But it is important here to take notice of this fact, because it is not, as at first sight you would suppose, the corolla, but the calyx, which encircles the base of the ovary. The tube of the corolla stops towards the middle of the latter organ, and nearly on a level with the linear divisions of the calyx. You must be careful not to confound with these calycine divisions the green foliola which lie around the base of the flower, and which are neither more nor less than abortive leaves. Now call to mind that the flower is an union of concentric whorls, or of rings set one within another. The _staminal_ whorl and the _carpellary_ whorl, surrounded by a double perianth (corolla and calyx), are here composed--the first, of five stamens, and the second of a bilobed ovary, surmounted by a twisted style. We may now examine more closely the reproductive organs. The stamens are inserted upon the top of the tube of the corolla, and if you look at the base of their filaments you would be inclined to pronounce it a foliaceous expansion, or, rather, a metamorphic doubling of the corolla. Suppose the stamens to be the result of the transformation of the petals, the filament would be the "claw," and the anther the "limb" of a foliole. At least, theory would tell you so. But observation will show that these are not petals changing into stamens, but, on the contrary, stamens changing into petals; as is seen in the sterile (or "double") flowers of many of our ornamental plants, and even of some of our fruit trees. How, then, shall we conciliate theory with observation? Look, and you shall find. * * * * * Observe the anthers which surmount the filaments. There is something peculiarly characteristic in them. As they open and spread abroad the pollen, they visibly coil themselves up in the form of a spiral. (Fig. 58.) Owing to this twisting, they are found more or less inclined upon their filaments, and are gifted with a considerable mobility. Thoroughly to understand the relation between the continuous anther and the filament, we must examine the stamens before their expansion, while they are still folded up within the floral bud, their matrix. The anthers are then quite straight, and, with a magnifying-glass, you can easily see how they are inserted, by the lower part of their back, upon the top of the filament, whose (so-called) _connective_ prolongation separates the two lobes of the pollen receptacle. The anthers are then _introrse_ (_introrsum_, inwardly)--that is to say, their face being inclined inwards, they look towards the centre of the flower, occupied by the style, a filiform prolongation of the ovary; the apex of the style (stigma) is thick, globulose, and of a glandular structure. The fruit, resulting from the metamorphosis of the ovary, is an elongated fusiform capsule, composed of two lobes, each containing a very large number of extremely small seeds. [Illustration: FIG. 58.--Anthers of the Centaury.]