CONTRIBUTION'S TO BOTANY. 
151 
velopment of the lowest grade among dicotyledonous plants. 
In Ephedra, on the contrary, we have regular spikelets of Howers, 
though small, consisting of opposite decussating involucels, with 
a perfect petaloid perigonium seated on each axil. There is ab- 
solutely no parallelism in the organization of these two families. 
Most botanists have considered the position of the Gnetacea 
among Gymnosperms to be close to Taxinece (a suborder of Coni- 
fera). In Taxus, however, if we consider the outer coating of 
its pistil to be a carpel, which grows into an osseous shell, then 
its ovule is perfectly devoid of any integumentary envelope, as 
in Pinus, and in like manner it produces a perfectly naked 
seed*; and, according to the masterly analyses of Taxus by 
Mirbel and Spachf, the facts of which have been confirmed by 
Schleiden, its ovule has constantly three foramina in its apex, 
leading into as many embryoniferous cavities in the amnios, 
after the manner of other true Coniferce. This is perfectly at 
variance with all that we find in the Gnetaceoe. 
With the Cycadacece the dissimilarity in these respects is still 
more striking. The existence, however, of a suspensor in its 
seeds has greatly favoured the idea of the close affinity of Gne- 
tum with this family : the coincidence is unquestionable ; but 
this circumstance is of little import upon its own merit, for we 
may conceive the possibility of its occurrence, as we find it in 
Gnetum, in any family of the highest order of development. The 
growth and structure of the seed in Cycadacece, as demonstrated 
by MiquelJ, present many peculiarities of which we have no 
parallel in the Gnetacece. If we regard in its proper light the 
outer covering of the pistil, and the thick fleshy and coriaceous 
shell of the seeds of the Cycadacece to be the growth of a true 
carpel, it will be evident that the erect nucleus enclosed within 
the pistil and the seed contained within the carpel are deficient 
of any proper integument ; and under this point of view both may 
be considered to be perfectly naked — a condition widely dilferent 
from that of the Gnetacece, where both the ovule and seed are 
covered by two distinct integuments. Miquel, who considered 
the peric^pial covering of the fruit to be the testa of the seed 
lined with an adherent inner integument, notices that the latter 
* From the analysis of Torreya taxifolia, as given by Sir \Vm. Hooker 
(Icon. 232, 233), it would seem that its carpel is at first pervious at its 
summit, and is furnished with an erect atropous ovule, provided with two 
very distinct free integuments, of which the primine afterwards becomes 
agglutinated to the carpel in the fruit, while the secundine remains co- 
herent with and enters deeply into the plicatures of the ruminated albu- 
men. Should this analysis of the ovary be confirmed, it would show that 
Torreya cannot belong to Taxinece, but is more allied to Myricacece. 
+ Ann. Sc. Nat. 2 ser. xx. 257. 
Ann. Sc. Nat. 3 ser. hi. 193, pi. 8. 
