10 
as that of Lindernla, from which Bonnaya differs only in the lower 
filaments being always abortive, simple, and club-shaped. 
Vandellia, Artanema, and Torenia are connected together by an 
apparently constant character, the form of the filaments of the lower 
stamina, which, arising from the mouth of the tube, are always 
either bifid at the base, or furnished with a tooth or angle, and from 
thence the fertile filament is arched over to the upper lip, under 
which are their anthers, adhering to each other. This character was 
originally assigned to Torenia ; and Tittmannia was separated by 
Reichenbach, to include those species which have a short calyx not 
plicate, and a globular fruit, — these were considered by Chamisso 
as the true Toreniae, those with a linear or oblong fruit, shorter than 
the calyx (T. asiatica, &c.) are referred by him to Nortenia, and 
such as have a siliquose fruit and short calj-x to Vandellia. The 
length of the fruit, however, in this case, as in Bonnaya, is very 
uncertain (its dehiscence being the same in all), and I have preferred 
taking the form of the calyx (in this case much more conformable to 
habit) as the best character. The name of Torenia having been 
originally given to the T. asiatica, must be retained for Dupetit 
Thouars, Nortenia, in which that species remains, and as Linnseus' 
specimen of Vandellia, bad as it is, leaves in my mind no doubt as 
to its identity with one of the Brazilian species of the other genus, 
I have adopted that name for it. Artanema, with a peculiar habit 
(that of Sesamum), is distinguished as well by the form of the co- 
rolla as by the presence of squamae in the tube of the corolla, ana- 
logous to those of Hydrophylleee, Labiatse, &c. w'hich I have not ob- 
served in any other Scrophularineous genus. It was mistaken by 
Persoon for Loureiro’s Diceros, which, from his description, appears 
to me to be something very different, although I can form no guess 
as to what it is. 
Lindernia, which I have by mistake placed among those in which 
the lower filaments are entire, has them on the contrary bifid. The 
L. pyxidaria scarcely differs in character from the short-fruited Van- 
delliee. The American ones are distinguished from Bonnaya by the 
bifid filaments, and from Vandellia by the lower stamina being sterile. 
Perhaps it might be better to consider these three genera as forming 
but one. 
Heteranthia, distinguished, according to Nees and Martius, by 
the singular disparity of the anthers, is unknowm to me, as also 
Mecardonia of ]VJ artius, which, according to his description of the 
capsule, belongs to the 3rd sect, of Gratiolese, although it evidently 
is allied to Matourea, by the imbricate many-leaved irregular calyx. 
Conobea of Aublet has precisely the habit of some of the Ameri- 
can Herpestides, but wdth a regular calyx. Not having an opportu- 
nity of examining the flowers and capsule in the only specimens I 
have seen (in the Banksian herbarium), I am unable to determine 
the precise situation of the genus. 
