UC-NRLF 
 
 TS3 
 .3 
 
 C 2 7^2 Db^ 
 
 PAYILAND CAVE : AN ADRIGNAGIAN 
 STATION IN WALES. 
 
 (THE HUXLEY MEMORIAL LECTURE FOR 1913.) 
 
 BY 
 
 W. J. SOLLAS, M.A., So.D., LL.D., F.KS., 
 
 Professor of (Jeolog// in the University of Oxford. 
 
 [WITU PLATES XXI-XXIV.] 
 
 rUBLlSHED BY THE 
 (Ro^af (^nt^ropofogtcaf ^mtiiixte of cSfrcat 09ritain an^ 3refan^, 
 
 no, TtREAt pjjssell street, London, w.c. 
 
LIBRARY 
 
 UNIVERSITY O^ j 
 CALIFORNIA/ 
 
 ANTHROPOLOGY LIBRAM 
 


 CK/f^AA^ 
 
 PAVILAND CAVE: AN AUEIGNACIAN STATION IN WALES. 
 
 The, Huxley Memorial Lecture for 1913. 
 
 By W. J. SoLLAS, M.A., Sc.D., LL.D., F.R.S., Professor of Geology in the 
 
 University of Oxford. 
 
 As a temporary habitation it would be difficult to find a more excellent cave than 
 Paviland ; situated on the face of the steep limestone cliffs of^ Gower, it looks out 
 over the changeful waters of the Bristol Channel ; behind it is a fertile land which 
 must have provided a rich hunting ground in early times ; it is roomy, well lighted 
 and dry, with a natural chimney to promote ventilation — serving also to carry off the 
 smoke of a fire kindled beneath ; in front of the entrance is a rocky platform with 
 natural seats where the hunter can sun himself in the open air. Add to this that it 
 is concealed from the landward view and difficult of access to those unfamiliar with 
 the way. Evidently in every respect a highly " desirable " hunting lodge ! How 
 its advantages appealed to the palaeolithic man of Glamorgan during the Aurigna- 
 cian age is shown by the great kitchen midden which forms its floor. Here, it is 
 plain, he fabricated his implements and weapons, here he roasted his meat, flesh of 
 the horse, the bison, the mammoth, and the bear, and here on one solemn occasion he 
 entombed his dead. 
 
 Chance visitors have entered it in later times, as Eoman coins attest, and 
 nearly a century ago it began to attract attention by the discovery of the remains 
 of extinct animals in its floor. Mr. L. W. Dillwyn and Miss Talbot commenced 
 excavations and obtained a large number of teeth and bones which were originally 
 deposited in Miss Talbot's collection at Penrice Castle but subsequently transferred 
 to Swansea Museum, where they now remain. News of these doings brought 
 Buckland down from Derbyshire to carry out that admirable exploration which is 
 described with the master's accustomed skill in the classic pages of the Reliquice 
 Diluviance} 
 
 Among the many facts of interest which were brought to light none approaches 
 in importance the " Eed Lady of Paviland," — the remains of a skeleton, coloured 
 red with iron ochre, and supposed by Buckland to have belonged to a woman. 
 
 * Buckland, Reliquice Diluviance, London, 1823, pp. 82 et seq. 
 
 331869 
 
2 ,{i i}J^'''.-if)'' ^i$^i'^^-TiJ*(f'^Mlaiul Cave: An Aurignacian Station in Wales. 
 
 Associated with the skeleton, and evidently, as Buckland remarks, interred along 
 with it, were numerous objects carved out of ivory as well as a heap of little sea- 
 shells, both shells and ivory being stained like the skeleton with red ochre. 
 
 The source of this ivory was admittedly the mammoth tusks of which remains 
 were found in the kitchen midden, and hence it might have been supposed that the 
 men who had fashioned it were contemporary with this extinct elephant. But this 
 is a conclusion which Buckland refused to accept, resting his position on the 
 
 FIG. 1. — ROMAN COINS (NATURAL SIZE). 
 
 The three coins are all included within the same period (a.d. 274-306), those on the 
 right and left belong to the time of Carausius (a.d. 287-293), the one in the 
 middle to Constantine the Great (a.d. 274-306). The Carausius on the left bears 
 the inscription imp caravsivs pf avg, with the head of the emperor radiate (ob- 
 verse) and a figure of Peace bearing an olive branch in the left hand and leaning 
 on a staff on the right with the inscription pax avg (reverse). The Carausius on 
 the right is not so well preserved ; the head of the emperor is radiate and the 
 visible inscription is caravsivs p(e?)avg (obverse) ; the female figure holds a staff 
 in the left hand and a cornucopia in the other. The Constantine shows a 
 helmeted female head and the inscription vrbs roma (obverse), and the wolf 
 suckling Eomulus and Remus, with the letters trs below and two stars above, of 
 which one has almost entirely disappeared (reverse). 
 
 assumption that the ivory was already in a fossil state before it was worked. On 
 a later page he cites instances of fossil ivory found in Scotland and at Bridlington 
 which even at the present day is still " fit for the turner's use."^ There is a British 
 camp on the hill immediately above the cave, and since ivory bodkins are some- 
 times found in British barrows, Buckland was led to believe that the " Ked Lady " 
 was an ancient Briton who inhabited the cave during the Koman occupation 
 of Wales. 
 
 The existence of " antediluvial " man in Europe was a question already ripening 
 for discussion in Buckland's time, and Paviland Cave was rich in material which 
 
 ^ Buckland, Reliquiw Diluviance, London, 1823, p. 179. 
 
W. J. SoLLAS. — Paviland Gave : An Aurignacian Station in Wales. 3 
 
 might have assisted in its solution. Worked flints of many kinds abound in the 
 deposits of the floor, but curiously enough, Buckland makes mention of a single 
 specimen only and this he compares to a strike-a-Hght. If, as is probable, he had 
 before him one of the short square scrapers, of which this cave affords numerous 
 instances, the comparison may be admitted, but it was unfortunate and misleading. 
 We may perhaps attribute to this the fact that Buckland completed his examination 
 of the cave without discovering its most important secret. 
 
 FIG. 2. — PAVILAND CAVE AS SEEN BY BUCKLAND, 1823. 
 
 A, mouth of the cave ; B, inner extremity of the cave, accessible to a dog but not to a man ; 
 C, body of the cave ; D, chimney with ledges and hollows (L) ; E, bottom of the cave, " to 
 which the sea water never reaches " ; F, mass of the same materials as E, but less 
 disturbed and overhanging E in a small cliff, 5 feet high ; G, loose sea pebbles washed up 
 during storms ; H, rock basins, produced by movement of large pebbles which still lie in 
 them ; I, naked limestone of the floor. (After Buckland.) 
 
 When McEnery a few years later (1825)^ while exploring Kent's Hole, a cave 
 much poorer in this kind of evidence, encountered flint implements along with the 
 bones of extinct animals beneath the stalagmitic floor, he at once drew the obvious 
 and, as it seems to us, inevitable conclusion ; but in this he was possibly aided by 
 the fact that some of these flints resembled, not a modern strike-a-light, but an 
 ancient arrow-head. 
 
 a 2 
 
4 W. J. SOLLAS. — Paviland Cave : An Aurignacian Station in Wales. 
 
 McEnery, who was thus the first to definitely establish the existence of 
 palaeolithic cave man, at once communicated his results to Buckland. Strange to say, 
 this acute observer, who was naturally of the most open mind — as is shown by the 
 abandonment of liis preconceived views about the deluge, directly he had become 
 acquainted with the work of the Swiss glaciers — received them with hostility and 
 sought refuge in impossible and almost evasive objections. 
 
 I can only attribute this to the power of preconceived ideas. Buckland, as the 
 result of his long continued and masterly researches into " diluvial" phenomena, 
 was convinced that man had not inhabited Europe in company with the " ante- 
 diluvial " animals, and in this conclusion he was fully supported by the famous 
 Cuvier. No one who is at all aware of the power of a prepossession will blame my 
 illustrious predecessor ; fortunate indeed is the man whose prepossessions happen 
 always to accord with the truth. 
 
 The age of the "Eed Lady" long remained a vexed question. Falconer 
 followed Buckland in attributing the interment to Eomano-British times ; but to 
 do so he had to withdraw the admission, made by Buckland, that the associated 
 ivory objects had been carved from mammoth tusks ; they were, he asserted, 
 imported from France, and that such imports were made during the Eoman 
 occupation he proves by a quotation from Strabo, which runs as follows : " They 
 (the Britons) pay but moderate duties on the imports . . . from Kettica, 
 which are ivory bracelets and necklaces, amber, vessels of glass, and small wares."^ 
 
 When, however, Lartet and Christy, fresh from Perigord, visited the remains 
 from Paviland which Buckland had placed in the University Museum of Oxford, 
 they at once recognized the complete correspondence of all the facts with those they 
 had observed at Cro-Magnon. There also skeletons had been found rouged over 
 with ochre and associated with mammoth tusks, ivory ornaments, and marine 
 shells. In both cases, as they believed, it was the men who were contemporary with 
 the mammoth who had made use of its ivory .~ 
 
 This, however, was not the opinion of Professor W. Boyd Dawkins^ who 
 approached the question a few years later. Laying special stress on the disturbed 
 condition of the cave deposits and on the presence of bones of sheep, he 
 reaffirmed Buckland's conclusion but in a more general form, stating that " the 
 interment is relatively more modern than the accumulation with extinct mammalia." 
 
 So the matter rested for nearly forty years, when'*f'5SB[^ M. Cartailhac being 
 on a visit to Oxford, partly as the recipient of our honorary degree, called at the 
 Museum and expressed a wish to see the bones of the " Eed Lady." He then 
 confided to me his belief that the interment was not only palaeolithic but more 
 precisely that it was of Aurignacian age. This led me to make a search for the ivory 
 
 ' The Geography of Straho, translated by H. C. Hamilton and W. Falconer, 1854, vol. i> 
 p. 298. 
 
 2 Lartet and Christy, Reliquice Aquitanicce, London, 1863, pp. 93, 94. 
 
 3 W. Boyd Dawkins, Cave Hunting, London, 1874, pp. 232-234. 
 
W. J. SoLLAS. — Faviland Cave : An Aurignacian Station in Wales. 5 
 
 implements in our collection, which I had been unable to show to M. Cartailhac,^ 
 and on discovering them I fonnd their agreement with the well-known Aurignacian 
 forms to be so exact that I had no difficulty- in accepting the opinion of my 
 distinguished colleague. M. Cartailhac has since developed his views in detail ; 
 they are contained in his masterly contribution to the famous monograph on the 
 caves of Mentone,'' where also he enters into a full discussion of the general subject 
 of paleeolithic interments, once a burning question among Continental anthropologists. 
 For G. de Mortillet had stoutly denied that palaeolithic man paid any regard to his 
 dead, maintaining, in opposition to M. Riviere, that the numerous interments in 
 the Grotte des Enfants and the other caves of Mentone were all of neolithic age, 
 and, further, that such a thing as a palseolithic interment nowhere exists. This 
 controversy is at length happily terminated ; the successive burials in the caves of 
 Mentone are now known to have taken place in Aurignacian times, for each was 
 found to be completely covered over by an intact layer containing hearths and an 
 Aurignacian industry ; and that interment was practised at an even still earlier 
 period is proved by several instances, of which the most notable is that afforded by 
 La Chapelle-avix-Saints, where a Neandertal skeleton was discovered in a 
 Mousterian tomb. 
 
 In the summer of last year (1912) Professors Boule and Breuil, who had come 
 to England for the purpose of investigating the alleged occurrence of flint 
 implements at the base of the Eed Crag, continued their journey to Oxford and 
 made a careful examination of the remains in the Paviland collection. I'rofessor 
 Boule was much impressed with the general resemblance of the human bones to 
 those of the Cro-Magnon race and Professor Breuil at once recognized the 
 Aurignacian character of the ivory implements. Professor Breuil and I then 
 agreed to visit Paviland together in search of further information. On the way we 
 called at the Swansea Museum and found there an extremely interesting collection 
 of flint and bone implements which had been obtained from Paviland by 
 Colonel Wood, Mr. Vivian, and Colonel Morgan. Among them Professor Breuil 
 identified numerous upper Aurignacian gravers and scrapers as well as some 
 examples of early Solutrian flaking. 
 
 We then proceeded to Paviland, which is situated on the coast about 
 15 miles west of Swansea, but our first attempt to enter the cave was unsuccessful 
 owing to the state of the tide and our ignorance of the approach from the land- 
 ward side. The next day we were guided by Mr. John Gibbs, a farmer at Rhossili, 
 to a path which leads down a steep valley to a recess on the right-hand side 
 (Plate XXI, Fig. 1) ; it is necessary to ascend this, walk round it and descend on 
 the other side, when all that remains is — in Buckland's words — some "dangerous 
 climbing along the face of a nearly perpendicular cliff" (Plate XXI, Fig. 2). 
 
 ^ " Quelqiies ossements humains sont parvenus, sans honneur, au museum d'Oxford. Je 
 n'ai pu savoir le sort des objets," E. Cartailhac, Les Grottes de Grimaldi, 1912, torn, ii, fasc. II, 
 p. 305. For "ossements humains " we should doubtless read "objets en ivoire." 
 
 ^ W. J. Sollas, Ancient Hunters, London, 1911, p. 213. 
 
 •' E. Caitailhac torn, cit., pp. 304 et seq. 
 
6 W. J. SOLLAS. — Paviland Cave : An Aurignacian Station in Wales. 
 
 This proved not to be so difficult as it looks,^ and at the top a scramble to the 
 right brought us face to face with the cave (Plate XXI, Fig. 3). 
 
 FIG. 3. — PLAN AND SECTIONS OF PAVILAND CAVE. 
 
 1. Plan of the floor : b. 6'., limestone boulders ; s., supposed position of the skeleton ; t. t., line 
 of transverse trench ; L. L., limestone not covered with cave earth ; c, projected position 
 of the chimney. 2. Longitudinal section : o., band of ochreous clay. 3, 4, 5. Transverse 
 sections taken at positions indicated by lines bearing corresponding numbers in 2. 
 Horizontal and vertical scale the same. 
 
 We were struck with the change that had taken place, since Buckland's visit, 
 in the aspect of the floor. The cliff, which he describes as cut into the cave 
 deposits, 5 feet high and running longitudinally up the middle of the cave, had 
 disappeared, and a smooth, pebbly surface swept uniformly like a beach with a 
 
 ' Later, during our systematic exploration of the cave, the climb was made several times by 
 a lady, Mrs. Cunnington, under the guidance of her son. The expert climber will find a 
 really dangerous way back from the cave by continuing past it to the west till he reaches a 
 deep gully ; the ascent of the cliff at this point is possible and interesting. I have to thank 
 Mr. Cunnington for showing me the way up. 
 
W. J. SoLLAS. — Pavilaiul Gave : An Aurignacian Station in Wales. 7 
 
 rapidly rising slope from the entrance to the extremity. It was evident that during 
 exceptional storms the sea had free access to the cave. 
 
 We commenced to dig and the Abbe soon obtained several Aurignacian imple- 
 ments, including a Gravette point and some doubtful Mousterian flakes, from no 
 great depth below the surface. From his observations it was clear that the cave 
 still contained much interesting matter and I was urged by Professor Breuil to under- 
 take its exploration with the view of writing a full account of this Aurignacian station 
 — the earliest discovered in Britain. I consented to do so, though reluctantly, for it 
 seemed to me that Professor Breuil, the virtual creator of the Aurignacian system, 
 was the proper person to perform tliis task. I was encouraged, however, by his 
 promise to examine the spoils, a promise which he has fulfilled most generously : 
 all the flint implements we discovered have passed through his hands, and it is by 
 him that they have been nalned and classified. 
 
