^{page 75 note 1} Archaeologia, lxii, 449 f., and lxiii, 203 fGoogle Scholar.

^{page 75 note 2} Mr. de Chrustchoff and Mr. Fleming Struthers deserve special credit for having, during the whole'course of operations in 1914 and 1915 respectively, occupied a small and lone cabin on the site itself where the finds had to be temporarily stored under the care of a guardian.

^{page 76 note 1} Just at one spot, n ft. in from the entrance and 7 ft. out from the western wall, a flat rock, standing 5 ft. to 6 ft. above floor-level, was found to be strewn with human refuse. See Archaeologia, lxiii, 205.Google Scholar

^{page 77 note 2} As measured from floor-level.

^{page 77 note 3} See Marett, R. R., Compte-rendu du Congres internat. d’ anthropologie prehistorique de Geneve, 1912, i, 359Google Scholar; also Marett, R. R. and de Gruchy, G. F. B. in Man, 1912, No. 93Google Scholar; reproduced in Bulletin de la Soc. Jersiaise, xxxviii, 326Google Scholar, and in Records of the Past (Washington), 1912, 270Google ScholarPubMed.

^{page 78 note 1} The implementiferous bed along the eastern wall was uniformly about 10 ft. thick, and never more, but its base was 2 ft. above floor-level from about 18 ft. from the entrance inwards.

^{page 78 note 2} See pl. XIII, fig. 2, in which this experimental trench is to be seen.

^{page 78 note 3} See also the detailed account of the work of 1914 in Report of the Brit. Assoc, 1914, 230–4Google Scholar, reproduced, with five plates added, in Bulletin de la Soc. Jersiaisp, xl, 63–9Google Scholar.

^{page 79 note 1} It is just possible, however, that the bed awaits discovery outside the entrance at a lower level than was reached by our excavation, which went no farther down than our conventional floor-level; for 40 ft. away on the other side of the ravine, where Mr. de Gruchy and I found a cave in 1912, Mousterian implements occurred at a point about 8 ft. below the floor-level of the main cave to the north. See Man (1912), No. 93, p. 117Google ScholarPubMed.

^{page 80 note 1} This was on the morning of 2nd September, and, as there was a dangerous fall of rock here a few minutes afterwards, and the roof finally caved in next day, there was no subsequent opportunity of finding out how far down the implementiferous bed went at this point.

^{page 82 note 1} I put forward this hypothesis originally in 1910 (see Archaeologia, lxii, 453Google Scholar) in opposition to a suggestion made in Man (1910), No. 102, p. 185Google ScholarPubMed, that, apart from blocks fallen from the roof, the cave-filling was due to a lateral thrust of rock-rubbish from the ravine to the south. In 1914 we sank an experimental shaft on the north side of the cliff forming the back of the cave, and conjecturally identified the top of the supposed funnel with the very spot that has since fallen in.

^{page 82 note 2} After a tentative excavation in January 1916, which resulted in the removal of about 50 tons of debris from the northern face of the cliff, Mr. Daghorn gave it as his opinion that it would not be feasible to clear out the cave from this side.

^{page 82 note 3} For an account of the operations in 1915, see Report of the Brit. Assoc, 1915, and also the separate and very full account in Bulletin de la Soc. Jersiaise, xli, 154 f.

^{page 82 note 4} All bone found in the cave, with the exception of a few duplicate specimens in the possession of the British Museum of Natural History, is at present in the museum of the Societe Jersiaise in Jersey.

^{page 82 note 5} In Archaeologia, lxii, 456Google Scholar, I ventured to term the owner of the teeth in question Homo Breladensis, seeing that Professor Keith in Journ. Anat. and Physiol. (1911), xlvi, 12 f.Google Scholar, discovered in them an extreme form of primitiveness falling somewhere between the ordinary type of the Neanderthal man and the type of Homo Heidelbergensis. I had no intention, however, of suggesting that we had here anything more than a new variety of man. Hence, if M. Boule insists (see L'Anthropologie, xxii (1911), 675Google Scholar, and xxvi (1915), 36) that such a denomination should be reserved for a specifically distinct kind of man, I am quite willing to withdraw the expression in favour of the term Homo Neanderthalensis, which undoubtedly applies in a broad sense to the Jersey specimen.

^{page 83 note 1} Dr. Keith in subsequent conversation seemed inclined to pronounce judgement somewhat less decidedly. After all, the skull of a child, even if it were one of the Neanderthal type, would be thin in any case; and, for the rest, the original character is bound to be somewhat obscured by the effects of cremation. It may be noted that two interments of Palaeolithic age seem to have been made over a hearth while the fire was still burning, namely, at Mentone and Solutre; cf. Sollas, W. J., Ancient Hunters and their Modern Representatives (2nd ed., 1915), 383Google Scholar.

