QERANIACEJS. 27 



They form in tliis case a sectiou cliaracterised by leaves reduced to 

 three or to ouc siuo'lc foliole.' 



Geranium and the neighbouring genera, after having been generally 

 connected with the Mallows, became before the middle of the 

 eighteenth century- the type of a separate group. B. de Jussieu,^ in 

 1759, and Adanson, in 1763, each established an order Gerania; but the 

 former very improperly by placing beside Geranium and Oxalis^ Mai- 

 pighia and two neighbouring genera, the Sapindaceœ known in 

 his time, the Viues, Menispermcœ^ Passion flowers, and 31alvacece : 

 Bomhax and Hcrmannia. Adanson' much improved the family 

 by removing the Mcnispermeœ ^ Passijlorccv.^ Bomhacew^ and Am- 

 j)hélidcœ^ but he introduced the Nastiu'tiums, Melianthus, and 

 Viola. In 1789, A. L. de Jussieu-^ only left in the Geranium order 

 Geranium and 3Io}isonia, and as genera affinia, 2'ropa'olum, Balsamina, 

 and Oxalis, which he again unfortunately separated in 1817.'' The 

 Neuradcœ^ whose close affinities with the Geraniums had been 

 known to most botanists, had unfortunately, imtil quite recent times, 

 been placed in the order RosaceceJ Biebersteinia^ considered as 

 Rutaceœ,^ were in, 1862, placed by Bentham and Hooker^ among 

 the Geranieœ. The same authors reintroduced the B'alsaminece, 

 IVopœoleœ, and Oxalideœ, in the Geraniaceœ family, as well as 

 FlœrJcea^ and Liinnanthes^ before considered as a distinct family. 

 The Balhisiea' and Vivianieœ, from the time they were first known, 

 have been considered as very analogous to the Oxalidecc.^" 



As actually now constituted the family comprises eighteen genera, 

 divided into eight series whose general characters are the foUow- 



i Le JDnpatiia raccniosa (KoRTH. in Ned. Geranium with certain Sapiiidacfic, etc. in his 



Kiuidk. Arch. (1854), 381.— Pl. in Ann. Sc. Order 50, Trihihitœ. 



Nat. ser. 4. ii. 266.— MlQ. Fl. Lid. Bat. Suppl. ' Ex A. L. Juss. Gai. Ixviij. 



i. 398.— B. H. aeii. 277, n. 18), a plant from ' Fam. des PI. ii. 388, Fam. 49. 



Sumatra, which, according to its described * Gni. 268, Ord. 13. 



characters, is nearly allied to Arerrfioa, hut " In Mem. Mtis. v. 230, 232. 



distinguished from it by its simple leaves and " Voy. ]i. 10, note 2. — M. Hy. in Adansoiiia,-:^. 



ovary cells (" carptila ^^iibcoalita "). It has also 361. 



been doubtfully allied to the Qiiassica: {Simnr- ^ Lindl. Veg. Kiiigd. 469. 



bma:). ' Gen. 270, 271. 



2 LiNN.EUs, in 1738, in Classes Phmt. ranges '" Juss. in Mém Mas. v. 231. 



