Through the Heart of Patagonia

by H. Hesketh-Prichard

I. Account of the Discovery. By Dr. Moreno.

In November 1897 I paid a visit to that part of the Patagonian territory which adjoins the Cordillera of the Andes, between the 51st and 52nd degrees of South latitude, where certain surveyors, under my direction, were carrying out the preliminary studies connected with the boundary-line between Chile and Argentina; and in the course of this expedition I reached Consuelo Cove, which lies in Last Hope Inlet. In that spot, hung up on a tree, I found a piece of a dried skin, which attracted my attention most strangely, as I could not determine to what class of Mammalia it could belong, more especially because of the resemblance of the small incrusted bones it contained to those of the Pampean Mylodon. On inquiring whence it came, I was informed that it was only a fragment of a large piece of skin which had been discovered two years before, by some Argentine officers, in a cavern which existed in the neighbouring heights. Immediately on receiving this news, I hastened to the spot, guided by a sailor who had been present when the original discovery had been made. As, at that moment, I had no means of making more than a few hurried excavations, which gave no further traces of the discovery, I left orders that the search should be continued after my departure; but this once more also failed to give any ultimate results. Nothing could be found but modern remains of small rodents, and these chiefly on or near the surface of the ground. From the most careful inquiries which I set on foot, it appeared that, when the first discovery was made, no bones were found, the skin being half buried in the dust which had accumulated from the gradual falling away of the roof of the cavern, composed of Tertiary Conglomerate. It was only in the broad entrance to the cavern that were found a few human bones, borne thence to the shore of the Cove and afterwards broken up.

As already stated, the skin here presented to you formed but a small part of a larger one. One small piece had been carried off by Dr. Otto Nordenskjöld, and others by officers of the Chilian Navy, who later on had visited the spot. The inhabitants of the locality looked upon it as an interesting curiosity, some of them believing that it was the hide of a cow incrusted with pebbles, and others asserting that it was the skin of a large Seal belonging to a hitherto unknown species.

In Consuelo Cove, I embarked on board a small Argentine transport, which had been placed at my disposal to carry out the study of the western coast as far as Port Montt, in lat. 42°. At this latter place I left the steamer, which then proceeded to make a series of surveys. These lasted until her return to La Plata, at the latter end of July 1898, when she brought back to me the fragment of skin in question.

This is an accurate and true version of the discovery of this skin, which gave rise to the publication of Señor Ameghino's small pamphlet,[33] in which he gave an account of the discovery of a living representative of the "Gravigrades" of Argentina, distinguishing it by the name of "Neomylodon listai".

I have an idea that Señor Ameghino never saw the skin itself, but only some of the small incrusted bones, of which he had obtained possession. The vague form in which he draws up his account compels me to believe this suspicion to be true.

My opinion is that this skin belongs to a genuine Pampean Mylodon, preserved under peculiar circumstances resembling those to which we owe the skin and feathers of the Moa. I have always maintained that the Pampean Edentates, now extinct, disappeared only in the epoch which is called the "historical epoch" of our America. In the province of Buenos Aires, buried chiefly in the humus, I have found remains of Panochthus, and others of the same Mylodon from the seashore, all of which present the same characteristic marks of preservation as the remains of human beings discovered in the same spot. In this identical layer of the sea-shore, close to the bones I have also found stones polished by the hand of man, and flints cut like those found in the Pampean formation. In 1884, in a cavern near to the Rio de los Patos, in the Cordillera, I discovered some paintings in red ochre, one of which, in my opinion, resembles the Glyptodon on account of the shape of the carapace.

Ancient chroniclers inform us that the indigenous inhabitants recorded the existence of a strange, ugly, huge hairy animal which had its abode in the Cordillera to the south of lat. 37°. The Tehuelches and the Gennakens have mentioned similar animals to me, of whose existence their ancestors had transmitted the remembrance; and in the neighbourhood of the Rio Negro, the aged cacique Sinchel, in 1875, pointed out to me a cave, the supposed lair of one of these monsters, called "Ellengassen"; but I must add that none of the many Indians with whom I have conversed in Patagonia have ever referred to the actual existence of animals to which we can attribute the skin in question, nor even of any which answer to the suppositions of Señor Ameghino according to Señor Lista. It is but rarely that a few Otters (Lutra) are found in the lakes and rivers of the Andes, as in the neighbourhood of Lake Argentino, in the "Sierra de las Viscachas," and in the regions which I believe Señor Lista visited, there are only a few scarce Chinchillas (Lagidium), which have a colouring more dark greyish than those found to the north, and are in every case separated from these by a large extent of country.

The Pampean Edentata have in former days certainly existed as far south as the extreme limit of Patagonia. In 1874, in the bay of Santa Cruz, I met with the remains of a pelvis of one of these animals in Pleistocene deposits, and also remains of the mammals which are found in the same formation, such as the Macrauchenia and Auchenia. It would not be astonishing that the skin of one of these should have been preserved so long, because of the favourable conditions of the spot in which it was found.

The state of preservation of this piece of skin, at first sight, makes it difficult for one to believe it to be of great antiquity; but this is by no means an impossibility, if we consider the conditions of the cave in which it was found, the atmosphere of which is not so damp as one might at first imagine it to be, although it is situated in the woody regions near to the glaciers and lakes. It is well to mention that in 1877, under similar conditions, and in a much smaller cave, scarcely five metres from the waters of Lake Argentino, situated sixty miles more to the north, I discovered a mummified human body painted red, with the head still covered in part with its short hair wonderfully preserved, and wrapped up in a covering made of the skin of a Rhea, and holding in its arms a large feather of the Condor, also painted red; this was all covered up with a layer of grass and dust fallen from the roof of the cave. In another cave in the neighbourhood I discovered a large trunk of a tree, painted with figures in red, black, and yellow. The sides of the rock close to the entrance of the cave were covered with figures, some representing the human hand, others combinations of curved, straight, and circular lines, painted white, red, yellow, and green. Now, this mummy, which is preserved in the Museum of La Plata, does not belong to any of the actual tribes of Patagonia. Its skull resembles rather one of those more ancient races found in the cemeteries in the valley of the Rio Negro—a most interesting fact, since they belong to types which have completely disappeared from the Patagonian regions, and it is well known that the actual Tehuelches may be considered to have been the last indigenous races which reached the territory of Patagonia. Many a time the Tehuelches have spoken to me of these caves as abodes of the evil "spirits," and of the enigmatical painted figures they contained: some attributed the latter to these same "spirits," others to men of other races, of whom they have no recollection. In another cave, four hundred miles farther to the north, in 1880, I discovered other human bodies, more or less mummified and in good preservation, but of a different type, and beside them some painted poles which served to hold up their small tents, the use of which had already disappeared more than three centuries ago; together with the upper part of the skull of a child perfectly scooped out like a cup. And yet the historical Tehuelches, the same as all the indigenous races in the southern extremity of South America, hold their dead in great respect, and never use such drinking-vessels.

