CORRESPONDENCE. 98 
nature, a two-celled ovary becoming a nearly TAER nut, one-celled by 
abortion, and perfecting only a single seed, which is exalbuminous, with large 
fleshy cotyledons. I pointed out long ago the aba of this structure 
with Platycarpum and Henriquezia, genera also referred to Bignoniacee. M. 
Bureau lays much stress upon the structure of the seed as being contrary to 
the usual development in that tee an inference which is true; but it is not 
less true that such a structure e ists, as I have shown i in Tanecium, Adenoca- 
nn. T 
that can militate against the admission of my tribe mL into the 
Bignoniacee, except the occasional suppression of one of the cells of the fruit, 
and the abortion of all the ovules es except one, a circumstance not to be wondered 
niacee (where you originally suggested they should be placed), but the € 
absence of albumen in the seeds, and their large thick cotyledons, are charae 
quite — with such a position. Another argument urged by pd 
Bureau against the admission of Henriquezia into the Bignoniacee is the pre- 
sence of coda which he seems to think is ineompatible with this family ; but 
here he is unquestionably mistaken, for, as you are aware, they are are not unfre- 
quently developed to a large size, being in opposite pairs in each node, like 
sessile rounded leaflets, without the smallest indication of a petiole. The last 
becomes quite superior; but this fact is not fatal to the reception of the genus 
among Bignoniacee, however abnormal it may be. I have before pointed out 
a "y regs instance i in true Styracee, where in most cases the ovary and 
fri erior ; but in the Zalesiec, the ovary, at first almost superior, 
in the Mame ut. its growth becomes nearly quite inferior, but Halesiee and 
Pterostyrax cannot, on that account alone, be separated from Styracee (see 
"Contr. Bot. i. plates 29, 30, 31). 
. Bureau has objected to my hypothesis that in Bignoniacee the carpellary 
leaves are ovuliferous on the midrib, and he evidently considers that I have 
is not sufficient to a the orthodox rule which M. Bureau bein advo- 
then why may die ey not bear ovules on their midrib? If we put aside a pre- 
judgment of the subject, the latter development will not appear more unnatural 
than the former. In a plurilocular ovary, it is not impossible that sometimes 
the midribs of the oepami carpels may unite in the centre, and, becoming 
