— 
THE ORIGIN OF ANGIOSPERMS. 65 
among Angiosperms, as for instance among certain members of the Magno- 
liaceæ. With the Angiosperms generally, just as with the Bennettiteze, there 
has been a constant tendency, by suppression of internodes, to derive a cyclic 
arrangement of the parts of the cone from the primitive spiral type. 
The first step in the immediate evolution of the Angiosperms was the 
transference of the pollen-collecting mechanism from the ovule to the carpel 
or carpels, with consequent localisation of the stigmatic surface. It was this 
act which called the Angiosperms into being, as we shall endeavour to 
emphasize more fully at a later stage in this consideration. 
We may therefore first consider the gynzeceum. 
The Gyneceum. 
We regard the Bennettitese, so far as the megasporophylls of the cone 
are concerned, as departing considerably from the main line of descent 
of the Angiospermez. The orthotropous ovule or seed, enveloped by what 
was probably a single integument *, may be regarded as a fairly primitive 
structure. In the anthostrobilus of the Аполозрегтег, the primitive condition 
of the ovule was undoubtedly orthotropous, and probably there was a distinet 
funicle, a feature whieh may, or may not, be homologous with the seed-pedicel 
of the Bennettiteze. The origin of the second integument does not appear to 
us to present any great difficulty. It is absent in many living Angiosperms, 
especially among the Gamopetalze, and several members of the Ranunculacez, 
an order which we regard as having retained a comparatively large number 
of primitive features in the strobilus f. Moreover, we regard an integument 
as a structure which may arise de novo, and one without close homologies 
among those plants which do not bear seeds. That this is the case is evident 
in such a seed as that of the Paleozoic Lycopod, Lepidocarpon 1, and in 
certain arils found among living Angiosperms. 
The seeds of the Bennettiteze show a close approximation to those of the 
Angiosperms in the fact that the embryo of Bennettites, and presumably of 
the Hemiangiosperme:e, possesses two cotyledons, and that, unlike the Cycads, 
and in all probability the Pteridosperms, these seeds germinated after a 
comparatively short resting period, both of which we regard as primitive 
features among the Angiospermez. 
The structure of the unfertilised ovule of the Bennettiteze is still practically 
unknown, for in all the specimens examined so far, the ovule has apparently 
already become a mature seed. We are, therefore, ignorant of the precise 
anatomy of the micropylar end of the ovule. Did it possess a pollen chamber, 
comparable to that of Lagenostoma, or was the pollen-collecting mechanism 
confined to the micropylar region of the integuments? On this point 
* Wieland (1906) p. 234. T Prantl (1888). 
t Scott (1901) p. 317. 
LINN. JOURN.—BOTANY, VOL. XXXVIII. F 
