APPENDIX. 171 
ened placentiferous dissepiment, its many-seeded capsular fruit, 
and its slightly curved embryo with an inferior radicle, are cha- 
racters quite opposed to its admission into that family. Nor can 
these be made to harmonize either with the Scrophulariacee or 
Solanaceae, to the latter of which they offer a nearer affinity. 
These characters are sufficiently prominent and distinct, and de- 
mand a more attentive investigation. 
y. Anthotroche.—This genus was placed by Prof. Endlicher 
in Scrophulariacee, among the Salpiglossidee, but it has been 
since excluded from the order by Mr. Bentham, and referred to 
Solanacee (DeCand. Prodr. x. p- 586). It appears to me how- 
ever to have as little relation with the one as with the other of these 
families. Upon examining a specimen belonging to this genus 
from Swan River, I find that in the structure of its anthers it 
agrees entirely with that just described as existing in Duboisia ; 
this consists of one reniform unilocular cell, fixed extrorsely on 
the filament, and dehiscing on the exterior face by one hippocre- 
pical suture. Here the tube of the corolla is short and straight, 
and the border is divided into five regular lobes, which are ro. 
tately expanded ; the stamens are 5 and equal. The ovarium 
has an epigynous prominent stylobasic gland as in Cacabus, ana- 
logous to that of Hyoscyamus. 
Respecting the Duboisiee it only remains to be observed, that 
the main points of distinction between it and the other tribes 
with which it is here associated, will be found to exist in the ex- 
trorse direction of the anthers and the singular zstivation of the 
corolla, peculiarities which, although very remarkable, are not 
of themselves of sufficient importance to claim for the plants that 
compose it the rank of a separate family, but they constitute a 
very distinct tribe of the Atropacee. It will consist of two sec- 
tions: 1. Euduboisiee, with baccate fruit, and 2. Anthotrochee, 
with capsular fruit, comprising Anthocercis and Anthotroche. It 
corresponds with the other tribes of the Atropacee in the ori- 
gin of the floral peduncles being lateral with respect to the point 
of insertion of the petiole. 
4, Schizanthee—The genus Schizanthus, from the lateral 
extra-axillary insertion of its pedicels and other characters, ap- 
pears evidently to belong to the Atropacee rather than to the 
Scrophulariacee, but it does not accord with any of the tribes 
above noticed. It differs from them in the structure of its an- 
thers, which consist of two parallel cells, quite distinct and sepa- 
rated from one another, but conjoined by a broad membranaceous 
connective, upon which they are dorsally attached: it. possesses 
five stamens, of which three are quite anantherous and rudimental ; 
the corolla is deeply cleft into numerous unequal segments which 
have an imbricate wstivation. Its stigma approaches the form 