 Permission to dig having been obtained from Miss Talbot, who owns the 
 cave, I invited and received the co-operation of the Cardiff Museum in the work 
 of exploration. The Museum was ably represented by Mr. Arthur Loveridge, who 
 was present on the spot during the whole time the excavations were in progress. 
 My friends. Dr. Marett and Mr. Henry Balfour, gave invaluable aid as volunteers, 
 so, too, did Mr. Ward, of New College, a university blue who proved a mighty 
 man with the spade. My assistant, Mr, C. J. Bayzand, was our photographer and is 
 responsible for all the illustrations of this lecture. We were extremely fortunate 
 in our workmen, Harry Long and Jack Gibbs, who entered fully into our plans, 
 and watched for specimens with eagle eye ; very little that was of value was 
 allowed to escape their hands. 
 
 Our first step was to construct a plan. This (Fig. 3, 1) shows how the cave, 
 which is 70 feet in length, runs fairly straight into the cliff in a direction which 
 is slightly south of south-west. The sides are undulating and give passage here 
 and there to tributary channels : many pocket-like holes and crevices open at 
 about a man's height, and as similar holes were often used by the palaeolithic 
 hunters to store their most treasured implements, these were ransacked by 
 Dr. Marett, every one, but in vain. Along an irregular line crossing from the 
 western to the eastern side of the cave the limestone is seen emerging from below 
 the cave deposits to form the eastern half of the floor. 
 
 In the longitudinal section (Fig. 3, 2) the floor of cave earth is seen to rise from 
 the entrance towards the blind end with a uniform slope of 3 in 20, How great 
 a change has taken place since Buckland's time will be seen by reference to his 
 section (Fig. 2), which represents the remains of the original floor as possessing 
 a horizontal surface and ending on one side in a vertical cliff. 
 
 Of the transverse sections that given in Fig. 3, 3 is especially instructive, it 
 shows how the form of the cave has been determined by two sets of master joints, 
 one dipping at about 45°, the other more nearly vertical : they cross the strike of 
 the carboniferous limestone almost at right angles. The sea has played a large 
 part in the work of excavation, probably at a time corresponding with the 
 
8 
 
 W. J. SoLLAS. — Paviland Cave : An Aiirignacian Station in Wales. 
 
 formation of the raised beaches which are found along the coast at heights of from 
 25 to 40 or even 50 feet above the existing sea-level. 
 
 Subsequent to the formation of the raised beaches the sea-level must have 
 fallen considerably, probably at least 120 feet, as supposed by Prestwich. 
 
 It was during this period of emergence that the cave was inhabited by man. 
 
 Still later the sea again rose and — possibly after some slight oscillations — 
 rested at its present limits. 
 
 Whether these movements occurred during an interglacial episode or were 
 wholly post-glacial is an open question. Eminent observers, particularly Tiddeman 
 and Strahan, have asserted that a part at least of the cave deposits met with along 
 the coast of Gower are older than the glacial drift of the district. The succession 
 of littoral deposits along the coast of Gower is as follows : — 
 
 Glacial bed, gravelly boulder clay. 
 Head, a talus of angular limestone fragments. 
 Blown sand, more or less cemented into hard rock. 
 Raised beach with erratics and modern marine shells. 
 
 The superposition of the l)oulder clay has been established by Mr. Tiddeman's 
 observations : nothing has yet been discovered to suggest that these are in error. 
 Mr. E. E. L. Dixon^ admits that in Caldey Island the glacial deposits appear to 
 overlie the raised beach, though the section there is obscure ; and more recently 
 Mr. A. L. Leach^ has described an exposure at Forth Clais in N.W. Pembrokeshire, 
 which supplies the strongest possible confirmation. The boulder clay is there seen 
 to sweep right over the head and the raised beach on to a rocky platform with a 
 glaciated surface (Fig. 4). 
 
 FIG. 4. — DIAGRAMMATIC SECTION OF THE RAISED BEACH AND BOULDER CLAY NEAR FORTH CLAIS, 
 
 ST. David's. (After A. L. Leach.) 
 
 R, stratified shingle and boulders of the raised beach ; H, head ; G, boulder clay ; Di, D2, 
 intrusive dykes ; s, the surface of D, is striated ; b b', striated boulders derived from the 
 raised beach. 
 
 1 E. E. L. Dixon, "Summary of Progress for 1906," Mem. Geol. Sitrvey, p. 63. 
 ^ A. L. Leach, " On the Relation of the Glacial Drift to the Raised Beach near Forth 
 Clais, St. David's," Oeol. Mag., 1911, vol. viii, p. 462. 
 
W. J. SoLLAS. — Paviland Cave : An Aurignacian Station in Wales. 9 
 
 The raised beach of Gower is the local representative of a widely distributed 
 feature, which is found on both sides of the English Channel and elsewhere. In 
 many places it contains at the base large erratics which are generally admitted to 
 have been deposited by floating ice, and thus to indicate an earlier glacial episode 
 (Eiss) than that of the boulder clay (Wiirm). 
 
 We may infer that when the raised beach was in process of formation the 
 <3aves of Gower were occupied by the sea, and consistently with this we find in 
 some of them (Bacon Hole and Mitchin Hole) a deposit of sand, containing marine 
 littoral shells, which lies immediately upon the rocky floor, and is covered over by 
 whatever other deposits may be present. To correlate this marine sand with the 
 raised beach seems only reasonable, but if we take this step it would then seem 
 to follow that the succeeding contents of the caves are older than the culmination 
 ■of the last glacial episode (Wiirm) and younger than the last but one (Eiss). 
 
 This conclusion is in accordance with the chronology of the upper and middle 
 palaeolithic industries which I had already suggested in 1911,^ and it is in general 
 agreement with the views of Penck and Brilckner. There is, at least, one doubtful 
 point, however, in my scheme : I have placed the Chellean on the remote side of the 
 third glacial episode ; whether this is its true position or not will depend upon the 
 result of a renewed examination of the bouchers found at Biddenham and Hoxne. 
 If these should all prove to be Acheulean my chronology may stand, but if, as is 
 most likely, some of them should be Chellean, then it is clear that this industry 
 would have to shift its position into the last genial episode (Eiss — Wiirm). 
 
 In a courageous attempt to reduce to order the raised beaches and other 
 quaternary deposits which occur along the south coast of England, Wales, and 
 Ireland, Mr. Henry Devvey^ has arrived at some interesting results, which will be 
 best understood from the following table. I may venture, however, to point out 
 
 Climate. 
 
 Industry. 
 
 Deposits of the South Coast. 
 
 The Valley of the 
 Somme. 
 
 Arctic 
 
 — 
 
 Boulder clay 
 
 — 
 
 Advancing cold 
 
 Mousterian ... 
 
 Combe rock and head 
 
 Upper loss. 
 
 Cold? 
 
 Acheulean . . . 
 
 — 
 
 Lower loss. 
 
 Warm temperate \ 
 
 Chellean 
 Strepyan 
 
 >• Cemented sands and cave sands... 
 
 30-metre terrace. 
 
 Ketreating cold 
 
 — 
 
 Erratics of raised beach ... 
 
 — 
 
 Arctic 
 
 -- 
 
 — 
 
 — 
 
 1 W. J. SoUas, Ancient Hunters, p. 404. 
 
 2 Henry Dewey, " The Raised Beach of North Devon,'' Geol. Mag., N.S., Dec, V, vol. x, 
 pp. 154-163, 1913. 
 
10 
 
 W. J. SoLLAS. — Paviland Cave : An Aiirignacimi Station in Wales. 
 
 that the correlation of the upper loss with the Combe rock and head is doubtful, 
 even improbable ; and, again, the Aurignacian and Solutrian industries, and perhaps 
 the Magdalenian, in addition to the Mousterian, should be assigned to this deposit. 
 The nearest approach to a standard succession of Quaternary deposits and 
 industries is that of St. Acheul, which has been established by the painstaking 
 genius of M. Comment.^ It is of as fundamental importance to the study of 
 palaeolithic history as Lapworth's classic section at Dobb's Lynn is to the study of 
 Ordovician geology. I give it here in a generalized form : — 
 
 Eecent. 
 Azilian. 
 Magdalenian. 
 
 Quaternary. 
 
 Upper Loss 
 
 IV. Glacial episode -= 
 
 (Wiirm). 
 
 Ergeron, weathered.. 
 
 Gravels 
 Ergeron 
 Gravels 
 Ergeron 
 ^Gravels, angular 
 
 End of the third 
 genial episode. 
 
 Lower Loss 
 
 Third genial 
 episode. 
 
 f Stratified sands and gravels 
 L (at Montieres-les- Amiens). 
 
 fLoss ... 
 I 
 i 
 
 I Gravels 
 
 r Gravels of low terrace 
 J Sands on river gravel of 
 I second terrace. 
 Gravels at base of second and 
 third terraces. 
 
 Solutrian 
 
 Aurignacian, Upper 
 „ Middle 
 
 Mousterian, Upper 
 
 Mousterian, Lower 
 „ Lowest- 
 
 Aclieulean, Upper 
 
 „ Lower 
 
 ii >> 
 
 Chellean, Upper 
 „ Lower 
 
 Strepyan... 
 
 Cold fauna. 
 
 Warm fauna. 
 
 Temperate fauna 
 (Oscillation ?). 
 Warm fauna. 
 
 Fauna with Plio- 
 cene affinities 
 at Abbeville. 
 
 The loss of the Ehine corresponds in all its subdivisions with the loss of the 
 Somme, and it presents a lower loss with a warm fauna which is succeeded by an 
 upper loss with a cold fauna. But the loss of the Ehine can be traced up the 
 valley till it is brought into intimate relations with the glacial deposits of the Alps, 
 thus, according to M. Comment, providing us with good grounds for assigning 
 its higher member to the last glacial and its lower member to the last genial 
 episode. 
 
 But the age of the loss is one of the most vexed questions in Quaternary 
 history. According to I^rofessor Haug^ the two sheets of loss in the Ehine Valley 
 
 ^ V. Commont, " Chronologic et stratigi'aphie des industries protohistoriques, n6olithiques et 
 paMolithique dans les d6p6ts holocenes et pleistocenes du nord de la France,' Congrh Internal. 
 cP Anthropologie, C.R., Geneva, 1912, pp. 239-254. 
 
 2 V. Commont, " Mousterian k faune chaude dans la vallee de la Somme a Montiferes-lfes- 
 Amiens," torn, cit., pp. 291-300. 
 
 ^ E. Haug, Traite le GMogie, Paris, 1911, tom. ii, 1815. 
 
W. J. SoLLAS. — PavilandCave: An Aurignacian Station in Wales. 11 
 
 lie one upon the fluvio-glacial deposits of the last (Wiirm) and the other on 
 those of the penultimate (Kiss) glaciation. If this be so it is evident that the 
 whole of the upper palaeolithic deposits with their industries and fauna are later 
 than the last boulder clay or glacial drift ; a conclusion which is directly opposed 
 to the results of Mr. Tiddeman's observations made on the coast of Gower. The 
 problem, therefore, remains unsolved. It would be useless to discuss it : what is 
 really wanted is further observation in the field.^ 
 
 As we shall learn later, Paviland Cave itself has furnished us with the fauna 
 and industry of the upper loss only, but in some other caves of the coast, Mitchin 
 Hole and Bacon Hole, the warm fauna, represented by Elephas antiquus and 
 Rhinoceros hemito&chus, was found by Falconer, unmixed with the cold fauna, in the 
 lower part of the cave deposits. In two other instances, however, Spritsail Tor 
 and Long Hole, the two faunas occurred mixed together ; at the same time Falconer 
 expressly mentions that in Spritsail Tor the molars of E. primigenius usually 
 " present a much fresher appearance and contain more animal matter than those 
 of E. antiquus." 
 
 On a review of these observations it would seem probable that Paviland Cave 
 was excavated before the great glaciation of the British isles ; it may, judging from 
 analogy with other caves along the coast, have been entered by the sea during the 
 formation of the raised beach, but its occupation by man was confined within 
 the limits of the last glacial episode, possibly following upon the deposition of the 
 glacial drift during the climax of the cold conditions. 
 
 We may now proceed to give an account of our exploration. The first step 
 was to obtain a transverse section of the deposits by cutting a trench across the 
 floor of the cave, 30 feet from the entrance and 8 feet in depth. The face of the 
 cutting was found, however, to be so badly illuminated that observation was 
 impossible, so we proceeded to clear away the material in front which excluded 
 the light. While doing so we came across a large limestone boulder, a rough 
 parallelopipedon in shape, but so large and unlike any we had seen before that the 
 workmen at once exclaimed a " monument ! " It lay between two and three feet 
 below the surface and measured I foot 10 inches in length, 10 inches in breadth 
 and height at one end and 6 inches at other. A little later, about 6 feet 
 further away from the entrance and on the same level we found two others, not 
 quite so large. In the earth between and round aboiit them we found plentiful 
 traces of ochre and many fragments of ivory, including bits of ivory rods such as 
 Buckland discovered in the vicinity of the " Eed Lady." From Buckland's plan of 
 the cave it is evident that this skeleton must have been somewhere near ; it is 
 represented as extended with its long axis nearly parallel to the left (west) wall of 
 the cave and its feet towards the entrance. It seems, therefore, extremely likely 
 
 ^ The question, in the case of the Rhine valley, depends on the identification of Professor 
 Steinmaun's middle terrace of the lower Rhine. Dr. L. van Werveke and others regard it as 
 the equivalent of the lower terrace of the upper Rhine, Professors Penck and Bruckner as the 
 equivalent of the upper terrace. 
 
 I 
 
12 W. J. SoLLAS. — Paviland Cave : An Aurignacian Station in Wales. 
 
 that the boulders had been placed in position — one at the feet and two at the head 
 of the corpse — at the time of interment ; at the same time it is difficult to under- 
 stand, if this were the case, how they came to escape the keen-sighted observation 
 of Buckland. 
 
 In the course of this excavation we obtained a considerable number of flint 
 implements and bones of extinct mammalia : the bones were almost all broken 
 and many of them burnt and blackened : all were irregularly scattered through the 
 deposit without any trace of order or arrangement. The earth being now cleared 
 away it was possible to examine the face of the trench. The result was disappoint- 
 ing. The implement-bearing deposit — a reddish-brown earth crowded with 
 angular and rounded fragments of limestone which extended down to a depth of 
 from 4 to 5 feet — had evidently been much disturbed : nowhere was any trace of a 
 definite industrial layer to be seen the hearths had been broken up and their 
 contents distributed pell mell. This accords with Buckland's statement that the 
 floor had been already dug over before his visit ; a similar observation was made by 
 Falconer. 
 
 Below this layer is a barren cave earth full of limestone fragments which 
 extends to the rocky floor, attaining in one place a depth of 14 feet. At a depth 
 of 7 or 8 feet from the original surface of the floor, part of the lower jaw of some 
 kind of deer was seen projecting from it and we imagined that this might indicate 
 the existence of an industrial layer, but on subsequent excavation this proved a 
 vain hope. 
 
 The barren ground was marked off from the implement-bearing deposit by 
 an irregular greyish white band which might perhaps have resulted from 
 weathering, and on the western side of the cutting this was succeeded by a thin 
 layer of well washed pebbles resting on a dark sand about 6 inches thick which 
 was speckled white with comminuted sea-shells and precisely resembled the sea 
 sand now driven in times of storm over the limestone ledges in front of the cave. 
 The layer of pebbles thickened out towards the wall of the cave and on reaching 
 the wall descended in a vertical sheet. 
 
 The restriction of these exceptional deposits to the western side of the section, 
 which in Buckland's time was exposed at the foot of the longitudinal cliff (Fig. 2), 
 led me to suppose that they might be explained on the supposition that the sea, 
 entering the cave during storms, had first cut out a narrow fissure in the cave earth 
 next the wall of the cave and that this was then filled with pebbles ; continuing 
 its work it enlarged the fissure, forming successive cliffs till the state of affairs 
 described by Buckland was reached. Since then the planing down of the cliff 
 would have led to the covering up of any deposits left by the sea. 
 
