The continuity of the enderon over the scales will be seen below to be more apparent than real. I have not been able entirely to satisfy myself, as to the exact relations of the parts, in the case of the eel, but in the other fishes which I have examined the surface of the scale is very partially covered by the enderon, being in its centre, at any rate, in:contact with the cellular ecderon.
The vascularity of the scale never extends to its most superficial layers, and rnay be ex plained in the same way as that of the test of an Ascidian, which however is unquestionably an ecderonic structure. The passne of its deep layers directly into the connective bundles of the enderon, which Leydig has observed in Poly pterus (and which I will not say,does not occur elsewhere, though I have not observed it), would appear to me only to indicate that this scale, and perhaps others, are composed of two portions, a superficial ecderonic part extend ing as far as the most superficial vascular canals, and a deep portion beneath these be longing to the enderon.
However, all these points can only be de cided by a much more extensive series of in vestigations, principally directed to the ascer tainment of the position of the protomorphic line and of the direction of growth of the constituents of every scale, than I have hither to had time or opportunity to carry out ; and as the attention of other observers does not appear to have been directed to these par ticular points, the question must for the present remain undecided.
Professor Williamson in his valuable and philosophical contributions to our knowledge of this subject (Phil. Trans. 1849-1852) laid the foundation for a comprehension of the mode of development of fish-scales, by pointing out that Agassiz's views, though essentially true, yet require a certain modifi cation. For though a fish-scale does really grow by the apposition of layers to its deep surface, as Agassiz asserted, yet it is not in cluded in a sac of the epidermis (if by that term we are to understand the ordinary cellular ecderon); and it is also true that its deeper portions grow by their superficial surface. Professor Williamson points out, in fact, that every fish-scale consists of at least two portions, a superficial homogeneous, or at most canaliculated, laminated layer, the ganoin (so called enamel or horny layer of authors), and a deeper, also laminated, fre quently fibrous or osseous portion commonly traversed by Haversian canals. Now these
two portions have a certain independence in their mode of growth, at any rate after their first formation, as may be easily understood by the accompanying diagram (fig. 307.), which represents a series of imaginary sections of scales from their first growth onwards ; a, is the protomorphic plane; b, b", the deep ecde ron ; b', the superficial cellular ecderon, and the line x, the centre of the scales from which development commenced.
Suppose A to be the youngest scale, con stituted merely by a thickening and calcifica tion of the deep ecderon, which in n has added several layers by apposition to its inner surface, all of which retain the ganoin struc ture except the deepest, which becomes fibrous in its texture, and forms the commencement of the " Lepidine" layers of the scale ;—these layers, however, being as much a part of the ecderon as the former. In c the scale widen ing, the edges of its " Lepidine " layer do not remain in contact with the ganoin layer; but it will be obvious that the re-entering angle thus formed by the protomorphic line between the two, is only, as it wt re, a fold of the deep surface. If the two layers go on increasing in this way, however, the ultimate effect will be that, although growing in reality by its deep surface as before, the " Lepidine" layer of the scale will appear to grow by its superficial surface, and that addition of layers to the upper surface of the scale observed by Pro fessor Williamson, will take place. If the ex planation here proposed, however, be correct, this will form no objection to, but a confirmation of, Agassiz' views.
It will be well, however, with this clue to turn from the theory to the facts of scale de velopment.
All that I have observed leads me to con firm Professor Williamson's conclusion, that there is no real line of demarcation to be drawn between placoid, ganoid, ctenoid, and cycloid scales ; all these forms passing into one another. Indeed, I conceive that the only method thoroughly to comprehend the cycloid and ctenoid scales is to examine, in the first place, the so-called placoid and ganoid forms.