52 
CONTRIBUTIONS TO BOTANY. 
find a striking evidence in favour of this conclusion in the fruit 
of Pithecoctenium, of which mention has been previously made. 
Here the seeds are affixed upon a broad, prominent zonal reflexion 
formed round each margin of the thick dissepiment, which, as 
we have seen, is separable into two plates; pjo-. 5 . 
and the thick valves at the line of their 
separation are seen to be respectively r ^ 
moulded around these seminiferous in- 
flexions, leaving the replum on each mar- 
gin free both from the valves and the dissepiment, though still 
connected with the latter at its base and apex (fig. 5). 
This conviction is more forcibly confirmed by the examination 
of the ovary in the Eubignoniece. Upon the fall of the corolla, 
we find the ovaiy grown to a length a little exceeding that of 
the calyx, and compressed as usual, with a deep groove along 
each edge coinciding with the margins of the dissepiment, which 
there bears the ovules close to that line, at the furthest possible 
distance from the axis ; upon each face is a prominent line, which 
is continued up the style : this structure is shown in fig, 2. The 
same appearance is observed in the ovary, in a very early stage 
of the bud, long before the anthers are perfected ; we have then 
the same deep marginal inflexion, showing the line of confluence 
of the component carpels, as in the manner shown. The stigma 
consists of two broad petaloid lamellae, which are smooth within, 
and apparently are not provided with the usual papillose stig- 
matic surface, which here seems to reside in the infundibuliform 
mouth of the style ; this is hollow for half its length, and on 
each side dark longitudinal lines are distinctly seen, which ap- 
pear to correspond with the stigmatic channels in communica- 
tion with the four placentiferous lines of the ovary : the longi- 
tudinal line seen upon each face of the ovary is continued along 
the style, in the form of a nervure, through the middle of the 
stigmata, and on each side of this nervure the dark lines just 
mentioned are distinctly seen. These appearances are all con- 
firmatory of the assumption that the ovary is composed of four 
carpels combined in the manner previously shown ; and we may 
also infer that each lamella of the stigma is composed of two 
confiuent stigmata. The occurrence of a bifid stigma, with four 
united carpels, at first thought might be considered to be an 
improbability ; but we must remember that it is no unusual 
combination, for it is constant in the extensive families of the 
Ehretiacea, Borraginacece, and Labiates, and also frequent in 
Verbenacees. 
In the Catalpecs, where the capsule is loculicidal, the dissepi- 
ment being contrary to the valves, which separate from it, there 
is evidently a difierent structure, the nature of which may pro- 