These proofs of the favourable conditions of the climate and of the lands near to the Cordillera, which are revealed to us by the preservation of objects undoubtedly dating from very remote epochs, strengthen my opinion that this skin of a huge mammal, which has long since disappeared, may well have been preserved till the present time.

I may add that a further careful search is now being made in the earth forming the floor of the cave, and I hope in due time to have the honour of communicating the results to this Society.

II. Description and Comparison of the Specimen.
By A. Smith Woodward.

(a) Description.

The problematical piece of skin discovered by Dr. Moreno measures approximately 0.48 m. in the direction of the main lie of the hair, while its maximum extent at right angles to this direction is about 0.55 m. The fragment, however, is very irregular in shape; and it has become much distorted in the process of drying, so that the anterior portion, which is directed upwards in the drawing, is bent outwards at a considerable angle to the main part of the specimen which will be claimed to represent the back. The skin, as observed in transverse section, presents a dried, felt-like aspect; but there is a frequent ruddiness, suggestive of blood-stains, while the margin exhibits distinct indications of freshly dried once-fluid matter, which Dr. Vaughan Harley has kindly examined and pronounced to be serum. Its outer face is completely covered with hair, except in the region marked C and above B, where this covering seems to have been comparatively fine and may have been accidentally removed. The inner face of the skin is only intact in a few places, the specimen having contracted and perhaps been somewhat abraded, so that a remarkable armour of small bony tubercles, irregularly arranged and of variable size, is exposed over the greater part of it. At one point there is an irregular rounded hole about 0.02 m. in diameter, which might possibly have been caused by a bullet or a dagger, but in any case was probably pierced when the skin was still fresh. Owing to its direction, this hole is partly obscured by the overhanging hair.

The skin in its dried state varies in thickness in different parts. The average thickness of the flattened portion, which must be referred to the back, is shown by the cleanly-cut right margin of the specimen to be 0.01 m. This is slightly increased towards the posterior (lower) end of the border, while above it the thickness becomes 0.015 m. The latter thickness also seems to be attained in the much-shrivelled corner marked C—a circumstance suggesting bilateral symmetry between at least part of the two anterior outer angles of the specimen. The thinnest portion preserved is the border above B; and the skin must also have been comparatively thin in the region of the accidental notch to the left, considerably below C.

The portion of skin above B is interesting not only from its relative thinness, but also from the occurrence of an apparently natural rounded concavity in the margin. This excavation, which measures 0.05 m. along the curve, is marked by the remains of a thin flexible flap, which is sharply bent outwards, and is covered with short hairs on its outer face. It is especially suggestive of the base of an ear-conch; and if this appearance be not deceptive, it is worthy of note that the dried skin hereabouts and in the region which would have to be interpreted as cheek (C) is much more wrinkled than elsewhere.

SKIN OF GRYPOTHERIUM, OUTER VIEW. ¼ nat. size.

SKIN OF GRYPOTHERIUM, INNER VIEW. ¼ nat. size.

As already mentioned, the outer aspect of the skin is completely covered with hair, which is very dense everywhere except on the left anterior corner. Here it seems to have been removed by abrasion. A small patch of hair has also clearly been pulled out near the gap in the left border of the specimen; and close to the middle (where marked D) there is a small hairless depression which may perhaps be interpreted as a wound inflicted and healed during life. The hair is only of one kind, without any trace of under-fur, and it is still very firmly implanted in the skin, without signs of decay. Its arrangement seems to be quite regular, there being no tendency towards its segregation into small groups or bundles. It is of a uniform dirty yellowish or light yellowish-brown colour, and, making due allowance for slight ruffling and distortion of the specimen, it may be described as all lying in one direction, vertically in the photograph, except at the two upturned anterior corners of the specimen, where there is an inclination from the right and left respectively towards the centre. The longest hairs, which usually measure from 0.05 m. to 0.065 m. in length, are observed in the half of the specimen in front of (above) the letter D. Those in the middle of the extreme anterior (upper) border measure from 0.03 m. to 0.05 m. in length, those at the hinder (lower) border about the same; while some of the comparatively small and delicate hairs on the supposed cheek are not longer than 0.01 m. The hairs are stiff, straight, or only very slightly wavy, and all are remarkably tough. Examined under the microscope, their cuticle is observed to be quite smooth, while the much-elongated cells of the cortex are readily distinguishable. Mr. R. H. Burne has kindly made some transverse sections, which prove the hairs to be almost or quite cylindrical, and none of the specimens examined present any trace of a medulla.

The dermal ossicles are very irregular in arrangement, but are to be observed in every part of the specimen, even in the comparatively thin region near the supposed ear. They form everywhere a very compact armour, and some of them are quite closely pressed together; rarely, indeed, there is a shallow groove crossing a specimen, possibly indicating two components which were originally separate. As shown by every part of the cut margin, and especially well in a small section prepared by Prof. Charles Stewart, they are all confined to the lower half of the dermis, never encroaching upon the upper portion in which the hair is implanted. It is also to be observed that, where the inner surface of the skin is intact, the ossicles are completely embedded and only faintly visible through the dry tissue. The exposure of a considerable number of them, as already mentioned, is due to the rupture and partial abrasion of this surface. No tendency to arrangement in parallel lines or bands can be detected; and large and small ossicles seem to be indiscriminately mingled, although of course allowance must be made, in examining sections and the abraded inner view of the skin, for differences in the plane of adjoining sections and varying degrees of exposure by the removal of the soft tissue. The largest ossicles are oblong in shape when viewed from within, and measure approximately 0.015 m. by 0.010 m.; but the majority are much smaller than these. They are very variable and irregular in form; but their inner face is generally convex, sometimes almost pyramidal, while the outer face of the few which have been examined is slightly convex, more or less flattened, without any trace of regular markings.

In microscopical structure the dermal ossicles are of much interest, and I have examined both horizontal and vertical sections, one of the former kindly prepared by Prof. Charles Stewart. The tissue is traversed in all directions by a dense mass of interlacing bundles of connective-tissue fibres, which exhibit an entirely irregular disposition, except quite at the periphery of the ossicle. Here they are less dense, and are arranged in such a manner as to form at least one darkened zone concentric with the margin in the comparatively translucent border. Occasionally, but not at all points, the fibres in this peripheral area may be observed to radiate regularly outwards. Numerous small vascular canals, frequently branching, are cut in various directions; and the bony tissue, which is developed in every part of the ossicle, exhibits abundant lacunæ. Nearly everywhere, except in the narrow peripheral area just mentioned, it is easy to recognise the bony laminæ arranged in Haversian systems round the canals; and most of the lacunæ between these laminæ are excessively elongated, with very numerous branching canaliculi, which extend at right angles to their longer axis. Near the margin of the ossicle, especially in its more translucent parts, the bone-lacunæ are less elongated, more irregular in shape, and apparently not arranged in any definite order. There is no clear evidence of bony laminæ concentric with the outer margin, though appearances are sometimes suggestive of this arrangement. A vertical section of an ossicle presents exactly the same features as the horizontal section now described. It is thus evident that the vascular canals with their Haversian systems of bone have no definite direction, but are disposed in an entirely irregular manner.

