CONTRIBUTIONS TO BOTANY. 
0 / 
linear hilum, by which it is attached to the inner face of the 
valves; upon this face, along the margins of each valve, there is 
seen a broad longitudinal seminiferous zone, leaving the inter- 
vening portion, which is as broad as the two zones united, co- 
vered with a polished yellow endocarp : these seminiferous zones 
are marked with a number of dark, rough, oblong cicatrices, 
placed alternately in three or four parallel series, which corre- 
spond with the hilar points of attachment of the seed, as in 
Pithecoctenium. From this it will he seen that the attachment 
of the seeds is strictly parietal — a feature hitherto novel in Big- 
noniacecR. This circumstance has probably led botanists to 
suppose that it belonged to the Crescentiacea, as it resembles 
Crescentia in its hard shell, of similar shape and size ; but as it 
is decidedly 2-valved and dehiscent, has a perfect and free dis- 
sepiment, with a distinct replum, and quite devoid of pulp, it is 
strictly Bignoniaceous and belongs to the tribe Bignoniece, dif- 
fering in no respect except in its parietal placentation. It should 
be observed, however, that the direction of the seeds and their 
attachment by a marginal hilum, precisely in the same position, 
occur in Pithecoctenium (see fig. 5, page 52) ; but in the latter 
case the zonal hands of placentation remain fixed at a right angle 
to the inflected margins of the free dissepiment, whereas in 
this fruit they are attached to the valves : in all other respects 
there is a perfect agreement between the two cases. The struc- 
ture of the seed and the form of its integuments in this instance 
are precisely as I have described them in Pithecoctenium, the 
testa being in like manner bimarsupiate at the base of its discoid 
cell by a short septum ; and in the recesses thus formed the 
lower portions of the embryonary lobes are sheltered. In Pithe- 
coctenium, the hroad expanded part of the testa which forms the 
pellucid winged margin around its discoidal portion is of the same 
extent as in the fruit under consideration ; but, in the latter case, 
it is filled to its extreme edge by a thick deposit of cellular tissue 
between its faces, so that it loses its transparent character, and 
becomes thick, hard, and coriaceous, as in the seeds of Hetero- 
phragma-. the inner integuments, which closely invest the em- 
bryo, are about half the length and half the breadth of the 
testa. 
I shall presently recur to the genus Tanaecium of Swartz, 
respecting which a strange degree of confusion has existed. 
The group of the Crescentiacece merits observation in this in- 
quiry: it was considered by Jussieu, Endlicher, and DeCandolle 
to be a tribe of the Bignoniacece. Gardner first proposed it as a 
distinct family, which view was adopted by Prof. Lindley ; and 
VOL. II. 
I 