 The implement-bearing deposit was removed in successive layers and no signs 
 of stratification were met with till we had advanced to within 23 feet of the inner 
 end of the cave, when we again came across a band vividly coloured with red ochre 
 and comparatively rich in clay ; this extended from the eastern wall only half-way 
 across the cave ; sloping backwards and downwards it reached the rocky floor and 
 
W. J. SOLLAS. — Paviland Cave : An Aurignacian Station in Wales. 13 
 
 disappeared at a distance of 8 feet from its outcrop. This was the most marked 
 line of separation we encountered anywhere, but it does not seem to have had any 
 significance since the same kinds of implements were found both above and 
 below it. 
 
 The Stone Implements. 
 
 For so small a cave Paviland proved remarkably rich in worked flint and 
 chert. In our excavations over 3,600 flakes and fragments were found, and of 
 these some 700 or 800 are implements. Many had previously been extracted by 
 other visitors, several hundred flakes and fragments are preserved in the Swansea 
 Museum, and I owe to the kindness of Mr. Cunnington an interesting series of 
 50 implements. Our richest finds were made at the inner end of the cave where 
 the deposits had been less disturbed than elsewhere. 
 
 The material. — The flint, all derived originally from the Chalk, is of many 
 kinds, owing probably to its having been collected from the adjacent glacial drift, 
 in which stones scattered over a wide area have been swept together by the 
 movement of the ancient ice. But if this explanation should prove to be correct 
 it becomes clear that the Aurignacian people cannot have inhabited the cave till 
 after the maximum of the last glacial episode. 
 
 A common variety of the flint is translucent and honey-yellow in colour, it 
 weathers readily ; another, rather common, is black, and a third, rather rare, is 
 grey. The specific gravity ranges from 2'57 to 2'61, as determined in different 
 specimens; in one and the same specimen it is extremely uniform. The flint 
 often contains patches of opaque material, grey in colour, without lustre, and of 
 comparatively coarse texture ; this is very resistant to the action of the weather 
 and is never patinated ; its specific gravity is 2*6. 
 
 Some of the chert also is Cretaceous, its original home being the Upper 
 Greensand ; it usually contains glauconite grains and weathers white. 
 
 The greater part, however, is Carboniferous, and comes from the Carboniferous 
 limestone. This is 'usually black, rarely lustrous, and scarcely ever patinated. 
 In thin slices under the microscope the black colour is found to be due to 
 diffused transparent brown material, doubtless carbonaceous. 
 
 Some of the Carboniferous chert is grey or greenish grey; one variety is without 
 colour and remarkably translucent. The grey flint is sometimes dotted through with 
 minute white spots. 
 
 The specific gravity of th6 Carboniferous chert ranges from 2*6 to 2-67. Under 
 the microscope it presents the same granular crystalline structure as flint, but 
 different varieties differ greatly in texture, some being remarkably coarse. The 
 specific gravity and resistance to weathering combine to show that the Carboniferous 
 chert is composed almost entirely of quartz. The white spots mentioned above are 
 granular aggregates of an undetermined mineral which has a very high refractive 
 index and strong double refraction. 
 
14 W. J. SoLLAS. — Pavilmul Gave : An Aitrignacian Station in Wales. 
 
 FIG. 5. — MOUSTERIAX IMPLEMENTS ( X |). 
 
 Nos. 1, 2, 11, 13, and 14 are flint, the rest are dark Carboniferous chert, except No. 15, 
 which appears to be some kind of igneous rock : it is very fine grained and of a greenish 
 black colour. 
 
W. J. SOLLAS. — Paviland Cave : An Aurignacian Station in Wales. 15 
 
 Belativc Age. 
 
 The only direct and trustworthy evidence of the relative age of the cave 
 implements is the order of succession of the industrial layers in which they occur. 
 But as we have seen, our excavations failed to reveal the existence even of any 
 industrial layers and consequently our classification of the implements rests solely 
 on their morphological characters. It will be understood therefore that an imple- 
 ment which is assigned say to the Mousterian class is not necessarily of Mousterian 
 age. Indeed it may be added as an independent statement that implements of 
 Mousterian workmanship are well known to occur in Lower Aurignacian deposits ; 
 as one among several instances we may cite the deposits of La Coimba-del-Bouitou, 
 so admirably investigated by the Abbds Bardon and Bouyssonie.^ 
 
 On the other hand we must not press our disclaimer too far ; the order of the 
 successive industries with their characteristic implements having been sufficiently 
 established by observation in France, Belgium, and Germany, the form of the 
 implements becomes a guide to the succession in those cases where this cannot 
 otherwise be discovered. The characteristic implements thus afford precisely the 
 same assistance here as characteristic fossils in geology ; and all the implements 
 classified as Aurignacian may confidently be accepted as belonging to that age : 
 indeed, we may go further still, for though we cannot certainly assign each individual 
 Aurignacian implement to its special horizon, yet we can, on the evidence of a group 
 of these implements, definitely assert the existence of certain subdivisions, 
 such as the Middle and Upper Aurignacian. 
 
 Mousterian. 
 The cave afforded numerous examples of Mousterian implements both in flint 
 and chert, more especially in chert of the black variety. The flaking is generally 
 rude and characterized by the removal of rather thick scales, which are short and 
 broad. The forms, though often irregular, include a great number of varieties, 
 some of which seem to be precursors of the more finished Aurignacian implements.^ 
 They may be grouped as follows : — 
 
 1. A rather large ovate flint flake, with a plain face below, retaining the bulb of 
 
 percussion ; the upper surface, which preserves some of the crust, boldly 
 flaked, and the edges serrated by minute Mousterian retouches (Fig. 5, 
 No. 1). 
 
 2. Large, fairly thick racloirs,^ trimmed all round and carefully retouched 
 
 1 L. Bardon, A. and J. Bouyssonie, " Station prehistorique de la Coimba-del-Bouitou, 
 pr^s Brive" (Corrfeze), Bull. Soc. sci. hist, et arch, de Correze, 1907-1908, 54 pp. 
 
 2 The number and variety of the Mousterian implements is so considerable and their 
 patina, according to the Abb6 Breuil, so distinctive that I am inclined to think a Mousterian 
 horizon must originally have been present. 
 
 3 I have tried in vain to find simple English equivalents for the French " grattoir " and 
 "racloir"; "scraper," without qualification, has too wide a meaning, and,' "scratcher" and 
 " rasper " are inappropriate ; I have, therefore, retained the French terms, which are now 
 sufficiently familiar to English anthropologists. " Scraper," however, may be used as a general 
 term. 
 
16 W, J. SOLLAS. — Paviland Cam : An Aurignacian Station in Wales. 
 
 along one edge. The bulb of percussion is generally removed (Fig. 5, 
 Nos. 2, 15 ; Fig. 6, No. 20). 
 
 3. Eough, rather thick, biconvex lenticular or disciform bodies (Fig. 5, 
 
 Nos. 5, 13), occurring both in flint and chert. Their purpose is obscure ; 
 for want of a better explanation they are sometimes spoken of as sling 
 stones. 
 
 4. Small thick irregular forms, generally rudely flaked all over, often terminat- 
 
 ing at both ends in a rude point (Fig. 5, No. 10). These are very 
 numerous, especially in black chert. 
 
 5. Thick but more regular forms than the preceding, flaked all over, trimmed 
 
 along both sides and ending in front in a point (Fig. 5, Nos. 11, 12). 
 
 6. Straight-edged grattoirs, with abrupt flaking at the scraping end (Fig. 6, 
 
 No. 23). 
 
 FIG. 6.— MOUSTERIAN IMPLEMENTS ( X f). 
 
 (These are all in black chert.) 
 
 7. Straight-edged grattoirs, with one or more concave notches (Fig. 5, No. 9). 
 
 These pass into the next variety (No. 8). 
 
 8. Rostrate grattoirs, sometimes resembling the Aurignacian implement of the 
 
 same name (Fig. 6, No. 21), sometimes with the beak more fully 
 developed (Fig. 5, Nos. 6, 8). 
 
 9. Short scrapers, not thick, with concave notches which when they occur side 
 
 by side give rise to points, such as might be used for scratching or boring 
 (Fig. 5, No. 7). It may be observed, however, that in this as in many 
 similar cases it is the concave notch which shows most evidence of wear 
 and tear. In one instance (Fig. 5, No. 17) we have a scraper with two of 
 these points. 
 
 10. Various forms of racloir (Fig. 5, Nos. 3, 16, 18 ; Fig. 6, No. 22). 
 
 PSEUDO-MOUSTEEIAN. 
 
 The implements included under this term closely resemble the Mousterian 
 but bear traces of Aurignacian workmanship and may be referred to that age. Of 
 the numerous examples in our collection we may especially mention the 
 following : — 
 
 1. Bacloirs. — Thin flakes often resembling in general form the Mousterian 
 
W. J. SoLLAS. — Pmiland Cave : An Aurignacian Station in Wales. 17 
 
 point, with minute marginal retouches, producing a finely serrated 
 edge (Fig. 7, Nos. 27, 30) ; some of these flakes bear a positive bulb on 
 one side and a negative bulb on the other, showing they were not the first 
 to be struck off the nucleus (Fig. 7, No. 31): some are abruptly retouched 
 over a part of the periphery (Fig. 7, No. 26). 
 
 2. Simple grattoirs. — Short thick flakes with a straight edge. 
 
 3. Bostrate grattoirs. — Boldly flaked into form with lamellar flaking on the 
 
 snout and a well-marked anterior notch. Sometimes abruptly flaked at 
 the base. 
 
 FIG. 7. — PSEUDO-MOUSTERIAN IMPLEMENTS ( X |). 
 
 (These are all in fiint.) 
 
 Universal tool. — A flake of medium thickness ending in a grattoir at one 
 extremity and a point at the other, with a conve.K edge retouched to form 
 a racloir on one side and a concave edge retouched to form a concave 
 scraper on the other (Fig. 7, No. 29), 
 
 Large fragments of flakes remarkably uniform in thickness and with equally 
 remarkable flat scaly flaking over the whole surface ; this flaking is as 
 tangential as the Solutrian but never gives rise to long narrow channels 
 (Fig. 7, Nos. 25, 28). 
 
 Aurignacian. 
 
 The implements of this age, here as everywhere, are distinguished by a rich 
 variety of form. Such a multifarious equipment of tools implies that the workman 
 exercised his skill in many different handicrafts. The worked bone and ivory to 
 
 h 
 
18 
 
 W. J. SoLLAS. — Pavilanil Cave: An Aurignacian Station in Wales. 
 
 be described later support this inference, tliey show that the Aurignacian hunter 
 was already familiar with the principle of the saw, the graver, the spokeshave, 
 racloir, grattoir, and drill ; but much of his work was accomplished on perishable 
 material, and he probably produced a whole host of objects — spears, bows and 
 arrows, digging sticks, thongs of hide, fur garments, basket work and nets — of which 
 no trace has been or could be preserved. Yov some of his implements we are 
 unable to assign a use. 
 
 The oldest of the Aurignacian implements from Paviland are : — 
 
 1. A variety of the carinate grattoir. The typical grattoir car6ne or grattoir 
 Tart4 is thick and short, with a flat surface below, and a regular flaking 
 on the sides, especially in front, produced by detaching comparatively long, 
 thin, narrow lamellae which ran parallel or with slight convergence from 
 below upwards. When the lower face is worked large flakes are taken off 
 to produce a flat base. The examples represented (Fig. 8, Nos. 32, 32a, 
 
 o2a ^-"-^ 35 ^'^«!^r34a 
 
 FIG, 8. — THE OLDEST AURIGNACIAN IMPLEMENTS ( X |). 
 
 (These are all in flint. The right-hand extremity of No. 34 is the top of No. 34a.) 
 
 33, 34, 34a) answer to this description, but are ruder in shape and less 
 elaborately flaked than in the true grattoir car^ne, to which they may 
 have given rise, or from which they may have been derived. 
 
 2. Nucleiform grattoir. — This is a thick scraper related to the preceding, but 
 
 not provided with a keel (Fig. 8, No. 35). 
 
 3. Atypical racloir. — This is a roughly chipped oval flake. 
 
 Middle Aurignacian. 
 The Aurignacian style of working in flint attained its fullest expression in this 
 period : the certainty and elegance of the retouch is admirable, and the extreme 
 
55 -.^.-^ ^^ ^-^^^^^^dS '''«^^59 
 
 FIG. 9.— MIDDLE AURIGNACIAN IMPLEMENTS ( X t). 
 
 (Nos. 35, 36, 42, and 43 are in Carboniferous chert, Nos. 37 and 55 in Cretaceous chert 
 coloured with glauconite, the i^emaining specimens are in flint.) 
 
 h 2 
 
20 W. J. SOLLAS. — Paviland Gave : An Aurignacia7i Station in Wales. 
 
 differentiation of the forms show that the workman appreciated the value of 
 specialized tools. 
 
 1. Nucleus grattoir. This is a nucleus which, as so often happens, has been 
 
 utilized as a grattoir. Similar forms occur in Neolithic deposits. 
 
 2. Straight grattoir (Fig. 9, Nos. 45, 53, 54, 58, 60 ; Fig. 11, Nos. 73, 74). 
 
 Eather thick, generally tabular, flakes, rectangular or square, not unlike 
 a gun flint, with wide abrupt lamellar flaking, perpendicular to the edge. 
 Professor Breuil adds that Nos. 54, 58, 60 and 73, present a salient lateral 
 angle which was certainly intended for a definite purpose. 
 
 3. Grattoir with oblique terminal retouches. This differs from the preceding 
 
 mainly in the direction of the lamellar flaking. 
 
 4. Terminal grattoir, or grattoir at the end of the flake {grattoir stir bout de 
 
 lame) (Fig. 9, Nos. 49, 55, and Fig. 10, Nos. 61 to 65. The scraping end 
 may be rectilinear and oblique, with a lateral angle as in Nos. 49, 55, or 
 curved in an arc as in Nos. 61 to 65. The flake may be a long thin 
 lamella (No. 63), or short and thick (No. 61). 
 
 5. Straight notched grattoir, (Fig. 9, Nos. 47, 52) a straight grattoir with a 
 
 concave excavation at the terminal edge. 
 
 6. Eostrate grattoir {grattoir a museau). There are several varieties of this 
 
 form ; a very common one seems to be closely related to the preceding 
 form, the terminal notch bringing into greater relief one of the corners 
 which forms the snout or rostrum (Fig. 9, Nos. 37, 38, 40, 41, 57, and 
 Fig. 11, No. 71). 
 
 7. Ogival rostrate grattoir. The snout is more developed, and the plan of the 
 
 rostral end has the form of an ogive arc (Fig. 9, Nos. 35, 36, 42, 43). 
 (This is a form which, so far as I remember, is rather peculiar to Paviland. 
 It passes insensibly into another type which closely approaches a kind of 
 burin : compare Nos. 34, 34«, oldest Aurignacian, and Nos. 37, 41, 44, 49, 
 and 50, middle Aurignacian. — Note hy Professor Breuil.) 
 
 8. Eostrate grattoir with inverse terminal retouches. The retouches are on 
 
 the underside of the rostrum, and slope in the opposite direction to those 
 on the ordinary form (Fig. 9, Nos. 39, 44, 48). 
 
