SOUTH AMERICAN PLANTS. 85 



small chalaza at the end opposed to the radicle ; the cotyledons are 

 large and fleshy, filling the entire cavity of the testa, flat within 

 and convex without ; the radicle is superior, very short and small, 

 and appears like an umbilicate disk. The plant in Gardncr^s 

 Brazilian collection, no. 760, is identical with the above*. 



SCHWENKIA. 



This 



gular structure of its corolla, which for a long while offered a 

 question difficult of solution. It was considered by Jussieu as 

 nearly allied to Browallia^ the two genera being placed by him 

 among the Labiatce. Linnseus, with much penetration, suggested 

 its affinity to the Solanace^By an opinion quite disregarded by 

 succeeding botanists. By Kunth it w^as classed, together with 

 Broivallia, in the Scrophulariace(^. Dr. Lindlcy, in his ^ Introd. 

 to Bot/ p. 224, arranged it among the Primulace^ej because the 

 stamens are opposite to the expanded segments of the border of 

 the corolla. Mr. Bentham subsequently pointed out what he 

 considered to be the true nature of those gland-like processes, 

 always seen between the divisions of the border, and which he 

 showed to be the true segments, w^hile the others were mere 

 appendiciform expansions, and under this ingenious point of 

 view, he was enabled to reconcile its structure with the opinions 

 of Jussieu : following the example of Kunth, he therefore ar- 

 ranged both Schwenkia and BrowaUia in the Scrophulariace<Xy 

 among his tribe of the Salpiglossidece (De Cand. Prodr. x. p. 122). 

 Martins suggested its affinity with the Acanthace<Sy a view not 

 confirmed by other botanists, and quite unsupported by facts. 

 It is now four years since I first explained in what respects this 

 genus differs from BrowaUia and other genera of the Salpiglos- 

 sideiB (huj. op. vol. i. 177), and I indicated the circumstances that, 

 in my opinion, point to its nearer affinity to Fabianay in the 

 Bolanace<B. This genus forms one of those instances, in which 

 it is difficult to determine, under the ordinary interpretation of 

 the respective ordinal characters, whether it belongs to Solanacece 

 or Scrophulariacea. In order to obviate uncertainty in similar 

 cases, I endeavoured to show {Joe. cit. p. 163), and again lately 

 {huj. vol. p. 46), how by separating certain anomalous genera of 

 these two orders, marked by peculiar characters, into a separate 

 family, a prominent and unerring line of demarcation may be 

 established between the former, — a difficulty that has hitherto 

 puzzled every botanist. Under this test, Schwenkia must be 

 referred to the SolanacecBy on account of the decidedly valvate 

 sestivation of the corolla, as I shall presently endeavour to show. 



* A figure of this species, with generic details, is shown in plate ^2, 



