THE PLATANO-PROTEAL DIVISION. 



Aquilariacece, Leguminosce, or Polygalacece, which in the accompanying Table are nearly approximated to each other. 1. The corolla in Tapura Guianensis and ciliata is unilateral, two of the petals 

 which are anterior being perfect, two which are lateral being ligulate, and the remaining one which is posterior being filamentous, so as to appear like a barren stamen, and this gives rise to an 

 irregular flower, in which the odd sepal is anterior, i.e., between the two perfect petals ; — the position of the petals and sepals is therefore the same as in Leguminosce, and as I find those of Krameria 

 and Trigonia* to have the same position, this family may have a near relation to them. The stamens also of Tapura have the same kind of irregularity, only two or three on the anterior side 

 being fertile, the posterior filament being shorter than the others. 2. The singular deeply bifid petal of Chailletia and most of the genera, may perhaps be accounted for by comparing it with the 

 barren stamens of Aquilaria, of which two belong to one stamen, so that supposing it to become a petal, and to retain its bifid character, the petal of Chailletia would be formed.f 3. The seed 

 of Moacurra gelonioides very closely resembles that of Aquilaria Agallocha, as described and figured by Dr. Roxburgh, (Trans. Linn. Soc. vol. XXL, p. 199,) both in its singular testa and embryo ; 

 for the inner integument of the seed in the former adheres so slightly (if in fact at all) to the external, that on dividing the seed the embryo covered with the internal integument can be readily taken 

 out, and it also agrees with the second integument of the latter, in being of a chocolate brown colour ; and the embryo of the former has, comparatively with the size of the cotyledons, the same 

 minute radicle (1. c. p, 199), which although attached to their bases is scarcely perceptible till they are separated. 



With Rhamnacece, Chailletiacem agree in habit, and with Celastracece in their inflorescence, but the irregular corolla of Tapura, and their two pendulous ovules with the raphe next the 

 placenta, prevent a close approximation to either of these families, but such coincidences may serve to show that the true station of this family is in this Division. 



And finally, they may perhaps be compared with Polygalaeece, especially with the § Moutabete, but like Polygalacece, Moutabece have the odd sepal next the axis, and their petals are not 

 bifid but quite entire, and although their stamens are more perigynous than those of Chailletiacece the approach between them is comparatively distant. 



Pen^sace^e. 



The genus Geissoloma differs so much from the other genera of this family, that I believe it should be kept separated as a distinct section of it, for besides other differences the two pendulous 

 ovules have the raphe dorsal. In Ad. de Jussieu's, and other figures which have appeared, the position of the foramen of the ovule is either doubtful or it might be supposed that it was 

 external, and the raphe consequently next the placenta, but having re-examined the ovules I am satisfied the raphe is dorsal, and that the appearances in the figures are due to a projection of cellular 

 tissue upwards and outwards from the raphe in the form of a wing-like process. 



Proteace^e. 



This family scarcely differs from Chrysobalanem except in the absence of petals and fewer stamens, and they seem very naturally to take their station as the apetalous form of that family and 

 Leguminosm, especially as the carpels when two are quite distinct (Ann. Nat. Hist. Ser. 2, vol. XI, p. 453), the gland also at the back of the carpel in Grevillea being, I believe, a second in a 

 rudimentary condition. With Rosacea they agree in the position of the raphe, but it may sometimes be lateral, as in some instances it is not distinct. The want of adhesion between the coats 

 of the seed in Persoonia may be regarded as showing a close approach to Chailletiacem v. supra. 



* Trigonia may also approach Chailletiacece in the 2 abrupt or irregularly emarginate anterior petals, and in the glands representing barren stamens which have much the same appearance as those of Chailletia, 



t Aquilaria agrees with Daphnacece in the scales representing barren stamens being double the number of the lobes of the calyx, but in Aquilaria as in Daphnacece they are not alternately opposite and alternate with them, but each lobe 

 has 2 of them opposite to it. Struthiola however among the latter has deeply, bifid glands alternate the sepals and these terminate in 2 apiculate processes which widely separate from each other which serves to explain the true nature of the barren 

 stamens in the other genera and in Aquilaria, 



