Downloaded from http://jgslegacy.lyellcollection.org/ at University of California-San Diego on December 19, 2016

456 H. G. SEEL~ ON A MAMMALIAN FEMUR AND

35. No~E on a Fsmra and a H~vs of a sx,~ MA,~AT. from the S~O~SF~LV S~r~. By H.G. S~.~Y, Esq., F.R.S., F.G.S. (Read February 26, 1879.) WHrLE examining, with the assistance of Mr. W. Davies, F.G.S., the unarranged and more fragmentary remains of Pterodaetyles from the Stonesfield Slate which are preserved in the British Museum, I was so fortunate as to detect two small slabs which contain, in good preservation, a small mammalian femur and a mammalian humerus of corresponding size. Mr. Davies has had as much of the matrix removed from these bones as was necessary to fully display their essential characters ; and I now offer to the Geolo~cal Society a brief account of the remains, for the opportunity of making which I would express my thanks to Mr. Waterhouse. It is perhaps impossible to determine with absolute certainty whether they belong to AmThi- therium, Phascolotherium, or to some new type, or whether the bones might not perhaps have belonged to two different ; but, on the chance of their being naturally associated, the generalized mar- supial characters which they manifest make a useful contribution to our knowledge of that type, which has hitherto only been known, from Stonesfield, by rami of lower jaws.

The right femur rests flat in the matrix, so as to expose the aspect Fig. 1.--Femur. of the bone, which during life was posterior and inferior. The ex- treme length of the specimen is 17~v inch. It has a straight aspect, is moderately expanded proximally, less enlarged at the distal end, and has the spheroidal articular head in- dined to the inner side and looking slightly forward. The most re- markable feature of the specimen is the expansion below the articular head of the two ~rochanteroid pro- cesses, which make the width of the bone in this position something less than ~ inch. The'inner trochanter is much the less massive of the two ; but its margin is imperfectly pre- served, and it was obviously thinner than the great trochanter. It ex- tends proximally to near the base of The outline shows the natural size the proximal articular head of the of the bone. bone, and had a narrow wing-like extension inward, curving some- what forward at its proximal margin ; but, as preserved, its internal Downloaded from http://jgslegacy.lyellcollection.org/ at University of California-San Diego on December 19, 2016

HUMERUS I~RO~ THE STONESFI~ELD SLATE. 457 extension beyond the proximal articulation is hardly more than inch. It is in no sense a process, but simply an expansion of the bone, and terminates distally in a short sharp ridge on the angle of the side of the shaft. The external trochanter is much thicker, and extends beyond the border of the articular head for T2~yinch. It is separated from this convex surface by a concave excavation about -~o inch wide ; and its outer rounded border, which is nearly T~ inch thick, is reflected backward, and terminated proximally in a rounded process, which at the proximal end curves a little forward. The rounded articular head has no neck connecting it with the shaft; but a constriction extends below the globular surface pos- teriorly, and another depression is apparently more marked on the anterior aspect. The transverse measurement of the head is about ~_a inch, and the antero-posterior measurement is somewhat more. Its depth on the posterior aspect is about ~ inch. Between the head of the bone and the two trochanters is a con- cave space, which runs down the shaft for about one third of its length, disappearing in a line as the bone narrows from side to side, and it is deepest towards the great trochanter. There is no trace of an ob- turator pit, nor of any excavation towards the external trochanter, such as characterizes this part of the bone in the majority of mam- mals, and especially ; so that at first sight there is a primd facie suggestion in this region of bats, or moles, or of a type allied to the , in which the trochanters were less deve- loped than in Ornithorhynchus; but there are slight concavely curved lines extending between the trochanters and indicating muscular attachment. The shaft becomes reduced to its least width of about .-f-r inch near the middle. Its posterior lateral outline is concave, so as to give the distal articulation the aspect of being produced slightly outward. The internal outline is nearly straight between the in- ternal troehanter and the distal articulation. The shaft of the bone is evidently flattened naturally, is straight on the posterior aspect below the trochanters, and becomes convexly curved from side to side. There is a slight groove extending down the length of the shaft in the median line from the external trochanteroid ridge to between the condyles, and gradually widening distally. The width of the bone across the distal condyles is rather less than a inch ; and though the specimen is not sufficiently developed from the matrix to show the thickness of the condylar end, it was evidently much less thick than is usual among . The outer condyle is the larger; it has a transversely ovate outline, and is divided on the posterior aspect by a moderately deep depression from the smaller inner condyle, in which the greatest measurement was vertical. The depth of these condyles is about T~ inch, and the shaft above them is concavely excavated. These are the characters on which the systematic position of th6 animal must be determined. The bone is altogether less robust and is rather smaller than the same bone in Ornithorhynchus, has the trochanters narrower and smaller, the distal condyles narrower, Downloaded from http://jgslegacy.lyellcollection.org/ at University of California-San Diego on December 19, 2016