 9. Short grattoir. Several varieties are included under this form. One of the 
 
 most striking (Fig. 11, No. 68) is a short thin flake retaining the bulb of 
 percussion on its plain under surface, and minutely retouched by regular 
 oblique flaking at the front end to form a circular arc with a finely serrated 
 cutting edge. The fine oblique flaking recalls some Solutrian and Neolithic 
 work. I doubt whether this was intended for a scraper, it would be much 
 more efficient as a cutting implement. Another variety (Fig. 11, No. 67) is 
 a thicker oval flake with rather abrupt retouching all round the periphery, 
 and a pointed extremity; a third is a short-beaked grattoir (Fig. 11,. 
 No. 70) with oblique lamellar retouches. 
 
W. J. SOLLAS. — Paviland Cave : An Aurufnacian Station in Wales. 21 
 
 63 ^^^^ 66 
 
 FIG. 10. — MIDDLE AURIGNACIAUT IMPLEMENTS ( X |). 
 
 (These are all in flint.) 
 
 FIG. 11. — MIDDLE AURIGNACIAN IMPLEMENTS ( X f). 
 
 (These are all in flint, except No. 72, which is in black chert.) 
 
22 W. J SoLLAS, — Paviland Cave : An Aurignacian Station in Wales. 
 
 FIG. 12. — MIDDLE AURIGNACIAN IMPLEMENTS (x |). 
 
 These are all in flint. The upper right-hand corner of No. 86 is the lower right-hand corner of 
 No. 86a. No. 85 is inverted. No. 91 is in the Swansea Museum.) 
 
W. J, SoLLAS. — Paviland Cave : An Aurignacia^i Station in Wales. 23 
 
 10. Flakes with terminal inverse oblique retouch (Fig. 9, Nos. 46, 50, 51). 
 
 These resemble burins, and Professor Breuil writes : " I am not sure that 
 they are not lateral burins, Nos. 50 and 51 at all events." 
 
 11. Double racloir (Fig. 10, No. 66). Eacloirs are extremely rare among our 
 
 Aurignacian implements : this is one of the best examples and, as 
 Professor Breuil remarks, it is entirely devoid of any Mousterian aspect. 
 
 One of the most characteristic implements of the later Aurignacian industry 
 is the burin, which is distinguished by its strong, short cutting edge. It was by no 
 means exclusively employed for engraving pictures or carving in the round,^ but 
 probably served a variety of purposes ; it would be very useful for cutting thongs 
 out of hides, and it was almost certainly the implement with which strips were 
 cut out of reindeers' horn or ivory. The process in this case was to score two 
 parallel grooves in the substance, and to deepen them till they met. The following 
 varieties were found : — 
 
 1. Simple burin (hurin en hec de Jldte). This is the classic burin of the 
 Magdalenian age, but it is common enough in the Aurignacian also. 
 The edge, which is perpendicular to the plane of the lamina, is formed 
 by two faces which meet in an acute dihedral angle. In its typical form 
 each of these faces is simple (Fig. 12, Nos. 75, 76, 78, 79), but not 
 infrequently one or both may be composed of several facets, and in this 
 case the form may differ sufficiently from the simple burin to deserve 
 a separate name ; as Mr. Henry Balfour pointed out 
 to me, it seems to have served some other purpose ; 
 perhaps we might call it a spatulate burin (Fig. 12 
 Nos. 77, 83, 83«). 
 
 2. Lateral burin (hurin d' angle or ImHn latcrale). 
 This is essentially a flake with a distal truncated 
 end, which is more or less transversely retouched. 
 Its sharp corner might be used for graving without 
 further modification, but in order to strengthen the 
 edge a spall is split off, leaving a longitudinal facet 
 which intersects the truncated end (Fig. 12, Nos. 84, 
 85, 86). When the graving edge is blunted by use 
 it can be renewed in various ways, as by taking 
 off a second spall parallel to the first (Fig. 13) 
 or by retoviching the truncated end, or even by 
 removing the whole of this end by a transverse 
 fracture. All these methods were employed, but 
 the last named had the disadvantage of unduly shortening the implement. 
 Some specimens show by the step-like outline of the burin face that 
 
 FIG. 13. — Diagram to illustrate 
 the manufacture of a lateral 
 burin and the method of 
 renewing its edge. (After 
 the Abbes Bouysonnie and 
 Bardon.) 
 
 ^ Note hy Professor Breuil.- 
 more than for engraving. 
 
 -The burin was used for working in bone and ivory mucli 
 
24 W. J. SoLLAS. — Paviland Cave: An Aurignacian Station in Wales. 
 
 the edge had been several tmies renewed. The little spalls struck 
 ofif are preserved in the cave deposits and may be recognized by their 
 form, retouching and worn edges. In some cases (Fig. 12, No. 80) 
 a pair of notches near the base seem to have been intended for a ligature 
 to secure the burin to a handle. (J^ote hy Professor Breuil. — Nos. 80 and 
 82 are not very typical lateral burins.) 
 
 3. Lateral burins with inverse retouch (Fig. 12, Nos. 81, 81a). Professor 
 
 Breuil remarks : " This is no doubt a spatulate burin, but more lateral than 
 usual. The greater part of the spatulate and beaked burins from Pavi- 
 land show a marked tendency towards inverse retouch." My friend, 
 Mr. Miles Burkett, has called my attention to the prevalence of a similar 
 retouch in the Aurignacian Station of La Bertonne (commune of Peujard, 
 Gironde) ; it is described by M. FranQois Daleau.^ 
 
 4. Beaked burins (hurins husquies). The graving end is formed by a flat facet 
 
 below on one side and an arched surface with parallel lamellar flaking on 
 the other (Fig. 12, Nos. 86, 87). The edge is renewed by flaking off the 
 face to form a fresh flat facet. In some examples the back has been 
 hammered down to give a comfortable grip. A specimen in black chert 
 (Fig. 11, Nos. 72, 72a) recalls one figured by the Abbes Bardon and 
 Bouyssonie," from Coimba-del-Bouitou, the chief difference being that this 
 is single-ended while theirs is double-ended. 
 
 5. Pound grattoir with opposed burin (Fig. 12, Nos. 88, 89, 89«,and 91). Three 
 
 of these beautifully symmetrical little implements have been obtained from 
 Paviland, one is in the Swansea Museum and the other two were found 
 during our excavations. One end is a simple burin, the other a circular 
 grattoir with a carefully retouched edge ; the under surface is flat, the upper 
 swollen and facetted. Similar forms have been found by the Abbes 
 Bardon and Bouyssonie in the upper hearths of Coimba-del-Bouitou.^ 
 
 6. Lames or flakes variously retouched. These include forms with oblique 
 
 terminal retouch, some with a notch on one side. 
 
 7. Finally, and quite atypical, are long flakes in flint and chert without retouch 
 
 or definite signs of use. 
 
 Upper Aurignacian. 
 
 Many of the forms already described from the Middle Aurignacian are 
 repeated here but they are usually distinguished by less careful retouching. 
 
 ^ F. Daleau, " Silex k retouches anormales," Actes de la Societe Archeologiqii,e de Bordeaux, 
 
 1910, torn, xxxi., pp. 18, pis. 
 ^ L. Bardon, A. and J. Bouyssonie. Loc. cit., p, 46, fig. 24. 
 3 ibid. Loc. cit., p. 49 ; Fig. 26, Nos. 60 and 61. 
 
W, J. SOLLAS. — Paviland Cave : An Aurignacian Station in Wales. 
 
 25 
 
 The characteristic implement of the sub-division is : — 
 
 1. The Gravette point (Fig. 14, iN"os. 92 to 97. Nos. 96 and 97 are the pointed 
 extremity broken off), a long, straight, parallel-sided flake, generally triangu- 
 lar in section, one edge of which has been completely removed by minute and 
 thorough retouching. It differs from the Chatelperron point of the Lower 
 Aurignacian, with which it may be confused,^ by its greater straightness, 
 
 ' FIG. 14. —UPPER AURIGNACIAN IMPLEMENTS ( X t). 
 
 (Nos. 106 to 110 are in Carboniferous cliert, all the others in flint.) 
 
 elongation and narrowness, as well as by its more acute point. There is 
 a difference also in the retouch, which is more regular and finer in the 
 Gravette point, and is almost constantly directed from below upwards. 
 In the Chatelperron point it sometimes follows this direction, but some- 
 times the opposite from above downwards. (The flat of the blade is 
 
 ^ H. Breuil, " Les Subdivisions du Palcolithique superieur et leur Signification," Compte 
 Rendu de la XIV^ Session, Geneve, 1912, Congres International d^ Anthropologie, torn. I, p. 165 ; 
 see in particular Fig. 1, Nos. 6 to 9. 
 
26 
 
 W. J. SOLLAS. — Faviland Cave : An Aurignacian Station hi Wales. 
 
 FIG. 15. — Minute imple- 
 ments or microliths (nat. 
 size). 1. resembling a grav- 
 ette point, but without a 
 cutting edge. 2. A spoke- 
 shave or borer. Both are in 
 flint. 
 
 regarded as the lower surface, the lojigitudinally facetted side as the 
 
 upper surface.) In all our specimens the direction is from below 
 
 upwards : most of them show signs of use on the 
 
 cutting edge. However acute the point, and in some 
 
 cases its sharpness is extreme, the retouch is always 
 
 continued along the back right up to the extremity. 
 
 As shown by the Abb^ BreuiP, the gravette point 
 
 passes into minute forms at Font Robert (Correze) and 
 
 this appears to be the case also at Paviland. Among 
 
 the specimens collected by Dr. Cunnington is one 
 
 beautiful example (Fig. 15, No. 1) ; small as it is, the 
 
 retouch is no less perfect and is carefully maintained 
 
 throughout the whole length of one side. 
 
 The Abbe Breuil calls my attention to the tendency to assume a, 
 
 triangular form which is presented by Nos. 94 and 96 (Fig. 14): a point 
 
 of interest, since it has also been observed at Grimaldi and in the 
 
 neighbourhood of Brive. 
 
 2. Some similar little thin knife-like flakes occur but with a truncated 
 
 extremity (Fig. 14, No. 98). 
 
 3. Borers {pergoirs). — Numerous flakes have been trimmed in such a manner 
 
 as to produce one or more sharp projections, which served probably for 
 perforating skins or boring holes in wood or ivory (Fig. 14, Nos. 100 to 105) ; 
 one of these borers (No. 102) resembles a form described from the Grotte 
 du Trilobite. 
 
 4. Spokeshaves {coches diverses). — A few of these notched flakes (Fig. 14,. 
 
 Nos. 106 to 110), resembling those from the Upper Aurignacian of the 
 Grimaldi caves, have been chipped out of chert.^ A microlithic form 
 occurs in Dr. Cunnington's collection (Fig. 15, No. 2). 
 Numerous small flakes occur, Aurignacian in age, which could never have been 
 
 retouched in their present state ; they are no doubt the worn ends of burins and 
 
 scrapers which have been struck off to renew the edge. 
 
 \\2 114 
 
 FIG 16. — SQUAMOUS FLAKES (Pieces esquilUes). 
 (All these are in flint.) 
 
 1 Loc. cit, Fig. 1, Nos. 18 to 23. 
 
 2 H. Breuil loc. cit., p. 179, fig. 11 Nos. 1 and 6 to 10. 
 
W. J. SoLLAS. — Paviland Cave : An Aurignacian Station in Wales. 27 
 
 Squamous flakes. — These curious forms (Fig. 16, Nos. Ill to 115), which are 
 found so commonly in the Aurignacian, extend upwards into the Magdalenian and 
 Azilian. Here they are no doubt Aurignacian and probably belong to the middle 
 subdivision. The Abb^ Breuil remarks that they are instruments broken by some 
 very violent usage which he does not understand. 
 
 SOLUTKIAN. 
 
 The approach of this age is heralded by a number of Proto-Solutrian imple- 
 ments distinguished by a new style and technique, markedly different from any 
 that preceded them, but destined to a long development, which reached its 
 highest perfection in some of the Neolithic achievements of Egypt or the still more 
 recent triumphs of South America. 
 
 1. Tanged stylet {pointe a sole). — Two fragments (Fig. 17, Nos. 116, 119) 
 
 represent one the tanged end (No. 116) and the other the pointed end 
 (No. 119) of a triangular stylet; these are not unlike some forms 
 described by the Abbes Bardon and Bouyssonie from the Grotte de la 
 Font Eobert,^ but they are on a larger scale. The flaking is Solutrian in 
 character but comparatively rough. 
 
 2. Laurel-leaf point {pointe en feuille de laurier). — This leaf -like lamina, in 
 
 dark green chert, is faintly channelled on one side by thin lamellar 
 flaking which just skims across the surface (Fig. 17, Nos. 118, 118a). It 
 is keeled on the other side, and this character distinguishes it from the 
 genuine Solutrian implement. 
 
 3. Parallel-sided point. — A well-pointed weapon (Fig. 17, No. 122), flat on one 
 
 side, convex and flaked all over on the other ; the proximal end is broken 
 off. This is preserved in the collection of the Swansea Museum. 
 
 4. Eazor blades. — These are triangular in section like a common razor ; one 
 
 side is without retouches, the other flaked all over in characteristic 
 Solutrian style ; the flakes were started from the back where it meets the 
 side and propagated across to the cutting edge. One example is complete 
 (Fig. 17, No. 117); another (Fig. 17, No. 115) is a fragment of a much 
 finer blade, thinner and with surprisingly beautiful channelling ; a third 
 (Fig. 17, No. 120) is not retouched except on the back, which is crossed 
 by channelled flaking. 
 
 5. A universal implement. — This (Fig. 17, No. 121), which at first sight seems 
 
 to be merely a fragment, is a complete implement. It may be regarded as a 
 razor blade which narrows to a dressed point at one end and expands and 
 thickens to form a grattoir at the other. 
 
 ^ L. Bardon, A. and J. Bouyssonie, Za Grotte de la Font Robert, Brive, 1908, Fig. 5 
 p. 11. 
 
28 
 
 W. J. SoLLAS. — Paviland Cave : An Aurignacian Station in Wales. 
 
 The Abbe Breuil states that the forms Nos. 115, 117, 120, and 121 are not 
 Proto-Solutrian ; this is shown by the direction of the retouch. 
 
 Hammer-stones of TJ-p-per Palceolithic age. — Among the stone implements 
 which cannot be assigned to any particular horizon are two small pebbles, 
 bearing the marks of use. They were selected not without judgment to serve as 
 hammers ; the material of which they consist is very hfird and resistant, in one case 
 
 l!9 ...^ ,^ 
 
 Fig. 17. — proto-solutrian implements (x|). 
 (These are in flint, except No. 118, 118a, which is in dark green Carboniferous chert.) 
 
 a silicified breccia, in the other a peculiarly compact hornblende schist ; their form 
 is such as to afford an easy and firm grip as well as good striking ends, which are 
 now much bruised by use. 
 
 In one of these specimens (Fig. 18) the rounded and bruised ends are extended 
 in planes at right angles to each other, and there is a depression in the middle 
 
W. J. SoLLAS. — Paviland Cave : An Aurignacimi Station in Wales. 29 
 
 whicli gives a convenient hold whichever end may be used : by simply turning it 
 round it can be made to strike with either edge, horizontal or vertical. 
 
 FIG. 18.- — HAMMER-STONE. (nAT. SIZE.) 
 
 A hammer-stone with the edges at each end in planes at right angles to each other. One 
 figure passes into the other by rotation about a longitudinal axis in the plane of the 
 paper through 90°. 
 
 There is a third hammer-stone of black flint which has been chipped into 
 shape ; it is much bruised on the striking surface. 
 
 Minerals. 
 