Taking into consideration all characters, and making comparisons with the aid of my friend Mr. W. E. de Winton, I am inclined to regard the fragmentary specimen as the skin of the neck and shoulder-region with part of the left cheek. The apparent bilateral symmetry between at least part of the thickened anterior outer angles of the specimen has already been noted; and if this observation be well founded, the middle line of the back extends vertically down the middle of the photograph, [p. 306]. If the rounded notch above B be the base of the external ear, as seems probable, the thick wrinkled skin (C) with fine short hair still further to the left must be the cheek. The ear and cheek on the right side have been removed; but at the base of the outwardly-turned angle on this side of the specimen there are the very long hairs which occupy a similar position on the left. It thus seems possible to estimate the transverse measurement between the ears as from 0.25 m. to 0.30 m., which corresponds with a tentative estimate of the same distance in Mylodon robustus based on a skull in the British Museum.

(b) Comparisons and General Conclusions.

The skin now described differs from that of all known terrestrial Mammalia, except certain Edentata, in the presence of a bony dermal armour. There can therefore be little doubt that the specimen has been rightly referred to a member of this typically South American order. Even among the Edentates, however, the fragment now under consideration is unique in one respect; for all the ossicles are buried deeply in the lower half of the thickened dermis and the hairs are implanted in every part of its upper half, whereas all the forms of bony armour hitherto described in this order reach the outer surface of the dermis and are merely invested with horny epidermis. This is the case, as is well known, in the common existing Armadillos, in which the hair is only implanted in the dermis between the separate parts of the armour. Even in the unique and remarkable skin of an Armadillo from Northern Brazil, described by Milne-Edwards under the name of Scleropleura bruneti[34] the bony plates and tubercles are still covered only by epidermis, although most of them are reduced to small nodules and might well have sunk more deeply into the abnormally hairy skin. There is also reason to believe that in the gigantic extinct Armadillos of the family Glyptodontidæ the same arrangement of dermal structures prevailed; for one specimen of Panochthus tuberculatus obtained by Dr. Moreno for the La Plata Museum actually shows the dried horny epidermis in direct contact with the underlying bone, and seems to prove that the numerous perforations in the Glyptodont dermal armour were not for the implantation of hairs (as once supposed), but for the passage of blood-vessels to the base of the epidermal layer. Similarly, among the extinct Ground-Sloths of the family Mylodontidæ dermal ossicles have been found with the remains of Cœlodon[35] and various forms (perhaps different subgenera) of Mylodon; but the only examples of this armour yet definitely described[36] exhibit a conspicuously sculptured outer flattened face, and it thus seems clear that Burmeister was correct in describing them as originally reaching the upper surface of the dermis and only covered externally by a thickened epidermis. It is, however, to be noted that Burmeister himself actually observed armour of this kind covering only the lumbar region of the trunk. He believed that the other parts of the animal were similarly armoured, because he had found "the same ossicles" on the digits of the manus, where they were "generally smaller and more spherical"; but he unfortunately omits to make any explicit statement as to the presence or absence of the characteristic external ornamentation on the latter.

The omission just mentioned is especially unfortunate, because on careful comparison it is evident that the irregular disposition of the small ossicles in the piece of skin now under consideration is most closely paralleled in the dermal armour of the extinct Mylodon, as already observed by Drs. Moreno and Ameghino. There is obviously no approach in this specimen to the definite and symmetrical arrangement of the armour such as is exhibited both by the existing Armadillos and the extinct Glyptodonts. There are, then, two possibilities. Either the dermal armour of Mylodon varied in different parts of the body, being sculptured and covered only by epidermis in the lumbar region, while less developed, not sculptured but completely buried in the dermis in the comparatively flexible neck and shoulder region—in which case Dr. Moreno may be correct in referring the problematical specimen to Mylodon; or the dermal ossicles of this extinct genus may have been uniform throughout, only differing in size and sparseness or compactness—in which case Dr. Ameghino is justified in proposing to recognise a distinct genus, Neomylodon.

To decide between these two possibilities, it is necessary to wait for additional information concerning the anterior dorsal armour of Mylodon as precise as that published by Burmeister in reference to the lumbar shield. Meanwhile it must suffice to compare the microscopical structure of the ossicles from the new skin with that of the small sculptured tubercles of undoubted Mylodon. It must be remembered that the specimen has been buried in the Pampa Formation for a long period, and that the oxides of iron and manganese have infiltrated the margin of the bone, rendering the structure of its outer border more conspicuous than that of its central portion. It must also be noted that some of the manganese has assumed its familiar "dendritic" aspect, in this respect presenting appearances not due to original structure. The calcified interlacing fibres of connective tissue are as abundant here as in the ossicle of the so-called Neomylodon; but in a very wide peripheral area they exhibit a marked radial disposition, nearly everywhere extending in bundles at right angles to the border. Rather large vascular canals, infiltrated with the oxides of iron and manganese, are observed in places, often bifurcated and usually bordered by a transparent zone free from the connective-tissue fibres. Well-developed bone-lacunæ are very abundant, many exhibiting short branching canaliculi, and most of the others very irregular in shape, evidently furnished with canaliculi which cannot be seen from lack of infiltration. The lacunæ are never much elongated, and are not arranged in distinctly differentiated Haversian systems in any part of the section; while the only regular disposition of the bony laminæ is traceable near the circumference, where the lacunæ are frequently arranged or clustered in parallel zones concentric with the border. A vertical section of one of the same specimens shows the connective-tissue fibres radiating outwards towards the lateral margins, but not directly towards the upper sculptured face. There are no bony laminæ clearly parallel with the latter face, and at least one vascular canal in transverse section seems to be the centre of a Haversian system.

The histological structure of the ossicles in the skin now under consideration thus resembles that of the sculptured tubercles of Mylodon in all essential features, but differs in two noteworthy respects. In the ossicles of the so-called Neomylodon, as already described, the fibres of connective tissue do not exhibit much definite radiation towards the lateral margin; while the bony tissue at most points is disposed in definite Haversian systems. There is thus enough discrepancy to justify the suspicion that the new and the old specimens do not belong to the same animal. In fact, so far as the differentiation of the dermal bone is concerned, the so-called Neomylodon is precisely intermediate between Mylodon and the existing Armadillo (Dasypus); sections of the scutes of the latter animal, both in the Royal College of Surgeons and in the British Museum, showing that in this genus nearly the whole of the osseous tissue is arranged in Haversian systems, although abundant interlacing connective-tissue fibres are still entangled in it, at least near the border.

