^ 



APPENDIX. 21 



panded. The stamens are equal, hairy and sHghtly arched at 

 their origin, slender, smooth, and erect above ; the anthers con- 

 nive around the included stigma, and arc oblong, cordate and 

 apiculated; like those oi Hyoscyamus, they are articulated upon a 

 prominence of the dorsal connective. The ovarium is seated 

 upon a short hypogynous gland with two prominent lobes, •op- 

 posite the furrows of the dissepiment ; these lobes remain after 

 the growth of the ovarium, but the gland itself soon disap}>cars. 

 The stigma is capitate, somewhat 3-lobed, and covered with nu- 

 merous viscose papillse. I observed the fruit of a specimen in 

 M. de Boissier's herbarium {M. microcarpa from ^Malaga) ; here 

 the persistent calyx preserves the same form, the tube growing 

 to a diameter of 7 Hues and a length of G lines, while the erect 

 lobes in addition are 9 hues long ; it is membranaceous, reticu- 

 lated, and incloses an oval berry crowned with the persistent 

 style, being 7 Knes long and 5 or 6 hues in diameter j the seeds 

 are flat, reniform, oval, and about 1^ line long *. 



Anthocercis. 



Th 



Duboisia in a separate division of Solamcea by Jlr. Bro^^^l 

 (Prodr. 448). MV. Bentham, first in Lindley's ' Introd.' p. 292, 

 and subsequently in the 'Prodromus' of DeCandolle, x. 191, 

 placed it among Scrophulariacea, in his tribe Salpujhssideie . 

 About four years ago [liuj. op. vol. i. p. 170), I offered several re- 

 marks, with the intent of showing that it possessed many peculiar 

 features not before observed, quite distinct from Salpiglossis and 

 its alhed genera, for which reason I suggested its association with 

 Duboisia and Anthotroche, in a separate tribe [Dubomea], form- 

 ing a section of an extensive group, distinct as well from true 

 Solanacea as from Scrophulariacece, and which group I proposed 

 as an intermediate family [Atropacea) between the large natural 

 orders just mentioned. The reason of its being placed m Scro- 

 phulariace^B by Mr. Bentham was obviously on account ot its 

 didynamous stamens, notwithstanding the presence of a rudi- 

 mentary fifth : at that time, however, the closely allied genus- 

 Anthotroche, with five regular fertile stamens, was not known. 

 I have since shown that nearly half the genera heretofore placed 

 in Solanacea present unequal stamens, with a strong tendency in 

 many to assume a didynamous character; while, on the other 

 hand several unquestionable Scrophulariaceous genera have hve 

 regular and equal stamens. The obliquity of the corolla am! 

 irregular dimensions of its segments, and the unequal size of the 



* An analysis of the generic features of this genus are given in one of 

 the supplementary plates at the end of this volume. 



