shaped median piece flanked by separate lateral ossifications as in Plesiosaurus, while this condition parallels the interclavicle and clavicles in all animals in which they are found. It has never been shown that any one of the bones in question in Plesiosaurus retains a surface which has the aspect of having been cartilaginous. On the contrary, every specimen which I have examined is more or less thin and squamous, with contours completely ossified to sharp edges, even in the most immature specimens; while the interclavicle, when preserved, unites with the clavicles either by a thin squamous overlap or by sagittal sutures. This condition seems to me to demonstrate that the bones are membrane bones. I submit it follows that they are clavicles, and therefore that the visceral position of the clavicular arch, although anomalous, is not inconsistent with clavicular homology. Bone for bone, the three clavicles in Plesiosaurus seem to me to correspond to those of Ichthyosaurus and Nothosaurus. In the former their union is usually squamous, in the latter it is sutural. In Sauropterygia both conditions are found. The proposal made to identify the three anterior bones in the shoulder girdle in Nothosaurus as omosternum and precoracoids introduces the precoracoid as a distinct bone,* which is not known to be paralleled in any allied group of animals except the Anodomontia, in which there is no omosternum, and where the precoracoids are differently conditioned, being in the closest union with the coracoids, with a well-developed clavicular arch. But when the supposed precoracoids of Nothosaurus are recognised as clavicles, which rest by squamous overlap on the visceral surfaces of the scapula, like the clavicles of Plesiosaurus, the clavicular arch is in harmony with that of the Sauropterygia, and the supposed differences in its composition dis appear. There are two family types in the Sauropterygia defined by differences in the shoulder girdle and other characters, known as Plesiosauride and Elasmosauridæ, though the organic differences which characterise them have not been fully set forth. II. FURTHER EVIDENCE OF THE NATURE OF THE CLAVICULAR ARCH IN THE PLESIOSAURIDE. § 1. Nature and Limits of the Family. There are four principal genera of Plesiosauridae, which are named Plesiosaurus, Eretmosaurus, Rhomaleosaurus, and Pliosaurus. The family is characterised by the cervical ribs being attached to the vertebræ by more or less completely-defined double facets and by the scapula being separated in the median line by the clavicular arch, by Hulke, loc. cit. which they are braced to the coracoids. In the British Museum Catalogue (Fossil Rept. and Amph.,' Part II), the Plesiosaurida is made to also include the Elasmosauridæ, and the genera are enumerated in the following order :-Pliosaurus, Peloneustes, Thaumatosaurus, Polyptychodon, Cimoliosaurus, Eretmosaurus, Plesiosaurus. I should restrict the family to the fossils indicated by the names Pliosaurus, Peloneustes, Thaumatosaurus, Eretmosaurus, and Plesiosaurus. Good skeletons of these genera are known with the exception of Thaumatosaurus, which was founded by von Meyer ( Palaeontographica,' vol. 6) upon remains which closely resemble those of Pliosaurus. And, after examining the type specimens, which are imperfect cervical vertebræ, dorsal vertebræ, teeth, and portions of the hinder region of the maxillary bone, I was unable to discover any character inconsistent with reference of the species to Pliosaurus. The head was evidently as large as in Pliosaurus; the teeth are circular in the crown, and show no trace of the area more or less flattened and free from carination defined by a lateral ridge on each side which characterises the anterior teeth of Pliosaurus grandis, resembling in this respect the posterior teeth. In the late cervical vertebra figured by von Meyer, the centrum has the same form and relative shortness from front to back as in Pliosaurus; the articular facet for the rib is similarly elevated, has a like transverse division forming a superior subtriangular part and an inferior transversely ovate part. The only characters in which there is not absolute agreement with the English species are that the articular faces of the centrums are more circular and more concave. These differences may be of specific value; and von Meyer's species may be classed as Pliosaurus oolithicus, till it is fully known. For similar reasons I am unable to separate Peloneustes from Pliosaurus. And if the type species was originally referred to Plesiosaurus, it was because I then regarded the subtriangular crowns of anterior teeth in Pliosaurus as a generic character, and that character now seems less important. It has been necessary thus to explain differences of nomenclature, because the genus Thaumatosaurus ('Brit. Mus. Cat. Foss. Rept.,' Part II) has been made to include six species in addition to the type, which, with one exception, are all from the Lias. They were previously named Rhomaleosaurus Cramptoni, Plesiosaurus arcuatus, P. megacephalus, P. carinatus, P. propinquus, P. indicus. I am unable to place any of these species in Pliosaurus or Thaumatosaurus, nor is there evidence that all are referable to one genus; and it does not appear that a genus based on characters drawn from this assemblage of species can displace the definite conception of von Meyer indicated in the type of Thaumatosaurus. Most of these species not included in Rhomaleosaurus appear to belong to Eretmosaurus. *Index to Aves, Ornith., and Rept. in Woodw. Mus.,' 1869, p. 139. § 2. The Clavicular Arch. (i.) Since the clavicular arch was figured in Pl. Hawkinsi (‘Geol. Soc. Quart. Journ.,' 1874, p. 444), v. Zittel has figured the clavicular bones in Pl. laticeps ('Handbuch der Paläontologie,' vol. 3, p. 489); but, while the clavicles are clearly shown, the interclavicle is named episternum. The most important evidence of this structure in Plesiosauridæ, however, is to be seen in Plesiosaurus arcuatus (Brit. Assoc. Rep.,' 1839, p. 76; and 'Cat. Foss. Rept. and Amph.,' Part II, p. 163), preserved in the British Museum. From that specimen, No. 2028*, the character has been attributed to Thaumatosaurus (loc. cit., p. 159): "Omosternum consisting of a large single plate, much expanded transversely, with a wide and shallow anterior notch."