BERNARD AND A. L. DE JUSSIEU 355 



recognised by every botanist; among these are the 

 Labiates, Umbelliferse and Leguminosse. He adds the 

 instructive remark that these families, which are so easy 

 to define, are hard to divide, the agreement of the whole 

 family in many particulars of structure leaving only 

 minute differences for the subdivisions. He shows that 

 in Eanunculacese the separation of the sepals and petals, 

 the insertion of the stamens and the attachment of 

 the anther-lobes are constant. On the other hand, 

 Kanunculacese vary in the character of the stem (her- 

 baceous or woody), in the insertion of the leaves (alter- 

 nate or opposite), in the number and function of the 

 petals, in the number of the stamens, in the number and 

 cohesion of the carpels, and in the number of seeds to a 

 carpel. The fruit may be an achene, a follicle, a capsule, 

 or a berry. Doubtful cases must be decided by the 

 aggregate of all the characters. Anomalous Ranun- 

 culaceae suggest to Jussieu instructive reflections. Thus 

 Actsea and Podophyllum have single carpels which ripen 

 to berries, yet they are undoubtedly near of kin to the 

 Ranunculaceae ; whether they are to be placed in that 

 family, or in the allied family of the berberries, signifies 

 little in Jussieu's opinion ; they make a transition from 

 one family to another, and such transitions, which are 

 faults in an artificial system, are merits in a natural 

 system.^ Elsewhere he traces a transition in the petals 

 of different Ranunculaceae.^ Such reasoning, supported 

 by well-chosen instances, justifies the respect in which 

 the Examen has been held by subsequent generations of 

 botanists. 



In his Exposition A. L. de Jussieu gives a table of 



1 «• Cotto transition, qui Horoit rogardiio oommo un d^faut daiM let tytUnat, 

 est une perfection dans I'ordre naturol." Excunen, p. 236. 



2 Examen, pp. 225-6. 