If the characteristic dermal armature does not suffice for the definite expression of an opinion as to the precise affinities of the specimen, a still less satisfactory result can be expected from a comparison of the hair. For, in the first place, no hair has hitherto been discovered in association with the skeleton of any extinct Ground-Sloth; while, secondly, the hairy covering of a mammal is perhaps that part of its organisation most readily adapted to the immediate circumstances of its life. So far as their endo-skeleton is concerned, the extinct Mylodonts and their allies are precisely intermediate between the existing Sloths and Anteaters; they combine "the head and dentition of the former with the structure of the vertebral column, limbs, and tail of the latter."[37] It might therefore be supposed that the hair of this extinct group would exhibit some of the peculiarities of that in one or other of its nearest surviving relatives. The epidermal covering of the piece of skin now described, however, entirely lacks the under-fur which is so thick in the Sloths; while the structure of each individual hair, with its smooth cuticle and lack of a medulla, is strikingly different from that observed both in the Sloths and Anteaters, and identical with that of the hair in the surviving Armadillos. The large hair in the Sloths and Tamandua exhibits a conspicuously scaly cuticle; while that of Myrmecophaga is remarkable for its very large medulla. All these animals now live in the tropics, either in forests or swamps, whereas the Patagonian animal must have existed under circumstances much like those under which the Armadillos still survive. Hence the characters of the hair of the so-called Neomylodon may be of no great importance in determining the affinities of the animal, but may represent a special adaptation to its immediate environment.

Finally, there is the question of the antiquity of the problematical skin. On two occasions I have examined the mummified remains of the extinct Mammoth and Rhinoceros from Siberia in the Imperial Academy of Sciences at St. Petersburg; I have also carefully studied the remains of the neck and legs of the Moa from a cavern in New Zealand, now in the British Museum. Compared with these shrivelled and dried specimens, the piece of skin from Patagonia has a remarkably fresh and modern aspect; and I should unhesitatingly express the opinion that it belonged to an animal killed shortly before Dr. Moreno recognised its interest, had he not been able to give so circumstantial an account of its discovery and strengthened his point of view by recording the occurrence of a human mummy of an extinct race in another cavern in the same district. The presence of an abundant covering of dried serum on one cut border of the skin is alone suggestive of grave doubts as to the antiquity of the specimen; but Dr. Vaughan Harley tells me that similar dried serum has been observed several times among the remains of the Egyptian mummies, and there seems thus to be no limit to the length of time for which it can be preserved, provided it is removed from all contact with moisture. I may add that I have searched in vain in the writings of Ramon Lista (so far as they are represented in the Library of the Royal Geographical Society) for some reference to the statement which the late traveller made verbally to Dr. Ameghino; and as the piece of skin now described certainly represents an animal almost gigantic in size compared with the Old-World Pangolin, I fear it cannot be claimed to belong to Lista's problematical quadruped, whatever that may prove to be.

The final result of these brief considerations is therefore rather disappointing. There are difficulties in either of the two possible hypotheses. We have a piece of skin quite large enough to have belonged to the extinct Mylodon; but unfortunately it cannot be directly compared with the dermal armour of that genus, because it seems to belong to the neck-region, while the only dermal tubercles of a Mylodont hitherto definitely made known are referable to the lumbar region. If it does belong to Mylodon, as Dr. Moreno maintains, it implies either that this genus survived in Patagonia to a comparatively recent date, or that the circumstances of preservation were unique in the cavern where the specimen was discovered. On the other hand, if it belongs to a distinct and existing genus, as Dr. Ameghino maintains—and as most of the characters of the specimen itself would at first sight suggest—it is indeed strange that so large and remarkable a quadruped should have hitherto escaped detection in a country which has been so frequently visited by scientific explorers.

[P.S.—At the reading of this paper Prof. Ray Lankester remarked that he should regard the characters of the hair as specially important, and would not be surprised if the problematical piece of skin proved to belong to an unknown type of Armadillo. This possibility had occurred to me, but I had hesitated to mention it on account of the considerable discrepancy observable between the arrangement of the bony armour in Neomylodon and that in the known Glyptodonts and the unique Brazilian Armadillo (Scleropleura), which happen to exhibit an incompletely developed (incipient or vestigial) shield. In each of the latter cases, the armour is not subdivided into a compact mass of irregular ossicles, but consists of well-separated elements which could only become continuous by the addition of a considerable extent of bone round their margins, or by the special development of smaller intervening ossicles.

Since the paper was read, I have had the privilege of studying Dr. Einar Lönnberg's valuable description of the pieces of the problematical skin mentioned by Dr. Moreno as having been taken to Upsala by Dr. Otto Nordenskjöld.[38] It appears that with the skin was found the epidermal sheath of a large unknown claw, which may have belonged to the same animal. This specimen proves to be different from that of any existing Sloth, Anteater, or Armadillo, and is considered by Dr. Lönnberg to belong probably to the hind foot of a Mylodont, which did not walk on the exterior, lateral surfaces of the toes to the same extent as Mylodon. In a section of the skin provisionally ascribed to the leg, he observes that the small ossicles are very irregular, and shows two instances in which two are placed one above the other. In microscopical sections of the ossicles, however, he does not find the distinct Haversian systems of bone so conspicuous in my slides; and hence he fails to remark the differences between the structure of the armour in Neomylodon and Mylodon, which seem to me to be particularly noteworthy. His so-called "pigment cellules" in Mylodon are the dendritic infiltrations of oxide of manganese and stains of oxide of iron, to which I have made special reference. His observations as to the absence of a medulla in the hair confirm my own; but I have not seen any evidence of the suspected loss or disintegration of the hair-cuticle. Finally, Dr. Lönnberg has boiled a piece of the skin, thereby extracting glue, "which proves that the collagen and gelatinous substances are perfectly preserved." The latter observation confirms the evidence of the serum recorded above, and indicates that if the specimen is "of any considerable age, it must have been very well protected against moisture and bacteria."—A. S. W.]

III. Description of Additional Discoveries.
By A. Smith Woodward.[39]

Last February, when presenting to the Zoological Society an account of the skin of a Ground-Sloth discovered in a cavern in Southern Patagonia, Dr. Moreno mentioned that further excavations were being made in the hope of finding other remains of the same animal. The task referred to was undertaken by Dr. Rudolph Hauthal, geologist of the La Plata Museum, who met with complete success.[40] He not only found another piece of skin, but also various broken bones of more than one individual of a large species of Ground-Sloth in a remarkably fresh state of preservation. Moreover, he discovered teeth of an extinct horse and portions of limb-bones of a large feline carnivore, in association with these remains; he likewise met with traces of fire, which clearly occurred in the same deposits as the so-called Neomylodon. All these remains were found beneath the dry earth on the floor of an enormous chamber which seemed to have been artificially enclosed by rude walls. In one spot they were scattered through a thick deposit of excrement of some gigantic herbivore, evidently the Ground-Sloth itself; in another spot they were associated with an extensive accumulation of cut hay. Dr. Hauthal and his colleagues, indeed, concluded that the cavern was an old corral in which the Ground-Sloths had been kept and fed by man.

As the result of these explorations, Dr. Moreno has now the gratification of exhibiting to the Society complete proof that the piece of skin described on the former occasion belongs to a genuine Pampean Ground-Sloth, not Mylodon itself, but a very closely related genus Grypotherium, of which skulls are already known from Pampean deposits in the Province of Buenos Aires.[41] The collection which we now have the privilege of examining distinctly supports his contention that the large quadruped in question belongs to an extinct fauna, though contemporary with man. The discovery is thus unique in the history of palæontology, on account of the remarkably fresh state of preservation of all the remains. Some of the new specimens exhibit no indication whatever of having been buried. Many of the bones retain their original whitish colour, apparently without any loss of gelatine; while both these and other bones, which have evidently been entombed in brownish dust, bear numerous remnants not only of the dried periosteum, but also of shrivelled muscles, ligaments, and cartilages. Very few of the bones are fossilised, in the ordinary sense of the term.