458 H. G. SEELEY ON A MAMMALIAN F]~MUR AND

and the articular head directed less upward and more inward. In all these points of difference there is a nearer approach to the femur of , which, however, is twice as long. In that the tro- chanters are relatively less developed, and the head of the bone is less well defined on the inferior aspect, while in the fossil the in- ternal trochanter has a greater proximal extension. Both these genera of living Monotremes agree in possessing a similarly unex- cavated condition of the posterior aspect of the external trochanter ; and there are enough intermediate steps between the two surviving genera to include such a form as this fossil bone, if there were any corroborative evidence to justify its location in the same order with them. Among the marsupials there are not many in which a sufficiently near correspondence can be detected to justify a comparison. It is only among some of the Phalangers, such as Phalm~gista Cookii, that there is a general resemblance seen in form and type, though the existence in that group of an obturator pit renders detailed com- parison impossible. The only other marsupial to which reference may naturally be made is the Myrmecobius, interesting from its resemblance to Ampl~itherium ; but the bone is relatively shorter than in Myrmecobiusfasciatus; the condyles at the distal end are re- latively much less thickened. The internal troehanter does not ex- tend so far proximally in the recent genus, and the fossil shows no trace of the excavation of the inferior aspect of the proximal end. It may therefore be concluded that the resemblance to Marsupials is, on the whole, less evident than the probable affinity of the bone to a . It is interesting that has presented characters in the inflected angle of the lower jaw which induces Prof. Owen to dwell on its resemblance to the mole and the hedge- hog, and to express a necessary caution against hastily concluding that the animal was not insectivorous. Dr. Gill, one of the most profound systematists of modern times, has sufficiently recognized, in his tabular view of the classification of the Mammalia, the near affinity of the Chiroptera, Insectivora, Rodents, and Edentates with the Marsupial and Bionotreme orders; and it is certainly remarkable that this fossil bone, like the jaw referred to, should present cha- racters which necessitate close examination as to its placental or its implacental determination. There are many animals in which the proximal end of the femur has two well-developed trochanters, and frQm which the obturator pit is absent ; but in none do the condyles attain so great a size as in the fossil. In the Little Ant-ester (Oycloth~,r,s didact!flus) the femur is 11-~ inch long, while the transverse measurement across the trochanters is only ~ inch, so that in this animal the shaft is 3] times as long as the transverse measurement over the trochanters, while in the fossil the shaft is not more than twice as long as the width over the trochanters. The femur of the Mole has only a general resemblance to the fossil ; the trochanters are narrower ; the great trochanter is more developed proximally, and the concave area between the tro- chanters is of an altogether different character. A similar remark Downloaded from http://jgslegacy.lyellcollection.org/ at University of California-San Diego on December 19, 2016

HUMERUS l~RO~f THE STON]~SFIELD SLATE. 459 applies to the proximal end of the femur of the Hedgehog, which at first sight approximates in form to the fossil. The Bats probably make a nearer approach than the Inseetivora, both trochanters being well developed; but in the Bat there is not the same concave channel extending down the shaft between the trochanters, which is so remarkable a character in the fossil. I am therefore disposed to believe that the Monotreme characters predo- minate, and justify us in believing that the specimen was impla- cental, although there may be some reason for believing that it was, to some extent, also marsupial in its characters.

HumERus. The right humerus rests flat in the matrix, so as to expose its pos- terior aspect. It is 1TLo inch long, Fig. 2.--tTumeru~. with a globose head, slender shaft, straight on the anterior or inner side, concave on the external or posterior side, and compressed and somewhat expanded at the distal end. It is much more marsupial ill its characters than the femur. The head is ~ inch wide, and is rather directed outward and backward ; it is about ~(~ inch deep, and presents a form which is seen in Phascolarc- tos, but which is unusual. The shaft narrows to ~ inch in the middle, where its section would ap- parently be subtriangular, owing to the inner side of the bone being flattened vertically. This inner side extends upwards so as to form a The outline shows the natural size small trochanter on the inner side of the bone. of the articular head ; but the poste- rior border of the bone is sepa- rated from the head by a well-de- fined transverse depression. On the proximal half of the inner margin rises a well-defined moderately elevated crest, which, so far as I am aware, .is not paralleled in position in any humerus which is com- parable with the fossil. The inner side of this ridge is continuous with the flattened inner side of the bone, and it extends lip to merge in the trochanter; but the outer side of the ridge is defined by a slight clean groove. The ridge attains its greatest elevation at ~ inch below the articular head of the humerus, and nearly ~1~ inch from its proximal articular surface. The distal end of the bone lies in the same plane with the proximal end. The external margin is expanded into a thin plate, which has a highly convex external outline. This plate on its posterior aspect is flat and smooth, and in no way de- fined from the rest of the bone. Its contour helps to form the bow- Downloaded from http://jgslegacy.lyellcollection.org/ at University of California-San Diego on December 19, 2016

