336 NATURAL HISTORY OF PLANTS. 



colouring. It is however certain, the mode of insertion, and the 

 organization of the ovules are very different in Cistacece and Dillcni- 

 acece ; if not, we might consider the former as representing a form 

 of carpels united edge to edge in a unilocular ovary, while the latter 

 would have, in general, independent and unilocular carpels, and 

 would be to Cistacece what Illicece is to Canettea, Anonece to 3fo?w- 

 dorece, Astrocarpcce to Resedece, &c. The false racemes of Helianthe- 

 mum, compared with the unilateral inflorescence of certain species of 

 ] libber tia, would singularly complete the analogy between the two 

 groups. On the other hand, Cistacece has been placed by most 

 authors near Capjjaridacece, Resedacece, and Bixacece. They have not 

 the habit, corolla, anatropous or campylotropous ovules, nor the 

 seeds of either of them. They cannot always be absolutely distin- 

 guished from all the Capparidacece, by the presence of an albumen, 

 since certain of them are also provided with it. But in Cistacece it 

 is either farinaceous or subcartilaginous. The orthotropy of the 

 seeds, and the more or less pronounced curvature of the embryo, 

 often convolute and conduplicate, serves, however, to distinguish 

 Cistacece from Bixacece. The latter have sometimes a calyx, with 

 unequal sepals, with two small bractiform and exterior leaves, as in 

 so many species of Helianthemum, such as Ryania, which is, moreover, 

 destitute of petals. The Violacece are correctly considered as nearly 

 allied to Cistacece ; but they have either irregular flowers, or when the 

 corolla is regular, a definite number of stamens and ovules, and seeds of 

 quite a different character. The polypetalous Canellece have almost 

 the same organization as the Cistacece as to perianth and placenta- 

 tion ; but their monadelphous stamens, fleshy fruit, and anatropous 

 seeds are totally different in character. There are also analogous 

 resemblances between Luxemburgiece and Cistacece ; but the former 

 has a characteristic foliage, an eccentric gynseceum, and anatropous 

 ovules. We may say, in short, that Cistacece, a syncarpous form of 

 Dilleniacece ?, is a connecting link also between Bixacece and Violacece. 

 Turnerea, which we have, moreover, connected with Bixacece, is also 

 very analogous to the Cistacece by its corolla, mode of placentation, 

 and capsular fruit ; it is distinguished from them particularly by 

 the definite number of its stamens, and often also, but not con- 

 stantly, by the difference in their mode of insertion. 