An admirable brief description of this collection has already been published (op. cit.) by Dr. Roth, who was the first to recognise the generic identity of Neomylodon with Grypotherium. Some of the specimens, however, are worthy of a more detailed examination; and Dr. Moreno has kindly entrusted them to me for study in connection with the collections in the British Museum and the Royal College of Surgeons. The following notes, supplementing Dr. Roth's original memoir, are the result of this further investigation.

1. Remains of Grypothrium listai.

Number of Individuals.

Among the fragmentary bones of the Ground-Sloth, it is easy to recognise evidence of three individuals, which do not differ much in size. There are three distinct examples of the occiput and fragments of the dentigerous portion of three mandibles. It is also noteworthy that the three malar bones preserved are all different in shape, while three corresponding fragments of the acromial process of the scapula differ in size. One portion of maxilla seems to represent a fourth individual, being probably too small for either of the skulls to which the occiputs belong. Finally, as Dr. Roth has pointed out, one shaft of a humerus, which appears to be the bone of an adult, belongs to a much smaller animal than is indicated by any other specimen in the collection.

Remains of three individuals are thus recognisable with certainty; two others can probably be distinguished; while some of the fragments may even belong to a sixth specimen. It must also be noted that other portions of jaws are said to have been discovered by E. Nordenskjöld.[42]

Skull and Mandible.

The largest portion of cranium (No. 1) is not stained in any way, and does not retain a trace of the material in which it was buried in any hollow or crevice. It does not appear to have been damaged during excavation, but exhibits fractures which were almost certainly made when the animal was freshly killed. The cranial roof near the occipital region is battered in four places, though the injuries do not affect the brain-case itself; while the right occipital condyle is partly removed by a sharp, clean cut. There can, indeed, be no doubt that the animal was killed and cut to pieces by man.

This skull is evidently that of an adult animal, all the sutures in the hinder region being closed. The inner wall of the temporal fossa is much flattened, without any irregular convexities, but marked with the characteristic reticulately-decussating, fine ridges of bone, and studded with adherent patches of muscle-fibre. The upper border of the fossa is a remarkably sharp edge, while the narrow flattened cranial roof is only marked by a faint longitudinal median furrow and by a diminutive tuft of fibre in a small median pit near the occipital edge.[43] The fractures exhibit the very large cancellated chambers surrounding the brain-case dorso-laterally; while a median longitudinal section shows both these cells and others in the basi-sphenoid. The basi-cranial axis is nearly straight, inclining a little upwards in front. The anterior condyloid foramina piercing the basi-occipital are remarkably large, as usual; the basi-sphenoid is very long and narrow, flattened mesially on its lower face, but with one slight median prominence near its hinder end; the pre-sphenoid forms a short acute rostrum, above which there are remains of the vomer. The hinder ends of the pterygoids are shown to be inflated with large cancellæ, but the sides of the base of the skull are somewhat obscured by the dried soft parts. The mastoid process of the periotic, with its articular facette for the stylohyal, seems to be rather smaller than in Mylodon. The tympanic bone is preserved on the right side, though wanting on the left. It is an irregular curved plate only slightly bullate, but forming a complete floor to the tympanic cavity. As usual in Edentata, it is not produced into an auditory meatus.

The right maxilla (No. 4) is in precisely the same state of preservation as the specimen just described, and probably belongs to the same skull. Its anterior margin is perfectly preserved, indicating that the facial region is very short in front of the anterior end of the zygomatic arch, which is pierced by a rather large suborbital canal. Its upper border proves that the nasal region was raised into a slightly convex dome; while its antero-superior angle is not rounded as in Mylodon, but curves upwards and forwards and ends in a point as in Grypotherium. At the oral border there are the shattered bases of four teeth.

A fragment of the nasal region (No. 13) may also have belonged to the same skull, but its state of preservation is a little different from that of the two specimens just described. It has clearly been buried in a powdery deposit, which has stained it brown; but the enveloping dust must have been extremely dry, for fragments of cartilage adhere to it, as well preserved as in the nasal chamber of the cranium itself (No. 1). It also bears traces of the integument.

Judging by the figures of the skull of Grypotherium published by Reinhardt (loc. cit.), this specimen seems to have occupied an anterior position in the nasal region. It is thus of great interest, because the three known skulls of Grypotherium leave the precise nature of the bony arcade separating the narial openings undecided. According to Reinhardt, the nasal bones terminate as in Mylodon, and the arcade is an element interposed between them and the premaxillæ. According to Burmeister, the nasals themselves extend forwards and constitute the greater part, if not the whole, of the problematical bar. The fragment now under consideration is clearly in favour of the latter interpretation. Its lower thickened end is a massive bone, not bilaterally symmetrical, and not showing any trace of a median suture. Its inferior face is irregular and roughened, and can scarcely be regarded as an articular facette. Its upper portion consists of a pair of bones separated by a very well-marked median longitudinal suture. These are not thickened at their contracted upper end, where they have evidently been broken, and are not quite bilaterally symmetrical. They doubtless fuse at their lower end with the problematical azygous bone already mentioned, but the arrangement is obscured by the enveloping soft parts. A pair of bones, which may be regarded as nasals, thus extend forwards in a narrow arch to a point just above the anterior end of the premaxillæ; while the massive bone effecting a union between the two normal pairs of elements is probably an ossification in the internasal septum. It is interesting to note that there is an incipient trace of a similar forward production of the nasals in the genus Scelidotherium; while there is sometimes an ossification of the internasal septum in Megatherium.[44]

The three specimens now described, when placed approximately in their natural positions, afford a very satisfactory idea of the form and proportions of the skull when complete. The malar bone is the only important part to be added; but unfortunately it is impossible to decide which of the three specimens of this element in the collection belongs to the individual now under consideration. As already mentioned, these three bones are all different in the shape and proportions of the hinder bifurcated end. They are all very fresh in appearance, but have been stained reddish-brown by the earth in which they must have been buried.

The hinder portion of the second skull already mentioned (No. 2) comprises the occiput and brain-case as far forward as the front of the cerebral hemispheres. It is much battered and broken, and in quite as fresh a state as the cranium already described, with a considerable investment of dried soft parts on its base. It is only very slightly smaller than No. 1, but is of interest as exhibiting some of the sutures, besides a roundness and smoothness indicative of immaturity. The supraoccipital is shown to be very large; a small median point of it enters the foramen magnum, while the suture separating it from the parietals and squamosals extends along the rounded lambdoidal ridge. The horizontally extended suture between the squamosal and parietal on the inner wall of the temporal fossa is seen in the position where Owen determined it to occur in Mylodon.[45] Both tympanics are preserved, but they are more obscured by soft parts than in No. 1.

