OF THE EOSSIL PLANTS OF THE COAL-MEASURES. 
245 
of this species. In it the membranes are much disturbed, still they are not difficult of 
identification. There is an external testa (a) of uniform thickness. At the lower part 
of the seed we find an outer membrane at f, separating from an inner one at g. I 
presume that the former is the nucular membrane, and the latter the perispermic one. 
The perisperm, n, has obviously been much shrivelled before mineralization. The 
upper or apical part of the seed is so important that I have given an enlarged view of 
it in fig. 83, magnified 60 diameters. The testa (a) surrounds this end of the seed. At 
/we find an inner membrane separated from the testa. Fig. 82 shows that this is 
the nucular membrane, if not that and the perispermic membrane, g, united into one. 
At f-f a portion of this membrane turns horizontally inwards, as in figs. 79 & 80, whilst, 
in a similar manner, another portion of it ( f'f ) arches round the apex of the seed in 
close contact with the inner surface of the testa. The parts forming the representative 
of the lagenostome perplexed me until I discovered the specimen (fig. 84), which is an 
oblique section passing along a plane corresponding with the line x in fig. 83. In 
fig. 84 we have the testa at a , but apparently splitting into two layers at the upper part 
of the figure ( a a'). At ff is a distorted membrane, of which the portions f'f obviously 
correspond to the parts indicated by the same letters in fig. 83. Above this the mem- 
brane rises into a dome-shaped structure at d, and is now very distinctly seen to consist 
of prosenchymatous cells. As these cells approach the central axis of the seed, they 
obviously bend inwards and downwards, forming a central funnel-shaped depression 
leading to the representative of the lagenostome (<?). It thus appears that the part 
represented by fig. 83 (d) is in its normal position, whilst the corresponding rounded 
lip of the funnel on the opposite side (c") has been accidentally lifted up from its true posi- 
tion. At fig. 83, d", we have the semblance of a horizontal continuation of the tissues of 
d, d' uniting the extremities of these two curved structures ; but I do not believe that 
this is actually the case. We see a similar condensed tissue occupying the same 
position in fig. 84, but without exhibiting any trace of the conspicuous prosenchymatous 
structure characteristic of the wall (d) of the lagenostome. AVithin the lagenostomal 
cavity we have both in figs. 83 & 84 a mass of delicate, thin-walled parenchyma (c), 
which I think must be identical with that seen in the lagenostome of Lagenostoma ovoides, 
and which, occupying, as it does in the seed under consideration, a central position, 
apparently closing up the entrance from the micropyle into the lagenostome ( c ), throws 
some light upon the similarly placed tissue (p) in fig. 79. At f" in fig. 84 we have a 
layer of membrane which appears to be identical with the layer f" in fig. 83. The 
length of this seed is about T, and its diameter about '05 to '066. Fig. 87 is a 
longitudinal section of a seed from Burntisland, which may possibly be a state of 
Conostoma ovalis, but which I am inclined to regard as distinct. It is longer and 
narrower than that species, besides which I cannot identify the tissues in the two forms. 
In its general contour it resembles Conostoma oblonga. There is a thin testa at a , 
within which are other membranes difficult of identification. Thus at V we have one 
thickish membrane ; at b" b" this membrane splits into two. At the apex of the seed 
MDCCCLXVII. 
2 M 
