of the sharks are with their shagreen; and so slight was the degree of imbrication, that the portion of each scale overlaid by the two scales in immediate advance of it did not exceed the one twelfth part of its entire area. In the scale of the _Cheiracanthus_ we find the covered portion indicated by a smooth, narrow band, that ran along its anterior edges, and which the furrows that fretted the exposed surface did not traverse. It may be added, that both genera had the anterior edge of their fins armed with strong spines,—a characteristic of several of the Placoid families. [Illustration: Fig. 6. a. _Scales of Osteolepis macrolepidotus._ b. _Scales of an undescribed species of Glyptolepis._[10] (The single scales mag. two diameters;—the others nat. size.)] In the Dipterian genera _Osteolepis_ and _Diplopterus_ the scales were more unequivocally such than in the Acanths, and more removed from shagreen. The under surface of each was traversed longitudinally by a raised bar, which attached it to the skin, and which, in the transverse section, serves to remind one of the shagreen footstalk. They are, besides, of a rhomboidal form; and, when seen in the finer specimens, lying in their proper places on what had been once the creature’s body, they seem merely laid down side by side in line, like those rows of glazed tiles that pave a cathedral floor; but on more careful examination, we find that each little tile was deeply grooved on its higher side and end, (for it lay diagonally in relation to the head,) like the flags of a stone roof, (fig. 6, _a_,)—that its lateral and anterior neighbors impinged upon it along these grooves to the extent of about one third its area,—and that it impinged, in turn, to the same extent on the scales that bordered on it posteriorly and latero-posteriorly. Now, in the Cœlacanth family, (and on this special point the foregoing remarks are intended to bear,) the scales, which were generally of a round or irregularly oval form, (fig. 6, _b_,) overlapped each other to as great an extent as in any of the existing fishes of the Cycloid or Ctenoid orders,—to as great an extent, for instance, as in the carp, salmon, or herring. In a slated roof there is no part on which the slates do not lie double, and along the lower edge of each tier they lie triple;—there is more of slate covered than of slate seen: whereas in a tile-roof, the covered portion is restricted to a small strip running along the top and one of the edges of each tile, and the tiles do not lie double in more than the same degree in which the slates lie triple. The scaly cover of the two genera of Dipterians to which I have referred was a cover on the _tile_-roof principle; and this is an exceedingly common characteristic of the scales of the Ganoids. The scaly cover of the Cœlacanths, on the other hand, was a cover on the _slate_-roof principle;—there was in some of their genera about one third