66 
CONTEIBUTIONS TO BOTANY. 
have 3-foliolate leaves, in which very often, and especially in the 
superior axils, the odd leaflet is transformed into a cirrhus, thus 
forming cirrhosely conjugated leaves; in rarer instances, the 
leaves are either simply pinnate or 2-3-pinnate, the leaflets 
being always petiolulated, and then generally the odd normal 
foliole is also converted into a cirrhus. Among the Catalpea, on 
the contrary, the leaves, with very few exceptions, are either 
pinnate or quinate ; and I am not aware of the existence of a 
cirrhus in any legitimate species of this tribe ; for I have shown 
that all the conjugate-leaved species of Tabebuia must be ex- 
cluded, not only from that genus, but from the tribe. A great 
portion of the simple-leaved species of the order have been 
arranged among the Eubignoniece ; but it will be seen that most 
of them must be removed into the CatalpecB. An exception 
to this rule might be cited in the genus Delostoma-, sub- 
stantial reasons, however, will be given for transferring it 
into the Catalpece. I have not seen the fruit of Astiantlius, 
nor are any satisfactory characters given of it : the hairy, pap- 
pose villosity of its seeds constitutes a feature quite unknown 
among the Eubignoniece, and the examination of its ovary leads 
to the belief that it belongs to the Catalpece ; indeed, in the form 
of its capsule, the hairy covering of its seeds, and its crowded, 
alternate or almost verticillate, linear, simple leaves, it is proba- 
ble that Catalpa longisiliqua is congeneric with it. I believe the 
general habit of the plant to be a constant feature, and that for 
purposes of generic distinction it offers a character equal in im- 
portance to that of the structure of the flower or of the fruit 
and seed. The presence, however, or absence of a cin-hus in a 
conjugate leaf, which frequently falls off at the point of its arti-' 
culation with the petiole, or the substitution of a third foliole in 
the place of a cirrhus, are not indications of much value ; for all 
these three conditions commonly occur in the same plant among 
Eubignoniece ; so that its adoption for a divisional character, as 
employed by DeCandolle, has been quite useless. The presence 
of simple and compound leaves in the same species, or even in 
the same genus, must not be held to be a feature of frequent 
occuiT-ence, as some botanists have inferred ; for I believe such 
instances to be extremely rare : they occur chiefly among the 
Catalpece, and then only in the few cases where the folioles are 
not petioled, or where the main petiole is winged, that is to say, 
where the leaf is rather pinnatifld than truly pinnate : in such 
cases the extent of division of a simple leaf may be varied, even 
in the same plant, as happens in many other families ; but this 
kind of division is only the modification of the simple serrated 
leaf. In regard to this rule among Eubignoniece, Chamisso re- 
lates that in his Bignonia pterocarpa and B. samydoides, the 
