180 APPENDIX. 
with which group they exhibit beyond all doubt a very close 
alliance. This is manifest in their general habit, their alternate 
leaves with glutinous pubescence, their fetid smell, their power- 
fully narcotic and other medicinal qualities, which are so charac- 
teristic of the Solanacee and Atropacee: to these may be added 
the particular structure of their stamens, which have their an- 
thers of a somewhat lunar form, and quite unilocular, curved 
round a large clavate termination of the filament, with an almost 
globular expansion of their connective, within the cell, that serves 
as the polliniferous receptacle, a character pomted out by Nees 
as being foreign to the Solanee and rare among the Scrophula- 
ring, and as claiming for them a distinct station in the system. 
On the other hand it should be borne in mind, that this peculiar 
character exists also in the genus Scrophularia itself, the flowers 
of which exhibit often declinate anthers and barbate filaments, 
together with a fifth sterile stamen, a feature rare in the Scro- 
phulariacee, and one that tends to show a very close connexion 
of this genus with the Verbasceew, with which tribe it had been 
before associated by all preceding botanists, until Mr. Bentham, 
in his admirable monograph of the order, has placed it among 
the Chelonee (DeCand. Prodr. x. 299). In most of the genera 
of this last-mentioned tribe, the anthers are formed constantly, 
I believe, of two distinct and divaricate cells, affixed at their apex 
on the slender summit of the filament, and quite wanting of the 
fleshy connective so manifest in Scrophularia and the Verbascee. 
Whatever may be determined in regard to the proper place of the 
Verbascee in the system, it is manifest that it is not by the 
number of the stamens that we can fix the limit between the Afro- 
pacee and Scrophulariacee: thus it is impossible to separate 
Celsia from Verbascum, and it would be equally as admissible to 
include Celsia with its didynamous stamens, or Janthe with its 
single pair, in Atropacee, as it is to place Verbascum, with its 
regular pentandrous flowers, in Serophulariacee: such disere- 
pancies cannot fail to occur in many solitary points of osculation 
between the genera of different tribes, in all our artificial modes 
of the classification of plants. We have also other instances not 
less strikingly contrary to the ordinary rule in the Xuaresia bi- 
flora of the ‘Flora Peruviana, which has a regular 5-partite co- 
rolla and 5 alternate equal stamens: this plant Mr. Bentham 
uh gly considers to be a true species of Capraria, a genus 
decidedly Serophulariaceous ; and in like manner the Bacopa of 
Aublet with its 5 equal stamens offers another exception, but 
here the plant has opposite leaves, and possesses so precisely the 
habit and general features of Herpestes, that its position must 
without doubt be fixed contiguous to that genus. The same rule 
will apply to another anomalous case instanced by Mr. Bentham 
