PROaRESS OF MICROSCOPICAL SCIENCE. 
217 
exterior of these seeds, but are casts of the interior of the sclerenchy- 
matous endotesta, the three thin, longitudinal, wing-like appendages 
being merely casts of the three slit-like extensions of that interior 
just referred to. These slits extend upwards into the prolonged 
micropyle, the interior of which displays a triangular section, each of 
the sides of which is convex, the convexity projecting inwards. The 
nomenclature of this type of seed is in great confusion, owing to 
specific differences being based on mere differences of size, many of 
which are probably nothing more than varieties due to age and 
development. Casts of seeds with six longitudinal wings are described, 
corresponding with Brongniait's genus Hexapterospermum. They are 
more oblong than Trigonocarpum olivceforme, but apparently identical 
with the T. Noggerathi of the ' Fossil Flora.' The author doubts the 
wisdom of Brongniart's estab ishment of a separate genus for these 
seeds. Several species of the important genus Gardiocarpum have 
been obtained displaying the internal organization of these remark- 
able seeds. They all agree in possessing a central endosperm which 
is remarkable for the very large size of its conspicuous parenchy- 
matous cells. This is invested by a perispermic membrane, the whole 
being enclosed within a testa composed of two very distinct and 
separate layers. A thin inner one, which may be identical with the 
nucular membrane of other seeds, is entirely composed of delicate 
prosenchymatous cells, and is prolonged into an elongated micropyle, 
into which the endosperm is not prolonged. Externally to this is an 
exotesta composed of a denser parenchyma. In some species this latter 
tissue is uniform throughout, in others it is separable into a dense 
endotesta and a more lax parenchymatous exotesta. The first species 
described is apparently identical with the G. anomalum of Carruthers, 
and has a trigonous endosperm invested by tlie two layers of testa (?), 
both of which are prolonged into a slender tapering beak, half the 
entire length of the seed, and which contains the elongated micropyle. 
Another sjDecies, designated G. compressum, has its ap^Darent testa 
composed (as just described) of two continuous layers. In it the 
micropyle is comparatively short, and its apical extremity is patulous 
or trumpet-shaped. To a third very beautiful little cordato-lanceolate 
species with a peduncle or funiculus equal in length to the seed, the 
author gives the name of Gardiocarpum Butterworthii, after its 
discoverer. These seeds exhibit no specialized organ corresponding 
to the lagenostome of Lagenostoma and other seeds described. The 
pollen has passed down the long narrow micropyle into the triangular 
space at its inner extremity, where it came into direct contact with 
the endospermic membrane. It thus appears that the seeds known 
by the name of Gardiocarpum have a very simple organization, ap- 
proximating somewhat closely to that of the ovules of Juniperus, 
Gallitris, and Welwitschia. Some small seeds, which appear to be 
identical with the Gardiocarpum tenellum of Dawson, found in great 
numbers on slabs of shale by Mr. John Smith, of Kilwinning, in 
Ayrshire, are described. They were found in the upper coal- 
measures near Stonehouse in Lanarkshire. The last form noticed is 
a very curious winged seed from the uppermost coal-measures of 
