to the middle 
line, which shows that the main supply has, in a phylogenetic sense, only 
recently undergone splitting. 
Summary and Conclusions. 
The present paper contains an elaboration of a theory developed by 
Professor Oliver in 1903, when he offered an interpretation of the seed- 
structure of Torreya in terms of the Palaeozoic seed Cardiocarpus , which 
was regarded as the ancestral type. It is suggested that the Palaeozoic 
seeds Cycadinocarpus , R hab do sperm it m , Mitrospermum , and Taxospermum , 
all of which probably belong to the Cordaitean phylum, illustrate the general 
tendency that may have operated in producing the types of seed known to 
occur in the modern genera Taxus , Torreya ) and Cephalotaxus . It is pro- 
posed to place these three genera in a distinct phylum, Taxales : the 
members of this phylum are regarded as the nearest existing relatives — 
apart from Ginkgo — of the Cordaitales, and like Ginkgo direct descendants 
of the Cordaitales. While accepting a Cordaitalean origin for the Coni- 
ferales, I regard the further question — whether the connexion was a direct 
one or whether the Conifers arose as a branch from the Taxales — as still 
sub judice , although I am inclined to favour a direct connexion. The 
problem as to the ultimate microphyllous or megaphyllous origin of the 
Cordaitales (and therefore also of the Conifers) is left untouched in this 
paper. 
If the seeds of the Taxales have been derived from those of the 
K 2 
