56 
CONTRIBUTIONS TO BOTANY. 
In Amphicoma the capsule appears ditFerently constituted from 
that of Incarvillea, as before shown ; and in several other respects 
the genus is at variance with Bignoniacea : it seems more con- 
formable with Cxjrtandracea, as Mr. Brown long ago indicated 
{loc. sup. cit.) ; in the structure of its flower it quite agrees with 
that family, especially in its stamens with large appendiculate 
connective, which in Bignoniacece is not so strongly developed ; 
its fruit is also in perfect accordance with CyrtandracecB, parti- 
cularly in its long comose seeds, which are pendent (not trans- 
verse), with a superior (not centrifugal) radicle. The genus, at 
one time placed in Cyrtandracece by DeCandolle, was afterwards 
removed, in great measure on account of its divided leaves j it 
must, however, be remembered that it agrees better with the 
last-mentioned family in its herbaceous habit and general aspect, 
and that if in CyrtandracecB the leaves are not pinnatisected, they 
are only one degree removed from this condition in their deeply 
and unequally serrated margins. 
Before I dismiss this inquiry into the carpological structure 
of the BignoniacecE, I will mention a novel and interesting form 
of development which I have noticed in a fruit brought from 
Jamaica, and now in the Collection of the British Museum, 
where there are two specimens, collected by different individuals 
at distant periods, from which we may infer that the plant which 
produces it is not of rare occurrence in that island. It is to be 
regretted that these fruits are not accompanied by any dried 
specimen of the plant from which they were gathered. I have 
no doubt that it is the fruit of Tanaecium albijloruni, DC. {Tanae- 
cium Jaroba of Swartz, and the Cucurbitifera fruticosa, Sloane), 
agreeing in every essential respect with the account given of it 
by Swartz. The specimen I have seen is 6 inches long (accord- 
ing to Swartz sometimes a foot in length), 2;^ inches in diameter, 
and nearly circular in its transverse section ; it splits into two 
valves of a solid ligneous texture, about ^ inch in thickness ; it 
has a rather thin coriaceous dissepiment, quite smooth on both 
sides, parallel to and quite free from the valves, and a bi partible 
compressed replum, which lies between the margins of the valves ; 
these valves partly split down the middle, as in 
Distictis and Dolichandra : the two cells are 
filled with a great many irregularly orbicular 
and compressed seeds, much resembling those 
of Adenocalymna, closely packed together and 
imbricated, apparently without the least trace 
of intervening pulp (fig. 15) : the outer tunic 
of each seed is hard and smooth, and is trun- 
cated on one side by a straight marginal edge, forming an oblong 
Fig. 15. 
