The polygonal cells of the medullary sub stance are produced from the indifferent tissue of the pulp in exactly the same manner as those of an ordinary homy, cel lular ecderon from that of the rete mucosum : that is to say, the periplast increases, and becomes marked out into polygonal areze ; it then acquires a horny consistence, and a stronger and stronger definition along the lines of demarcation, until polygonal " cells " (as in fig. 317. B, a) are formed. The walls of the latter now thicken and become granular ; the endoplasts disappear, and at length no thing is left but the honey-combed perfect me dullary substance. The mode of formation of the cortical su,bstance is the inverse of this. On examining the line of junction (fig. 317. B) of the pulp (c) with recently formed cor tical substance (b), it is observable that the endoplasts do not become surrounded by cell cavities, but that the periplast acquires a granular, longitudinally fibrous, appearance ; while the endoplasts, though they are oc casionally visible in the striated mass, soon completely disappear.* The elongated ca vities or tubuli do not at first exist in the cortex, but are the result of a secondary va cuolation, and so far as I have been able to observe, have no relation with the pre-existing endoplasts. In fact, these canals, like those in the hair-shaft, the clefts in the fenes trated rootsheath, and the canaliculi of bone, must be regarded as the results of a second ary vacuolation. The feather sac resembles that of the hair in all essential points of struc ture, except that the relations of the layers of the inner rootsheath are different. As in the hair, two layers may be distinguished in the inner rootsheath, an outer, strong, dark, horny membrane corresponding with the fe nestrated membrane, and an inner delicate flexible layer, corresponding with the inner horny rootsheath. The former has a structure intermediate between that of the two layers of the inner rootsheath in the hair, consisting of irregular polygonal plates, which retain the remains of their endoplasts (fig. 317. B), as in the inner layer of the horny rootsheath, and do not become separated by fissures ; while they resemble the plates of the outer horny rootsheath in their thickness, complete cor nification and striated appearance.
The inner layer of the horny rootsheath is a delicate, often granular membrane, which closely invests the outer surface of the feather, and from presenting a cast of its elevations and• depressions, has been called the outer "striatedmembrane" ofthe feather sac (suprd, § 2.) It is a sheet of horny matter, in which traces of closely-set endoplasts are discover able. The inner (fig. 317. it, d) " striated membrane " is a membrane having a similar structure, possessing similar relations to the inner surface of the feather, and which is con tinuous with the so-called " pith " in the quill of a fully formed feather. The mode of de velopment of these rootsheaths is identical with that of those in the hair, and therefore requires uo further elucidation here.
Tegumentary glands. — The other con versionary productions of the ecderon which we have to consider, are the glandular ap pendages, which are always diverticula of the cellular ecderon inwards.* Under this head I include only those small glandular organs which, so far as we know, have no reference to any other functions than that of cutaneous transpiration or fatty secretion, referring to the articles on special divisions of the animal kingdom for an account of those organs, such as the " water vessels " of Echinoderms and Trematoda, the nidamental glands of Mollusks, the genital glands of Vertebrata and Insecta, which might strictly be regarded as productions of the integument.
Tegumentary glands in this limited sense are somewhat rare among the Invertebrata. They have, however, been observed in the Annelids, where they consist of delicate tubes, terminating internally by a blind extremity containing a single nucleated cell. Such glands exist on the ventral surface of the head and foot discs in Piscicola, and are scattered all over the body in Clepsine and Nephelis. Similar glands are found opening upon the ventral surface of Argulus foliaceus.
Simple meal glands are scattered over the whole surface of the body of the Procession Caterpillars, opening at the points of the hairs ; on the sides of the body in Myriapods, on the joints of the legs in Beetles and Bugs.
In Molluscs a peculiar, probably glandular, canal exists in the foot of certain Lamelli branchs, and glandular cceca have been ob served in the lower surface of the foot in Paludina. A ciliated canal runs in the foot of Pulmonata, and receives glands on each side. The existence of cutaneous glands in the Cephalopods appears doubtful—at least, H. Miiller could only find them as shell glands in the expanded arms of Argonauta.
Among the Vertebrata, Fishes, Ophidia, Chelonia and Birds, appear to possess no proper cutaneous glands* ; in Sauria they attain a very slight and local, but in Batrachia and Mammalia, an immense de velopment. In the frog, the whole surface of the ecderon is beset with minute trifid apertures, so disposed between three epi dermic cells, as to present a singular resem blance to the stomata of plants (fig. 318. n). These lead directly into spherical sacs (fig. 318. A. d.), which are lined by a continuation of the cellular ecderon, and lie in the superficial part of the enderon above its stratified layer (fig. 318. A. g.) (vide infra). Nerves (f) and vessels penetrate the latter to reach the superficial layer of the enderon, and rarnify among these close-set glandular sacs. The sacs usually contain only a clear fluid * ; they are contractile, and may be made to expel their contents by irritation of the nerves dis tributed to them.-l In Mammals, we rneet with two kinds of cutaneous glands, sebaceous and sudoriparous. The former are almost invariably developed in connection with the hair sacs, consisting in fact of diverticula of the Malpighian layer of the cellular ecderon of the upper portion of these sacs, whence their position is always superficial. The innermost cells of the solid process become filled with fat— break down, and pour their contents into the hair sac itself, by whose aperture they make their exit. Sometimes, as in the hairs of the head in man and in the pig's bristles, the sebaceous glands are very small and simple, while in other lo calities they throw out processes, and assume the appearance of complex racemose glands, disposed like rosettes around the hair-sac, from which they are developed.