4] THE EXTINCT RACES OF THE MASCARENES.
In the introductory part of my paper in Phil. Trans. 1875, p. 253, it has been stated that the extinct races of the Mascarenes have a flat cranium, truncated beak, a broad bridge between the obturator foramina, and (it might have been added) the fourth cervical vertebra biconvex. ‘They were thus sharply and structurally differentiated from the Tortoises of the Aldabra group; but there was no common character apparent by which they could be distinguished from the Tortoises of the Galapagos. Such a character can now be pointed out since we obtained more or less complete carapaces from Mauritius and Rodrignez: it consists in the apparently insignificant absence of the suture which divides, in most Land-Tortoises, the gular plate of the sternum into two longitudinal halves. I have besides been able to ascertain that the nuchal plate is invariably absent in all Mascarene Tortoises. ‘Thus, then, what for many years past and at the beginning of these researches seemed to be an almost hopeless task, and what, without the recent explorations instituted in Mauritius and Rodriguez, would have remained an insoluble problem, has found a most simple solution, and there will be in future no difficulty in determining the origin of those carapaces which for generations have been kept in museums, and whose history was veiled in obscurity or forgotten long ago. We may state :— 1. That the specimens with a nuchal plate (and with double gular) came from Aldabra; 2. That the specimens with single gular (and without nuchal) came from the Mascarenes;and 3. That the specimens without nuchal and with double gular are Galapagos Tortoises. With the aid of this key we can now not only refer various species based by older authors on more or less perfect specimens to their proper geographical division, but also identify them with such as have been distinguished from the osseous remains only. A. Tue Gieantic LAnD-TortToIses OF MAURITIUS.
The materials in the British Museum on which the following remarks on the Tortoises of Mauritius are based are the following :— 1. A series of limb-bones and portions of the cranium from the ‘“ Mare aux Songes,” G
42 THE RACES OF THE MASCARENES. chiefly due to transmission by Mr. Grorcr CuarK, C.M.Z.S. *, and obtained by the Trustees of the British Museum in 1865 and following years. 2. The carapace of a Tortoise found at Grand Port, a few years ago, in the same place where the bones of the Dodo were found (‘Mare aux Songes”), sent by our esteemed correspondent, L. Bouton, Esq. [This appears to be the carapace mentioned in Mr. Clark’s statement; but no plastron was received with it then or afterwards. | 3. Limb-bones and parts of the pelvis from the district of Flacq, sent by the same gentleman. 4, A perfect pelvis of a very large individual, obtained through C. E. Brwsuer, Esq., of Port Louis. 5. Skulls, sternum, and limb-bones from the ‘‘ Mare aux Songes,” presented by the Hon.KE.Newton. Besides, I have had the advantage of consulting the magnificent collection made by the Hon. E. Newton, who most kindly submitted it to my examination. It is now in the Museum of the University of Cambridge, and consists of several fragmentary carapaces, perfect sterna, and a great number of skulls, scapularies, pelves, humeri, and femora. Surpassing in the number of bones the collection in the British Museum, it does not offer the same amount of evidence as regards the variety of races in Mauritius, the objects having been collected at one locality only, the ‘“‘ Mare aux Songes,” whilst the remains in the British Museum were obtained at various periods from at least two distant districts of the island. There can now be no doubt that Tortoises from the Mauritius have been enumerated in the system under distinct names for many years, although zoologists were ignorant of their origin. Indeed the very first gigantic Land-Tortoise described proves to be a Mauritian species. I allude to the Tortoise described by PERRAULT in the year 1676 (Mém. pour servir a l Hist. Nat. p. 193, c. tab.) under the name of ‘“‘ La Tortue des Indes,” and stated by him to have come from the coast of Coromandel. That no gigantic Land-Tortoise ever came from that coast is certain; but a ship returning thence to Europe would most probably touch at Mauritius. ‘This Tortoise, then, was named by succeeding systematists Zestudo indica, and finally, by Duméril and Bibron, * The circumstances under which these bones were found will be readily understood from the following abstractofMr.Clark’s“Statement”(Trans.Zool.Soc.vi.p.51):—“Ontheestatecalled‘Plaisance,’about three miles from Mahébourg, in the island of Mauritius, there is a ravine of no great depth or steepness, which apparently once conveyed to the sea the drainage of a considerable extent of circumjacent land, but which has been stopped to seaward, most likely for ages, by an accumulation of sand extending all along the shore. The outlet from this ravine haying been thus impeded, a sort of bog has been formed, called ‘La Mare aux Songes,’ inwhichisadepositofalluvium,varyingindepth,onaccountoftheinequalitiesofthebottom,whichis formedoflargemassesofbasalt,from3to10or12feet.Theproprietoroftheestate,afewweeksago,conceived the idea of employing this alluvium as manure; and shortly after, the men began digging in it.
