264 ON THE GENUS VILLARESIA. 
at an early stage the cell of the ovary in Bursinopetalum stands above 
the line of the insertion of the stamens, and that it is owing to the sub- 
sequent downward growth, that it becomes inferior in the fruit. I cited 
the instance of Halesia, where the ovary in the first instance is in like 
manner superior, but where, from a similar cause, the fruit is rendered 
inferior; but no one for this reason would separate Halesia from Sty- 
racee. The two cases are quite parallel, and favour the view then sug- 
gested. In regard to the position of the genus in Cornacee, we have yet 
no fact on record to show that any of its genera possess a dorsal raphe ; 
the only evidence on this subject is that furnished by Mr. B. Clark 
(Kew Journ. Bot. v. 129, pl. 5), where the position of the raphe in 
Cornacee (Cornus albus) is shown to resemble that in Alangiacee (Mar- 
lea) ; these two Orders, hitherto placed widely apart, might perhaps 
with propriety be united, or at least approximated,* especially as, in 
both, the raphe is neither ventral nor dorsal, but singularly lateral, as 
in all the Colletiee. The position of the raphe will be found a very 
useful index in the determination of plants, and notwithstanding it has 
been neglected or despised by botanists, it remains a character of 
much value, as it indicates a fundamental difference in the development 
of the ovules, resulting probably from some unknown modification in 
the origin of the carpels; it shows that in all the Celastral group of 
families, the ovules, whether erect or suspended, have a peculiar mode 
of origin, which I have called hypopylar, to distinguish it from the 
epipylar development, which usually occurs in Phanerogamous plants. 
. Baillon (l. c. 99) objects, with great reason, to the doctrine of 
the authors of the new ‘Genera Plantarum,’ enounced in p. 342, that 
“the Cornacee differ from the Olacinee only by their inferior ovary.” 
We must infer that they had only their tribe Zcacimiee in view in 
that declaration, for we can hardly believe they could possibly include 
Olacacee proper in the same category. Upon this opinion regarding 
* This identity was long ago shown by Mr. 
y Bennett, in describing Polyosma 
(Fl. Jav. var. 194), the affinity peel Cornus and Marlea having been first pointed 
out by Mr. Rob. d wn i arr. ejns e Aie ee ed x r. metes 
o no 
Dr. 
amag con (Diacaca rpiu H assk. in Bonplandia, vii. 172) with vrbes 
artling (* Ordines Plantarum,’ 238) e bi racing 
dera, €: separate Order, which he re 27 Pisza ombines Cornus, Marlea, 
+ my — on the development of the 6 rale, Aun. Nat. Hist. ser. 8, iv. 
24; Contrib. Bot. i. 120. 
