CONTRIBUTIONS TO BOTANY. 
43 
points of great importance, furnished by his eminent son. Prof. 
A. DeCandolle. Few additions to our knowledge of the family 
have been made since that time : among these may be mentioned 
those of Dr. Seemann on the CrescentiacecB and some of the 
genera of the Catalpea : the attention of the same botanist has also 
been for some time directed to the study of the family ; and it is 
to be hoped that he will publish the results of his investigations. 
The order is divided by DeCandolle into two tribes, distin- 
guished in great measure by the presence or absence of winged 
margins to the seeds ; in the former case [Bignoniece] the fruit 
is capsular and dehiscent, with winged seeds ; in the latter case 
[Crescentiece] it is either fleshy or ligneous and indehiscent, with 
apterous seeds, often imbedded in pulp. I have already pointed 
out, in the former tribe, the anomalous instance of Oxycladus *, 
where the fruit is an indehiscent nut, normally 2-locular and 
pluriovular, but by abortion unilocular and monospermous, the 
seed being perfectly apterous, with two thick fleshy cotyledons 
conjoined by a very small terminal radicle. I indicated also an- 
other instance in Adenocalymna-\ , where, although the fruit is 
capsular and dehiscent, the seeds have no wings, their integu- 
ments being coriaceous and hard, and their cotyledons very thick 
and fleshy. In Platycarpum and Henriquezia the seeds are like- 
wise fleshy and wingless. In Argylia, also, some species have the 
seeds quite apterous, in others the wing is only rudimentary. 
Thus it appears that this feature is not sufficiently constant to 
serve as a basis for tribual distinction. I propose in the follow- 
ing remarks to search for other characters , and with this view I 
will first notice several modifications I have observed in the 
structure of the seeds, and then inquire into the modes in which 
the carpels are combined. 
In the Bignoniece the seed usually consists of a coating, con- 
siderably flattened, with a coriaceous centre surrounded by a 
broad, delicately membranaceous wing, generally broader than 
long : it has no funicle, but near the margin of the coriaceous 
portion, contiguous to its base, a small linear hilum is seen, 
which corresponds with a similar cicatrix on the dissepiment, 
where it was attached. As an example of the general structure 
of the seed in the Bignoniece, we may take that of Pithecoctenium 
squalus. Here it is surrounded by a very broad wing, of extreme 
tenuity, delicately reticulated, perfectly hyaline, with a number 
of strong nervures radiating fr,om the coriaceous discoidal centre. 
After sufficient maceration, it is easy to introduce a blunt needle 
into the substance of the wing, when it can be separated into two 
very distinct laminae, even to the utmost margin, proving that the 
* Linn. Trans, xxi. 145; Linn. Proc. ii. 270. 
t Linn. Proc. ii. 272. 
