32 ME. BENTHAM's notes on irOMALIUM. 



they do each other, and aa tlie whole aeries have a remarkable 

 conformity in the general structure of their flower as well as in 

 foliage and in habit, it appears more appropriate to consider them 

 all as one genus, which can be artificially divided into two sections 

 on the old character. 



There is one point, however, in which the American species 

 appear slightly to difler from the Asiatic and African ones : the 

 fruit, at least in H. densiflm^um and H. pedicellatum, becomes very 

 hard, and in our specimens shows no disposition to open in valves, 

 whilst it does so most readily in some of the Asiatic ones, splitting 

 the styles so as to cause them to have been occasionally described 

 as double their real number. But I have not seen the fruit quite 

 ripe in any species : the ovary, shortly after flowering, appears to 

 harden in the Asiatic H. grandiflorum as in the above-mentioned 

 American ones ; and in the original American R. racemosum and 

 H. Macoubea, the fruit is said to open in short valves at the top, 

 although I see no tendency to it in our specimens. This cha- 

 racter cannot therefore be made available for generic distinction, 

 being accompanied by no corresponding differences in habit or in 

 other organs. 



Among old genera not really differing from Homalium, Hacouhea 

 (Aubl.) has long been united with it, and Astranthus (Lour.) with 

 Blachwellia ; Miquel has also correctly joined with it the Gondy- 

 locarpus lately established by Blume. On the other hand, Na- 

 pimoga (Aubl.), which does not appear to have been examined since 

 his time, can scarcely be a congener, not having the characteristic 

 glands : the analyses given, rude as they are, are not to be de- 

 pended on for correctness, and afford no evidence of the plant 

 belonging even to the same natural order. 



In the distinction of the species, besides the artificial sectional 

 character above mentioned, the inflorescence, either a close panicle 

 with short divaricate branches, or long racemes, either solitary or 

 few, forming a long loose panicle, as indicated by De Candolle, 

 divides well the Blackwellias into two groups ; and the species are 

 mostly w^ell characterized by the relative size and shape of the 

 calycine segments and petals taken especially w^hen slightly en- 

 larged after the flowering is over. In some of the section Black- 

 wellia they are all nearly of a size, linear or oblong, slightly nar- 

 rowed at the base, ciliate at the edges, so as to give the young 

 fruit precisely the shape of elegant little shuttlecocks ; in H. 

 senaHtmij H. pedicellatvm, and otlier.<, they are all broader at the 

 base, spreading or reflexed and radiate round the base of the 