^{page 84 note 1} In the case of ox, horse, reindeer, and ? rabbit considerable discrepancies occur in the size of teeth and other bones. Thus it may well be that bison coexists with urus, Prjewalski's horse with the ‘forest horse’, caribou with a smaller reindeer, and Arctic hare with rabbit.

^{page 84 note 2} Of the mammals, No. 14 was found in 1914 only; Nos. 1, 8, 13, 16, 17, 18, 19 in 1915 only. Of the birds, No. 26 was obtained in 1914, the others in 1915. Amongst the mammals, reindeer, horse, ox, and lemming are exceedingly common, the rest being more or less rare. The birds are in each case represented by no more than an odd bone or two. Not much stress, however, can be laid on the argument from relative frequency. A mere remnant has survived of the rich store of bone that must obviously have existed formerly; while, of the fragments that remain, comparatively few are determinable, thanks partly to the disintegrating effects of time, and partly to Mousterian man's systematic search for marrow.

^{page 85 note 1} The reindeer and banded lemming, to judge by their present habits, belong on the whole to the tundra, though of course they are migratory. Indeed, the latter is the most characteristic land mammal of the treeless north (cf. Wright, W. B., The Quaternary Ice Age, 222Google Scholar). The extinct mammoth and woolly rhinoceros were steppe and tundra forms, with a predilection for distinctly cold weather. Elephas trogontherii, though the contemporary of E. antiquus, would seem to have had a more Continental and eastern, and therefore a colder, range (cf. Soergel, W., ‘Elephas trogontherii Pohl. und Elephas antiquus Falc’ in Palaeontographica, lx (Stuttgart, 1913).Google Scholar See especially ad fin). The voles belong rather to the steppes. Microtus ratticeps, a large long-skulled mouse with relatively heavy teeth, is still found in Northern Eurasia, but is locally extinct. According to Mr. Hinton, it reached Britain from France during the latter part of the ‘middle terrace’ stage ( Hamilton, G. E. H. Barrett and Hinton, M. A. C., A History of British Mammals, Part xvi (November, 1914), 469Google Scholar). Various small unidentified species of Arvicola occur in the Middle Pleistocene of Grays Thurrock and Ilford, but are not encountered again in Britain until the Ightham horizon, when A. abbotti is numerous (ibid., 477). As regards the shrew (identified by means of a single imperfect skull), it is true that Sorex araneus proper first appears in Britain only in the latest Pleistocene deposits, but an extinct species, almost of the same size, occurs in the ‘middle terrace’ brick-earth of Grays, in Essex (ibid., Part viii (September 1911), 80, 86). The known range of Hyaena spelaea in Pleistocene times accords perfectly with the theory of a prevailing sub-arctic climate, even if it be true that it is but a variety of H. crocnta, now confined to Africa. Of the remaining species of mammals none, in respect either to its habitat or to its place in the time-series, is suspect, with the possible exception of sheep or goat, and rabbit, seeing that their bones likewise occurred high up in the debris, and may consequently be intrusive when found at a lower level. Yet as regards the latter species, at any rate the rabbit is known from Pleistocene deposits both in Britain and on the Continent (ibid., Part x (February 1912), 173), while possibly the present remains are those of the Arctic hare. Of the birds, the two grouse are no longer found in Jersey; the pink-footed goose and the barnacle, though probably visitors, are not definitely recorded; the dipper is rare; and the brent goose, moorhen, and kestrel common. The geese and waders at present breed in the Arctic regions; the ptarmigan is an Arctic or mountain species, while the blackcock has a more temperate though northerly range; the dipper and the kestrel range far north in summer; and the moorhen extends at least as far north as Scandinavia. I owe the information about the birds to my co-worker, Mr. de Gruchy.

