238 
PEOEESSOE W. C. WILLIAMSON ON THE OEGANIZATION 
are traces of an indentation (x) into the substance of what I presume has been the 
nucleus, and which contains some dark-coloured granules. These have most probably 
been derived from the lagenostome, yet the position they occupy may be merely an 
accidental one. 
Fig. 70 represents the upper extremity of another very instructive specimen, enlarged 
80 diameters. A portion of the testa remains at a. The nucular membrane (f) is in 
near approximation to the testa, whilst the perispermic membrane ( g , g) is now so much 
distended as almost to occupy the entire cavity of the seed. w r e again see, as 
in figs. 60 & 61, the nucular membrane/ 1 , splitting into three layers. On the left hand, 
one of these (fl) goes to the upper surface of the perispermic membrane ( g ), the second 
(f 2) goes to the base of the lagenostome (c), whilst the third and uppermost {f%) forms 
the canopy. On the right hand, the innermost layer ( f 1) first comes into contact with 
the perispermic membrane, but soon leaves it again, bending upwards so as to reach the 
base of the lagenostome. The second (f 2) goes direct to the lagenostome, as on the 
opposite side ; whilst, in like manner, the outermost layer (f 3) again forms the canopy, 
additional shreds of which are seen on both sides, yet more externally. Here, again, the 
membrane of the canopy appears to be thick, but it is not so. Its true thickness is 
represented by the dark boundary line at the inner margins of f 3 and f 3 ; the rest is 
merely a continuation of one of the folds seen perspectively, as in fig. 69. 
The lagenostome is here very clearly shown to be open inferiorly so as to bring the 
base of its internal cavity into intermediate contact with the upper part of the peri- 
spermic membrane g\ But this fine section further demonstrates that the membranous 
wall of the lagenostome ( c , c) is composed of prosenchymatous cells like those constituting 
the canopy and nucular membrane, whilst it differs altogether, as I shall be able to 
demonstrate, from the structure of the perispermic membrane, g'. 
But the most interesting feature of this section appears in the interior of the lageno- 
stome c. Its central portion is occupied by the usual delicate parenchyma, but in 
the vacant space remaining between this parenchyma and the prosenchymatous wall of 
the lagenostome are twelve or thirteen round clearly defined bodies ( l , T), w 7 hich are 
unquestionably identical with those seen by M. Brongniart, and regarded by him as 
pollen grains. These little spherical bodies have a mean diameter of about '0023, and 
exhibit every appearance of being true pollen grains. We shall find them again in the 
next specimen to be described. 
Satisfactory, and in some degree conclusive as all these several features of the vertical 
sections proved to be, there yet remained some important unsolved problems ; but my 
materials were exhausted. I had no satisfactory evidence respecting the horizontal 
arrangements of the lagenostomal region. To relieve me from this dilemma my friend 
Mr. Butterworth sacrificed his only specimen of this seed ; unfortunately it proved to 
be so crushed that it failed to give me all the information I sought, though, as we have 
seen from figs. 53-57, it revealed what none of my other specimens had done, the hori- 
zontal arrangement of the parts near the mouth of the micropyle. One other seed 
