CONTRIBUTIONS TO BOTANY. 
65 
are admitted into the Bignoniacea, and must be associated with 
Jacaranda, many of the species of which form large trees with 
ecirrhose pinnated leaves, we cannot refuse to admit Tourretia 
into the Sesamea because of the presence of a cirrhus. The 
seeds of Tourretia are alike in shape and position, and have the 
same kind of cristate margin as in Sesamopteris. 
In the tribe Bignoniem, the cells of the ovary are anterior and 
posterior to the axis of inflorescence ; but the four lines of pla- 
centation stand laterally right and left of the same line of axis. 
In the CatalpecE, on the contrary, the cells of the ovary are right 
and left, while the lines of placentation are upon the transverse 
dissepiment, which has a direction radiating from the axis of 
inflorescence. In the Platycarpece, where the ovary is also 
2-celled, the placentations being on the dissepiment, the cells 
as well as the placentae preserv’e the same dextral and sinistral 
aspect. In the Bccremocarpece, where the ovary is unilocular, 
the two component carpels and their placentae stand right and 
left of the axis of inflorescence. The Cyrtandracea present pre- 
cisely the same cai’pellary characters. In the Pedaliacece, where 
the sutural lines of a 2-carpellary ovary are also anterior and 
posterior, the two bifurcate placentae, as in Gesneriacece, have a 
parietal origin on the right and left of the axis of inflorescence. 
The above character, founded upon the origin of the placentae 
upon the midrib of the normal carpellary leaves, is, I believe, 
universal among all the tribes of the Bignoniacece, the Crescen- 
tiacece, the Cyrtandracece, the Pedaliacea, the Gesneriacece, and 
perhaps also the Orobanchacece, which might all be associated in 
one general alliance. This is somewhat at variance with the 
Bignonial alliance of Prof. Bindley, which comprehends also the 
Acanthacece, Sci'ophulariacece, and Lentibulariacece, which ought 
to be rejected — the latter because of its free central placentation, 
the two former on account of the different nature and position 
of their carpels. The Scrophutariacece have a most intimate 
relationship with the Solanacea, as I have elsewhere demon- 
strated ; they, as well as the Atropacece, Gentianacece, and Acan- 
thacece (forming a Solanal alliance), are distinguished by an ovary 
composed noi’mally of two carpellary leaves, which are placen- 
tiferous on their margins (not on their midribs), which margins 
are more or less deeply inflected and conjoined into a dissepi- 
ment : the lines of placentation here are antical and postical in 
regard to the axis of inflorescence, not right and left of it as in 
the Bignonial alliance. 
There is much evidence in favour of the conclusion that, ex- 
cept in the few instances where they form stunted shrubs, the 
Eubignoniece are climbing plants, and that the Catalpece invari- 
ably form standard trees, or erect shrubs. Most of the former 
VOL. II. K 
