380 NAl URAL HISTORY OF FLANIS. 



differential traits, striking in the extreme types where we couki not 

 fail to aj^preciate them, diminish insensibly in numei ons intermediate 

 types and sometimes by an uninterrupted series of gradations, the 

 observation of which will lead to the conclusion that there is only the 

 most artificial liue of demarcation between the tribes with regular and 

 those with iiTcgular flowers formerly separated by Bltjme ; between 

 the series of A and B distinguished for the same reason by J. D. Hooker 

 in the group of Sapindew ; between those of the Soapworts and the 

 Pancovias^ which, following his example, we have preserved, as more 

 convenient in pi-actice, when it is a qiiestion of a family of which we 

 have so much yet to learn. But it will not be impossible to improve 

 on this in future as regards the natural characters. "We do not think, 

 for example, that Diphglottis ought to be relegated to a different 

 series from Cupania to which it is attached by so many common 

 characters, simply because its flower is iiTegular. We see the closest 

 affinities between Harpullia and Cossignia, the first being regular, the 

 latter irregular. We do not affirm that the Ilemigyrosas of Bltjme, with 

 their irregular flowers, are not however quite as closely allied to the 

 irregular types Anomosanthes and Scorodeyidron^ from which J. Hookee 

 was bound to separate them, as to Pancovia near which he was 

 obliged to place them. A genus such as DUtelasma, separated con- 

 siderably from Sapindus on account of its ii'regular floral type, has 

 however such affinities with it that it was for a long time considered as 

 congeneric. These are, it is plain, all questions which still requii-e 

 to be thoroughly examined into. 



As to the organs of vegetation, the Sapindacece are very rarely herbs, 

 suflfrutescent at their base ; and in that case they are climbers holding 

 on by teudi'ils, as occurs so frequently when they become enormous 

 woody bindweeds often described in standard works on account 

 of their abnormal structure.' In Pcmllinia, Serjania, and Urvillea, 

 for example, these stems are often characterised by the presence of a 

 central woody body siu-rounding a pith and a medullary sheath, but 

 itself surrounded, and sometimes with great regularity by three or more 



1 GKvmm. Rech. Sur I'Oiy. Ve'i/et... (1841), 1861), 343, t. 161, 162— Ckxieg. in Bot. Zeit. 



t. 13, 18; in G«i7fe(«. ^cc/i. "^o(. ii. 501, t. 19. (1851), 481. — ISchacht. Lcrbiich, ii. 57.— 



— A. Rich. .B/em. ed. 10, 77, fig. 45. — Mart. Metten, in Zinncea (1847) 682. — A. Juss. 



Oclehrt. Anzciff. (1842), 390.— Tbevir. in Hot. Momgr.Mulpigh. (1843), p. i. 110.— Oliv. Stem 



Zeit. (1847), 393. — Schleid. Orwidz. (ed. Dicot. 10. 