^{page 86 note 1} The tooth on which the determination depends is almost entire, thanks to the care with which it was rescued from the bed by my friends Messrs. R. de J. Fleming Struthers and P. H. Brodie, the former of whom afterwards treated it most successfully with gelatine. It is probably the first upper molar. Eleven plates are preserved, the number lost being uncertain. Three wear-surfaces are complete, two partly worn. The width at the wear-surface is 50 mm. The height (at vii) is 156 mm. The enamel is about 2 mm. thick. The length as preserved is 139 mm. Into this length the number of the plates goes about thirteen times. Such a proportion seems too low for Elephas antiquus, but, on the other hand, is too high for E. primigenius. Hence Dr. Andrews is inclined provisionally to assign the tooth to E. trogontherii (Pohlig). The characters of its molars are carefully differentiated by Soergel, W. in Palaeontographica (1913), lx, 6, 7.Google Scholar He notes that this elephant seems to pass into the primitive form of the mammoth from the second interglacial period onwards. For an illuminating monograph on Pleistocene elephants in general, with references to the literature of the subject, see Zuffardi, P. in Palaeontographica Italica (1913), xix, 121.Google Scholar I owe these two references to Dr. Andrews and Prof. Sollas respectively. The exact spot at which the tooth was found on 23rd August 1915 was 18 × 23 × 2 (i.e. feet from the entrance, the W. wall, and floor-level respectively, cf. p. 112). The mammoth teeth, on the other hand, were found on 20th April 1914 at 18 × 39 × 12.

^{page 87 note 1} Thus, at Sirgenstein there are two layers of this character—the lower one occurring immediately over the Mousterian, and the upper corresponding to an Early Magdalenian horizon; while the latter consists of Myodes (= Dicrostonyx) torquatus exclusively, the former yields M. torquatus in association with M. obensis. See Schmidt, R. R., Die diluviale Vorzeib Deutschlands (Stuttgart, 1912), 20, 160.Google Scholar Mr. de Gruchy suggests to me that, since the closely-allied lemming of Norway has the habit of emigrating in huge swarms, herein may lie the clue to the peculiar distribution of the remains of its Pleistocene congener. As regards the place where the lemming remains were found, one cluster occurred at 8 × 39 × 10 and another at 22 × 34 × 12, that is, in each case on the very top of the bed, which rose 2 ft. as it went farther in.

^{page 88 note 1} See Martin, H., Recherches sur l'evolution du Mouste'rien dans le gisement de La Quina (Charente), 1907-1910, vol. i. Industrie osseuse.Google Scholar

^{page 88 note 2} Commont, V., Les Hommes contemporains du Renne dans la Valle'e de la Somme (Amiens, 1914), 134Google Scholar; and id., Compte-rendu du Congres internal, d'anthropologie prehistorique de Geneve, 1912, i. 298Google Scholar.

^{page 88 note 3} All flint or other stone used by man from this cave, with the exception of a type-series of something over 100 specimens presented with the approval of the Société Jersiaise to the British Museum, is at present in the museum of the Société Jersiaise, where full facilities for studying it are provided. The figured specimens are labelled S.J. or B. M. accordingly. Those in Jersey were drawn by Mr. A. H. Barreau, and I cannot thank him enough for the trouble he has taken. Mr. Reginald A. Smith had the others drawn for me at the British Museum, and likewise was kind enough to furnish me with their descriptions, which I reproduce with a few slight modifications.

^{page 90 note 1} For this distinction, see Smith, R. A. in Archaeologia, lxii, 526Google Scholar.

^{page 91 note 1} Commont, , op. cit., 139Google Scholar. The place of origin of the Jersey flint remains an obscure point. On this subject see my remarks in Archaeologia, lxii, 458.Google Scholar

^{page 91 note 2} Cf. Archaeologia, lxii, 457n.Google Scholar

^{page 92 note 1} Cf. Smith, R. A., Archaeologia, lxii, 528Google Scholar.

^{page 94 note 1} The scale of relative sizes is as follows: 2 specimens of 130-140 mm. in length, 1 of 120-130 mm., 1 of 110-120 mm., 5 of 100 no mm., 12 of 90-100 mm., 39 of 8090 mm., 38 of 70-80 mm., 32 of 60-70 mm., 25 of 50-60 mm.

^{page 95 note 1} See Archaeologia, lxii, 464–6Google Scholar, where I assign the industry of La Cotte de St. Ouen to at least a slightly earlier period than that of La Cotte de St. Brelade. It is true that but one coup de poing was found in the former cave, but there are other archaic features, such as the ‘cordiform’ pattern of the flake-implements and the complete absence of later forms.

^{page 97 note 1} In categories A to H all implements are above 40 mm. in length.

^{page 98 note 1} ‘Square’ means merely that length and breadth are approximately equal. See p. 101.

^{page 98 note 2} Below 40 mm. in length.

^{page 100 note 1} See p. 78, and cf. p. 113.

^{page 102 note 1} Compare Sollas, W. J., Ancient Hunters and their Modern Representatives (2nd edit., 1915), 167Google Scholar

^{page 102 note 2} Commont, , op, cit., 173.Google Scholar

^{page 102 note 3} See p. 106.

