10 KANUNCULACE^E. 



pointed out by Mr. Brown (and which may also be observed in several 

 other instances), serves to confirm the general rule. The analogous 

 eversion of the raphe, in many Rhamneae, is shown by Mr. Bennett to arise 

 from the lateral torsion of the funiculus : but " the object of this displace- 

 ment," lie concludes, " it is difficult to conjecture."' * Perhaps some light 

 may be thrown upon it by the present eases, in which the design of this 

 arrangement may, I think, be distinctly perceived ; namely, to facilitate the 

 fertilization of the ovule, by placing its foramen in juxtaposition with the 

 placenta, or that portion of the carpel (the confluent edges of the infolded 

 metamorphosed leaf) which is in the ovary a direct continuation of the stig- 

 matic surface or lines of the style, f and through which impregnation is ef- 

 fected. A glance at the analyses on Plates 2 to 5, especially those of He- 

 patica (Plate 5, fig. 4) and Myosurus (Plate 8, fig. 6, &c.), will render this 

 evident, and show that, if the ovule were brought into the normal position, 

 its orifice would be thrown to the side of the cell farthest from that through 

 which the fecundating influence is communicated. 



In the case of Rhamnus, where a solitary anatropous ovule arises from 

 the very base of each cell, a broad ventral funiculus, interposed between the 

 foramen and the placental surface, may readily be conceived to offer an ob- 

 struction to fertilization, while the subsequent lateral torsion of this funicu- 

 lus would bring these parts into the most favorable position. Where there 

 is a pair of ovules, as in Celastrus, no displacement is needful for the attain- 

 ment of this end ; since the raphe is originally lateral in such cases, that 

 is, the two raphes are applied face to face, or very nearly. This equally 

 occurs in horizontal collateral ovules, as in Magnolia, Plate 22, and no less 

 so where they consist of many pairs, as in Helleborineae generally, or even 

 where the numerous ovules are not collateral. Indeed, this may be assum- 

 ed as the typical condition ; the ovules, which are a growth from the pla- 

 cental margins of the infolded leaf, being themselves likewise folded inwards, 

 thus bringing their raphes next the suture. 



In no instance do we find the pericarp of the monospermous species co- 

 herent with the integument of the seed, as described by De Candolle and by 

 Endlicher. 



Conspectus of the Tribes and Genera. 



Tribe I. CLEMATIDEiE. — Sepals valvate-induplicate in iestivation, 

 colored, deciduous. Petals none or stamen-like. Ovaries numerous, form- 

 ing a head of achenia in fruit. Ovule solitary, suspended ; the raphe dorsal. 

 — Chiefly frutescent vines, climbing by their petioles. Leaves opposite. 

 Atragene. (Plate 1.) Petals, or staminodia, shorter than the calyx. 



Achenia caudate with the plumose-hairy persistent style. 

 Clematis. (Plate 2.) Petals none. Persistent style plumose or naked. 



" In Horsefield's Plonta Jaron. Rar. p. 131. 

 t Brown, in PI. Juran. Rar. pp. 108- 110, note. 