When theyhadgottoadepthof3or4feettheyfoundnumerousbonesoflargeTortoises,amongwhichwerea carapaceandaplastronprettynearlyentire,asalsoseveralcrania....Thesewerefoundimbeddedina black vegetable mould, the lighter-coloured specimens being near the springs.”
THE GIGANTIC LAND-TORTOISES OF MAURITIUS. 43 Testudo perraultit. The carapace of the specimen from which Perrault took his original description is still in the Paris Museum; it is that of a male, about 32 inches long. Unfortunately the sternum is lost, so that the French herpetologists could not be aware that this species belongs to that division of the genus Testudo which they distinguished by the presence of eleven sternal scutes only. However, quite apart from the consideration of the sternal characters, the detailed description in the ‘ Erpétologie générale’ offers sufficient evidence of the close affinity of this 7. indica or perraultii with our Mauritian species: “la suscaudale simple, ¢rés-élargie; la dernicre de la rangée vertébrale bombée,” are two of the most striking features of the carapaces from Mauritius.
A second carapace, with epidermoid scutes, 16 inches long, preserved in the Paris Museum for more than a century, and without known history, was described by Scuweiccer as Testudo tabulata, var. africana, in the year 1812*, and recognized by Duméril and Bibron as a distinct species, Testudo grayi. This is likewise a Mauritian Tortoise; and as it has the same undulated vertebral profile as the carapace subsequently described in this work as possibly belonging to 7’. triserrata, these two individuals may eventually have to be referred to the same species. For the present I must be satisfied with having thus indicated the natural affinity of Testudo indica and T. grayi with the Mauritian species described from osseous remains ; but a much more complete series of carapaces (and, indeed, of carapaces with their skulls and other parts of the skeleton) will be required before we can venture to decide whether or not those two species are identical with the species distinguished in this work. A carapace with so straight a vertebral profile as that delineated and described of 7. perraultii is not represented amorg the specimens collected by Messrs. BoutonandNewton.
The Mauritian Tortoises were inferior in size to those of Aldabra and Rodriguez, the majority of the remains belonging to individuals with carapaces from 2 to 3 feet long ; specimens above this size must have been exceedingly rare. In making the collections mentioned above, no care was taken to keep separate such of the bones as were found in juxtaposition, and which, therefore, might reasonably have been presumed to belong to the same individual. Hence it would be most difficult (nay, impossible) to avoid errors, if we were to attempt to specifically group the modifications which can be observed in almost all the different kinds of bones; and it will be the safer plan to adopt in their description an analytical method—that is, to take up the various kinds of bones in their natural succession and to describe the modifications of each. Perhaps, in a few instances, it will be possible to indicate the specific unity of the types thus distinguished 7.
*Konigsb.Arch.Naturwiss.i.p.322.
+Inapreliminarynoticepublishedin‘Nature,1875,p.297,beforeIhadaccesstoMr.Newton’scollection,Ihadnotfullyrealizedthedifficultyofreferringeachmodificationofbonetoitsproperspecies;and, more especially, I fell into the error of referring to 7’, bowtonti a humerus which proves to belong to 7’.
inepta.
