OF PLANTS CALLED COMPOSITE. 311 



completely differ from Compositae in the disposition of 

 vessels of the corolla, while both orders agree in the no less 

 remarkable structure of the jointed filament; a character 

 which had been observed in a very few Compositse ^ only 

 before the publication of M. Cassini's second Dissertation, 

 ^Yhere it is proved to be nearly universal in the order. 



In the opposite parietes of the ovarium of Brunonia two 

 nerves or vascular cords are observable, which are continued 

 into the style, where they become approximated and parallel. 

 This structure, so nearly resembling that of Compositse, [is* 

 seems to strengthen the analogical argument in favour of 

 the hypothesis advanced in the present paper — of the com- 

 pound nature of the pistillum in that order, and of its type 

 in phsenogamous plants generally ; — Brunonia having an 

 obvious and near affinity to Goodenovim, in the greater part 

 of whose genera the ov^arium has actually two cells with 

 one or an indefinite number of ovula in each ; while in a 

 few genera of the same order, as Bavipiera, Biaspasis, and 

 certain species of Scavola, it is equally reduced to one cell 

 and a single ovulum. 



Sir James Smith, in establishing _5r^<?^o?^^« as a genus, is 

 disposed to refer it to Bipsacece. To certain species of this 

 order it, indeed, bears a striking resemblance in habit; it 

 also very nearly agrees with them in its remarkable inflo- 

 rescence ; and one great objection to its union with them 

 may be supposed to be removed in adopting M. Decandolle's 

 account of their ovarium. 



But as Brunonia differs from the whole order in the 

 following characters, all of which are of primary import- 

 ance ; — namely, in the origin and sestivation of corolla ; in 

 the insertion and whole structure of stamina ; in the indu- 

 sium of the stigma ; in the ovulum being inserted at the 

 base of the cavity of the ovarium ; in the erect embryo and 

 want of albumen; — I continue to think that its proper 

 place in the natural method is between Goodenovm and 

 CompositcR. 



' Batsch Anal. Elor. p. 107 ; et Solikulir Handb. tab. 236 et 2il. 



