CONTRIBUTIONS TO BOTANY. 
69 
gether in one group, as in Spaihodea, Tabebuia, &c. It will be 
seen, in the descriptions I here propose to give of sundry Big- 
noniaceous plants, that in the same genus the size, shape, mar- 
ginal dentations or fissures of the calyx vary to a considerable 
extent, and that the great peculiarity of form, which was thought 
to characterize only the genus Spaihodea, exists also in the 
genera Macfadtjena, Mansoa, Dolichandra, Tabebuia, and some 
others. The same may be said of Cuspidaria, where but few of 
the species enumerated by DeCandolle possess the long cuspi- 
date teeth that suggested this generic name, while similar long 
setaceous teeth are found in Mansoa, Tynanthus, and several 
other genera. I have placed the Bignonia gluiinosa, DC., and 
other kindred species, in a genus near Dolichandra, notwithstand- 
ing that its calyx (as also its corolla) becomes nearly as much 
enlarged and coloured as in CaUichlamys •, but this feature is only 
due to the extreme increment of those parts, which may be traced 
in all its various gradations. 
One of the most important features that serve to mark the 
genera of the Bignoniaceee exists in the form and development 
of the fruit ; but unfortunately this is rarely available, as few 
cabinet specimens present this test ; indeed, in some genera the 
fruit is quite unknown, and I am glad to be able to supply this 
desideratum in several cases. 
I’he anther-lobes, in most instances, are divaricated to their 
utmost extent, when, from the mutual incurvature of the fila- 
ments, the lobes stand in a vertical position and at right angles 
with the filaments ; the two lobes of each pair are thus brought 
into juxtaposition, as in the Gesneracea: sometimes, as in Tyn- 
anthus, the anther-lobes, fixed at right angles upon the apex of 
the filaments, are suddenly curved upwards. In several in- 
stances the filaments are nearly straight, and the anther-lobes, 
although free in their whole length, are parallel and pendent 
from the summit of the filament; this occurs in the genera 
Pyrostegia, Dolichandra, Cybistax, Salpingophora, Astianthus, 
Calosanthes, Millingtonia, Catophractes, and Rhigozum, and is a 
constant and valid generic character. In some few genera the 
anther-lobes are sagittately divergent. The glandular summit 
of the filament (connective), to which the anther-lobes are at- 
tached by their apex, is often excurrent and mucronate, some- 
times e.xtended into one or two membranaceous appendages, and 
at other times pilose ; but these expansions are not always 
constant in the same group, and I have not considered their 
deficiency to be of any generic value. The anther-lobes, as a 
general rule, are glabrous, but in some instances they are cilio- 
late or pilose in the same group where others are glabrous ; this 
feature, therefore, cannot be held to be of sufficient importance 