460 H. G. SF,ELEY ON & M&MMALT&N FEMUR AND shaped outline of the external side of the shaft. The inner border of the distal end also expands distally, though not to the same ex- tent. It is less perfectly preserved, but was obviously much thicker and apparently flattened, and makes the width of the distal end in a line with the olecranon pit a trifle over .7~ inch. The details of the distal articulation are not well seen; but the bone appears to have been compressed and rounded, and the extension below the elecranon-pit of the condyles is rather less than ~a inch. The ole- cranon-pit is not very deep, and is transversely ovate, placed a little obliquely, and nearer to the external than to the internal border of the bone. The principal condyle was below the inner half of the olecranon depression, and is partly broken away ; but a small second condyle appears below its external corner, and between these there is a narrow smooth impressed intercondylar area, which extends obliquely forward and upward, and indicates the mode of flexure of the fore limb. The nearest approach to this form of humerus is probably pre- sented by some of the Phalangers, though the ridge, which is lateral on the side of the upper part of the shaft, would need to be turned, together with the head, through a considerable angle before the two forms could be brought into harmony. On the oblique inner distal border is the usual supracondylar foramen seen in Marsupials and several other orders, which appears to have a vertically ovate outline, and to pass obliquely forward and downward and very slightly inward; it is placed just above the border of the olecranon-pit, and is separated from it by a width of inch. Its inner border is a narrow arch of bone, which is angular, being co]npressed from the front. So far we have compared the humerus as though it were an isolated bone ; and have ignore~t the possibility that it might present affinities of the same character as the femur. The difficulties which have pre- sented themselves hitherto in its interpretation are :--(1) the con- cave anterior or external margin; (2) the ridge running down the superior aspect of the bone from the proximal articular head. But when our comparisons are turned to the Ornithorhynchus, singularly modified as the posterior or internal margin of the bone in that genus is, we recognize at once the significance of the characters which have hitherto been so difficult to interpret. In no other genus is the an- terior concavity of the bone between the proximal and distal ends so pronounced as in this Monotreme ; and although this right humerus presents a form which is more suggestive of a Marsupial in its general aspect, yet in these important characters, as well as in the outward direction of the articular head, it makes an approximation to the Ornithorhynchus which there is no reason for believing to be accidental, and every reasoll to regard as an indication of organic grade. It is not impossible that the Stonesfield animal may have had burrowing habits ; but it is highly probable that a humerus so modified as this must have been connected with a pectoral arch like that of a Monotreme in construction, and unlike the ordinary form of the arch in placental Mammals or Marsupials. And if any diffi- culty should suggest itself in this comparison on the ground of the Downloaded from http://jgslegacy.lyellcollection.org/ at University of California-San Diego on December 19, 2016

ItU~[ERUS FROM THE STONESFIELD SLATE. 461

difference of the bone from the same element in Ornithorhynchus, it should be borne in mind that the Monotremes are at present known to us very imperfectly, and that between the Duck-billed and the Echidna there is a wide interval of organization. But per- haps the most interesting evidence of the affinity of the humerus is obtained when we bear in mind that it is found with the femur, and that the two bones correspond in presenting predominant indi- cations of Monotreme characters under the external forms usual among Marsupials. The only history of these specimens is that they were collected by Mr. S. Peace Pratt, formerly one of the Secretaries of the Geological Society, and presented by him, with four other bones from Stonesfield, to the British Museum. Mammalian remains occur so rarely that there is a strong predisposition to regard these bones as naturally associated ; and the somewhat smaller size of the humerus, as compared with the femur, would favour such judgment, however strong the temp- tation might be to refer the bones to different genera on account of their resemblances to such different types of existing animals. At least, in the absence of evidence to the contrary, it may not be in- convenient to consider them to have been portions of a single indivi- dual. There then arises the more difficult question as to whether they pertain to any of the genera of the Stenesfield-Slate mammals hitherto established or not. And necessarily a judgment on this point must be largely founded on probabilities. The comparatively large Ste- reognathus is so generally regarded as placental, that its claims to the humerus and femur may be dismissed. There then remain only Amphitherium and Phascolotherium with which they may be associated. The size of the bones is not very different from what might be anticipated from the length of the jaws already described in those genera. In many small animals the length of the lower jaw is not greatly different from that of the femur, and the humerus is usually somewhat shorter, though there are many animals in which the skull is much longer than the femur. In Ornithorhynchus the lower jaw measures 3~ inches, the femur 1~ inch, the humerus ]l-a6 inch. In Echidna the skull is 4T140 inches long, the femur 2~ inches, the humerus lS~ inch. In Didelphis virginiana the lower jaw is 31-~-a inches, femur 3~o inches, humerus 2~ inches. In Myr- mecobius fasciatus the lower jaw is 1Ts~ inch, ~cmur 1T~ inch, hu- merus 1~ inch. In Phalangista Coolcii the lower jaw is 1~-7~ inch, the femur 21%(~inches, the humerus 2T~ inches. In Phascolarctos the lower jaw is 2]~0 inches, femur 3~o inches, humerus 2~6 inches. In the Little Ant-eater the lower jaw is 1Ta0. inch, the femur 1]~ inch, the humerus 1-~6.ineh. In the Hedgehog the lower jaw is 1~ inch, and the femur and humerus are of the same length. In the Mole the skull is 1~-6 inch, femur ~ inch, and humerus r inch. In the Beaver the lower jaw is 2~ inches, femur 2~-a6 inches, humerus 1~ inch. In the Marmot the lower jaw is 2;'-~o inches, the femur 3 inches, the humerus 2~ inches. Now in Phascolotherium the length of the jaw is about 1~- inch, and in Am phitherium the length of the jaw is about 1~2~ inch, while Downloaded from http://jgslegacy.lyellcollection.org/ at University of California-San Diego on December 19, 2016

