If we examine a hair sac above the level of the bulb, it will be clear that these inner root sheaths are not generated from the contiguous surface of the external rootsheath, as would at first seem probable. No transitional forms, in fact, are visible in the direction of the transverse diameter of the sac. Traced to wards the base of the sac, however, it is ob vious that opposite the lower portion of the bulb the inner layers of the outer rootsheath become metamorphosed into horny cells ; and that of these cells, the inner are converted into the imperforate layer, while the outer un dergo a more complete cornification, and lose all trace of their primitive endoplasts. The clefts which ultimately exist between these cornified plates are not present in the young state, but are the results of a secondary va cuolation. They have nothing to do with the disappearance of the endoplasts ; for traces of the latter may be observed in the centre of horny plates, at whose edges the clefts are commencing (fig. 315. r). It would appear, therefore, that the rootsheaths grow like the shaft of the hair itself, not by addition to their surface, but by growth of their deep-seated inner ends.
Such is the composition of the growing hair ; but the completely formed hair (see § 2. Morphology) presents very great differences in the minute structure of its inner termina tion. In the first place, the shaft runs out into an irregularly conical mass, like a worn out painter's brush. It consists, at its ex tremity, entirely of cortical substance, and the cornification runs in irregular lines into the indifferent tissue, which occupies the bottom of the hair sac and represents both pulp and outer rootsheath. The inner root sheaths terminate above this point, in an irre gularly horny layer, which unites with, and is in a manner reflected into, the cuticle of the shaft, which ceases above its brush-like ex pansion. Finally, the outer rootsheath in the immediate neighbourhood of the inner, is me tamorphosed into large horny cells, like those of the cellular ecderon. The development of these from the indifferent tissue of the outer rootsheath, may be very clearly traced. The periplast first becomes enlarged and marked off into definite granular arere around each endoplast, and the limits of each area are metamorphosed into clear homy walls. The cavity which these inclose enlarges, and the endoplast, with its surrounding granular mat ter, remains attached to one wall, and then eventually disappears, while the cavities en large, and their walls thicken into clear horny " cells," which may eventually be detached from one another.
The whole process of the completion of the root of a hair, then, is simply a return of the diverticulum of the ecderon,— the meta morphosis of whose elements, so long as the hair was in course of formation, was guided and determined into distinct forms along cer tain fixed lines, to its general tendency to undergo the ordinary cellular metamorphosis over its whole surface. With this return to its primitive tendencies, the increase of the hair of course ceases, and sooner or later it is pushed out and falls away.
The spines of the Porcupine, of the Hedge hog, and of the Echidna*, present in their histological, as in their morphological relations, an interesting approximation to feathers. Ex ternally, they are coated by a cuticle, while the principal mass of their walls consists, at the ends, of a fibrous horny substance; in the middle, there is added to this a medullary substance composed of polyhedral horny cells.
The section of the shaft of a fully-formed feather presents exactly these constituents except the cuticle; the centre is occupied by medullary substance (fig. 317. Bp a), composed or a coarsely granular horny substance exca vated by polygonal cavities of about To'oc, inch in diameter, frequently if not invariably containing air, which adds to the dark hue (by transmitted light) arising from the granular opacity of the horny matter. At its edges, this tissue passes into the coitical sub stance, which, in a transverse section (fig. 317. ri) appears as a clear, hornogeneous or slightly granular mass, dotted over by minute aper tures, about 1-00 in. in diameter, and 1.-0--0 in. apart. In a longitudinal section, on the other hand (fig. 317. c, h), the general mass appears obscurely striated in a longitudinal direction ; and in the place of the circular apertures, we see elongated fissures, some what narrowed at each extremity, whose transverse sections constituted these aper tures. The pointed ends of the fissures were continued by a line which could fre quently be traced into some other fissure above or below, so that I conceive the fis sures are in reality more or less complete canals.
The quill of the feather is entirely com posed of cortical substance ; the barbs have the same structure as the shaft ; the barbules present both cortical and medullary sub stances in a rudimentary condition. Each barbule in fact (fig. 317. E, e) exhibits along its axis a series of oval cavities, the remains of cells like those of the medulla, while its lateral portions are composed of striated horny matter like that of the cortex, and are produced into the curved and hooked lateral processes (f).