To this cranium probably belongs a detached portion of the left side of the facial region (No. 5), in a similar state of preservation and slightly smaller than the maxilla (No. 4). The suture between the frontal and the maxilla still persists, while the oral border is preserved farther forward than in the last-mentioned specimen, showing a fragment of the much-reduced premaxilla united with the maxilla by a jagged suture.

The third imperfect occiput is about as large as the immature specimen No. 2, but does not exhibit any features worthy of special note.

The largest and most important portions of the mandible are Nos. 9 and 11, which evidently belong to the right and left rami of one and the same jaw. They are much broken and are in the same fresh condition as the skulls, with traces of the periosteum and even considerable portions of the soft parts of the gum. The right ramus is preserved sufficiently far forwards to show that there was no caniniform tooth in front of the series of four ordinary molars. Judging by the extent of the latter series, the specimen probably belongs to the same individual as the skull No. 1.

Another portion of a mandibular ramus (No. 10) of the left side is slightly smaller than the last and may well have belonged to the immature individual No. 2. It is similarly quite fresh in appearance, and bears the shrivelled remains of the gum. It is interesting as exhibiting the two posterior molars slightly different in shape from those of the former mandible. In this specimen the longer axis of the third molar is oblique, whereas in No. 9 it is coincident with the axis of the mandible; while in the former the fourth molar is not so long in proportion to its width as in the latter. Such slight differences, however, cannot be regarded in the Edentata as more than individual variations.

Brain-cavity and Cerebral Nerves.

By the kind permission of Dr. Moreno, the cranium No. 1 has been vertically bisected to display the character of the cranial cavity and the nerve-foramina. An instructive plaster-cast of the cavity has thus been made by Mr. C. Barlow, the Formatore of the British Museum.

The olfactory lobes are shown to have been well developed, projecting a little in front of the cerebral hemispheres. These hemispheres are together somewhat longer than broad, slightly broader behind than in front, and a little constricted in the middle. They do not overlap the cerebellum, which is relatively large. The origins of the nerves are very imperfectly shown in the cast; only their exits from the cranial cavity are clear. The most interesting are the optic and trigeminal nerves, which pass out of the cranial cavity at first by a common exit, which is soon subdivided by a bony partition into two canals, the former no less than 0.08 m., the latter 0.045 m. in length. The fourth, seventh, eighth and twelfth nerves are also recognisable on the cast; and one prominence of plaster has filled the foramen lacerum posterius.

Compared with the brains of Mylodon and Scelidotherium, so far as known from casts of the cranial cavity,[46] that of Grypotherium is observed to be more elongated, with less divergent and prominent olfactory lobes, less constricted cerebral hemispheres, and a larger cerebellum. In the form and proportions of the cerebrum and cerebellum, it similarly differs from Megatherium.[47] The cerebral hemispheres of the existing Cholœpus didactylus and Bradypus tridactylus[48] are more tapering forward, and their cerebellum is relatively smaller than in Grypotherium.

Auditory Ossicles.

The auditory ossicles were preserved in the tympanic cavities of both skulls, Nos. 1 and 2, being retained by the dried soft parts. They were detected by Prof. Charles Stewart, who kindly extracted them, with great skill, from both sides of each skull. Comparing these ossicles with the fine collection in the Royal College of Surgeons, they prove to be closely similar to those of all the existing Sloths, but most nearly resembling those of Cholœpus didactylus. The malleus is bent exactly as in the latter species, and is of similar shape. As observed by Prof. Stewart, it is remarkable in articulating with the incus not only by the head, but also by a diminutive lower facet, which is in contact with a small facetted process on the anterior arm of the incus. A feeble indication of the same secondary articulation is also observable in Cholœpus; but it is curiously absent in the second specimen of Grypotherium. The two divergent arms of the incus are equal in length, as usual in the Sloths. The stapes is only very slightly perforated in both specimens; while a small circular disc firmly fixed to the incus represents the orbicular bone in the second skull. The auditory ossicles of Grypotherium, therefore, are very different from those of Myrmecophaga, in which the malleus is less sharply bent, the incus has divergent arms of unequal length, and the stapes exhibits a large perforation.[49]

Vertebræ and Limb-bones.

Nearly all the remains of vertebræ and limb-bones are in the same state of preservation as the portions of skull and mandible already described, with adherent cartilage and traces of muscles and ligaments. With some of the ungual phalanges there are also well-preserved examples of the epidermal sheath. As already remarked by Roth, the edges of one sheath probably belonging to the fourth digit of the manus, are quite sharp, and indicate that if the animal walked on its fore feet it resembled Myrmecophaga in the peculiar twist of the manus.

All the specimens in this series seem to have been accurately determined and sufficiently described by Roth. It is only necessary to emphasise the fact that the two shafts of humerus with abraded, not sharply broken, ends have a much more fossilised appearance than any other specimen in the collection, and are deeply stained throughout by ferruginous matter. The small shaft, No. 22, certainly seems to have belonged to an adult animal, as remarked by Roth, and it was probably much smaller than any individual indicated by the other remains.

Skin and Hair.

The new piece of skin, which is stated by Hauthal to have been found in the deposit of excrement, is not quite so well preserved as the original piece. It is much folded in an irregular manner; and the hair, which is yellower than in the previous specimen, is preserved only in patches on the outer face. It must have been stripped from the body of the animal by man; but the only distinct marks of tools, which were evidently made when the skin was fresh, are a few indents and small pits on the outer face. The indents must have been made by oblique thrusts of a stick, or a small, blunt, chisel-shaped instrument. The small pittings are nearer the middle of the specimen and less conspicuous. A vacuity in the skin seems to be due to accidental tearing or to a thrust after it was dry: it may even have been caused by the fallen blocks of stone found lying upon it.

The specimen, as preserved, measures about a metre across in one direction by 93 centimetres in another direction. As already observed by Roth, its irregular folding makes the determination of its position on the trunk very difficult; but I am convinced that its state of preservation is not sufficiently good to justify an attempt to unfold the skin by the ordinary method of steaming. Taking all facts into consideration, Roth seems to be correct in ascribing it to the right flank and the postero-superior part of one of the limbs. It most probably belongs to the fore limb, as Roth supposes; but there is no clear proof that it is not referable to the hind-quarters.

The original situation of the piece of skin being thus determined, it is interesting to observe the disposition of the ossicles in the lower layer. Owing to abrasion, contraction, and partial disintegration, they are conspicuous in most parts of the specimen. They are very irregular in shape and size, and closely compacted together, as in the previous specimen. It is, however, to be noted that in some parts there is a distinct tendency to arrangement in regular, straight, parallel rows. The long axes of the elongated ossicles are nearly always coincident with the direction of these rows. They are especially well shown on the middle of the flank; and, as might be expected, the rows are here disposed vertically, parallel with the ribs.

