Gnetum africanum , with Notes on Gnetum scandens. 1127 
of an inner fibrous layer which is similarly distributed in the peduncle, 
main body, and apex of all three seeds, except that in Gnetum it forms at 
the tip of the seed a more definite palisade, which in G. africanum is 
lignified very early. B. Morierei and Gnetum have a middle palisade layer. 
In the seeds of the Bennettitales the stony layer tends to be formed from 
the outer layer of the integument; in Gnetum the inner layer shows the 
earliest signs of bonification. 
(ii) The comparison of the mtcellus and inner integument of the seed of 
Gnetum with the main body of the seed in Bemiettites involves more difficult 
questions. In both seeds there is a projecting micropylar tube , which in 
Gnetum is derived from the inner integument and projects beyond the 
outer, but is fastened firmly on to it by means of the downward projecting 
flange. Lignier’s figures of Bennettites are suggestive of the same arrange¬ 
ment in that seed, the relation 1 of the micropylar tube to the expanded 
portion of the integument being identical with that found in Gnetum. 
Lignier attempts to trace continuity between ‘ the integument ’ (outer 
integument?) and micropylar tube; he states, however, that the micropylar 
tube is homogeneous in structure, sclerified in the fertile seeds, not sclerified 
in the aborted, and shows no resemblance to the complex structure of the 
rest of the integument; continuity with the rest of the integument could 
not be demonstrated. This description would equally apply to Gnetum. 
The difficulty of investigating the apex of the seed may possibly be due to 
a similar withering of the apex after fertilization, and if it be true that the 
micropylar tube in Bennettites is also fastened by a flange to the top of the 
outer integument, as in Gnetum , it is easy to understand the confusion as to 
the boundary between the two. The description of the micropylar tube 
agrees with that of Gnetum , the canal being cylindrical below, with an 
epidermis composed of small thin-walled cells, irregularly slit-like above, 
with an epidermis which consists of narrow radially elongated cells, and is 
cuticularized (cf. Figs. 31, 32, PI. Ill, Lignier, with Fig. 16, PI. LXXXVII, of 
this paper). Opposite the expanded portion of the outer integument the 
canal is closed as in Gnetum. 
The micropylar tube of Cycadeoidea Wielandii is not clearly comparable 
with either Bennettites or the mature seed of Gnetum. It is open, and it 
may be, as Miss Berridge has suggested, that the seeds figured are at an 
earlier stage and should be compared with unfertilized ovules of Gnetum. 
A greater difficulty is that the outer thickened layer of the (outer?) integu¬ 
ment extends upwards and clothes the micropylar tube. It is, of course, 
possible that the seed is so young that the inner integument does not yet 
protrude beyond the outer, but this does not seem likely in view of the 
differentiation which has already taken place in the integument. It is more 
1 This comparison was made in detail by the author at the Sheffield meeting of the British 
Association, 1910; and a similar one has also been drawn by Miss Berridge, 1911. 
