The vertebral characters are well marked. The vertebræ are thin osseous rings slightly constricted in the middle; the concavity of both surfaces is shallow, and involves nearly the whole breadth. The diameters are unequal, their difference ranging from one-fifth to one-eighth of the longest measurement.

The notochordal foramen is also not a complete circle; its longest and shortest diameters coincide with those of the vertebræ, of whose area it occupies from one-fourth to a half.

The extreme examples measured in tenths of an inch are as follows:

In the majority the length (or antero-posterior measurement) is between two-tenths and three-tenths at all points; but the deeper examples show an inequality in this respect-the length at opposite points of the ring varying considerably, as 2.5 to 2.25 in one case, as 4.75 to 3 inches in another.

Straight bones, similar to the neural spines of Dendroptychius, occur in the same slab; they measure 21 inches in length. Associated with scales showing the above-mentioned characters, there is in the Andersonian Museum, Glasgow, a pectoral arch with which are connected several bones giving attachment to long undivided fin-rays.

Locality: English and Scottish fields, from the Coal-measures to the Lower Limestones.

RHOMBOPTYCHIUS, Huxley. Figs. 1, 2.

These scales vary from cordate-ovate to rhombic: the length in the former case exceeds the breadth; in the latter the reverse holds. In general symmetrical, the elongated scales sometimes show undue prominence of one anterior angle. The free and overlapped surfaces are nearly equal in the rhombic scales; but in the cycloidal the former is much the smaller. In both forms the punctate free surface is ornamented with coarse straight ridges which follow the outline of the sides, meeting at an angle posteriorly, and thus giving a rhombic aspect to the scale. The intervals of these ridges are crossed, chiefly near the margin, by finer radial ones. The pattern may lose its regularity-only two or three marginal ridges being complete, and the remaining surface set with irregular tubercles. This occurs in those scales which most nearly resemble the rhombic scales of Megalichthys; but the resemblance stops here; for though in some specimens of Megalichthys the bony surface exposed by the absence or non-development of the enamel shows irregular tuberculation, or even a faintly rhombic pattern like that just described, the scales presenting these characters have their anterior area bounded by straight lines, and do not show the cycloid outline of those of Rhomboptychius, those straight lines, however, which bound anteriorly the free area being common to both. The under surface is smooth;

growth lines are confined to the margin; the oval boss is subcentral. The outer surface of these scales is finely punctate, dense, and semilustrous.

The Andersonian Museum of Glasgow contains a large slab on which are displayed the vertebral column and upper surface of the cranium of this genus. About fifty vertebræ are seen, of which the anterior are the shortest and broadest, the caudal being longer and narrower. The cranium is flattened superiorly; its surface is tuberculated and traversed by mucous grooves, and lines which seem to be the sutures of the coalesced bones.

The cranial and facial portions occur as separate masses, the components of each being intimately united. The facial portion, consisting of the bones anterior to the frontals (those, namely, which compose the muzzle), very closely resembles that of Megalichthys and Diplopterus. Its crescentic anterior, or intermaxillary and vomerine, portion bears within and close to either outer extremity a large tooth; and on either side of the middle line is a similarly socketted tooth. The small marginal teeth are continuous with two curved rows of equally small teeth which pass in front of the outer pair of tusks, and curving to their inner side meet in the middle line at the anterior part of the basilar bar, whose surface is closely set with fine denticles.

The cranial shield is solid; its elements, intimately united by suture, correspond in number with those found in the cranial roof of Megalichthys (Poiss. Foss. pl. 63). But the well-ossified basilar region includes a massive basioccipital which projects behind the vertical posterior wall of the cranium, and sometimes has its length increased by the coalescence with it of at least the first vertebral ring, whose neural processes remained distinct. The anterior part of the cranial is sometimes deficient, the sphenoidal (and prootic?) portion becoming detached. In a lateral view, the ascending alisphenoidal (?) plates and an incomplete interorbital osseous septum are well seen.

The lower boundary of the elongated orbital space is nowhere preserved, the only surviving parts of the lateral surface of the head being the large opercular plates and the maxillæ.

