THE PLATANO-PROTEAL DIVISION. 



37 



POLYGALACEiE & MORINGACE^E. 



The foliage, fruit, and seed of Moringa are so much like those of Papilionacece, as to make a near affinity to that family very probable, and one of the species shows a more decided approach 

 in the stamens being ten, five of them fertile and five barren, the posterior one being barren, very short, and looking separate from the rest, so that a vacuity is left on the posterior side of the 

 otherwise closely packed stamens, like the fissure in the posterior side of the staminal tube of that family. This condition of the stamens, however, occurs equally well marked in Polygalacem, with 

 which I find they agree in having the odd sepal posterior, it being exactly opposite the vacuity on the posterior or upper side of the decimate stamens, and therefore exactly opposite the shortest 

 barren filament, the non-development of which produces the vacuity. The odd petal is therefore anterior, and the two posterior petals are evidently alternate with the vacuity on the posterior side 

 of the stamens, as in Polygalacece. The two posterior petals, in a part of the species, are of a different form from the other three, and are tufted with hairs on their lower half, while the other three 

 are smooth, so that the former are obviously a pair, as in Geraniacece and other families. 



The stamens of Moringa are in most, if not all the species, more or less decimate, being very unequal in two or three of them, the two posterior, when there are nine, being scarcely more 

 than half the length of the anterior one, which is the longest.* The anther is one-celled from consisting of only half an anther, as it is in a part of the species unequal-sided, one side being larger ; 

 and on the larger side is a furrow, having the appearance as if it were produced by an attempt to extend, in an oblique direction, the fissure which takes place in its dehiscence. 



In the absence, perhaps, of any proof that the one-celled anthers of Polygalacece are only half-anthers, the simplicity of their appearance and the single fissure, as in Poly gala vulgaris, 

 makes it most probable that they are so ; and if so, Moringa might take its place as a section of Polygalacea. Phaseolus vulgaris when dicarpous has all the appearance of having parietal placentae.f 



KRAMERIACE^. 



In Krameria the sepals are much and unequally imbricated, besides which they are unequal in size, which makes it difficult to decide their position, so that I believe it would be scarcely 

 possible to say whether the odd sepal is next the axis or the contrary. But, if the irregularity of the flower is taken as a guide, we find three of the petals separated as it were from the other parts 

 of the flower, and, as regards the irregularity of the flower, they are distinctly and certainly posterior. Of these three petals one, i.e., the intermediate one, is consequently directly posterior, which 

 shows the irregularity of the flower to be the same as in Leguminoste. Krameria further agrees with Leguminosce in the ovary consisting of a single carpel (the cavity at least belongs to one carpel), 

 and the raphe also in the two ovules is obviously quite lateral, which is not known to occur in Polygalacea. As the position of the odd petal, (known by its different form,) directly anterior, may be 

 the only separating character between Polygalacece and Leguminosce, I prefer placing Krameriacem near the latter. If the flower is inverted, then the carpel would be always posterior. 



Trigoniace^. 



The same remarks with regard to the position of the sepals, and especially of the petals which apply to Krameria, may be made with equal certainty with reference to Trigonia ; it differs 

 however, in having numerous horizontal ovules, and in this character especially it agrees with some of the genera of Leguminosa, to which it may possibly belong, notwithstanding the considerable 

 quantity of albumen its seed contains, as the embryo is large, having thin flat cotyledons, which are as broad as the seed. 



Chailletiace^e. 



The affinities of this family are perhaps more obscure than those of any other in this Division, but there are three of the more remarkable characters which bring them near either to 



* I should suppose there are 4 or 5 species, unless they are liable to much variation. f I have met with 3 or 4 instances all exactly alike and also produced it by cultivation ; below the last ovules the ventral sutures are scarcely apparent. 