+ The anterior margin of the interclavicle in this specimen resembles in contour that attributed to Eretmosaurus (' Geol. Soc. Quart. Journ.,' 1874, p. 445) in its wide open curvature; but there is no evidence to show whether the shoulder girdle, pelvis, and limbs in Plesiosaurus arcuatus were constructed on the same plan as in Pl. rugosus. There is no doubt that the bone consists of three distinct elements united by sutures. These are a median interclavicle and two lateral bones which I regard as clavicles. On the visceral aspect the triangular clavicles are separated from each other by the wide short posterior median bar of the interclavicle, but the clavicles extend forward so that only a narrow transverse bar of the T-shaped interclavicle is exposed in front of them, extending across the entire width of the bone. The interclavicle is 10 inches wide, concave on its anterior margin, 1 inch from front to back at the widened extremities of the cross-bar, and inch in the same measurement towards the oblong middle portion of the bone. The right anterior transverse limb of the cross-bar is 4 inches wide; the left limb is 3 inches wide. The middle portion of the bone is 34 inches wide and 24 inches in anteroposterior measurement. The sutural line which defines the interclavicle is sagittal, and consequently irregular. On each side of this T-shaped interclavicle (fig. 2), in contact with the posterior margin of its transverse bar and the lateral margin of its short wide median stem, is a large triangular clavicle which is directed backward and outward. In harmony with the dimensions of the transverse bar, the right clavicle is the wider. Anteriorly it is 4 inches wide; it is nearly 6 inches long. The external border, which is slightly convex, is continuous with the truncated lateral termination of the interclavicle in front of it. These external margins diverge outward as they extend backward, so that the transverse measurement over the posterior extremities of the clavicles is 14 inches. The postero-internal + Compare Sollas, 'Geol. Soc. Quart. Journ.,' vol. 37, 1881, p. 457. FIGS. 2 and 3.-Ventral and visceral aspects of clavicular arch of Plesiosaurus arcuatus, showing the median interclavicle and lateral clavicles. Ic is placed on the anterior margin. contours of the clavicle are irregularly concave, and as they extend inward are continuous with the posterior border of the interclavicle, and as they extend outward approximate toward the external contour of the bone without meeting it posteriorly in a point. The ventral aspect of the clavicular arch is different (fig. 3) owing to variation in the positions of the sutures between the bones. The interclavicle no longer shows the T-shaped contour of the visceral surface, but is a wide curved bar with an irregular sagittal termination on its postero-lateral extremities. This is owing to the method of its squamous interlocking with the clavicles, which overlap its visceral surface more in front, and overlap its ventral surface more behind, where their pointed extremities nearly meet each other in the median line behind the interclavicle, and in the inch of space from which they are absent there is a slight distortion of the bone, and some evidence of a median posterior notch. The triangular forms of the clavicles are more marked on this aspect of the bones than on the other. The most remarkable character here shown is the squamous sutural interlocking of the three bones by which their shares in forming the clavicular arch is definitely established. It is also shown by different directions of the lines of growth in the clavicles and interclavicles. An isolated clavicular arch in the British Museum, R. 1322, presents a similar character and form, and shows in its sutures similar evidence of composite character. It has been assigned to the species named Plesiosaurus megacephalus (Stutchbury) in the British Museum Catalogue. It has a similar resemblance to the anterior contour of the interclavicle in Eretmosaurus, but I am aware of no evidence by which the species is identified from this bone, beyond a general resemblance to some specimens in the Bristol Museum. The correspondence of structure in these clavicular arches with that figured in Plesiosaurus Hawkinsi and Plesiosaurus laticeps is a coincidence of plan, though the difference may indicate a sub-genus, and shows, I submit, that the original definition of the bones was not a conjectural suggestion, as stated by Professor Sollas, but a recognition of sutures which separate the interclavicle from the clavicles. And it seems to me a sound induction that whenever the margins of the clavicular arch are concave in front and behind, those concavities border the interclavicle, and whenever there are wings produced outward and backward, as in the specimen now figured, those wings are formed by the clavicles in all Plesiosaurida. (ii.) Sir R. Owen, in 1841 (‘Brit. Assoc. Rep.,' p. 64), remarks on the shoulder girdle of Pliosaurus :-" The pectoral arch owes its chief strength to a pair of immensely expanded coracoids, having a broad and short entosternal bone on their anterior interspace, and supporting the clavicles or acromion productions of the scapula."* Subse * I have examined the specimens in the Museum of the University of Oxford |