PROF. OLIVER ON THE STRUCTURE OF THE ANTHER. 12/ 
usual form of anther, and we find that an exceptional mode of dehiscence, by valves 
opening from below, characterizes the order*. With regard to Perls, referred to by 
Treviranus {supra, p. 425), I have examined several species (P. quadrifolia, P. hcxaphylla, 
P. obovata, P. incomplete*,, and P. polyphyllaf) ; and I certainly cannot agree with the 
opinion expressed by him, that the valves of each anther-cell answer to upper and 
lower surfaces of the leaf. The dehiscence is, it is true, apparently almost quite mar- 
ginal, while in Trillidium Govenianum it is distinctly extrorse; but here I find no diffi- 
culty further than may be explained by a slight dilatation of the inner face of the con- 
nective, or by the development of the lobes answering morphologically to the inner pair oi 
thickenings of parenchyma upon the outer, or lower, face of the metamorphosed lamina, 
instead of upon the inner or upper surface. In Smilax it would seem probable that the 
anther-cells originate from but two thickenings of parenchyma, one on each side of thi 
^ 
* !• 
midrib of the stamen-leaf. Dehiscence is along this median line, on the face of i he ant h«'i 
(vide PL XLIV. fig. 18). I consider, indeed, the structure of the anthers in Paris to be 
as easily explicable as in Geranium itself; in fact, the anthers in the latter present, from 
their versatile attachment and inconspicuous connective, th<- more specialized struct un 
of the two. And it is from this circumstance especially that I think the evidence which 
they afford on the question of anther-morphology to be important, since in Papa* r, 
Nymphcea, and some other genera, in which the anther-cells have been ol iserved to originate 
upon the upper surface of a petaloid lamina, the anthers are innate or adnata, and tin- 
connective sometimes quite conspicuous. And further, as Neumann (/. c., p. 3"»8) re- 
marks, the anther-sutures of Geranium present the dark line which Roeper noted in some 
Euphorbiaceee, and which he considered to be confirmatory of his view, that the suture 
corresponded to the leaf-margin. Robert Brown, in his paper on R<'fflesia%, in discuss- 
ing the structure of anthers generally, says, in a note, " the principal point in which 
the anthers* and ovaria agree consists in their essential parts (namely, the pollen and 
ovula) being produced in the margins of the modified leaf." But though the exterior, 
or upper, anther-lobes originate as thickenings of the parenchyma, parallel and coinci- 
dent, or nearly so, with the margin of the leaf, the inner and lower lobes have no such 
necessary marginal relation. "With regard to the origin of anthers from portions of t issue 
morphologically corresponding to the outer lobes only, or inner lobes only, of the anther- 
It is true that the upper pair of anther-cells (in Cinnamomum at least) have their bases slightly included between 
the upper points of the lower cells, and that the portions of the stamen-leaf thickened, in their ette, cannot j.rec.-iy 
correspond in position with those prevailing in the abnormal forms which I figure ; but this I believe to be quite mam- 
portant. In Berberide* (Berberis, Epimedium, Leontice, Bongardia), besides dehiscence along the line of junction 
of the posterior and anterior lobes, the walls of the former separate round their base and, dorsally, along each side 
the connective. The inner margins of the anterior cells remain attached, as is usual. 
In P. 
, JEgiceras, and some species of Rhu-ophora, the pore-dehiscence of several groups, distractile, appendaged, and 
" one-celled " anthers, the anthers of Conifer* mid Cycadese, do not offer for explanation greater difficulties on this 
theory than on that of Cassini, Roeper, and others; indeed some, I think, are clearly more easily. explicable on this view. 
t The amount of depression of the line of suture previous to dehiscence may vary slightly in the species. 
hexaphylla I have seen it scarcely perceptible, excepting towards the extremities of the lobes ; but this was in a «— 
specimen. In P. incompleta the connective is scarcely produced, as is often the case in P. polyphylla. Conf. Hooker, 
HI. Himal. Plants, pi. xxix. 
X Linn. Trans, xiii. 211. 
