156 Mr. J. Miers on the Bignoniacese. 



foliaceous, and lying in adpressed pairs, free all round their mar- 

 gins, but united together at a small point to the very apex of the 

 terete accumbent radicle, which is of the length of the lobes, and 

 lies recondite in the centre, concealed by the free margins of these 

 lobes : this radicle, which points to the basal hilum, is placed trans- 

 versely and centrifugally in regard to the axis of the capsule, as 

 in most genera of the family. The cavity of the coriaceous in- 

 tegument which encloses it is nearly orbicular or transversely oval, 

 though sometimes much compressed, without any tendency to the 

 formation of a semiseptum at its base, as occurs in many genera. 

 The reason of this is that the neck of the intermediate integument 

 is not agglutinated to the sides of the outer tunic, but remains 

 free ; and in Argijlia this neck is extremely elongated, indeed 

 longer than the integument itself, so that it is coiled up into a 

 shorter space, much in the same way that Dr. Wight represents 

 it in Calosanthes (Icon. tab. 1339, upper figure). 



There are some exceptions to the occurrence of this deep 

 emarginature of the cotyledons, — for instance, in Calampelis, 

 where the embryo assumes the ordinary form of two simple co- 

 tyledons, nearly orbicular, compressed, and foliaceous, with a 

 short, terete terminal radicle, all united together at their slightly 

 cordate base. In OxycJadus, and probably also in Monttea and 

 Reyesia, the embryo is similarly developed, the cotyledons being 

 simple, oval, very thick, and fleshy, with a short terminal radicle. 

 In Platycarpum and Henriquezia, the cotyledons are simple, 

 thickj fleshy, and transversely elongated, with a deep hollow 

 upon their inner face, at their true base (or apparent side), in 

 which the radicle lies concealed. In Crescentia, Sesamum, Peda- 

 lium, and their congeners, the embryo is constructed as in Oxy- 

 cladus and Calampelis. In the Cyrtandracece the embryo is said 

 to be terete, with cotyledons shorter than the radicle. 



In Anemopcegma, however, where the embryo is of the form 

 described in Pithecoctenium, we find not only the broad mem- 

 branaceous wing, but the coriaceous discoidal portion also, cleft 

 at its summit ; so that the tunics become bilobed, and as it were 

 2-celled at that extremity, and the apices of the cotyledons 

 nestle in these spaces : but although the base of these tunics 

 has no such corresponding cleft, we find there a short transverse 

 septum, as before explained, connecting the dorsal and ventral 

 faces of the integument ; and within the two marsupial pouches 

 so formed the lower lobes of the cotyledons are isolated and 

 enclosed. 



This structure is carried to the utmost extreme in Adenocalymna, 

 where the seeds are not highly compressed, as usual, but are 

 very thick, nearly circular, with a hard, polished, crustaceous 

 covering, without any membranaceous wing, and with only the 