G2
44 THE RACES OF THE MASCARENES.
Skull.—There are altogether thirteen skulls in a more or less fragmentary condition, and five mandibles, in the collections of the British and Cambridge Museums. They can be referred to two species—the first having three serrated dental ridges along the lower jaw (Zestudo triserrata), the second having only two ridges (Testudo inepta). a. Testudo triserrata (Plate XXIII. fig. A).—1. Frontal region perflectly flat, broad, its greatest width being a little more than one half of the distance of the tympanic condyles. 2. Only the anterior half of the parietal forms a flat surface, the posterior being compressed into a trenchant crest passing into the very long occipital spine. 3. The tympanic case with the mastoid is produced backwards, so that the paroccipital margin appears as a deep semicircular excision. 4. The impression on the lower surface of the occipital in front of the condyle is very shallow. 5. On the front margin of the temporal fossa, corresponding to the suture between parietal and tympanic, immediately in front of the foramen carotidis externe, there is a very large and projecting flat condyle-like tuberosity for the insertion of a portion of the temporal muscle. 6. Tympanic cavity large, with a néarly circular outer tympanic rim, which is posteriorly interrupted by the notch for the Eustachian tube. 7. This notch is not very remote from the columellar foramen, which is on an elevated ridge crossing transversely the tympanic cavity. 9. The front margin of the intermaxillaries projects but slightly beyond that of the frontals; so that the plane of the nasal opening is nearly perpendicular and as broad as high. 10. The inner nostrils are advanced, not very distant from the end of the snout, and, on the palatal view of the skull, hidden for the greater part below the alveolar lamella of the maxillaries. 11. The intermaxillaries are short, about one third the length of the maxillaries; their foremost portion is deeply hollowed out below, and vertically bent downwards to form the beak. The suture between the intermaxillary and vomer is immediately behind the inner angle of the alveolar edges of the maxillaries. The alveolar maxillary surface is provided with three denticulated ridges, the inner of which runs parallel to, but at a distance from, the inner alveolar edge. 12. The palatal region is much less concave than in the Aldabra Tortoises, and without median longitudinal crest. 13. Anterior surface of the tympanic pedicle deeply excavated. 14. Lower jaw with a triple serrated alveolar ridge, the two inner ridges being rather close together and belonging to the same raised osseous tract.
b. Lestudo inepta (Plate XXIII. fig. B).—The skull of this species is so similar to that of Testudo triserrata that only the following important points in which it differs need be mentioned. 11. The alveolar surface of the maxillary shows three denticulated ridges, one being confluent with the outer, and one with the inner alveolar margin. 14. Lower jaw with two serrated ridges only.
To judge from the skulls, both these species must have had smaller heads than the Galapagos Tortoises. The largest of our series of skulls belong to Testudo triserrata, one of them being 50 lines long (to the end of the occipital condyle) and 88 lines broad
THE GIGANTIC LAND-TORTOISES OF MAURITIUS. 45 (between the tympanic condyles) ; the largest skull of Testudo inepta measures 46 lines in length and 34 lines in width.
Carapace.—Among the carapaces two very distinct forms may be recognized, which, however, have this in common—that the centre of the last vertebral scute is raised in a very convex hump, which is separated from the preceding vertebral by a deep transverse depression. The shell of both forms is thin and fragile (though less so than in the Rodriguez Tortoise); but towards the margins it is much thickened, especially in the second form, and four or five times as thick as on the convex portions. The skulls belonging to the carapaces not having been found, or having been mixed with others, it remains uncertain to which of the species distinguished from the skulls the carapaces shouldbereferred.| a. Carapaces possibly belonging to Testudo triserrata (Plate XX. fig. D).—The greater portion of the upper part of a carapace (32 in. long and about 20 in. broad) and the hinder third and the foremost part (including first vertebral and marginals) and five sterna, of adult males, are in Mr. Newton’s collection. This carapace is more depressed than that of any other gigantic Tortoise. The upper profile is much undulated, the second, third, fourth, and fifth vertebrals being raised in the middle, and the humps separated by deep transverse impressions. The first vertebral is flat, its median profile gently sloping downwards towards the front margin. The decline of the upper profile from the hump of the last vertebral to the hind margin of the caudal is steep and straight. Also the costal plates participate in a slight degree in the peculiar formation ofthevertebrals.
Most of the marginal plates are lost: the first barely touches the antero-inferior angle of the first costal; the last two are as high as broad; and the caudal is much broaderthanhigh.
The margins of the shell are not visibly indented; but indentation may have disappearedwiththeepidermoidscutes.