 The cave men were evidently not unobservant of the properties of stones and. 
 minerals. In their choice of stones for implements they were guided by expert 
 knowledge ; they knew where to look for red ochre and had already recognized its 
 value as a pigment ; but they also showed a curious interest in minerals which they 
 could not apply to useful purposes. Thus among the minerals brought into the cave 
 are numerous fragments of psilomelane, one weighing nearly a kilogram; a hydrate 
 of manganese allied to pyrolusite is also present, and there are besides specimens of 
 lignite, fibrous hematite or kidney iron ore, pseudomorphs of hematite after calcite, 
 mammilated limonite, and a broken crystal of pure quartz. The pyrolusite might have 
 furnished a black pigment, but not the psilomelane, which probably excited attention 
 by its strange heaviness, as the lignite by its equally strange lightness ; the iron ores 
 were probably found in association with the red ochre and were collected as curiosities. 
 
 The quartz crystal was probably treasured for its magic powers : it recalls the 
 practice so common among the medicine men of most existing primitive tribes, 
 especially the Australians, of conjuring such crystals out of the bodies of their 
 patients with the object of effecting a cure. 
 
 Patination. 
 We have already called attention to the remarkable dead-white patination of the 
 flint implements, a character which they possess in common with a great number of 
 others found elsewhere, especially in cave deposits. 
 
30 
 
 W. J. SoLLAS. — Paviland Cave : An Aurignacian Station in Wales. 
 
 The question of patination has not as yet, I believe, been so closely studied as 
 it deserves. 
 
 The first step in the process is the removal in solution of a part of the flint 
 from a more insoluble remainder, which is thus rendered porous. When the pores 
 are filled with air the difference between the refractive index of this medium and 
 the flint is so great that the incident light is scattered and absorbed. Thus the 
 flint and its patina differ in their reaction to light in precisely the same manner as 
 ice and snow. 
 
 It is easy to obtain very thin laminte of flint, suitable for microscopic examina- 
 tion, by striking off flakes with a hammer: even flakes affording a transverse 
 section across the flint and its patina together may be obtained in this way. If 
 these be immersed in an alcoholic solution of some aniline dye and then well 
 
 ^^ 
 
 •< 
 
 ¥ 
 
 d 
 
 ^ ^ 
 
 n 
 
 1 2 3 
 
 FIG. 19. — PATINATION OF FLINT ( X 40). 
 
 The white patina is represented by the dark, cross-hatched, areas. 
 
 No. 1. Surface of a flint after digestion with concentrated ammonia sohition at 
 
 a temperature of about 140° C. for twenty-four hours. 
 No. 2. Section of flint and its patina crossing the junction of the weathered 
 
 and unweathered flint. 
 No. 3. Surface of a flint with a " blue " patination. 
 
 washed with alcohol the weathered crust or patina will be found to retain a 
 permanent stain and the flake, after replacing the alcohol by xylol, may be 
 mounted in balsam for observation under the microscope. Commencing our 
 examination at the junction of the crust with the flint (Fig. 19, No. 2) we shall 
 see the stained material fraying off into the flint in the finest of little threads ; 
 a little further in, away from the flint, the stained material forms a close network, 
 enclosing in its meshes the unaltered flint in the form of little globules from about 
 O'l mm. to 0*05 mm. in diameter. These strongly recall the chalcedonic globules 
 observed by Messrs. Jukes-Browne and HilP which may be traced in the half- 
 formed flint nodules of the chalk-marl through all stages of transformation from 
 the state of opal to flint. 
 
 On finely powdering the weathered crust of our flint implements these globules 
 may be isolated for separate examination. By applying a stain they can be shown 
 
 1 A. J. Jukes-Browne and W. Hill, " On the Occurrence of Colloid Silica in the Lower 
 Chalk of Berkshire and Wilts," Quart. Journ. Geol. Soc, 1887, vol. xlv, pp. 403-420. 
 
W. J. SoLLAS. — Pavilcind Cave : An Aurignacian Station, in Wales. 31 
 
 to carry with them, often coating their exterior, some of the weathered material, 
 which in rare instances is found, by examination between crossed Nicols, to be 
 isotropic, and, therefore, in all probability, some form of silica hydrate. 
 
 The amount of flint which may be removed in the natural process of patination 
 is considerable. A flake which had weathered through its whole thickness, as was 
 proved by breaking it across, was first weighed and then placed in melted paraffin 
 till the contained air had been entirely replaced, it was next removed and allowed 
 to cool ; after cleaning the surface, to remove all adhering paraffin, it was again 
 weighed and the difference in weight afforded a means of determining the volume 
 •of the pores which had been produced by weathering. The result in one instance 
 was 25 per cent., in another as much as 31 per cent, of the whole flake. 
 
 The white crust is evidently a residual effect of solution but the patina is 
 something besides. Its outermost part is an extremely thin impervious film or skin, 
 perforated here and there by a few minute holes only. The presence of this skin 
 may be most readily proved by breaking a weathered flake across and immersing it 
 first in a solution of ferric chloride and then, after washing, in a dilute solution of 
 ammonia ; red ferric hydrate is at once precipitated on the fractured surface, but 
 •elsewhere, under the natural skin, the yellow ferric chloride is seen to remain 
 unchanged, except at a few pin-holes, where a little ammonia enters and forms a 
 slowly expanding circular area of ferric hydrate. Similarly when a fragment is 
 placed in melted paraffin, the emission of air as the paraffin enters is confined to 
 the pin-holes and the broken extremity. 
 
 It is to the presence of this skin that the patina owes what little lustre it 
 possesses. 
 
 Its formation is the second step in the process of patination. The solution of 
 silica which is formed within the crust is concentrated by evaporation at the surface, 
 •deposition follows and continues till the superficial pores are so completely obliterated 
 that nothing can enter or leave the flint except by a few remaining apertures. 
 
 We have here much the same phenomenon as that presented by the well 
 known " desert varnish." 
 
 The existence of a superficial layer has already been recognized by Eenard 
 and Klement^ who speak of it as a " more compact " patina composed of crystalline 
 silica. The inner patina is said to be opaque, the outer a colourless silica with little 
 ramifications of the inner substance extending into it. This is not quite the same 
 structure as that here described for I have not been able to distinguish the " skin " 
 under the microscope, its existence can only be shown by indirect methods. 
 Kenard and Klement, on the other hand, were able to separate their superficial layer 
 from the inner patina and to analyse it. Such a thick outer layer, sometimes more 
 than 5 mm. in thickness, forms, however, the exterior of the weathered crust of 
 many chalk flints, and is, of course, immensely more ancient than any patina on 
 human implements. 
 
 ^ Eenard and 'KlQvci&a.t, Bull. Ac. roy. Belgique, 1887 (3), xiv, p. 773. 
 
32 W. J. SoLLAS. — Paviland Cave: An Aurignacian Station in Wales. 
 
 The whiteness of the patination in our Paviland flints is conditioned, not only 
 by the primary cause already pointed out, but also by its thickness and the 
 absence of pigment. There are other patinas which Dr. Allen Sturge distinguishes 
 as " blues " and " reds." The blue is obviously an incipient patination, in which 
 the merest film of weathered material is drawn over a background of black flint. 
 The red patinas are due to the presence of ferric oxide or hydrate : this has been 
 denied ; I therefore placed some flints bearing the red patina^ in a solution of 
 nitro-hydrochloric acid, this completely removed the colour, and on neutralizing 
 the solution with ammonia, ferric hydrate was precipitated. To complete the 
 proof the bleached flints were introduced first into a solution of ferric chloride and 
 then of ammonia, when the original red tint was restored to the patina. 
 
 Archaeologists have sometimes hoped or imagined that it might be found 
 possible to estimate the age of an implement from the thickness of its patina, and 
 my friend Professor Joly has even made experiments with this object. It is 
 possible that a method may be devised which will be applicable in special cases 
 but not, I think, universally. 
 
 Patination proceeds very slowly. We may gain some general idea of its rate 
 by determining the thickness of the patina produced in an estimated period of 
 
 time. The thickness may easily be found by 
 breaking our implements across and measuring 
 the white crust under the microscope (Fig. 20) ; 
 observations on seventeen of our specimens 
 gave an average of 1'36 mm. The age of the 
 Aurignacian is unknown, but I doubt if any 
 archaeologist would regard 13,600 years as an 
 excessive or even adequate estimate ; accepting 
 it for the sake of illustration we obtain 
 FIG. 20.— Two Aurignacian flint injple- 1*36/13,600 = O'OOOl mm. per year as the 
 ments broken across to show the average rate of formation. This, however, is 
 weathered crust, and the difference ^^^^-^ ^^ average based on thicknesses which 
 in its thickness in different examples. 
 (Nat size ) vary over a wide range, from 4"1 to 0" 15 mm. or 
 
 even to zero, for there are some varieties of flint 
 which do not weather perceptibly at all : this is particularly the case with a dense 
 black flint which traverses the white surface of some deeply weathered specimens 
 in a conspicuous band and does not present even a blue patination nor any loss of 
 lustre. Thus the rate of patination depends in part on the constitution of the flint, 
 a variable factor. It also varies with external conditions ; thus it is not the same 
 even for different sides of the same specimen, in one case, for instance, the thickness 
 of the patina on one side is 0-3 mm., on the other 0-9 mm., in a second 0*75 mm. and 
 1-8 mm. Here the difference depends upon aspect; A. de Mortillet has given 
 
 1 I have to thank Dr. Allen Sturge for these and other interesting examples of weathered 
 flints. His views, even when we differ from them, are always stimulating and suggestive. 
 
W. J. SOLLAS. — Paviland Cave : An Aurignacian Station in Wales. 33 
 
 another example which shows how the rate is affected by the different conditions 
 which prevail in different parts of the cave ; he describes a laurel-leaf point which 
 had been broken into four parts and scattered through the cave earth, so that 
 when found they were lying remote from one another ; they present different 
 patinas, three of them are grey, the fourth is white and deeply weathered. 
 
 The solvent agent on which the weathering depends has not yet been identified. 
 Water alone may produce some effect, but evidence is lacking ; observations on old 
 flint walls exposed to the rain might throw some light on the subject, but the walls 
 would need to be very ancient. The oldest I have examined was one in Winchester 
 College facing the rainy quarter and built a.d. 1394, or 520 years ago. In this inter- 
 val weathering at the average rate already supposed would have extended inwards for 
 0'05 mm., and have produced a perceptible result; but considering how great a number 
 of flints were exposed to observation we might perhaps take the maximum rate as 
 applying to some of them and this would give a thickness of 0"15 mm., yet no trace of 
 patination could be anywhere discerned : there was no sign of a bluish film, nothing 
 beyond a slight loss of lustre. This would seem to suggest that water alone is inade- 
 quate to account for the whole of the phenomenon, and we are led to assume the 
 influence of some dissolved ingredient. In the water percolating through the cave 
 deposits calcium dihydric carbonate occurs in considerable quantity and it is not im- 
 possible that it may play the part required. The only other likely substance is 
 ammonia. At Paviland ammonia is present in the cave earth ; when digging I was 
 often surprised at the rich ammoniacal odour which filled the air ; it is evolved 
 from the animal matter of the bones which are still in course of decomposition. The 
 effect of ammonia on silicates may be observed on old glass windows situated over 
 stables or on the glass of ammonia bottles in a laboratory ; it is very marked and 
 strengthens the suspicion that the patination of flints may be produced by the same 
 reagent. 
 
 To determine this point I exposed three fragments of flint to the action of a 
 concentrated solution of ammonia in a sealed tube which was heated to a 
 temperature not exceeding 140° C. for 24 hours. On removal from the tube they 
 were washed first with hydrochloric acid and then with water. After drying they 
 were all found to have suffered some loss of lustre and one in particular had been very 
 definitely etched. On examining this specimen under the microscope the whitened 
 part was seen to have the form of a delicate broken network with circular meshes 
 filled with the unaltered flint (Fig. 19, No. 1). The same kind of effect is produced 
 by natural weathering, as may be seen by examining the surface of a flint with a blue 
 patina in the region where the patina fades away into the unaltered part, but in 
 this case the network is more complete and more diffuse (Fig. 19, No. 3). The 
 appearance is indeed just such as our examination of sections of the fully weathered 
 crust would lead us to expect, the circular areas of unaltered flint corresponding 
 with the globules we have already mentioned. Chalcedony treated in the same 
 way is also etched, the part attacked being interstitial to the crystalline fibres. 
 
 The experiment affords proof of corrosion, but not necessarily by ammonia, 
 
 c 
 
34 W. J. SoLLAS. — Paviland Cave : An Aurignacian Station in Wales. 
 
 for water alone heated up to 140° C. exerts a corrosive action on silicates. It 
 was therefore necessary to make a control experiment, and some fragments of flint 
 were heated in water up to this temperature for 48 hours in a sealed tube. The 
 glass tube was corroded by this treatment, but not the flints. The experiment, 
 however, is not decisive, since some flints yield so much more readily to solution 
 than others. 
 
 I have also exposed freshly fractured flint to the action of a concentrated 
 solution of ammonia for the space of six months under atmospheric conditions of 
 temperature and pressure, but without result : a similar experiment, made with a 
 decomposing organic solution which evolved ammonia freely, was equally fruitless. 
 
 If we wish to probe the question of patination deeper we must endeavour to 
 obtain some clear ideas on the constitution of flint. As early as 1833 Fuchs 
 proposed to regard flint — and chalcedony, which only differs from it by possessing a 
 fibrous, instead of a granular structure — as a mixture of crystalline silica or quartz 
 and colloidal, or, as he termed it, " amorphous " silica, i.e., opal. Later Rose 
 controverted this view, alleging, erroneously, that flint and chalcedony are indentical 
 in all their most important characters with quartz. The subject was then 
 reinvestigated by Rammelsberg, who confirmed the observations and supported the 
 conclusions of Fuchs, but at the same time showed that the amount of silica 
 dissolved by alkalies out of flint is far in excess of the quantity of amorphous 
 silica which can be present in it, a result which was afterwards emphasized by 
 Renard and Klement, who dissolved as much as 86 per cent, out of a chalk 
 flint containing 97'5 per cent, of silica. These observers also investigated the 
 problem with the aid of the microscope and came to the conclusion that flint 
 consists of amorphous silica intercalated in infinitesimal particles between crystal 
 grains, some which have the character of " chalcedony." 
 
 That one of the constituents of flint is a hydrosol of silica can, I think, admit 
 of no doubt. It is not common opal, for that would be readily detected by staining 
 reagents ; in all probability it is some form of hyalite. 
 
 The effects of heat, long ago observed by Ehrenberg, afford decisive evidence 
 on this point. A simple experiment is to place some thin flakes of flints in 
 platinum foil and then expose them to the hottest part of a Bunsen flame. They 
 do not decrepitate under this treatment, nor is any " crackling " produced, such as 
 results when larger fragments are placed in a fire ; but they completely lose their 
 translucence and acquire a uniform dead- white colour. Chalcedony treated in the 
 same way affords the same results. As a control experiment quartz may be 
 calcined, it will be found to remain absolutely unaffected. The whitening is, of 
 course, accompanied by the loss of water. 
 
 The calcined flakes may be mounted in aniseed oil and examined under the 
 microscope ; they retain their whiteness when illuminated by reflected light, but by 
 transmitted light the whiteness gives place to brown, indicating the presence 
 of a finely porous material like opal. Add to this the significant fact that between 
 crossed Nicols the flake presents precisely the same crystalline granular appearance 
 
W. J. SOLLAS. — Paviland Cave: An Aurignacian Station in Wales. 35 
 
 that it had before ignition,^ and we perceive at once that the flint must consist 
 of two substances, one which is decomposed by heating with the liberation of water 
 — probably hyalite, and another, which suffers no change. 
 
 It is tempting to regard this less alterable substance as quartz. A mixture 
 consisting of one part of hyalite (having a specific gravity of 21 and containing 
 6 per cent, of water) and five parts of quartz, would have a specific gravity of 2*57 
 and contain 1 per cent, of water ; one consisting of one part of similar hyalite and 
 12 parts of quartz would have a specific gravity of 2*61 and contain 0"46 per cent, of 
 water. But the specific gravity of flint ranges from 2"57 to 2*62, and it contains 
 about 1 per cent, of water. Such a mixture would also explain the optical characters 
 of flint so far as they can be observed, and its remarkable solubility in alkali 
 solutions is stated by Eose to be no greater than that of powdered quartz. 
 