In some parts of the skin the ossicles are exposed on their outer face; but appearances render it almost certain that this exposure is due to the disintegration and abrasion of the specimen. In one patch thus uncovered by the removal of the soft parts, the ossicles are seen to form a closely arranged, flattened pavement; and their outer face is much more conspicuously marked by pittings than that of any ossicle extracted from the first discovered piece of skin. In fact, as Roth remarks, the pitting is here quite similar to that observable on many ossicles dug up in association with the fossil skeletons of Mylodon; though it does not form so regular a reticulate pattern as that of the dermal ossicles of Mylodon in the British Museum figured on the former occasion.[50]

Another interesting feature of the new piece of skin consists in the dwindling and even total absence of the ossicles towards the ventral border. A section along the edge exhibits only two diminutive nodules of bone in a length of 0.1 m.; while another similar section taken vertically from the skin of the limb shows no trace of ossicles, except perhaps two little specks. It must, however, be noted that the limb was not entirely destitute of armour; for on the border the bones are as well developed and conspicuous as on the middle of the flank. In the newly-cut sections the skin has a translucent aspect, showing that it is merely dried and not tanned in any way.

The hair on the new specimen varies in length from 0.07 m. or 0.10 m. to 0.15 m. or 0.22 m. It is thus longer than that of the previous piece of skin. Masses of still longer hairs—some 0.30 m. in length—were found detached among the excrement, and these are also believed by Roth to belong to the same animal. His determination is probably correct; for, when examined microscopically, these long hairs are observed to have a perfectly smooth cuticle, while some transverse sections (kindly made by Mr. R. H. Burne) demonstrate the complete absence of a medulla, exactly as in the short hairs. The latter feature proves that they cannot be referred either to the horse or to the guanaco.

Excrement.

The large cylindrical pieces of excrement, which may be referred to Grypotherium without any hesitation, have already been described and figured by Dr. Roth. They consist of irregular discoids of herbaceous matter closely pressed together, the largest measuring no less than 0.18 m. in diameter. Mr. Spencer Moore has kindly examined them from the botanist's point of view and reports that they are composed "in large part apparently of grasses, as the haulms, leaf-sheaths, fragments of leaves, &c., of these plants are frequent in the mass. A spikelet, almost entire, of what seems to be a species of Poa, and the flowering glume of another grass, probably avenaceous, have also been found. Besides these there are at least two dicotyledonous plants, one herbaceous and the other almost certainly so, the latter having a slender greatly sclerotised stem. Unfortunately, as no leaves have hitherto been observed attached to the fragments of stem, their affinities are altogether doubtful. There are numerous silicious particles in the excrement, and there are many pieces of the underground parts of the plants, suggesting that they have been pulled out of the ground. A few pieces of stems are sharply cut, not bruised or torn at the end." The latter fact is especially important in connection with Dr. Hauthal's discovery of cut hay in the cavern, and his theory that the Grypotherium was kept in captivity and fed by man.

Generic and Specific Determination.

The fortunate discovery of all parts of the skull and dentition renders the generic determination of this Ground-Sloth now quite certain. The teeth show that it belongs to the family Mylodontidæ; the presence of only four instead of five upper molars separates it from the genera Mylodon, Lestodon, and Scelidotherium; the forward production of the nasals and the ossification of part of the internarial septum place it definitely in the allied genus Grypotherium, as originally diagnosed by Reinhardt. The only question needing consideration is, whether the fragment of cranium described by Owen in 1840 as the type of the genus Glossotherium[51] is really identical with that subsequently described by Reinhardt under the name of Grypotherium darwini, as now seems to be commonly believed.

Darwin's original specimen, on which the genus Glossotherium of Owen was founded, is preserved in the Museum of the Royal College of Surgeons. It has thus been possible to compare it directly with the undoubted cranium of Grypotherium from the Patagonian cavern. The specimen is merely the left half of the hinder part of the cranium, and is therefore very inadequate for discussion; but several features seem worthy of note. Compared with the new skull No. 1, the fragment named Glossotherium has (i.) the inner wall of the temporal fossa less flattened, (ii.) the digastric fossa deeper in proportion to its width, (iii.) the hinder border of the inflated pterygoid vertical, instead of sloping downwards and forwards, (iv.) a much larger and deeper pit for the articulation of the stylohyal, and (v.) a longer canal penetrating the base of the occipital condyle for the passage of the hypoglossal nerve. In all these respects the so-called Glossotherium agrees much more closely with the typical Mylodon; and Owen was probably correct in 1842 when he expressed the opinion that the two are at least generically identical.[52]

I am therefore of opinion that Grypotherium is the correct generic name for the Ground-Sloth from the Patagonian cavern, while Glossotherium must be relegated to the synonymy of Mylodon.

The specific determination of the new specimens is more difficult. As remarked by Roth, only two species of Grypotherium seem to be already known from the Pampa formation—G. darwini by three skulls[53] and G. bonaerense solely by a maxilla.[54] The portions of skull and dentition now under discussion indicate an animal much larger than G. bonaerense (assuming the original maxilla to be that of an adult); while they are considerably smaller than any known specimen of G. darwini. Moreover, the nasal arcade now described is narrower and more concave on its outer face than that of G. darwini, as already observed by Roth. It thus seems very probable that the animal from the Patagonian cavern represents a distinct species, which must bear the name of G. listai. This specific name was given by Ameghino to a fragment of the first-discovered piece of skin, and the curious argument which leads Roth to propose the substitution of a new name for it does not affect its validity.

It may be added that Dr. Erland Nordenskjöld has recently compared his specimens from the Patagonian cavern with the skull of Grypotherium darwini at Copenhagen, and finds no specific difference.[55] No particulars however, have yet been published.

2. Associated Mammalian Remains.

Felis, sp.

A feline carnivore larger than the existing Jaguar (Felis onca), but about the same size as an average Tiger (F. tigris), is represented in the collection by the distal half of a right humerus (No. 44), a left fourth metatarsal (No. 46), and the distal end of another metatarsal (No. 47). These bones have evidently been buried in dust, but are in the same fresh state of preservation as those of Grypotherium.

Careful comparison of these bones shows that they are undoubtedly feline; and there is no difficulty in determining that they belong to Felis rather than to the extinct Machærodus. A humerus of M. neogæus, from a Brazilian cavern, now in the British Museum (No. 18972 b), is readily distinguished from the new Patagonian humerus by the remarkable lateral compression of its shaft and the much greater downward extension of its prominent and sharp deltoid ridge. The humerus in all the large species of Felis, on the other hand, only differs from the fossil now under discussion in very small particulars. In fact, the humerus and metatarsals of the existing Felis onca are essentially identical with the bones from the Patagonian cavern, except that they are rather smaller. I am therefore inclined to regard the newly discovered remains as indicating a comparatively large variety of F. onca, which once lived in the temperate regions of Patagonia, beyond the present range of this species. Such an occurrence would be a precise parallel to that of the Cave-Lion in Europe. It is well known that nearly all the remains of F. leo found in the Pleistocene formations of the temperate parts of the Old World indicate animals of somewhat larger size than any surviving in the warmer regions to which the species is now confined.[56]

It may be noted that bones of the Jaguar of ordinary dimensions have been recorded from the Pampa formation of the Province of Buenos Aires.[57]

Arctotherium, sp.