None of the sterna collected by Mr. Newton can be fitted to the individual carapaces or fragments of carapaces. They are those of adult males, deeply concave, and callous on the sides. ‘Two of the largest are 18 and 193 in. long, and respectively 143 and 15 in. broad. The lobes (Plate XX. fig. E) are short, especially the posterior, which is about twice as broad as long and broadly truncated behind. ‘Lhe single gular plate is triangular, with rounded front margin-and acute posterior angle, wedged in between the postgulars. Pectorals extremely narrow, almost linear. Abdominals very large, their length being two fifths of that of the entire sternum.
b. Carapaces possibly belonging to Testudo inepta.—A nearly entire specimen (Plate XX. figs. A-C), but without plastron, has been sent by M. Bouton; it was discovered in 1865, in the ‘Mare aux Songes.” A large fragment, representing the posterior third of the carapace, and perfectly agreeing with the other specimen, is in Mr. Newton’s collection. ‘This form of carapace is much higher and narrower than that of J. tri-
46 THE RACES OF THE MASCARENES.
serrata (?). Our specimen has a length of 22 inches and a greatest depth of 83 inches. The thickness of the shell scarcely exceeds 2°5 millims. along the central portions, but gradually increases towards the margins. It is that of a fully adult animal, as all the sutures have disappeared ; nevertheless we cannot draw a conclusion from this single individual as regards the maximum size to which this species attained, because in most forms of Testudinata we find examples of smaller and larger size, independently of age. The general shape of the carapace is very peculiar, and very different from any of the other gigantic Tortoises. The anterior three fifths of the shell are very convex and much bulged out in every direction; the posterior two fifths slope towards the muchexpanded hind margins; but this slope is rendered very uneven by the fourth and fifth vertebral plates being much raised in the middle, and the two elevations being again separated from each other by a depression. The margins of the shell also of this species are not conspicuously indented. ‘The sutures between the first, second, third, and fourth vertebrals are of nearly the same length (3” 11”, 4” 9", and 4” 6’”), whilst the suture between the fourth and fifth vertebrals is only one half that length (2” 3”). The first marginal plate is entirely limited to the anterior vertebral, without extending to the adjoining costal. The undivided caudal gently slopes backwards and downwards, is much expanded, not turned in or reverted behind; its hind margin is strongly curved. It measures 55 in. in its greatest width, and 3 in. 2 lines in length. Vertebral column.—A single example of the fifth vertebra of the cervical series is preserved (Brit. Mus. no. 39930); there is no means of ascertaining whether it belongs to 7. triserrata or T. inepta. As regards form, it is nearly intermediate between those of 7. ponderosa and T. vosmert; there is scarcely a trace of a crest on the dorsal surface. The individual may have been somewhat inferior in bulk to those whose vertebree are used in the subjoined table for comparison.
T.ponderosa.T.vosmeri.Mauritian.
millim. millim. millim.
Length of centrum ...... PN er ee oe 65 75 61 Horizontal width of middle of centrum ............ ye 14°5 14 Width of anterior ¢lenoid cavity, vea7ce enero 28 29 26 Breadth of posterior condyle 2. .ccsss.ssmaeraes 30 20°5 22 Scapulary.—Among the numerous specimens three types may be distinctly made Outs a. The majority (Plate XXIV. fig. B.) have the distal portion of the Fig. 1. acromium, as well as the proximal half of the scapula, distinctly trihedral, the anterior side of the latter being convex, so that a transverse section through that part of the scapula represents the following outline (fig. 1). The acromium is straight, the coracoid lamina sometimes deeply grooved. The whole bone is much more slender than in the Aldabra Tortoises, but less so than in the Rodriguez species. Although these specimens differ
THE GIGANTIC LAND-TORTOISES OF MAURITIUS. AT considerably in size (the scapula proper of the largest individual being 63” long, and that of the smallest 4}’’) it is remarkable that all have the coracoid firmly ankylosed to the scapula, the suture having entirely disappeared. This form of scapula, which may be referred to Testudo triserrata, does not differ from two other scapularies in the British Museum, obtained together with a femur and fragments of pelvis, for which I have proposed the name of Testudo leptocnemis. b. In the second type, of which I have three specimens (Nos. 39957 and 39938 in the British Museum and one in Mr. Newton’s collection) before me, the distal portion of the acromium is compressed and its end is curved; the proximal half of the scapula is trihedral; but the anterior and acromial sides have a deep longitudinal impression, so that a transverse section through this part of the scapula represents a figure like the following (fig. 2). Otherwise the bone ~~ is of the same build as in the first type; but, although these specimens / again differ considerably in size (the scapula proper of the larger being 64" long, and that of the smaller 44”), the coracoid never became ankylosed to the scapula, and is lost. This is a most singular distinction o from 7. triserrata. The species to which this type of scapula belongs ispossiblyTestudoinepta.
