suddenly narrowing near the extremity to an acutely pointed apex. Each tooth is also first inclined a little inwards, then bent outwards at an obtuse angle; finally, by another curvature, the apex conies to point upwards in the mandibular teeth, downwards in those of the maxilla. The teeth are also closely set in one row of nearly uniform size: there are certainly no larger teeth inside this row; nor have I seen any trace of smaller ones outside. Mr. Walker describes the teeth of the lower jaw of the Pitcorthie fish as "placed alternately one close to the outside margin; the next to it is fully half its own thickness further in, and so on the whole length of the bone." Of this arrangement I have seen nothing more in the Wardie specimens than occasional indications. By Agassiz the teeth of G. punctatus were described as being "en cônes obtus," an appearance due to their being only seen in antero-posterior vertical section in the specimen he examined, their peculiar flexures and pointed conical apices being there invisible. Nor do I find any evidence that they were arranged "sur plusieurs rangées," at least as far as the maxilla and dentary of the mandible are concerned, though it is probable that additional teeth were present in the palate.
The dorsal fin in Gonatodus is placed rather further back than in Elonichthys, the middle of its base being opposite the commencement of the anal: both dorsal and anal are large and triangular; the base of the ventrals is short. All the fins are composed of very numerous rays, which are closely jointed, including in that respect also the principal rays of the pectoral. The suspensorium is not so oblique as in most genera of Palæoniscidæ, but more so than in Amblypterus.
A new species of this genus (G. macrolepis, Traq.), characterized by its very large and nearly smooth scales, has recently turned up in the Blackband Ironstone of Gilmerten. As yet the genus is only known from the lower division of the Carboniferous formation in Scotland.
I regret that I have had no opportunity of forming any independent opinion as to the affinities of the Triassie species Amblypterus Agassizii of Münster, A. ornatus and latimanus of Giebel, or of that from the cupriferous sandstones of Kargala in Russia, named A. orientalis by Eichwald, and can only say regarding them that, from the descriptions given, they do not seem to belong to the genus Amblypterus as restricted by Troschel. A. Olfersi has been already referred by Agassiz himself to the Teleostean genus Rhacolepis.
Palæoniscus.
The genus Palæoniscus has been made to include an immense number of species, which are, indeed, referable to more than one generic type, some of them actually not being Palæoniscidæ at all. Authors have, in point of fact, been only too apt to refer nearly every small rhombic-scaled fusiform-shaped ganoid fish from Upper Palæozoic rocks to Palæoniscus, without inquiring too narrowly into