The maxilla is slender in front, and sends off near its truncated extremity a strong process directed inwards and forwards, for articulation, probably, with the prefrontal or ethmoidal region. It gradually expands posteriorly, and bears on the greater part of its inferior margin, which is rounded and slightly deflected at the lower posterior angle, a row of strong, sharp, abruptly conical teeth of uniform size; their bases are strongly plicate, their upper portion usually finely reticulated. Between the maxillæ and the basilar bar the roof of the mouth seems to have been closed in by a pair of plates like those of Megalichthys, whose surface is set with fine rasp-like teeth, bounded by a row of small, stout, conical teeth. The connexions of these parts are nowhere seen.

The mandible is a strong bone, deep in front, tapering posteriorly on both margins; the transversely elongated, saddle

shaped articular surface looks obliquely upwards and backwards, giving, when the great length of the bone is considered (for it reaches to nearly opposite the hinder cranial margin), an enormous gape. The symphysial extremity is imperfect; but the arrangement of the teeth is probably similar to that in the other genera of the Crossopterygida; that is to say, a row of small conical teeth borders the outer dentary margin. About the middle of the bone several large distant teeth occur, having their sockets internal to the marginal row. These teeth are strong straight cones, plicate at the base, surrounded at or above the middle by one or more rings of short longitudinal furrows or pits. The inner or splenial margin bears numerous rasp-like teeth. The splenial and dentary margins are thus separated by an interval, considerable in front (where the large teeth are socketed), but narrowing behind, and ceasing a little in front of the articular end of the mandible. The outer surface of the bone is covered with fine tubercles and less prominent connecting ridges. The aspect is thus rather granular than reticulate.

The vertebræ of Rhomboptychius are osseous rings, two of which measured respectively 1 and 18 of an inch in diameter, and in length. The area of the notochordal foramen is three-fourths of the whole surface, greater, therefore, than in Strepsodus, but less than in the typical annular vertebræ of Megalichthys and Rhizodopsis. The spines associated with them on specimens in the British Museum consist of a cylindrical recurved shaft, an articular head with convex inferior margin, and a compressed distal extremity *.

Portions of jaws are found in the same beds with the remains just described, whose outer surface is similarly ornamented, and whose teeth are disposed in the same way. But these teeth are remarkable for the great breadth of their bases as compared with the rapidity with which they taper to a fine point, often incurved.

MEGALICHTHYS, Agassiz.

For the cranial and vertebral characters of this genus, and the advance which the latter show upon the Saurodipterines with which it is associated, I would refer to p. 12 of Prof. Huxley's Essay, already quoted.

The resemblance between Megalichthys and Rhomboptychius is very close, both in cranial structure and in the arrangement of the dentigerous bones. Mention has been made above of palatal tooth-bearing plates. These are two in number, triangular, with rounded anterior apex; their posterior extremity is not seen, two inches being the greatest length to which they are exposed. Each plate bears

In this and other genera it must be remembered that only those vertebræ are spoken of which occur in actual association with characteristic portions of the fishes to which they are ascribed. This caution is necessary, and, at the same time, perhaps not always protective against error. For, on the one hand, there are good grounds for suspecting in some genera a difference in dimensions and even proportions of the vertebræ in the same column, and, on the other, the presence of vertebræ belonging to more than one individual is possible on the same block. The number of vertebræ still unaccounted for is large; but it would be both premature and out of place to enter on their description here.

a marginal row of short, very stout conical teeth with fluted bases; the rest of their surface is set with similar but smaller teeth, which are more distant over the anterior portion, but posteriorly pass into a dense rasp of minute denticles. The connexions of these plates are not seen; but they probably fitted into the angles formed on either side by the maxilla and sphenoid. Since 1861, specimens illustrating the form of the fins have been acquired by the Museum; but the description and illustration of these parts are reserved, the purpose of this paper being merely to give the diagnostic characters furnished by the scales and teeth.

