486 NATURAL HISTORY OF PLANTS. 



of its affinities in ranging with it, Liidwigia^ Epilohium, Circoeaj 

 and Trapa, some Mijrtacece^ Melastoma, Alangium, and even some 

 Bubiacece, A. L. de Jussieu ^ extended it still further, but rendered 

 it altogether heterogeneous by including in it some Ternstrcemiacece 

 as Visnea, some Saxifragacece as Vahlia and Escallonia, a part of the 

 OomhretacecB,^ some Santalacece as 0]phiraj SantcUum and Siriiim, 

 with several MyrtacecBj Melastomacece and the Loasece as allied genera. 

 But he did not separate Ealoragis (Cercodea)^ as did R. Brown ^ in 

 1814, recognizing at the same time that the Harolagece ought to be 

 placed close beside the Onagrariece, His opinion was adopted by 

 De Candolle,* who divided the Onagrariece into six tribes, one of 

 which comprised the Macrce, and the Halorageoe., regarded as a dis- 

 tinct order,^ itself divided into three tribes. The whole embraced 

 twenty-four genera, from which must be withdrawn Gallitriclie^ two 

 doubtful genera, Pleiirostemon^ and Onosurisj^ 2nid five types doubly 

 represented. The number of genera retained at this epoch numbered 

 therefore only sixteen. In a series of observations, models of patient 

 and exact analysis, Spach,® recognizing the little homogeneity of the 

 genera retained by De Candolle, thought it necessary to make a 

 great number of divisions which might well be considered as so many 

 distinct genera, but which we prefer to make only subgenera or 

 sections, following the example of Torkey and A. Gray,^ Bentham 

 and J. Hooker, ^° and most others'^ who have recently been occupied 

 with this group. 



We retain twenty-four genera, comprising about eight hundred 

 species, distributed over the whole world, but especially in the tem- 

 perate regions. Only two of them, Trajpa and Loudonia^ are limited 

 to the old world. Halcn^agis would be exclusively Asiatic and 

 Oceanic had not one species been observed in the isle of Juan 

 Fernandez. Mne other genera are common to both worlds ; but 



* Gen. (1789) 317, Ord. 6 ; Ann. Mus. in. 315. yet to be certain ; " there are some genera (ex 



2 From -wliicli unfortunately the apetalous Torr. et Gr.) which cannot be recognized " 

 types have been removed as far as possible. (B. H. Gen. 787). 



3 Flind.Voy.ll; Mise.Works (ed BEXN.),i. 21. " Monographia Onagreanmi, Ann. Mks. ser. 3, 



* Memoire sur la Famille dcs Onvgrariees iv. 321 ; Suit, d Biiffon, iv. 340 (1835) ; Ann. 

 (1829) ; Frodr. iii. 39, Ord. 74. Sc. Nat. ser. 2, iv. 161, 270 (1835). • 



» Op. cit. 65, Ord. 77. 9 Fl. N.-Jmer. i. (1840) 486-531. 



« Rafin. Jrmrn. Phys. Ixxxix, 258. ^o Gen. 785, Ord. 70 {Onagrariece). 



7 Rafin. loc. cit.—PIeurandra'Rxvm, Fl. Lud. i' Lindl. Veg. Kingd. (1846) 724, Ord. 278 



95 (not Labill.). These are perhaps ffino^A^ms (0;?aymc(?«).— Exdl. (?e;?. 1188, Ord. 265 {(Eno- 



inaccurately observed, but it is impossible as th^recB). 



