338 QUESTIONS IN RELATION TO HE 
better explanations, and even if there is some truth in the principle, it 
is peculiarly liable to abuse in its applications. Dr. Gray follows 
Brown in believing that the Gynoecium of Brassicaceae consists of only 
two carpels, a view which has been already sufficiently commented 
upon. Though particular in describing the glands, and employing 
them as characters of genera and species, he does not refer to them in 
judging of the symmetry of the order, and he relies on the arguments 
of Moquin Tandon and Webb, to prove that the six stamens represent 
one circle of four. These arguments then I must review :— 
Ist. In some species, as Clypeola cyclodontea, the filaments of the 
solitary stamens are furnished with two teeth, one on each side, whilst 
those of the double stamens have but one on their outer side. If we 
join these two stamens together, so that they form but one, a bidentate 
filament will result entirely similar to the solitary stamens. This is 
without doubt plausible, but we must recollect that the two anthers 
of a stamen represent the two sides of the lamina of the leaf, their 
presence therefore shows the completeness of the organ, whilst the 
tooth-like projection on the filament is only representative of a wing 
to the petiole, or an angle at the bottom of the leaf; since then each 
of the pairs of stamens has its two anthers, we must conclude that the 
development of the tooth at the inner side in the pair of stamens is 
prevented by the two organs being so near to each other, which causes 
a pressure unfavourable to such development. | 
2nd. In other species a longer or shorter portion of the filament 
remains simple, thus in Stertgma tomentosum the division takes place 
as far as the middle; and in Anchonium Billardieri in a third part 
only of the upper portion of the filament. Here the position of the 
longer stamens, double only m their upper portion, is exactly the 
same as that of the solitary stamens—these facts I reply afford no 
argument, because they are easily explained by partial coherence (an 
exceedingly common occurrence) of organs really distinct, and the two 
anthers tend to prove this distinctness. _ 
3. In Vella pseudo-cytisus we find in the place of the double stamens, 
a single one, its filament being frequently rather broader, sometimes 
divided only at its summit, sometimes entirely undivided, but bearing in 
that case an anther wholly or partially gemminated. I have not examined 
this case, but the description indicates a more complete coherence of 
two organs. Instances however which occur, of only one stamen being 
found in the place of the pair, are only cases in which that circle, as 
