into the usual cavities, which cavities communicate with each other by means of the ordinary sphinctral openings in these sarcodic expansions. The skeleton-spicules are of two forms, namely a subcylindrical one, which is curved at the distance of one third of its length, sometimes found pointed at the long end so as to form a curved acuate, and a long fine straight acuate spicule not only associated with this, but found also in considerable numbers in the dermal sarcode. The subcylindrical spicule (fig. 12) measures 0.0068 inch in length by 0003 at its thickest part; and the long fine acuate (fig. 13) is 0.01 inch long by 0-0002 inch in diameter. The fleshspicule is of one form only, viz. a minute birotulate, each umbrella-like extremity of which is divided into twelve rays or ribs connected with each other and with the shaft by the usual falciform expansions (figs. 14 and 15), measuring 0.00053 inch in length, the diameter of the heads being 0.00016 inch and the diameter of the shaft one tenth of that of the heads. This minute flesh-spicule is liable to be passed over and its beauty and form unobserved; for the composition of the umbrella-like head is not distinctly seen with a lower power than ath objective. Mr. Laurence Hardman of Rock Ferry, who kindly undertook to verify the counting of the number of rays or flukes, was fortunate enough to discover on the slide submitted to him a few rotulate extremities broken off from their shafts and lying flat on the cover, the form of which was beautifully seen under ath. The finding of these heads in this convenient position rendered the counting of the rays easy, and enabled a correct drawing of an end view of one of them to be made (fig. 15).

Size. The specimens from the Spanish main are of the massive and pyramidic form. In the latter the erect growth is not more than 3 inches in height, with a base of from 2 to 3 inches in diameter; while among the massive forms, which cover pieces of coral, the largest specimen has a basal attachment of 6 to 7 inches, and extends laterally in an irregular lobe 5 to 6 inches. The branched form, known to us only by the specimen in the possession of Dr. Allen of Jamaica, is stated by him to extend to the distance of 2 feet from its root or base, the diameter of the branches not exceeding 1 inch by toinch.

Loc. Puerto Cabello, Caracas, and Bay of Kingston, Jamaica.

In the Argo' collection there are some specimens of a branched sponge from Nassau resembling the Jamaica example in outward form, colour, skeleton, and structure, possessing a skeleton-spicule of slender cylindrical form, but lacking

altogether the flesh-spicule. In them the horny element is rather more developed than in Dr. Allen's sponge, and it yet remains to be considered how far they are related to H. birotulata; they will therefore be more particularly referred to when the rest of the collection comes to be described.

EXPLANATION OF PLATE XIV.

Fig. 1. Higginsia coralloides, half actual size, after a photograph by Mr. John Chard, Liverpool Museum.

Fig. 2. Smooth bent acerate skeleton-spicule of same, scale 0.001 to 0.0625 inch.

Fig. 3. Smooth straight acerate subskeleton-spicule of same, scale 0·001 to 0.0625 inch.

Fig. 4. Spined bent acerate surface-spicule of same, scale 0·001 to 0·0625 inch.

Fig. 5. Same spicule, scale 0.001 to 0.125 inch.

Fig. 6. Smooth conically spined stellate spicule of Donatia parasitica, scale 0.0002 to 0-083 inch.

Fig. 7. Spino-capitately rayed spicule of same, scale 0-0002 to 0-083 inch. Fig. 8. Subterminally inflated spinulate spicule of same, scale 0.0004 to 0.0416 inch.

Fig. 9. Entirely spined quadriradiate spicule of Hymeraphia unnamed, scale 0.0002 to 0.0416 inch.

Fig. 10. Bent acuate spicule of same sponge, scale 0.0004 to 0.0416 inch. Fig. 11. Halichondria birotulata, short branch, actual size, from a drawing by my daughter, Eva Higgin.

Fig. 12. Subcylindrical skeleton-spicule of same, scale 0-0004 to 0.0625 inch.