With the bones of Felis just noticed, Roth provisionally associates the imperfect distal end of a remarkably large right femur. He is thus induced to suppose that the carnivore represented by the fragments will prove to be a new genus and species of the Felidæ. He suggests for it the name of Iemisch listai, on the assumption that it is the mysterious quadruped which Ameghino states is known to the natives as the Iemisch.

A comparison of the distal end of the femur in question with the femora of Felidæ in the British Museum seems to prove conclusively that it cannot be referred even to the same family. Its width across the condyles is much greater, compared with its antero-posterior diameter, than that observed in any feline. Moreover, the pit for the tendon of the popliteus muscle below the external condyle is unusually deep. In both these respects the bone closely resembles the distal end of the femur of a Bear. I have been therefore led to compare it with the corresponding part of the extinct Bear of the Pampean formation, Arctotherium.

Fortunately, the fine and nearly complete skeleton of Arctotherium bonaerense in the Bravard Collection in the British Museum comprises the right femur and enables direct comparison to be made. The fragment lacks the inner condyle; but enough of the trochlea remains to show its broad and gently-rounded form, with a wide and deep intertrochlear notch, precisely as in Arctotherium. It has the same development of the external condyle as in the latter, while the fossa for the popliteal tendon is equally deep, only slightly differing in shape. In fact, there is very little discrepancy, except in its smaller size; and species of Arctotherium smaller than A. bonaerense are already known both from the Pampa formation of Argentina[58] and the caverns of Brazil.[59]

The fragment just described has evidently been severed from the rest of the bone by a sharp, clean cut made by man; and Dr. Hauthal is quite certain that this was not done by one of his workmen during excavation (op. cit. p. 59). At least one medium-sized species of Arctotherium must therefore have survived until the human period in Southern Patagonia.[60]

Onohippidium saldiasi.

A horse is represented in the collection by an upper molar, a fragment of premaxilla with two incisors, an imperfect atlas and two well-preserved hoofs apparently of a fœtus or perhaps of a newly-born animal. Of these remains only the upper molar is capable of satisfactory determination.

This tooth is the second upper molar of the left side, and has been exhaustively compared with corresponding teeth by Dr. Roth, who gives a good series of figures. It is readily distinguished from the homologous molar in the genus Equus by the peculiar form of its two inner columns—a fact which I have been able to verify by the examination of an extensive series of specimens, both recent and fossil, in the British Museum. Further comparison, indeed, shows that it must be referred to the extinct Pampean genus Onohippidium.[61] Roth assigns it, apparently quite rightly, to the same species as a maxilla from the Pampean formation of the Province of Buenos Aires, for which he proposes the name of Onohippidium saldiasi.

Large Extinct Rodent.

The proximal end of the femur of a large rodent has already been recognised by Roth, and compared with the extinct Megamys. It cannot be more exactly determined.

Existing Species.

One imperfect fragment of pelvis and sacrum seems to belong to a puma (Felis concolor) of rather large size; but it is not sufficient for precise determination.

The small mandibular ramus of a musteline referred by Dr. Roth to Mephitis suffocans, does not pertain to this genus and species. Mr. Oldfield Thomas determines it to belong to the rare Lyncodon patagonicus, which still lives in Patagonia and has not hitherto been found fossil. A slightly larger extinct species of the same genus has been described by Ameghino on the evidence of a skull from the Pampean formation near Lujan, in the Province of Buenos Aires.[62]

A cranium, some vertebræ, and a tibia and fibula appear to represent the existing Ctenomys magellanicus, as noted by Roth.

The remains of the Guanaco (Lama huanacos) do not present any features worthy of special remark.

Man is represented by a diseased scapula and by two bone awls, which are clearly made from the tibia of a species of Canis intermediate in size between C. jubatus and C. magellanicus.

3. Relative Age of the Remains.

As the result of Dr. Roth's researches, supplemented by the additional observations now recorded, it is evident that the majority of the mammalian remains from the cavern near Last Hope Inlet belong to the extinct fauna which occurs in the Pampean formation of more northern regions. To this category are referable the genera Grypotherium, Onohippidium, Megamys, and Arctotherium; also Macrauchenia, which is said to have been discovered in the same deposit on the floor of the cave by Dr. E. Nordenskjöld. The large Felis likewise probably belongs to the same series. Remains of mammals of the existing fauna, on the other hand, are comparatively few and insignificant, referable to the genera Ctenomys, Cervus, Lama, Lyncodon, and Felis.

Although Dr. Hauthal's explorations were rather hurried and Dr. Nordenskjöld's results have only been published hitherto in abstract,[63] their account of the deposits on the floor of the cavern seem to confirm the suspicion that the remains of these two faunas were introduced at two successive periods. According to Hauthal, the remains of the Guanaco were found along with fragmentary bones of Deer, shells of Mytilus chorus, branches of trees, and dried leaves, in the superficial dust of the cavern near the outer wall. The skin of Grypotherium and all the other remains of this and the associated Pampean genera were discovered in the deeper layer of excrement and cut hay between the mound and the inner wall of the cavern. According to Nordenskjöld, three distinct strata can be recognised on the floor of the cavern as follows:

A. A thin surface layer, containing ashes, shells, and bones of recent animals broken by man.

B. A middle layer, containing numerous branches of trees and dried leaves, with remains of Lama and the extinct horse, Onohippidium. Said to be probably the stratum in which the original piece of skin was found.

C. A bottom layer, usually about a metre in thickness, without any traces of branches or leaves, but only dried herbs. Remains of Grypotherium numerous and confined to this stratum, associated with its excrement and hair, also with remains of a large variety of Felis onca, Macrauchenia, and Onohippidium.

It is unfortunate that the question of the contemporaneity of the various bones cannot be tested by the ingenious method of chemical analysis which has been applied with success to similar problems by M. Adolphe Carnot in France. The French chemist has shown that when bones are buried in ordinary sediments they undergo changes which gradually cause the percentage of contained fluorine to increase. According to him, the longer a bone has been buried, the greater is the percentage of fluorine found in it on analysis. In one case[64] he examined the scapula of a deer and a human tibia, discovered together in fluviatile sand near Billancourt (Seine); he found that the former had seven or eight times its usual percentage of fluorine, while the human bone did not differ in any respect from the normal in this constituent. He therefore concluded that the latter bone was not of the same age as the former, but had been introduced comparatively recently by burial. In this and the other recorded cases, however, it is to be observed that the sediment was of a uniform character and admitted of free percolation of water. In the Patagonian cavern, on the contrary, the bones occur partly in dust, partly in dried herbage, partly in dried excrement, and partly in the burnt residue of the same. Moreover, they must always have been subjected to intense dryness, and the usual process of chemical alteration cannot have taken place.

Considering all circumstances, I think that, even without chemical evidence, zoologists and geologists cannot fail now to agree with Dr. Moreno and his colleagues of the La Plata Museum, that the remarkably preserved Grypotherium from the Patagonian cavern belongs to the extinct Pampean fauna of South America, and need not be searched for in the unexplored wilds of that continent. If we accept the confirmatory evidence afforded by Mr. Spencer Moore, we can also hardly refuse to believe that this great Ground-Sloth was actually kept and fed by an early race of man.