 Yet I am by no means convinced that the second substance is actually quartz. 
 Michel Levy and Meunier Chalmas,^ in a masterly study of the crystalline 
 constituent of chalcedony, have shown that it differs from quartz by possessing two 
 optic axes and a birefringence of between 0*009 and 0"01. I have repeated their 
 observations using chalcedony which had been ignited in the blowpipe flame, and 
 obtained similar results. I have also tried to isolate this constituent, hoping to free 
 it from hyalite by treatment with potash or hydrofluoric acid, but none of my 
 attempts have met with success. The effect of hydrofluoric acid on chalcedony is 
 not so much to isolate the crystalline fibres as to corrode the zones of growth, some 
 of which it dissolves more rapidly than others. 
 
 There is room for further investigation, but we may conclude from the fore- 
 going that flint consists of two constituents, one a hydrosol of silica and the other 
 a crystalline form of silicon dioxide. In the process of patination the hydrosol 
 is first attacked, afterwards the crystalline constituent. It is possible that the 
 latter is first hydrated and then dissolved ; in any case it is so finely divided as to 
 yield comparatively rapidly to solvents. 
 
 The Carboniferous chert differs considerably from flint, it rarely becomes 
 patinated, and some of the purer almost colourless varieties are more coarsely 
 grained and possess a higher specific gravity (2*65). They probably consist entirely 
 of quartz. 
 
 Implements of Bone and Ivoky. 
 
 Bits of worked ivory were turned up by the spade in great numbers, but most 
 of them were so soft and full of water that they crumbled at a touch ; some, 
 
 1 This fact ig so important that I have repeatedly heated flint flakes to the highest 
 temperatui'e attainable with a Herepath blowpipe, but in all cases the crystalline constituent 
 remained unchanged. 
 
 2 Michael L6vy and Meunier Chalmas, Bull. Soc. Frangaise de Mineralogie, 1892, torn, xv, 
 pp. 159-190, pis. These distinguished investigators have not taken sufficiently into account 
 the influence of associated silica hydrate on some of the properties of lut6cite. The same 
 oversight has led M. F. Wallerand to make some very curious remarks on the relation of 
 quartzine and quartz. Op. cit., 1897, p. 52. 
 
 c 2 
 
36 W. J. SOLLAS, — Paviland Cave : An Aurignacian Station in Wales. 
 
 however, especially where the cave earth was comparatively dry, were better 
 preserved, and the total number of ivory and bone objects which have been 
 obtained by various explorers from time to time is considerable. 
 
 Rods. — Numerous fragments of the ivory rods described by Buckland as lying 
 adjacent to the skeleton, and some found by ourselves in the same part of the cave, 
 are preserved in the University Museum ; the longest (Plate XXII, Fig. 2c) measures 
 99 mm. in length and 9 mm. in diameter, some of the shorter fragments are thicker 
 (13 mm.), some thinner ; all that I have seen are abruptly truncated by fracture at 
 both ends, but Mr. H. Balfour found one fragment in Miss Talbot's collection with 
 one end preserved : it is slightly swollen and well rounded off. 
 
 The more complete ivory rods found in Aurignacian deposits on the Continent 
 are also, as a rule, broken off at one or both ends ; some taper to a point at one end, 
 others to a chisel edge, and these may have been used as arrow or spear points, but 
 there are many which cannot have served this purpose, they may have been bag 
 handles, sinew twisters, netting pins or bow drills, it is impossible to say. 
 
 Considering the implements with which they were shaped, these ivory rods 
 are remarkably true to a cylindrical form, and the surface is often so smooth as to 
 suggest that the final touches were given by grinding and polishing. They were 
 worked parallel to the grain of the ivory and consequently have now a great 
 tendency to split longitudinally (Plate XXII, Fig. 2d, 2e). 
 
 Awls. — The cave has not yielded any bone needles, which indeed are extremely 
 rare before Magdaleniau times, but several well-shaped and finely pointed bone 
 awls have been found from time to time. One (Plate XXII, Fig. 6) is preserved 
 in the Swansea Museum and two (Plate XXII, Figs. 7 and 9) were obtained during our 
 excavation ; one of these was unfortunately broken by accident immediately after 
 its discovery. It is probable that the old-fashioned method of sewing with an 
 awl, which survived in recent times among the Australians, the Indians of 
 the North American plains, and the Bushmen, was generally practised by the 
 Aurignacian people. 
 
 Ar7)ilet. — The fragments of an ivory armlet found by Buckland are preserved 
 in the University Museum ; their curvature on the inside corresponds with a circle 
 62 mm. in diameter, or about the same as an average-sized silver bangle from 
 Ceylon. The surface is smooth ; the transverse section varies in form and thick- 
 ness, from point to point as shown in the illustration (Plate XXII, Fig. 12). 
 
 Some needless astonishment has been expressed at the industry and patience 
 exercised in carving so slender a ring out of solid ivory. The ancient workman 
 
 FIG. 21. 
 Diagram to sliow how ivoiy lings may be obtained fiom an elepliant's tusk. 
 
W, J. SoLLAS. — Paviland Gave : An Aurignacian Station in Wales. 37 
 
 performed no such feat. The base of the mammoth's tusk is not a solid mass but 
 excavated for the pulp cavity in a hollow cone, and all that is necessary to obtain 
 a ring is to make two parallel cuts with a saw across the tusk in this region (Fig. 21). 
 To remove the edges of this ring with a hollow scraper and to smooth its surface by 
 fine grinding would be a comparatively simple task. 
 
 Smoother. — The Paviland Collection at Oxford contains a beautiful example 
 (Plate XXII, Fig. 1) of a class of implements which are well known under the name 
 of " lissoirs," in the Aurignacian cave-deposits of France and Belgium. 
 
 Buckland describes this as a fragment " nearly of the size and shape of a 
 human tongue . . , ," and continues, " its surface is smooth as if it had been 
 applied to some use (by) which it became polished, and by which 
 the scratches of the coarse knife from which it received its shape f^'\ 
 
 have been nearly obliterated." It should be pointed out that the 
 implement is not a fragment ; a few bits have been chipped off its 
 proximal end, but otherwise it is complete. It measures 87 mm. 
 in length by 40 mm. in breadth, 15 mm. in thickness at the base 
 and 5 mm. at the distal extremity. On each side of the base is an 
 oblique facet, sloping from before backwards and possibly intended 
 to fit the implement to a handle. 
 
 Marroio scoops or spatulas. — Two incomplete implements 
 (Plate XXII, Figs. 4, 4a) carved out of a long bone (radius ?) are 
 exhibited in the Swansea Museum, one, 124 mm. long, retains the 
 pointed end, and the other, 142 mm. long, the handle end 
 uninjured. A third and perfect example is preserved in 
 Miss Talbot's collection at Penrice Castle (Fig. 22). These 
 implements, which were possibly used as marrow scoops, afford 
 the only instance of decorative carving from Paviland Cave.^ 
 
 Perforated teeth. — Three canine teeth of the wolf (Plate XXII, 
 Figs. 11a, lib, lie), perforated for suspension at the base, are in 
 Swansea Museum, and two additional specimens were found by 
 us. Besides these two perforated canines of the reindeer were 
 identified by Mr. Henry Balfour in Miss Talbot's collection. 
 They doubtless contributed, along with perforated shells, to form collection, from 
 a sort of necklace, such as was commonly worn in Aurignacian ^ drawing made 
 times and is still in use among primitive tribes. Balfour ( x i ) 
 
 Truncated hears canine. — We found in our excavation a large 
 bear's canine, from which the cusp seems to have been sawn off so as to expose the 
 pulp cavity. It is just possible that this might have been used as a handle, the 
 pulp cavity serving to receive one end of a splinter of flint. Messrs. Capitan and 
 
 ' The Abb6 Breuil is inclined to think that these objects may be less ancient than the 
 other bone implements. He points out that the bone (cetacean ?) of which they consist is not 
 much altered, and that their form recalls some ancient Magdalenian spatulas (described in his 
 paper, already cited, " Les Sub-divisions dii Paleolithique auperieur," Fig. 25, No. 4). 
 
 FIG. 22. 
 
 Spatula from 
 
 Paviland Cave 
 
 preserved in 
 
 Miss Talbot's 
 
38 W. J. SoLLAS. — Faviland Cave : An Aurignacian Station in Wales. 
 
 Peyrony^ have described a bear's tootli from la Ferrassie, the root of which is 
 incised all over but especially just below the enamel, and they mention a lion's 
 canine from the Euth bearing similar marks. 
 
 Saivn rib. — The bones obtained during our exploration were sent to Oxford ; 
 they were for the greater part in fragments, of which there were many hundreds ; 
 they weighed about 300 lbs. After washing each was separately examined for 
 signs of carving or engraving, but without result ; the most interesting object which 
 rewarded our search was a bit of a rib about an inch long which had been sawn 
 off the parent bone and whicli still retains marks of the tool. The saw was first 
 
 J^"^. 
 
 
 
 FIG. 23. 
 
 No. 1, a bone showing longitudinal striations, due to scraping, and bruised cuts 
 at the upper end, as though it had been used for an anvil ( x 4). 
 
 No. 2 and 2a, a fragment of a sawn rib ; the sawn surface is seen " en face " in 
 the upper part of No. 2, in 2a almost in profile ; a scratch left by an 
 erratic movement of the saw is to be seen just below it ( x f ). 
 
 applied to the face and when it had sunk halfway through, its thickness prevented 
 it proceeding further ; attempts were then made to saw across the edges but without 
 success, for the rib then broke across along the original notch (Fig. 23, No. 2, 2a). 
 
 Bruised hone. — We also found fragments of bones which had been scraped by 
 flint implements, and one of these is marked with bruises (Fig. 23, No. 1), 
 recalling those which have been observed on " compressors " from Mousterian 
 deposits and attributed to pressure applied in the flaking of flints. 
 
 Ivorypendant. — A singular history attaches to this object (Plate XXII, Fig. 3a). 
 In the course of his explorations Buckland found part of a mammoth tusk "in which," 
 he says, " has been formed an irregular cavity about two inches in diameter (similar 
 
 * Capitan and Peyrony, "Station pr6historique de la Ferrassie," Rev. Anthro., 1912, xxii, 
 p. 96, Fig. 34. 
 
W. J. SoLLAS. — Haviland Caoe : An Auricpiacian Station in Wales. 39 
 
 to those effects of ossific inflammation which are produced in recent ivory by gun- 
 shot wounds) and encircled with concentric laminae of bony matter placed obliquely 
 to the grain of the ivory : it is probably the effect of a blow or fracture received 
 while this part of the tusk was yet in its pulpy state and within the socket."^ 
 
 While our digging was in progress an incursion of the sea washed out of the 
 cave earth a curious egg-shaped body (Plate XXII, Fig. 3a), the nature of which 
 was not obvious to inspection. Its general form and appearance suggest a natural 
 growth but it bears evident traces of human workmanship. At one end it is 
 produced into a little process, evidently natural, and below this two plane surfaces 
 have been pared away, so as to reduce its thickness as a preliminary to boring a 
 hole through it. The boring was driven from both sides, two conical cavities being 
 produced which meet in the middle. 
 
 The substance of this pendant looks like ivory, but it does not show the ivory 
 grain. For further information its specific gravity was determined ; it weighs 
 114:"3 grams in air and 5 3 '5 grams in water, its specific gravity as a whole is 
 therefore 1-88, while that of the ivory found in the cave ranges from 2-52 to 2-83, 
 with a mean of 2*6. This discrepancy suggested the presence of some internal 
 cavity ; some minute fragments were therefore removed and their specific gravity 
 determined by means of a diffusion column, it was found to range from 2*52 to 
 2-7, the greater part having a specific gravity of 2*6. This object is therefore of 
 the same nature as the ivory of this cave, a conclusion which a chemical examina- 
 tion confirms. An internal cavity is present with a volume of 16 or 17 c.c, or 
 about one quarter of the volume of the whole (60"8 c.c). 
 
 The outer surface is smooth and polished, but rises here and there into tubercles 
 and tear-shaped ridges or exostoses, some of which have been planed away by a 
 stone implement. The perforated end is stained red with iron oxide. 
 
 On puzzling over this object I concluded that it must be an osseous growth 
 produced by a wound in the pulp cavity of a mammoth's tusk, and then on referring 
 to Buckland's work I came across his similar explanation of an abnormal growth 
 previously quoted. When discussing this with my friend and assistant, Miss 
 Byrne, she informed me that the specimen described by Buckland was in the 
 museum collection (Plate XXII, Fig. 3). We were therefore able to compare them, 
 and found that they tallied to a nicety, the egg-shaped body obviously fitting into 
 the cavity of the injured tusk. Thus, after the lapse of many thousands of years, 
 we are able to bring these objects once again into their natural relations. 
 
 This curious coincidence affords additional evidence of the contemporary 
 existence of Paviland man and the mammoth, and it shows in particular how sound 
 Buckland was in his original judgement that the worked ivory had been obtained 
 from the tusks of the mammoth. 
 
 As to the purpose of the pendant, which would have made an excellent 
 plummet, we may, I think, dismiss the idea that it was used as a net sinker or for 
 
 ' Buckland, torn. cit. 
 
40 W. J. SoLLAS. — Paviland Cave : An Aurigna,cian Station in Wales. 
 
 any mechanical purpose. The walls of the perforation are too fragile to resist 
 any but the slightest strain. Magic powers were probably attributed to so rare and 
 remarkable an object and it might have been suspended in the cave or slung round 
 the neck of the hunter to bring him good luck. 
 
 Composition of the Ivory. 
 
 The ivory and bones of the cave earth generally retain a considerable quantity 
 of organic matter which, on dissolving away the mineral constituents in dilute 
 hydrochloric acid, remains behind as a transparent, fairly consistent mass. 
 
 The specific gravity of this organic residuum, as well as of the bone or ivory as 
 a whole, can be determined on very small quantities by suspension in a diffusion 
 column, which thus affords a simple and elegant means of discovering how far the 
 decomposition of these substances has proceeded. 
 
 The organic matter is only slightly denser than the elastin or ossein from 
 which it IS derived ; thus the fresh elastin of recent ivory has a specific gravity of 
 1-28, the decomposing elastin of the mammoth ivory of Paviland of 1*29. 
 
 If we attribute a specific gravity of 3 to the bone earth, which is the mineral 
 constituent of ivory, then the composition of the ivory will be about 40 per cent, of 
 elastin and 60 per cent, of bone earth, this ratio giving a specific gravity of 
 1'953 while the observed specific gravity is 1'94. 
 
 The specific gravity of Egyptian ivory from tombs of the First Dynasty, i.e., 
 5,000 years old, is 2-4, of some of the mammoth ivory of Paviland 2-545 ; of 
 JS. antiquvs ivory 2-62, and of U. meridionalis 2-87, and the corresponding 
 quantities of elastin are for Egyptian ivory 20 per cent., mammoth ivory 13 per 
 cent., E. antiquus ivory 10 per cent., and U. meridionalis 3 per cent. These results 
 suggest the possibility of determining the age of a specimen of ivory from its specific 
 gravity, but on repeating my observations I find variations of so large an amount 
 that it will be necessary to make further observations. 
 
 The Skeleton. 
 
 The famous " Eed Lady " was found by Buckland on the left (west) side of the 
 cave under 6 inches of earth, but judging from the eroded extremities of some of 
 the bones (radius and ulna) it would seem probable that a part must have been 
 exposed to the air and worn away by the action of water. 
 