Fig. 13. Acuate subskeleton-spicule of same, scale 0·0004 to 0.0625 inch. Fig. 14. 12-rayed birotulate flesh-spicule of same, five rays only at each end shown, to avoid confusion of lines; scale 0.0005 to 1 inch. Fig. 15. End view of one of the umbrella-shaped extremities of same spicule, scale one 1900th to 1 inch.

XXV.-On the Structure of the Lower Jaw in Rhizodopsis and Rhizodus*. By R. H. TRAQUAIR, M.D., F.G.S., F.R.S.E., Keeper of the Natural-History Collections in the Museum of Science and Art, Edinburgh.

AMONG the detached and broken-up remains of the Coalmeasure fish known as Rhizodopsis sauroides, one of the most frequently observed is a bone of a somewhat narrow and elongated form, truncated and somewhat expanded at one extremity, which may be assumed to be the anterior, and pointed at the other or posterior. One margin, nearly straight,

* Read before the Royal Physical Society of Edinburgh, Feb. 21, 1877.

save just in front, where it shows a slight convexity, is set with a single row of small pointed teeth of nearly uniform size; but the anterior extremity bears in addition a single more or less incurved laniary tooth, much larger than the others, and also more internal in its position; the opposite margin, thin and sharp, displays a gently flexuous contour. Seen from the inner aspect, the anterior extremity of the bone presents a conspicuous thickening, in which the large laniary is socketed, and which at the dental margin passes into a delicate ledge, which runs back for some distance along the roots of the smaller teeth.

This bone, whose external form has been well described by Messrs. Hancock and Atthey*, was considered by them to be the præmaxilla of Rhizodopsis, being obviously distinct from another well-known dentigerous bone, which is indisputably the maxilla, and closely resembles in form the maxilla of Megalichthys. To all appearance it would also seem to be distinct from the mandible, the margins of which "are nearly parallel," and which displays, besides a large laniary tooth in front, "three or four others placed along the ramus, in a line within the small teeth."

66

With the bones described by Messrs. Hancock and Atthey as the præmaxilla, maxilla, and mandible of Rhizodopsis, every student of carboniferous ichthyology must be familiar. The interpretation of the first of these as præmaxilla" has been accepted by the Messrs. Barkast, and, so far as I am aware, has remained hitherto unquestioned. Nevertheless the accuracy of its determination as such was to me a matter of doubt from the first. It is true the bone in question does in some measure remind us of the elongated præmaxilla of Teleostei of the most specialized type, in which that element, loosely articulated with the front of the skull, extends backwards so as to shut out the now edentulous maxilla from the edge of the mouth (Perca, Gadus, &c.). But as Rhizodopsis is a Crossopterygian ganoid of the type possessing two dorsal fins and subacutely lobate pectorals, one would naturally expect that its præmaxillary bones would resemble in form and relations those of its natural allies, whether rhombiferous or cycliferous, in all of which, whose cranial osteology is sufficiently known, each præmaxilla is comparatively small and short, firmly fixed to the front of the cranial shield, and, in fact, very unlike the bone of Rhizodopsis which has been

Ann. & Mag. Nag. Hist. 1868, ser. 4, vol. i. pp. 350, 351.

Manual of Coal-measure Paleontology,' by T. P. Barkas (London, 1873), p. 24, pl. ii. fig. 61; W. J. Barkas in Monthly Review of Dental Surgery.'

so interpreted. How to fit this bone into the præmaxillary region was to me somewhat puzzling; and, accordingly, to find it in situ in the head of the fish was an object to be attained, before giving-in adherence to the views usually maintained regarding it.