The teeth are conical, more or less incurved, and of very elegant proportions. Both jaws, the premaxillary region above, and the anterior part of the mandible, contain teeth larger than the numerous small ones which occupy the edge of the bones, and, as in Rhomboptychius, form an angulated row across the roof of the mouth. The surface of the majority is smooth; many, however, are covered with very fine ridges, which involve merely the outer portion of the enamel, and, as Mr. Davis has pointed out to me, disappear by attrition. These lines are either parallel or anastomose to form a fine reticulation; but nothing approaching the pattern of Strepsodus, either in position or symmetry, is found. In a note read before the Society in February, I proposed the abolition of Centrodus, M'Coy, on the ground that that tooth is avowedly a fragment, that it is unaccompanied by either bone or scale, and that it is identical in character with teeth unmistakeably associated with Megalichthyic remains. I have since had the satisfaction of finding that Mr. Davis had arrived at the same opinion, a tablet of teeth from Carluke bearing, as synonyms, Megalichthys and Centrodus.

The scales and head-bones of Megalichthys are covered by a layer of smooth, porous ganoin. When this is detached by accident, the subjacent surface of bone is also smooth, but perforated by the larger and smaller tubes passing up from the interspaces in the osseous substance of the scale. Such is the typical appearance; but in a large number of examples (fig. 6) the ganoin both of scales and bones is patchy and incomplete, covering nearly the whole of the surface, and only interrupted by circular spaces, which are distinguished by their concentrically sloping margins from the mucuspores, which are surrounded by a thin raised ring; or the ganoin only occurs as scattered points. A carefully prepared section of a scale through such an isolated patch shows that, beneath the ganoin, the structure is the same as in the typical Megalichthysscale, of which Williamson has given an excellent description. The ganoin ceases on either side with an abruptly rounded margin, which dips down to and stops at the average surface of the scale; for it coats a low eminence made up of the kosmin or non-corpusculated bone, the tufts, capillary tubes, and the upper series of Haversian canals. Adjoining the patch the surface of the scale is smooth, and has, as in Rhomboptychius, a close and semilustrous aspect. Under the microscope it is seen that a thin layer of kosmin coats this part of the scale, whose irregularities are formed of the upturned laminæ of bone.

The Haversian system is in direct communication with the surface, giving rise to the wide pores. Sections of Rhomboptychius show the same structure, as do those of Rhizodopsis (H. sauroides) in Williamson's memoir. The absence of ganoin is therefore not the result of accident, but is due to the same cause in Megalichthys as in the other two genera named-non-development.

Though the main facts regarding these genera are sufficiently well ascertained, much still remains to be done in the working out of their details. The statements I have made will, in all probability, require modification and correction as better-preserved specimens

turn up.

From comparison of a large suite of specimens it appears that, while the typical scales of Megalichthys and Rhomboptychius are sufficiently distinct, the examples are very many which may, in the absence of other proof, be referred to either genus. Great differences of the scales of the same individual might account for the apparent confusion; but no specimens have been found sufficiently complete to make this certain.

13. Note on SUPPOSED BURROWS of WORMS in the LAURENTIAN ROCKS of CANADA. By J. W. DAWSON, LL.D., F.R.S., F.G.S., Principal of McGill University, Montreal.

AMONG other indications of fossils in the Laurentian rocks, mentioned in my paper on the structure of Eozoon*, are certain perforations resembling burrows of worms, found in a calcareous quartzite or impure limestone from Madoc, in Upper Canada. They occur in specimens in the Museum of the Geological Survey, and also in specimens subsequently collected by myself at the same place.

The beds at Madoc, containing these impressions, underlie, unconformably, the Lower Silurian limestones, and are regarded by Sir W. E. Logan as belonging to a somewhat higher horizon in the Laurentian than the Eozoon Serpentines of Grenville. They are also less highly metamorphosed than the Laurentian rocks generally. They are described in Sir W. E. Logan's Report on the Geology of Canada, 1863, at p. 32.

The impressions referred to consist of perforations approaching to a cylindrical form, and filled with rounded siliceous sand, more or less stained with carbonaceous and ferruginous matter, more especially near the circumference of the cylinders. These superficial portions being harder than the containing rock, and of darker colour, and also harder than the interior of the cylinders, project as black rings from the weathered surfaces; but in their continuation into the interior of the mass, they appear only as spots or lines of a slightly darker colour, or stained with iron-rust.

When sliced transversely and examined under the microscope, they appear as round, oval, or semicircular holes drilled through the

* Quart. Journ. Geol. Soc. Feb. 1866.