 Nearly the whole of the left side of the skeleton was preserved, but as 
 Buckland expressly states, the skull, vertebree, and the extremities of the right side 
 had disappeared. It is worth while to call attention to this statement because a 
 belief prevails that the skull was obtained and was deposited in the University 
 Museum but subsequently lost : so that one of the first inquiries made to a new 
 professor of geology is " Have you found the Paviland skull ? " For this legend 
 
W. J. SOLLAS. — Paviland Cave : An Aurignacian Station in Wales. 41 
 
 the diagrammatic illustration given by Buckland^ is probably responsible, since it 
 represents the skeleton as complete, skull and all. 
 
 The bones are said not to have been disturbed, so that this part of the 
 cave, at least, had not been dug over before Buckland's visit; they were 
 extended " in their natural order of contact " and included the humerus, radius, 
 and ulna of the left arm and the bones of the left leg with the foot " entire to the 
 extremity of the toes," as well as a part of the right foot and many ribs. These, 
 with unimportant exceptions, are to be found in the Museum, which contains, in 
 addition, the greater part of the right fibula, the distal extremity of the right tibia, 
 and a fragment of the left scapula ; but of ribs there are only four fragments, and 
 some of the bones of the left foot are missing. 
 
 Buckland notes the curious fact that in the midst of the bones of the 
 ankle there was a small quantity of a yellow wax-like substance resembling 
 adipocere. 
 
 There is a remarkable parallelism between this interment and those of 
 Mentone ; here as there the body was laid out approximately parallel to the axis 
 of the cave, with no attempt at orientation ; red ochre was profusely scattered over 
 and about it and this pigment still adheres to the bones ; large stones were placed 
 at the head and feet; and various funerary objects, such as ivory rods and 
 ornaments, and little seashells {Natica ncritalis) were buried along with it. 
 
 Miss Byrne found in the material from our excavation the distal end of a left 
 humerus which differs greatly in character from that of the *' Eed Lady." 
 
 Description of the Bones. 
 
 Eor the measurements of the bones given in the following account, as well as 
 for numerous descriptive notes, I am indebted to Mr. F. H. S. Knowles of Oriel 
 College. 
 
 The humerus (Plate XXIII, Figs. 2, 3) is a comparatively slender bone and the 
 muscular attachments are far from well marked ; perhaps this may help to account 
 for Buckland's belief that the skeleton was that of a woman. But as Mr. Knowles 
 points out the articular head is large, the maximum diameter of the articular 
 surface measuring 47 mm., and this points strongly to the male sex,^ a suggestion 
 which is strengthened by the magnitude of the head of the femur (49 mm.), 
 and still more by the head of the tibia (76 mm.). There is thus the strongest 
 presumption that the " Red Lady " was not a woman but a man. 
 
 1 Loc. cit., PI. 21. 
 
 - G. A. Dorsey, Boston Medical Journal, 1897, vol. cxxxvii, p. 80. Dorsey's observations on 
 38 Red Indians of the north-west coast of North America showed that the diameter of the 
 head of the humerus ranged from .39 to 43 ram. in the women and from 44 to 51 mm. in the 
 men ; of the femur from 39 to 43 mm. in the women and from 44 to 54 mm. in the men ; of 
 the tibia from 65 to 70 mm. in the women and from 72 to 86 mm. in the men. For civilized 
 and European races the resuhs were similar. 
 
42 
 
 W. J. SoLLAS. — Paviland Cave : An Aurignacian Station in Wales. 
 
 The dimensions of the humerus are given in the following table, which includes 
 those of the Mentone examples for comparison : — 
 
 Humerus. 
 
 Pav. 
 
 Cav. 
 
 Enf. 
 
 B.G.I. 
 
 B.G.2. 
 
 
 L. 
 
 R. 
 
 R. 
 
 L. 
 
 E. 
 
 L. 
 
 R. 
 
 L. 
 
 Total length 
 
 338 
 
 342 
 
 369 
 
 365 
 
 374 
 
 379 
 
 354 
 
 350 
 
 Breadth of upper end ... 
 
 49 
 
 — 
 
 54 
 
 53 
 
 61 
 
 — 
 
 57 
 
 62 
 
 „ „ lower „ 
 
 56 
 
 — 
 
 66 
 
 66 
 
 70 
 
 68 
 
 64 
 
 65 
 
 „ ,, lower articular surface ... 
 
 47 
 
 — 
 
 49 
 
 47 
 
 53 
 
 47 
 
 — 
 
 — 
 
 „ at the level of the deltoid ... 
 
 20 
 
 — 
 
 27 
 
 23 
 
 33 
 
 26 
 
 20 
 
 29 
 
 Thickness in the middle 
 
 19-4 
 
 — 
 
 26 
 
 22 
 
 27 
 
 23 
 
 24 
 
 28 
 
 Index of thickness 
 
 5-9 
 
 — 
 
 7-3 
 
 6-3 
 
 8-8 
 
 6-9 
 
 5-7 
 
 8-3 
 
 Note. — In this and subsequent tables the following abbreviations are used : — Pav., 
 Paviland ; Oav., Grotte du Cavillon ; Enf., Grotte des Enf ants ; B.G.I, Grotte de la Barma 
 Grande, 1 ; B.G.2, Barma Grande, 2. 
 
 The epiphyses are ossified to the shaft, but the line of demarcation is still 
 evident and the age is therefore not much over twenty-five yeare. 
 
 The ulna (Plate XXIII, Fig. 1), of whicli only the proximal three-quarters is 
 preserved, is comparatively slender, but with well-marked muscular impressions ; 
 the summit of the olecranon shows the commencement of a circular groove for 
 the insertion of the triceps and there is a deep pit for the short supinator muscle 
 (just below the little sigmoid cavity). 
 
 The radius (Plate XXIII, Pig. 1), of which also only the proximal three- 
 quarters is preserved, is a rather slender bone and not much curved ; the bicepital 
 tubercle is large and well marked. The sagittal flattening of the shaft just below 
 the attachment of the round pronator is much less than in the examples from 
 Mentone. In both radius and ulna the epiphyses are ossified to the shaft, an 
 indication that the age was above twenty years. 
 
 The OS innominatum (Plate XXIII, Fig. 4) is described by Mr. Knowles as 
 follows : — 
 
 " The general size and shape of this bone, the pronounced muscular attachments, 
 the large obturator foramen, and the form of the great sacro-sciatic notch, which is 
 narrow and longer along the inferior than the superior margin show that it 
 belonged in all probability to a male subject. The pelvic outlet and inlet, so far as 
 can be judged from one side alone, resemble the male rather than female type." 
 
 The femur (Plate XXIV, Figs, 1, 4) is a strong stout bone with pronounced 
 muscular attachments. The linea aspera forms a prominent ridge ; below and 
 immediately outside its upper and outer bifurcation is a large but shallow depression 
 which corresponds with the hypo trochanteric fossa of I)r. Houze, it is about 
 
W. J. SoLLAS. — Paviland Cave : An Aurignacicui Station in Wales. 43 
 
 60 mm. in length, and 11 mm. in breadth. This is a feature which is very- 
 characteristic of the Cro-Magnon race, though not confined to it. The lesser 
 trochanter is prominent ; but, again agreeing with the Cro-Magnon femur, there is 
 no third trochanter ; as Professor Verneau points out it is not till we enter the 
 neolithic age that a third trochanter becomes at all frequent. 
 
 The epiphyses are ossified with the shaft, but the line of demarcation is still 
 evident, from which we may infer that the subject was not much over twenty-five 
 years of age. 
 
 The dimensions of the femur are as follows : — 
 
 Maximum length in the oblique position 
 Subtrochanteric diameter, sagittal 
 
 ,, „ transverse 
 
 index ' 70-27 
 
 Pilastric diameter, sagittal 
 „ „ transverse 
 
 index 11818 
 
 476 mm. 
 
 27 „ 
 36 „ 
 
 32-5 „ 
 
 27-5 „ 
 
 Popliteal diameter, sagittal 
 „ „ transverse 
 
 index 
 
 73-33 
 
 33 mm. 
 45 „ 
 
 It will be seen from the platymeric and popliteal indexes that the shaft 
 presents a considerable amount of transverse flattening both above (platymery) 
 and below. Comparison with the Cro-Magnon race may be made from the following 
 table : — 
 
 Indices. 
 
 Pav. 
 
 Enf. 
 
 B.G.I. 
 
 B.G.2. 
 
 Men.* 
 
 Mean.t 
 
 
 L. 
 
 R. 
 
 L. 
 
 R. 
 
 L. 
 
 R. 
 
 L. 
 
 
 
 Subtrochanteric 
 
 70-27 
 
 77 
 
 76 
 
 69 
 
 75 
 
 54 
 
 67 ? 
 
 76 
 
 70-6 
 
 Popliteal 
 
 73-33 
 
 73 
 
 79 
 
 81 
 
 83 
 
 71 ? 
 
 83 ? 
 
 — 
 
 78-3 
 
 According to Trofessor Manouvrier platymery commences at 80, is well 
 marked between 75 and 65, and very strong below 65. Thus all the Cro-Magnon 
 femurs are obviously platymeric and the index of the Paviland femur is almost 
 identical with the mean obtained from the Mentone examples. The popliteal 
 flattening is more pronounced in the Paviland than in most of the Mentone femurs, 
 but, as Professor Verneau remarks, these do not present so great a popliteal 
 flattening as some modern Europeans. 
 
 * Men. Femur in Museum of Mentone. 
 t Mean for Grimaldi femurs. 
 
44 
 
 W. J. SOLLAS. — Paviland Cave : An Aurignacian Station in Wales. 
 
 The tibia (Plate XXIV, Fig. 2) shows the same vigorous development of the 
 muscular attachments as the femur. Its dimensions and index are given along 
 with those of the other Cro-Magnon in the table below tibia : — 
 
 
 Pav. 
 
 Cav. 
 
 Enf. 
 
 B.G.I. 
 
 B.G.2. 
 
 Cr6-M. 
 
 
 L. 
 
 R. 
 
 R. 
 
 L. 
 
 R. 
 
 
 L. 
 
 R. 
 
 
 L. 
 
 
 Length 
 
 398 
 
 404 
 
 448 
 
 450 
 
 436 
 
 
 432 
 
 402 
 
 
 398 
 
 — 
 
 Sagittal diameter 
 
 40 
 
 43 
 
 42 
 
 43 
 
 48 
 
 
 47 
 
 52 
 
 
 47 
 
 — 
 
 Transverse „ 
 
 24 
 
 28 
 
 29-5 
 
 29 
 
 28 
 
 
 28 
 
 32 
 
 
 29 
 
 — 
 
 Index of breadth 
 
 60 
 
 651 
 
 70-2 
 
 67-4 
 
 58-3 
 
 
 59-6 
 
 61-5 
 
 
 61-7 
 
 63 
 
 
 68-84 
 
 58-95 
 
 
 61-61 
 
 
 
 The transverse flattening or platycnemia exceeds the mean (63'4) of the 
 Men tone examples, but falls well within the series. 
 
 Professor A. Thomson calls attention to an accessory facet on the anterior 
 side of the lower articular surface. He regards this as associated with the habit of 
 sitting in a squatting posture ; it is common among primitive races, both prehistoric 
 and modern. 
 
 The Jibula (Plate XXIV, Fig. 3) is 388 mm. in length. It resembles in general 
 characters the fibula of the Cro-Magnon race, the longitudinal furrows are deep, 
 the ridge for the insertion of the interosseous ligaments is prominent, but there 
 is no marked torsion in the distal quarter of its length. 
 
 The astragahis and calcaneum (Plate XXIII, Figs. 5, 6) do not present any 
 noticeable peculiarities. 
 
 The Stature of Paviland Man. 
 
 In determining the stature of Paviland man we have the choice of two 
 methods, one that of a distinguished anatomist — Professor Manouvrier,^ the other 
 of a distinguished mathematician — Professor Karl Pearson.^ The latter is becoming 
 increasingly used in England, the former is adopted by Professor Verueau in his 
 monograph on the Grimaldi or Mentone skeletons. We will make use of both 
 and for the sake of comparison we will extend the application of Professor 
 Pearson's method to Professor Verneau's material. • 
 
 * L. Manouvi'ier, " La D6termination de la Taille d'aprds les grands os des membres," 
 Me'moires de la SoeiitS d/ Anthropologic de Paris, 1893, s6r. 2, vol. iv, pp. 347-402. 
 
 ^ K. Pearson, " Mathematical Contributions to the Theory of Evolution. V. The 
 Reconstruction of the Stature of Palfeolithic Races," Phil. Tram., 1899, vol. 192, pp. 169-244. 
 
W. J. SOLLAS. — Pavilaml Gave : An Aurignacian Station in Wales. 45 
 
 The lengths of the long bones which provide us with data are given in the 
 following table : — 
 
 
 Pav. 
 
 Cav. 
 
 Eiif. 
 
 B.G.I. 
 
 B.G.2. 
 
 Cr6-M. 
 
 Ba. 
 
 
 L. 
 
 R 
 
 R. 
 
 L. 
 
 R 
 
 L. 
 
 E. 
 
 L. 
 
 R. 
 
 L. 
 
 
 Humerus 
 
 338 
 
 342 
 
 369 
 
 365 
 
 374 
 
 379 
 
 354 
 
 350 
 
 322 
 
 320 
 
 363 
 
 Femur 
 
 476 
 
 470 
 
 523 
 
 522 
 
 532 
 
 526 
 
 491 
 
 — 
 
 475 
 
 
 — 
 
 Tibia 
 
 398 
 
 404 
 
 448 
 
 450 
 
 436 
 
 432 
 
 402 
 
 398 
 
 395 
 
 — 
 
 — 
 
 The statures inferred are given in the next table : — 
 
 Estimated Stature op Cro-Magnon Men. 
 
 Pav. 
 
 Cav. 
 
 Enf. 
 
 B.G.I. 
 
 B.G.2. 
 
 Cr6-M. 
 
 Ba. 
 
 
 L. 
 
 R. 
 
 R. 
 
 L. 
 
 R. 
 
 L. 
 
 R. 
 
 L. 
 
 R. 
 
 L. 
 
 
 Humerus ... 
 
 1-696 
 
 1-709 
 
 1-774 
 
 1-774 
 
 1-794 
 
 1-814 
 
 1-730 
 
 1-732 
 
 1-637 
 
 1-644 
 
 1-756 
 
 Femur 
 
 1-714 
 
 1-703 
 
 1-802 
 
 1-800 
 
 1-819 
 
 1-807 
 
 1-742 
 
 — 
 
 1-712 
 
 — 
 
 — 
 
 Tibia 
 
 1-732 
 
 1-751 
 
 1-851 
 
 1-861 
 
 1-823 
 
 1-818 
 
 1-742 
 
 1-737 
 
 1-725 
 
 — 
 
 — 
 
 Fe.andTb.(e) 
 
 1-732 
 
 1-732 
 
 1-842 
 
 1-845 
 
 1-839 
 
 1-830 
 
 1-752 
 
 — 
 
 1-725 
 
 — 
 
 — 
 
 » „ (/) 
 
 1-729 
 
 1-728 
 
 1-840 
 
 1-843 
 
 1-838 
 
 1—1 
 
 1-751 
 
 
 1-724 
 
 — 
 
 — 
 
 Means of (e) 
 
 1-8425 
 
 1-8345 
 
 
 
 — 
 
 Manouvrier 
 
 1-735 
 
 1-74 
 
 1-89 
 
 1-88 
 
 1-77 
 
 1-716 
 
 1-80 
 
 Verneau ... 
 
