BANUNCULACEJE. 67 



number of very different characters, and to group those genera tliat 

 we retain around a few well-marked centres which tliey approach 

 more or less closely. Hence certain genera happen to be on the 

 peripheral limits of two or more groups at once, and indicate by 

 what features these groups are bound together. Or, indeed, if we 

 draw up each of these groups in a line, with the typical species at 

 the head, we obtain a certain number of series which are parallel or 

 nearly so for some part of their course, but afterwards diverge 

 in various directions, and hence must intersect, their intersec- 

 tions indicating the characters common to the different sections.' 

 The prototypes that we have chosen provisionally" are Jqf//7f'pa, 

 Ranunculus, Clematis, and PcBonia, from which we afterwards derive 

 the other genera by the modifications found in the number and 

 direction of the ovules, the number of pieces and whorls in the 

 perianth, the symmetry and a)stivation of the flower, the position of 

 the leaves, &c. 



The Banimcnlacece are almost alsvays herbs, far more rarely 

 annuals than perennials. In the latter case we have seen how they 

 are propagated in the different genera by buds, nourished while 

 developing by the accumulation of juices, either in their own bases, 

 or in those of the neighbouring organs.^ The herbaceous stems 

 usually possess a normal or nearly normal organization. The pith 

 of branches which grow fast sometimes contracts so as to render 

 them more or less fistular." In several of these same species the 

 fibro-vascular bundles, dispersed with little apparent order through 

 the cellular mass, have the same distribution as in the stems of 

 Monocotyledons, and the medullary rays may lose their usual recti- 

 linear course so as to render doubtful their existence.* These fibro- 

 vascular bundles (often numerous in herbaceous stems that have 



' Thus we have shown (Adansonia, iv. 41), how the pra?floration of the calyx ; and so purely pro- 



Ficaria, by its close analogy to Caltha, connects visional is the grouping we propose, that we have 



Trollius and Manuncndus ; how the Hellebores, already s\id {Adansonia, iv. 55) that instead of 



closely allied to Trollius, lead back to the keeping the Clematis series distinct, we should 



Nigella which are Columbines with deduphcated perhaps do better in joining it to that of the 



nectaries. Trautvetteria is allied by its habit to Anemones. In fact we have seen (pp. 50, 51) that 



Actaa and Thalictrum, by its flower to Eanun- at a given moment the estivation of a Clematis 



cuius. Thalictrum, only separated from Actcea may become that of Anemone. 



by the lesser number of its ovules, is at the •* See pp. 5, 31, 38. 



same time closely allied to the Anemones by * In several aquatic species of Hanunndus 



Syndesmon, and Xanthorhiza, formerly left un- Delphinium, Aconitum, Anemone, Thalictrum 



separated from . 4c/ff a and Pffowia is, says Payeb, (on these last, see T)e G^-rvy-t:, Xylologische 



merely Aquilec/ia with but few staminal whorls. IStudien, Bull. Soc. Mosc, 1861, 423). 



2 From what we have stated before it will be » Haetig, Beitr.z. Vergl. Anat. der Eolzpfl. 



seen that Clematis and Anemone only differ in {Bot. Zeit. (1859), 93, 96). 



V 



