256 NATURAL HISTORY OF PLANTS. 



more closely allied to the Ternstrcemiacea*. Besides, these latter have 

 some genera (all the Bonnetiece except MarUa) where the corolla is 

 contorted as in Dipterocarpacece. But in this case the fruit is cap- 

 sular, polyspermous, and the habit is different as well as the foliage, 

 and it is very rarely that the calyx is accrescent round the fruit of the 

 Ternstrcemiacea? ; it never forms round or above it one or several 

 large membranous or ligneous wings. But we should not really de- 

 ceive ourselves as to the value of these latter differential characters. 1 

 We ought to say as much of the character presented by the indu- 

 sium of ChJaenacea. These (which might possibly be admitted as a 

 series in this family) are always distinguished by the three follow- 

 ing points : — within the sac of the indusium traces of the true 

 calyx are alwa} r s found ; the pieces are smaller in number than 

 those of the corolla; the stamens are inserted within a tube of 

 variable length, which in the monadelphous Ternstrcemiacete is 

 formed by the base of the staminal filaments themselves. 2 We shall 

 see subsequently that the Ternstrcemiacece with opposite leaves are 

 very difficult to separate by absolute characters from Hypericacea; and 

 ClusiacecB, with which they have usually been considered as very 

 nearly allied, and that they also present incontestable affinities, 

 although more distant with some Sapotacea and Ericaceae. By Acti- 

 niclia, extremely analogous to Saurauja, they approach Billeniacece by 

 Dilleniea? series ; and there are some Ochnacece of the Luxemburgia 

 series which ranged sometimes among the Ternstrcemiacece, may be 

 confounded with them by their habit, their foliage, and their inflo- 

 rescence, so that we should not be able to distinguish them without 

 having recourse to the observation of their tolerably developed sti- 

 pules, to their more or less oblique ovary, and to the characters pre- 

 sented by the insertion of the style, the organization of the anthers, 

 and the direction of the ovules. Let us conclude, that the limits of 

 this family are extremely artificial. 



The Ternstrcemiacea? are about 'two hundred and sixty-eight 3 in 

 number. They are scarcely ever found farther north than the Asiatic 



'Certain Dipterocarpacece, like Pachyno- asked if Rhodolcena is not one of the Chlcenacecs 



carpus, have the fruit destitute of wings, and the (B. H., Gen., 195). 



calyx is accrescent after anthesis in Visnea and 3 Lindley ( Veg. Kingd., 397, Ord. 142) 



Anneslea. admitted a hundred and thirty in 1846, without, 



' 2 Chlanacece belongs evidently to a region it is true, counting Marcgravice and Caryocarece, 



where there are scarcely any Ternstrcemiacea; which, according to him, comprised thirty-four 



there are only one or two species, little known species, 

 and doubtful, in Madagascar. But it has been 