c. The third modification, of which I have seen two examples (No. 39936 in the British Museum and one in Mr. Newton’s collection), approaches closely to the preceding in having the same compressed acromium, with the end slightly curved. The coracoid, likewise, was not ankylosed to the scapula; but the scapula itself is pie. 5. strongly compressed, in no part trihedral, and a transverse section would be of the following shape (fig. 3). This bone is the only one on which our | knowledge of a supposed fourth species, Testudo boutonii, rests (* Nature,’ 1875, p.297). But having found a perfectly analogous instance in the scapula ( of the Rodriguez Tortoise, I must regard this modification as a case of Se anomaly only.
Measurements.
T. triserrata, T’. leptocnemis. T. inepta. T. boutonii. N*.No.39939.No.389940.(B.6.) No.39987.No.399388. millim. millim. wmillim. millim. millim. millim. millim. Length of scapula (measured fromthesuture,with GORACOIG)) ease ce elas ds. 162 140 114 150 161 114 128 Circumference in its middle 72 62 50 59 65 39 53 Longitudinal diameter of glenoid cavity ............ 43 36 32 42 Length of coracoid............ 74 61 54. damaged ven Greatest width of coracoid 6] 59 51 Length of acromium ......... 70 54 52 62 75 55 61 * From Mr.
Newton’s Collection.
48 THE RACES OF THE MASCARENES.
Humerus.—The two collections contain together 18 specimens, among which again three types may be distinguished.
a. The first type (Plate XXV. fig. A) is represented by six humeri with a slender and strongly bent shaft, and with a deep hollow behind the head of the humerus and between the tuberosities. The head is globular, or, in large specimens, ovoid, and nearly entirely raised above the level of the summit of the radial tuberosity. A groove indicating the course of vessels becomes distinct on the convex side of the distal third of the bone close to the radial edge, and perforates the lower end of the bone very near to the margin of the cartilage of the joint. The slight build of the humeri of this type would seem to indicate that they belong to the species with the thinner carapace, viz. 7. triserrata.
b. In the second type the hollow behind the head is absent or scarcely indicated. Otherwise it does not differ from the first, and is, perhaps, nothing but an individual variation.
c. The third type (Plate X XV. fig. B), represented by eight specimens, has a much stouter and straighter shaft; it is trihedral nearly in its whole length, a strong ridge generally running down from the tuberosities to the radial margin of the condyle. On account of its greater strength, this form of humerus may be associated with the heavier carapace of Testudo inepta.
Measurements.
T. triserrata. T. inepta.
No.39947.
No.39946.
No.39942.
No.39944.
No.39943.
millim. millim. millim. milim. millim.
Lengthofhumerus(measuredinastraightline from summit of head to the middle of tPOCUICA)) 5.0, iccicaateann« on eee eran e eanenemeperer 137 110 184 119 160 Circumference of the narrowest part of shaft ... 55 47 82 63 88 Wargest diameter ot head) 2. ues) en ae we eeeee 28 24 46 29 39 Shortest diameter of head ......................00805 25 20 30 20 “Oo” Extreme breadth between condyles ............... 45 38 76 43 58 Ulna.—Two examples of ulna are preserved in the British Museum series: the smaller is that of an animal still growing; the larger (No. 39948) belongs to the left side and is 110 millims. long; it has a comparatively narrow shaft, not more than 15 millims. wide in its narrowest portion. The shaft is considerably twisted round its longitudinal axis, so much so that the transverse diameters of its proximal and distal dilatations stand nearly at right angles to each other. The olecranon is well developed. Radius.—¥our specimens are preserved in the British Museum series, differing considerably from one another, but leaving it quite uncertain whether these differences are merely accidental and individual or specific.
a. One radius of larger size (No. 39950) (Plate XVIII. fig. D) and another smaller