C0NNARACE2E. 13 



and the woody consistency of the stem. Other characters again are 

 both very valuable and nearly constant — namely, the pinnate leaves, 

 the orthotropous or nearly orthotropous ovules, the seeds possessing 

 an aril of variable thickness and localized or generalized. In the 

 third place come two characters, each present in about half the 

 Order and absent in the rest — a valvate calyx and an albuminous 

 seed. To these, however, an unequal value has been assigned, as we 

 shall now see. 



The character of the pignoration of the calyx has been held of 

 sufficient importance to serve to divide all the known Connaracea 

 into two tribes or series : the one, Connarea, in which the sepals are 

 imbricated in the bud, the other, Ctiestidea, in which they are valvate 

 If this clear demarcation came out in accordance with the lads, 

 this division of course would be most convenient in practice ; and we 

 have retained it for its convenience. But we cannot regard it 

 as being also absolutely natural. This position may be illustrated 

 by the fact that Troostwicki/a was placed by Bentham and Hookeb 

 among C/iestidete, because of its valvate calyx ; now this name is 

 exactly synonymous with Hemiandrina, a genus now suppressed, and 

 rightly considered a mere section of Agelcea, whose calyx is usually 

 imbricated, as befits the Connarea. Again, many species of Tricho- 

 lobus have altogether the flower of Omphalobium or Connarus, with the 

 gymeceum finally unicarpellary ; and a large number of them have 

 also the same vegetative organs; still, of these two types, so closely 

 allied in all their characters, TricJwlobus is referred for its valvate 

 calyx to Cnestidea, and Omphalobium, for its imbricate calyx, to 

 Connarete. Never was there artificial classification more con- 

 venient, we must allow; but at the same time, never was there 

 one that took less account of the generality of common characters. 



The character derived from the albumen is of even less import. 

 True, albumen is never found in any known member of the Connarea . 

 but while in half the genera of Cnestidea the seeds possess albumen, 

 in the other half they lack it. 



The other characters serve only to distinguish the several genera. 

 They are as follows:— 1. The prolongation of the receptacle above 

 the perianth into a column bearing the s-xual organs: this pecu- 

 liarity occurs only in Manotes. 2. The stalk to the base oi eacli 

 carpel; wanting in Eourea, present in Connarus. The absolute 

 number.of elements to the gynseceum : the specimens of '' 