^{page 104 note 1} The average length and breadth of the four classes are as follows: 268 oval and subtriangular flakes, 31×22 mm.; 243 long flakes, 36×21 mm.; 206 square flakes, 33×32 mm.; 78 sharpened flakes, 35 × 15 mm.

^{page 104 note 2} The broken implements of dwarf size are included in this number.

^{page 105 note 1} Figs. 30 and 31 are average examples. Six of these notched pieces are figured in Bulletin de la Soc. Jers., xl, 68, plate vGoogle Scholar.

^{page 106 note 1} See p. 91.

^{page 106 note 2} This piece is figured in Bulletin de la Soc. Jers., xl, 68, plate vGoogle Scholar.

^{page 106 note 3} See p. 90.

^{page 107 note 1} See p. 103.

^{page 107 note 2} See Archaeologia, lxii, 458Google Scholar.

^{page 107 note 3} Figured, but not very satisfactorily, in Bulletin de la Soc. Jers., xl, 68, plate vGoogle Scholar. It was found on 20th April 1914, at 18 × 30 × 9.

^{page 109 note 1} A petrologist would doubtless distinguish several varieties of stone in what is here classed as greenstone. One rough scraper, for instance, seems to consist of a hard sandstone such as is found at Alderney. Diorite, too, is not uncommon.

^{page 109 note 2} This was found at 22 × 36×8, i.e. near the eastern hearth.

^{page 110 note 1} See Archaeologia, lxii, 465Google Scholar. The stone in question, being both large and flat, was clearly one on which the rubbing was done, corresponding therefore to the quern of a later age rather than to a muller.

^{page 110 note 2} A nest of similar small pebbles was found associated with an interment in the dolmen at Les Monts Grantez, Jersey (see Bulletin de la Soc. Jers., xxxviii, 322Google Scholar), and such a discovery is not without its archaeological parallels in prehistoric Egypt and elsewhere. But it would be rash to found on such a fact a theory that the Mousterian pebbles had a ceremonial value.

^{page 111 note 1} See datum lines given in the plans facing p. 77.

^{page 112 note 1} Exception ought perhaps to be made in the case of the rearward parts of Working A, where along the western wall the bottom of the bed was not easily traced (the finds being somewhat rare and scattered), and may have been in places as much as 4 ft. above floor-level. Cf. Archaeologia, lxiii, 205Google Scholar.

^{page 112 note 2} The sand was very hard, having almost the consistency of sandstone. The late Dr. A. Dunlop, an expert geologist, examined it on the spot, and, having detected in it signs of fine stratification, was of opinion that it had been deposited by an intermittent agency, possibly wind.

^{page 114 note 1} The same succession of a black topped by a brown stratum was noticed in 1914 (see Bulletin de la Soc. Jers., xl (1915), 66Google Scholar), and, since the bottom of the brown layer seemed almost sterile, the finds of the two levels were segregated in the hope that a succession of forms might be detected; but little difference, if any, in the workmanship was to be noticed.

^{page 114 note 2} The contrast will be realized if the two groups of implements drawn by Mr. Barreau are compared. The five lumpish pieces occurred within the same cubic foot at the very bottom of the bed, here about 6 ft. thick, namely, at 20 × 22 × 2, not far from the place where the tooth of Elephas? irogontherii was found, namely, 23 × 22 × 2. The five more elegant and finished specimens likewise occurred together within the same cubic foot at 3 × 37 × 10, namely, at the top of the bed, at this point 10 ft. thick.

^{page 117 note 1} This piece is figured in Bulletin de la Soc. Jers., xl, 68, plate vGoogle Scholar.

^{page 117 note 2} This conjecture was put forward in our report for 1914 printed in Bulletin de la Soc. Jers., xl (1915), 67.Google Scholar I was glad to hear a suggestion to just the same effect made quite independently by Sir Arthur Evans, President of the Society of Antiquaries, when commenting on a specimen submitted by me to the Society on the occasion of reading the present paper.

^{page 117 note 3} See p. 95.

^{page 117 note 3} See p. 86.

^{page 118 note 1} I had the good fortune, when assisting Professor Sollas to excavate the Paviland Cave (see Sollas, W. J. in Journal of the Royal Anthropological Institute, xliii, 325 f.Google Scholar), to handle Aurignacian implements in the mass; and, for comparative purposes, the composite impression thus gained tends to be more helpful than the study of endless books or museum cases where the selected instance predominates.

^{page 118 note 2} I was able to pick out in all a dozen pieces showing a slight tendency towards perpendicular chipping, most of them being thick flakes that could not well be made to yield an edge in any other way. Fig. 10 affords a fair example of such chipping. Figs. 23, 24, and 40 may also be cited as in other respects approximating to Aurignacian types.