 1-785 
 
 1-79 
 
 1-94 
 
 1-93 
 
 1-82 
 
 1-766 
 
 1-85 
 
 The first six lines in the table give the results obtained by Professor Pearson's 
 formulae; the seventh the results from M. Manouvrier's tables, as calculated 
 by M. Verneau, except for the man of Cro-Magnon calculated by M. Rahon 
 and of Paviland by myself, the last line gives the results obtained by 
 adding 5 cm. to those in the seventh line, as is done by M. Verneau, who justifies 
 this procedure by the following statement : " Je possMe, a I'heure actuelle, 17 
 observations de sujets de grande taille dont on pu mesurer les os longs apres la 
 mort. La taille calculee a I'aide de ces longeurs est toujours infeHeure a la taille 
 prise sur le vivant. L'ecart, assez variable suivant les individus, atteint en 
 moyenne 5 centimetres." 
 
 Without this addition, however, most of the heights obtained by M. Manouvrier's 
 formulae are noticeably greater than those obtained by Professor Pearson's, which we 
 
46 
 
 W. J. SoiiLAS. — Paviland Cave : An Aurignae/ian Station iii Wales. 
 
 have provisionally adopted. If it should prove that we must add the 5 cm, 
 claimed by M. Verneau, our estimates given below will be increased by 2 inches. 
 
 It will be seen that the results calculated by Professor Pearson's formula3 
 from the humerus are, in all instances but one, lower than those obtained from the 
 other bones ; it is the same with those based on the femur when compared with 
 those based on the tibia. We shall discover the explanation of this when we come 
 to consider the proportions of the limbs ; and we shall be led to conclude that the 
 values which make the most probable approach to the truth are those obtained 
 from the femur and tibia taken together and from Professor Pearson's formula {e) 
 rather tlian (/). These give us in round numbers as the mean height of the race 
 5 feet 10 inches, with a maximum just over 6 feet and a minimum of 5 feet 
 8 inches. A tall people even on this estimate, still taller if we accept M. Verneau's 
 results, and offering in any case a striking contrast to the Neandertal men who, 
 from Professor Pearson's calculations based on the Neandertal and Spy skeletons, 
 attained a height of only 5 feet 3-6 inches (1616 mm.). The Chancelade man, with 
 an Eskimo skull, has the stature of an Eskimo, 1575 mm. (Pearson) or 1592 mm. 
 (Manouvrier), and not of a Cro-Magnon man. 
 
 PllOPOKTIONS. 
 
 As we learn from the investigations of M. Verneau the proportions of tho 
 limbs in the Cro-Magnon race differ notably from those of modern Europeans. 
 
 In the lower limb the length of the leg as opposed to the thigh is greater than 
 in most existing races. The relative length of the two segments as expressed by 
 an index (length of tibia x 100/ length of femur) is as follows : — 
 
 Barma Grande 1. 
 
 Barma Grande 2. 
 
 Paviland. 
 
 Gr. des Enfants. 
 
 Cavillon 
 
 81-20 
 
 81-54 
 
 83-83 
 
 85-44 
 
 85-96 
 
 The mean of the Grimaldi examples is 83-53 or almost identical with the 
 index for the Paviland man. In Europeans the index is 79-72 and in Negroes 
 81-33. 
 
 The unfortunate defect of the radius, which has lost its distal extremity, renders 
 it impossible to compare the upper and lower limbs in their entirety; we are 
 reduced, therefore, to a comparison of the femur with the humerus. Making use, 
 as before, of an index (humerus x 100/femur) we have the following: — 
 
 Cr6-Magnon. 
 
 Gr. des Enfants. 
 
 B.G.I. 
 
 Pav. 
 
 B.G.2. 
 
 Cav. 
 
 67-79 
 
 R. 
 
 70-56 
 
 L. 
 69-92 
 
 R. 
 70-3 
 
 L. 
 
 72-05 
 
 L. 
 
 71 
 
 R. 
 72-1 
 
 L. 
 
 R. 
 
 72-77 
 
 
 71-24 
 
 70-17 
 
 
W. J. SOLLAS. — Paviland Cave : An Aurif/nacian Station in Wales. 47 
 
 The mean of the Grimaldi examples is 70'8, again almost identical with the 
 index for the Paviland man. In modern Europeans the index is 77 ; thus in the 
 Cro-Magnon race the humerus is unusually short, and this explains tlie discrepancy 
 between the stature estimated from it (1*696 Paviland), and from the femur and 
 tibia (1-732). 
 
 We have seen that the femur is shorter than usual relatively to the tibia, and 
 this might lead us to suspect that the humerus may be shorter than usual relatively 
 to the radius ; that this is actvially the case is borne out by Prof. Verneau's study 
 of the Grimaldi skeletons. 
 
 Prof. Verneau has also shown that the ratio of the arm as a whole to the leg as 
 a whole is less in Cro-Magnon men than in modern Europeans, though it approaches 
 that in Negroes. 
 
 On surveying our results it is clear that the Paviland man — already shown to 
 be of the same geological age, associated with the same fauna, and to have arrived at 
 the same stage of industrial development, as the best-known examples of the Cro- 
 Magnon race — presents also in all the discoverable characters of his skeleton the 
 the same racial peculiarities. The " Ked Lady of Paviland " is a Cro-Magnon man. 
 He was a little over twenty-five years of age, tall of stature, probably a little 
 taller than the man of Cro-Magnon itself, and about the same height as the man of 
 Cavillon ; but these three examples are the shortest yet discovered within the limits 
 of the race. His lower limbs were longer than usual relatively to his arms, and 
 his leg was disproportionately long compared with his thigh. 
 
 The man of Paviland represents the most westerly outpost of a race which is 
 known to have extended to the east as far as Lautsch and Pf edmost in Moravia and 
 from Belgium on the north through the Dordogne in Prance to the margin of the 
 Mediterranean at Mentone. 
 
 The Fauna. 
 
 The fauna is such as our previous study might lead us to expect ; it is charac- 
 teristically upper palaeolithic, and by the abundance of the horse points especially 
 to the Aurignacian age. 
 
 I have to thank Miss Byrne of Somerville College for identifying the numerous 
 bones which were received at Oxford, and for making a long journey to Swansea in 
 order to study the interesting collection which is preserved there in the museum of 
 the Philosophical Institute. The following list has been drawn up from her 
 catalogue : — ^ 
 
 Equus cahall'us, very common, 
 Ursus spelcBus, „ „ 
 
 Bos primigenius, common. 
 
 1 This last agrees on the whole with that* given by Falconer (H. Falconer, Palcemitological 
 Memoirs, 1868, vol. ii, p. 525), who, however, mentions Cerviis elaphus, but not Megaceros 
 hibernicus, and Bos priscus, but not B. primigenius. He marks some animals as common which 
 we find to be rare, the mammoth, for instance. 
 
48 W. J. SOLLAS. — Paviland Cam : An Awriynacian Station in Wales. 
 
 Rhinoceros tichorhinus, common. 
 Ranujifer tara.ndtis, „ 
 
 Megaceros hibernicus, not common. 
 Ganis liqms, „ „ 
 
 Mephas primiyenius, rare. 
 Hycena spelcea, „ 
 
 To these may be added recent bones of fox, badger, and sheep, together with 
 one tooth of a pig. There are some recent bones of birds, and some which 
 appear to be ancient. Arvicola amphibia, mentioned by Falconer, is also repre- 
 sented. 
 
 The majority of the bones of the Pleistocene mammalia occur as broken 
 fragments, a fact explained by Buckland as due to the repeated diggings of 
 explorers. This is plainly not the case, and there can be no doubt that the bones 
 were broken by the inhabitants of the cave, for the most part with the object of 
 extracting the marrow. Many are covered with scratches such as would be 
 produced in scraping off the flesh with a racloir. There are, besides, some flakes 
 and fragments of bone which have the appearance of having been used as 
 implements, for after a thorough scraping to remove the periosteum they have 
 been subjected to some treatment which has marked the surface with bruised cuts 
 (p. 38). 
 
 As already observed by Buckland none of the bones have been gnawed by 
 hysenas, and this fact, taken in conjunction with tlie presence of broken hysena 
 bones in the cave., would seem to indicate that the ancient hunters were reduced 
 in times of scarcity to feeding on the flesh of this disgusting animal. 
 
 Our account of the Paviland Cave is now concluded. On reviewing our results 
 derived from the industries, the skeleton, and the fauna, we shall not fail to be 
 impressed with their consistency, both among themselves and with the conclusions of 
 archaeologists in other lands. 
 
 The hunters who found shelter in the cave were men of large stature, members 
 of that tall Cro-Magnon race, which occupied the greater part of habitable Europe 
 during the Aurignacian age. They were men of capacious brains, and had made 
 great progress in such simple mechanical arts as are essential to society in its most 
 primitive stage, when subsistence depends wholly on the natural products of the 
 earth, chiefly on roots and fruits which it is the allotted labour of the women to 
 collect, and next on the flesh of animals killed in the chase and contributed by the 
 men to enliven existence by an occasional feast. 
 
 For the fine arts the sojourners in the cave seem to have had little love, they 
 have left no recognizable drawings on ivory or bone, and the red stripes discovered 
 by the Abbe Breuil and myself in the neighbouring cave of Bacon Hole are the 
 only attempts at mural decoration which the race is known to have left behind in 
 Wales. Their persona! adornments were scanty, that their garments were made 
 of prepared furs is probable, but we know nothing of their taste in dress ; they 
 
W. J. SoLLAS. — Paviland Gave : An Aurignacian Station ii,i 'Waif^.: ' ', . _;49 1 
 
 wore necklaces made of wolves' teeth and little sea-shells, as well as bracelets of 
 mammoth's ivory, and they possibly dressed their liair with a pomatum made of fat 
 and red ochre, a primitive but certain protection against vermin. 
 
 They doubtless exercised some magic arts, and they respected their dead 
 suflficiently to provide them with a ceremonial burial. 
 
 Whether the Aurignacians were preceded in their occupation oC the cave by 
 the Mousterians is at least doubtful, such Mousterian implements as have been 
 found being also known in Aurignacian deposits, nor is it certain that they were 
 succeeded by the Solutrians, for the weapons of that war-like race, though found in 
 the cave, prove nothing more than the advent of Solutrian influence. Of any remains 
 which might indicate the presence of the Magdalenians there is no sign. 
 
 Since the disappearance of the Aurignacians the tide of civilization has ebbed 
 and flowed more than once over these islands ; all that it has left stranded in the 
 cave are three copper coins to mark the Eoman occupation and, for the greater glory 
 of our own age, many broken fragments of whiskey bottles. 
 
 But the cave is not the place where the future historian will search for relics 
 of the British Empire, and in fairness to the vanished Aurignacian hunters we must 
 admit that the information we obtain by rummaging their kitchen midden does 
 not exhaust their story, the best part remains untold : could we know the whole, we 
 might have occasion to admire their ingenuity, to applaud their courage and to 
 sympathise with their aspirations towards the ideal. 
 
 DESCRIPTION OF PLATRS 
 Plate XXI. 
 
 Fig. 1. — Genera] view of the approach to the cave from the landward side. The cliff bounds on 
 the right, a steep valley here entering the sea. It is descended by a slope leading to 
 a triangular grassy patch, from which a path, now well trodden, is seen as a black 
 band descending steeply to a rocky ledge. The ledge is well defined from the clift 
 by a chasm which is filled below by the sea. 
 
 Fig, 2. — Final part of the climb from the ledge at a point which may be identified in the first 
 photograph by measuring 0'3 inch to it from the figure seen against the sea. 
 
 Fig. 3. — View of the cave from in front. The foremost figure is standing in the trench which 
 was excavated across the cave in the middle of its length. 
 
 Plate XXII. 
 Objects in bone and ivory from Paviland Cave. 
 A II the figures in this plate are rediiced about |. 
 
 Fig. 1. — Ivory smoother or lissoir. la, seen from above ; lb, seen edgewise, showing one of 
 
 the lateral facets ; Ic, lower surface. 
 Fig. 2. — Fragments of ivory rods found by Buckland, 2a and 2b are comparatively thick, 2c is 
 
 the longest fragment preserved, 2d is split lengthwise, 2e shows the split face of 
 
 another fragment. 
 
 ' d 
 
i *».0V' ' '' '; ,Wv J. ^OlrLAS. — Paviland Cave: An Avriqnacicm. Station in Wales. 
 
 , I '} ,, i > .»<,..- ■. 
 
 Fig. 3. — Part of a niammotli's tusk, found by Buckland in 1821, showing an irregular cavity 
 
 resulting from a wound. Fig. 3a, egg-shaped growth of ivory or dentine formed in 
 
 the cavity of Fig. 3, found in 1913 : it has been pared down on opposite sides at one 
 
 end and perforated for suspension. 
 Fig. 4. — Bone spatulas or marrow scoops. Fig. 4 has lost its point and Fig. 4a its handle. 
 Fig. 5. — Part ot an Aurignacian point with a solid base — non fendu. 
 Fig, 6. — Ivory awl. 
 Fig. 7. — Ivory rod tapering towards both extremities, at one of them it is split lengthwise, but 
 
 there is no evidence to show whether this is by accident or intent. 
 Fig. 8. — Fragment of an ivory point. 
 Fig. 9. — [vory awl : when found this was complete, the elongated shar]) point was broken otl' 
 
 by accident. 
 Fig. 10. — Fragment of an ivoiy plate, scored by a graver. 
 Fig. 11. — Wolves' teeth perforated for suspension. 
 Fig. 12. — Fragments of an ivory amulet placed on a circle of the same curvature : cross sections 
 
 of the fragments are far from constant in form, as is shown by their outlines in the 
 
 interior of the circle. 
 
 Plate XXIII. 
 
 Bones of the " Red Lady " of Paviland. 
 
 The figure.'t are a liule less than half-size ( X about yVo)- 
 
 Fig. 1. — Proximal two-thirds of left radius and ulna. 
 
 Fig. 2. — Left humerus, posterior aspect. 
 
 Fig. 3. — Ditto, anterior aspect. 
 
 Fig. 4. — Os innominatum. 
 
 Fig. 5. — Calcaneum and metatarsals of right foot. 
 
 Fig. 6. — Calcaneum, astragalus, naviculars and cuboid, and four metatarsals of left foot. 
 
 Plate XXIV. 
 
 Bones of the left leg of the " Red Lady " of Paviland. 
 
 A little less than half-size ( x abo^it Yo^i)- 
 
 Fig. 1. — Left femur, seen from behind. 
 Fig. 2. — Left tibia, from behind. 
 Fig. 3.— Left fibula. 
 Fig. 4. — Left femur, seen from in front. 
 
 {^Reprinted in the Journal of the Royal Anthropological Institute, Vol. XLIII, 
 July-Decemher, 1913.] 
 
 'Unrrixon and Sons, Printer-'! in (Jrdinarii to Hix Mnje-ity, St. .Uartms J^nne, London. 
 
Journal of the Royal AnthropoidgidaVl'ri^itute, Vol. XLtTl\ ldl3, Plate XXI. 
 
 O 
 
 Q 
 
 i-i 
 I— t 
 
 (I4 
 
Journal of the Royal Anthropological Institute, fk'l. XLfll^ JP;i3, PJfitf J^^IJ^. 
 
 6 ^ 9 
 
 PAVILAND CAVE. 
 

Journal of the Royal Anthropological Institute 
 
 itute, VoLXI^m, inhVMj,pi:XflI,, 
 
 PAVILAND CAVE. 
 
Journal of the Royal Anthropological Institute, Vol. XLTIV, 191^, "Plate AT^/F. j[J 
 
 PAVILAND CAVE. 
 
MAY 2 
 
 1974 
 
 7 DAY USE 
 
 RETURN TX) 
 
 ANTHROPOLOGY LIBRARY 
 
 This publication is due on the LAST DATF 
 and HOUR stamped betow 
 
 MAY 24 1989 
 
 RB17-40to-8.'72 
 (Q4186S10)4188— A-32 
 
 .General Library 
 
 University of California 
 
 Berkeley