462 H. G. SEELEY ON A MAMMALIAN FEMUR AND

the femur described is l~0- inch and the humerus 1~ inch. Thus it is evident that there is not much ground for choice in affiliating the specimens to one genus or the other on the ground of proportion between the jaws and the limb-bones, though the presumptive evi- dence is strong, so far as size goes, that they might be referable to one of them. 0n the whole I am inclined to believe that the some- what insectivorous and stroug character of the lower jaw of Phas- colotherium may perhaps be taken into consideration in counexion with the resemblance which the head of the femur described offers to the bone in some Insectivores, as indicating a habit which would justify us in suspecting that the remains are rather to be referred to that genus. If the three specimens may all be referred to the same genus, they indicate a generalized ~[arsupial, which certainly suggests evolution from a Monotreme stock, and cannot be placed in any division of the existing Marsupial order. If the affiliation of these limb-bones to the jaw of Phascolo- therium be accepted as an identification sufficiently probable to excuse me, in the absence of evidenee of their generic distinctness, from the task of founding a new genus and for the fossils, then a cerC~n appositeness may be recognized in the association of a Marsupial jaw of generalized character with limb-bones indicating a generalized Insectivorous type, modified from a Monotreme stock in the direc- tion of the Marsupial plan, as one of those harmonies which the general truths of the doctrine of evolution would lead us to antici- pate ; for the researches of Profs. 0wen, Gaudry, and Marsh have gone far to prove that every type of mammalian life becomes more generalized when it is ~Lcer backwards in time, and presents affi- nities towards less specialized orders; from which we infer that these, or similar orders, were the parent stock from which the sur- viving types were derived. Similarly I am inclined to suggest that these remains from Stonesfield justify us in inferring the exis- tence of a generalized order of animals, which were not ]iIarsupials, but from which the Marsupials became evolved.

DIscussioN.

The PxmrDJJJ~T eon~atulated the author upon the discovery of the importance of these specimens, whioh had been so long over- looked. Prof. PRm~wIo~ remarked on the interest attaching to the dis- covery of these remains among old specimens from Stonesfield, since at present the slate-pits at that place are little worked. He suggested that if other store-collections of Stonesfield fossils (like those of Oxford) were overhauled, they might yield equally valuable results to so careful a worker as Prof. Seeley. Dr. ]~VRn~ expressed doubts whether the humerus was a fight one, and also as to the validity of conclusions derived from a com- parison of the lengths of the limb-bones and the lower jaws. Downloaded from http://jgslegacy.lyellcollection.org/ at University of California-San Diego on December 19, 2016

HUM~.RUS FROM TILE STONESFIELD SLATE. 463

He thought that Prof. Seeley's views as to the predominantly monotrematous character of the animal were hardly supported by the evidence adduced, and that if the jaws of Phascolotherium had not before been discovered the identification might not have been made. Mr. Ct][ARL]]SWORTKinsisted on the important discovery of bones other than mandibular in the Stenesfield Slate, a discovery which had been looked forward to in vain since 1814. He cited and ex- pressed doubt on the explanation of Prof. 0wen, that the lower jaws had fallen away from the floating skeletons. Mr. B~AxE argued in favour of both the bones belonging to one specimen, and pointed out that the greatly modified bone (the hu- merus) was the one most likely to vary with the habits of the animal. Prof. S~EL~v, in reply, expressed his thanks for the compliments which had been paid him. He defended his identification and in- terpretation, his comparisons of the lengths of the limb-bones and jaws in different orders of mammals, and his conclusion that the bones had monotreme affinities. He should have arrived at the same conclusion ff no jaws had ever been found.

Q. J. G. S. No. 139. ~"