A short time ago my friend Mr. Ward of Longton, to whose liberality in lending specimens from his magnificent collection I am on this, as on other occasions, so largely indebted, sent me a number of unusually good examples of the head of Rhizodopsis preserved in nodules of hard ironstone from the Coal-measures of Fenton in Staffordshire. One of these displays the entire extent of the gape on both sides of the head. Each maxilla measures here 1 inch in length; the upper margin is injured; but the lower, bearing one row of small teeth, is quite intact; the anterior extremity shows the little articular process projecting upwards and forwards as in the similarly shaped maxilla of Megalichthys. Now, placed between and articulating with the anterior extremities of the right and left maxillæ, while they are joined with each other in the middle line, are two small dentigerous bones forming the front edge of the mouth below the snout. Each of these two bones is nearly as high as long, these measurements being respectively

and inch; they are firmly fixed to each other and apparently also to the front of the cranial shield: the teeth, which in this specimen are seen attached to them, resemble those of the maxilla; but in another example there are traces of others somewhat larger. That we have here the true præmaxilla is beyond all doubt; some other signification must therefore be found for the bones hitherto considered such. Turning now to the mandible, both rami of which are displayed in the specimen under description, we find that over a considerable area the bony matter of the outer aspect has flaked off, leaving behind it a pretty sharp cast with sutural lines. On close examination a suture is now seen commencing near the posterior extremity of the upper margin of the jaw, and, passing gradually downwards and forwards, marks off as dentary an element precisely the counterpart in shape of the reputed præmaxilla. The pointed extremity is placed backwards, the enlarged one forwards, the toothed margin upwards. The rest of the outer surface of the mandible is composed of at least three additional bony plates, separated from each other by sutures which pass obliquely forwards and upwards. The posterior and largest of these, covering over the articular region of the jaw, may be perhaps equivalent to the angular element, though it also occupies very much the place of a supraangular; the other two, in

front of the latter and below the dentary, may be called infradentary; and there is also some evidence of a fourth, small plate on the lower margin of the jaw, separating here the angular from the first infradentary for a little distance*.

In another specimen, compressed vertically and showing the top of the head, both maxilla are seen, forming the upper margin of the mouth, while, forming its lower margin, both dentaries are seen on the edge of the nodule, here retaining their bony substance and external sculpture. Their contour proves beyond a doubt that the dentary element of the mandible of Rhizodopsis is undistinguishable from the bone hitherto reckoned as præmaxilla, but which I have already shown cannot possibly be so. The very same thing is most clearly shown in a shale specimen belonging to Mr. Plant of Salford, in which a vertically compressed head is seen from below; so that I have no hesitation in affirming the identity of the bones in question.

Here, however, an objection to this view may be raised. The mandible of Rhizodopsis when perfect, as in most of the specimens from Fenton now before me, shows not merely one large tooth in front, but two or three additional ones behind it and internal to the series of small teeth, though, as stated by Messrs. Hancock and Atthey, these additional larger teeth "are seldom present. What has become of these in the detached dentary, if such be the real nature of the reputed præmaxilla?

A ready explanation of this is found in the structure of the lower jaw of certain Old Red Sandstone "Dendrodonts," in which the laniary teeth are not attached to the dentary bone proper, but to a series of accessory "internal dentary" pieces articulated to its inner sidet. Should this be also the case with the posterior laniaries of the mandible of Rhizodopsis, then, in cases where its elements are broken up and separated, these additional pieces will also get detached, and the absence of all but the anterior laniary in the isolated dentary bone will thus be amply accounted for.

The material at hand not furnishing me with absolute proofs of this condition in Rhizodopsis, I now turned to its

That these sutures on the outer surface of the mandible in Rhizodopsis have not been previously observed is fully accounted for by the difficulty of tracing the line of demarcation between constituent and closely united osseous elements, in cases where we have to deal with a granulated or otherwise ornamented external bony surface. Such lines of demarcation are more easily determined where the bones are seen from the inner surface, or where a sharp cast in hard ironstone of that inner surface has been preserved.

See Pander's Saurodipterinen, Dendrodonten, &c. des devonischen Systems,' pp. 41-43, tab. x. figs. 2, 3, 4, 14, 22.