THE CASUAEINAL DIVISION. 



33 



The second pair are alternate with the first pair, from which they do not differ much in appearance, except that they are smaller and are separated at their bases by the stamens, which appear as if 

 inserted between them, but on a close inspection they prove to be slightly internal. The stamens are two, alternate with the second pair of sepals and consequently are anterior and posterior. 

 The female flower differs from the male only in the absence of stamens. The style and stigmas resemble those of some genera of Euphorbiacea, e. g. Caturus, in being much larger than usual 

 compared with the size of the ovary. The two stigmas are sometimes more or less bifid, and between them are four or five rudiments in which it agrees with Buxus semper vir ens. 



Corema is very different from Empetrum, besides the ovary being only 3-celled with only three stigmas, but may be compared with Ceratiola, if it is admitted that the inner pair of floral 

 envelopes are sepals, which appears to be sufficiently obvious, as it consists of two opposite scales placed right and left of the axis as in that genus. In C. alba the male flower has three pairs of 

 scales resembling sepals. The external pair I believe must be bracts, because being lateral they are opposite the innermost pair, which are, by a comparison with Ceratiola, sepals ; and they are 

 moreover inserted by a comparatively broad basis around a minute peduncle, which the succeeding pairs form by the contraction of their bases. The intermediate pair are doubtless sepals, being 

 anterior and posterior, but are not distinguishable from the internal pair except by their distinctly external position and inequality, the posterior being frequently much larger than the anterior. The 

 innermost pair which are the sepals before noticed as being lateral, are alternate with the intermediate pair. The stamens are four opposite the sepals, the posterior being most commonly the larger 

 and the anterior the smaller, so that they correspond in their inequality with the sepals, and the anterior is sometimes deficient, which occasions the two lateral to appear as if alternate with the 

 anterior and two lateral sepals. The anthers in aestivation appear as if opening inwards, but may perhaps be versatile. Corema is the only genus in which the male flower has no rudiment of 

 an ovary. The female flower differs from the male only in the absence of stamens. 



Dr. Asa Gray's character for this genus agrees with the foregoing so far, as that he describes the floral envelopes taken together as consisting not uncommonly of six scales. (Mem. Americ. 

 Acad. New Ser. vol. III.) 



Oakesia may be regarded as a reduced form of Corema, but differs so much that it may be a distinct genus, there being in the male flower neither bract nor sepal posterior, whereas in 

 Corema the posterior sepal is the larger of the two. In O. Conradii the male flower consists of three scales having the appearance of sepals, (but not forming a symmetrical whorl like the sepals of 

 Empetrum nigrum,) two of them being lateral and one anterior, within which are three stamens, occasionally four, and a rudimentary ovary having three stigmas. The two lateral scales are 

 evidently an opposite pair, like the two lateral bracts of Corema alba, and they also curve slightly backwards in the same manner, so as to leave scarcely a doubt of their identity. The anterior 

 scale is inserted distinctly within the two lateral, which are on the same plane, and is consequently a sepal ; and this inference is further supported by Corema alba, in which the anterior scale is a 

 sepal ; and also by its being more deciduous than the two lateral, which remain attached after it has fallen off. Of the three stamens, two are lateral and one posterior, to which occasionally a 

 fourth is added anteriorly, when it agrees precisely with Corema alba in its stamens,* but it, however, differs from it further in the presence of a rudimentary trigynous ovary. The female flower 

 differs from that of Corema alba only in the absence of the anterior sepal, which in C. alba is the smaller, but this is sometimes present when the identity becomes complete ; so that, if it were not 

 for the well-known accuracy of the observers who have described this genus, it might be supposed that the female flowers of C. alba had been mistaken for those of O. Conradii, as yet unknown. 



That such is the structure of the male flower is certain, as it is the same in unopened flowers, and it is also sufficiently evident from Dr. Asa Gray's figures (Mem. Americ. Acad. New Ser. 

 vol. III., Tab. I., figs. 3 & 4), that the floral envelopes consist only of three scales ; and as the anterior scale is a sepal both in Ceratiola and Corema, it is not in any probability otherwise in a genus 

 so closely approaching Corema. The margins of the bracts approach each other more closely on the posterior side, where they overlap each other, and have been regarded as adherent and compared 

 to a monopetalous corolla, but in 10 or 12 flowers examined, they were in each quite distinct to their bases. Dr. Gray's description of the female flower also agrees with mine as far as the number 

 of the scales is concerned (v. loc. cit. figs. 10 and 11), for he figures two as external, which I regard as lateral bracts and three as internal (loc. cit. fig. 10). 



May not this genus be a dimorphous form of Corema, occasioned by a tendency in the male flower (which has no rudiment of an ovary) to become hermaphrodite ? Dr. Gray (loc. cit.) describes 

 and figures hermaphrodite flowers with 1-celled anthers. May not the 1-celled anthers and the disappearance of three of the sepals in the male flower, be analogous phenomena?! 



In this view of their structure the main distinctions of the genera are as follows : — 



Sepals 6 .......... Empetrum. 



f Bracts 2 . . . . . . . Corema. 



Sepals 4 . [ Bract Y . . . . . . . . . Ceratiola. 



Sepal 1 ............ Oakesia. 



* When only three stamens are present they appear as if alternate with the bracts and sepals, but when the fourth is present they are of course opposite them as in Corema alba, 

 f In Corema alba I met with two instances of an anther reduced to one cell, the other side of the stamen having become membranous. 



