OF PLANTS CALLED COMPOSITE. 315 



rium inferam." This distinction, however, is neither uni- 

 versal, nor I believe absolute in any case. 



M. Auguste Saint Hilaire in his excellent Memoir on 

 Primulacece} while he admits the correctness of M. Decan- 

 dolle's account with respect to great part of Dijjsacece, has 

 at the same time well observed, that in several species of 

 Scabiosa the ovarium is entirely united with the tube of the 

 calyx. But neither of these authors has remarked the 

 curious, and I believe peculiar, circumstance, of the base of 

 the style cohering with the narrow apex of the tube of the 

 calyx, even in those species of the order in which the dilated 

 part of the tube is entirely distinct from the ovarium. 



This kind of partial cohesion between pistillum and calyx 

 is directly opposite to what usually takes place, namely, the 

 base of the ovarium being coherent, whilst its upper part is 

 distinct. It equally, however,, determines the apparent 

 origin or insertion of corolla and stamina, producing the 

 unexpected combination of " flos superus " with " ovarium 

 liberum-" 



In the vascular structure of the corolla Boojndece may be 

 considered as essentially agreeing with Composite, in many 

 of whose genera the middle nerves of the tube and seg- 

 ments are equally manifest. In stating the character derived 

 from this source in either of these orders, it is not sufficient 

 to describe the nerves of the lacinise only as M. Mirbel has 

 done in his character of Composite/ and M. Cassini in 

 that of Boopidese : but it is also necessary to give their [iso 

 disposition in the tube or undivided part of the limb ; there 

 being instances in both families where the lateral nerves of 

 the segments do not unite at top ; and, as has been formerly 

 remarked, several examples in other families of a nearly 

 similar disposition in the segments, accompanied by a 

 different disposition in the tube. To the examples of this 

 kind formerly given, Globularia cordifolia may be added, in 

 the segments of whose lower lip there are three simple 

 nerves, of which the lateral do not unite at top, and con- 



1 Mem. du Mus. d'LIist. Natur. ii. p. 17. 



- Elements de Physiol. Veget. ct de Botan. ii. p. S85. 



