University of Groningen

Early Artefacts from Pakistan? Some Questions for the Excavators Hemingway, Martin F; Stapert, Dick; Dennell, R W

Published in: Current Anthropology

IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from it. Please check the document version below. Document Version Publisher's PDF, also known as Version of record

Publication date: 1989

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA): Hemingway, M. F., Stapert, D., & Dennell, R. W. (1989). Early Artefacts from Pakistan? Some Questions for the Excavators. Current Anthropology, 30(3), 317-322.

Copyright Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).

The publication may also be distributed here under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license. More information can be found on the University of Groningen website: https://www.rug.nl/library/open-access/self-archiving-pure/taverne- amendment.

Take-down policy If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately and investigate your claim.

Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the number of authors shown on this cover page is limited to 10 maximum.

Download date: 30-09-2021

Early Artefacts from Pakistan? Some Questions for the Excavators Author(s): Martin F. Hemingway, Dick Stapert and R. W. Dennell Source: Current Anthropology, Vol. 30, No. 3 (Jun., 1989), pp. 317-322 Published by: The University of Chicago Press on behalf of Wenner-Gren Foundation for Anthropological Research Stable URL: https://www.jstor.org/stable/2743525 Accessed: 29-10-2018 11:07 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at https://about.jstor.org/terms

Wenner-Gren Foundation for Anthropological Research, The University of Chicago Press are collaborating with JSTOR to digitize, preserve and extend access to Current Anthropology

This content downloaded from 129.125.148.19 on Mon, 29 Oct 2018 11:07:31 UTC All use subject to https://about.jstor.org/terms Discussion and Criticism1

Early Artefacts from Pakistan? are very doubtful. A claim for stone tools dating before 2 million years from this area such as that of Dennell, Some Questions for the Rendell, and Hailwood (CA 29:495-98; see also Ren- dell and Dennell I987, Rendell, Hailwood, and Dennell Excavators I987) must be evaluated more stringently than an equiv- alent report from East Africa. From this demanding per-

MARTIN F. HEMINGWAY spective Riwat comes out poorly: (i) The pieces do not Department of Archaeological Sciences, University of occur in a concentration of probably worked pieces but Bradford, Bradford, U.K. 17 VII 88 were extracted from blocks of conglomerate, some more than Io m2, scattered for an undisclosed distance along a From the sediments associated with Tertiary rifting in gully. The worked pieces represent an undisclosed pro- East Africa have come hominid remains and acceptable portion of similar-sized pieces from the locality (Car- concentrations of stone artefacts dating before i.5 mil- mona [Bordes and Viguier 19711 in Spain is comparably lion years. Fruitful localities occur from Barogali in Dji- dubious). (2) The sites occur in a conglomerate in which, bouti in the north (Chavaillon et al. I987) via the Hadar according to Dennell et al., they are not out of sedimen- (Johanson and Edey I98I), Omo (Merrick and Merrick tary context. (3) The sites are not associated with any 1976, Coppens 1977), East and West Turkana (Leakey other evidence of the presence of hominids. and Leakey 1978, Brown et al. I985, Harris et al. I988), In the absence of the three criteria that facilitate ac- Senga V (Harris et al. I987), Olduvai (Leakey I97I), ceptance of the East African localities, Dennell et al. fall Chesowanja (Gowlett et al. I98I), Peninj (Isaac and Cur- back on criteria of technique (cortex as percentage of tis 1974), and Laetoli (Leakey and Harris 1987) to the probable original surface, number of flakes removed, Chiwondo Beds of Malawi (Kaufulu and Stem I987) in number of flaking directions, presence of bulbs of per- the south. The footprints from Laetoli and and the cussion, and subjective evaluation), using these to grade hominid remains from several of these sites are clear the artefacts they have initially extracted from an un- testimony to the presence of bipedal hominids, and scat- known sample. These criteria were established not by ters of stone artefacts at all except Laetoli bear witness study of accepted collections of very early material but to their tool-making ability. by study of other pieces collected in Pakistan from The stone artefacts fulfill several criteria that permit Lower Pleistocene horizons and regarded as ranging from us to accept them as such: (i) Pieces that appear worked "the extremely convincing to the almost certainly natu- occur in concentrations (floors). (2) The claimed worked ral." The use of such a comparative collection will inevi- pieces occur out of sedimentary context in finer-grained tably lead to circularity in attribution. The authors have fluvial and lacustrine sediments, contexts to which they also failed to test their criteria by studying a sample could not have been carried by any natural processes collected from an active high-energy fluvial environ- that can be considered to have been operating at the ment such as can be provided-quite convincingly if one time. (3) The pieces occur in clear association with other wants to be convinced-by the beds of most Pyrenean clear evidence of hominid presence: butchered bone, mountain streams. Without such tests, the criteria em- hominid fossils, probable structures. ployed cannot be accepted as defining a distinct early Because these sites and these associations exist it is hominid conceptual or manipulative set. Some artefacts easier to accept new claimed early occurrences from the from Olduvai would fail on the combination of criteria region, although each must be expected to justify itself. used (see Leakey 1971:26, 36, 74-78), while the major- Beyond sub-Saharan Africa there is no clear evidence ity are much more complex. for any hominids earlier than Homo erectus and no site The Riwat "artefacts," failing on the three external that fulfills the three criteria earlier than i million years criteria, cannot be justified by the use of a priori and b.p., at 'Ubeidiya (Stekelis I966, Stekelis, Bar-Yosef, and untested technological criteria. Schick I969) and Isernia (Salvatori I984), for example. Finally, the evidence cited indicates that the sedi- Other sites in the time range or perhaps earlier, for ex- ments had been subjected to folding and erosion before ample, Vallonnet (de Lumley et al. I963) and Soleilhac i.6 million years, and a date of 2.1-I.9 million years for (Bonifay I986), may fulfill one or two of them, but most the folding event may well be acceptable. Even so, the accumulation of the conglomerate must have pre- i. Permission to reprint items in this section may be obtained only ceded not only the folding event but also the accumula- from their authors. tion of the 65 m of overlying sediments. The normal

3I7

This content downloaded from 129.125.148.19 on Mon, 29 Oct 2018 11:07:31 UTC All use subject to https://about.jstor.org/terms 318 | CURRENT ANTHROPOLOGY

event beneath the conglomerate is not Other in types itself of evidence dated, crucial so for establishing the one must suppose a date for the conglomerate artificial character that of "flaked not stones" are not presented, only predates 2 million years but may for example,predate data iton flakecon- angles (between the ventral siderably. face and the striking-platform remnant or, in the case of The Riwat pieces fail to satisfy the test requirements "cores," between the flake negatives and the striking employed here, and unless such requirements can be platform). Geological processes may produce flakes with satisfied, in the absence of any evidence of hominid pres- angles less than or more than go9, while in man-made ence in Asia so early, the claim that they are artefacts flakes these angles are always greater than go9 (often must be doubted. I00-I 30). Do the striking-platform remnants on flakes show signs of preparation, and what are their metrical attributes? If several flake scars are present on the same DICK STAPERT core (several examples of this are mentioned), what are Biological-Archaeological Institute, Groningen State their metrical attributes? Are they similar in dimen- University, Poststraat 6, 97I2 ER Groningen, The sions and form, as is often the case with Palaeolithic Netherlands. 25 x 88 cores? Do they all have the same surface modifications, proving that they originated at the same time? Dennell, Rendell, and Hailwood present some data to "Ripple marks" are also mentioned as indicating a support their view that five or six unambiguous quartz- hominid origin, but these marks can also be produced ite artefacts dating to ca. 2 million years B.P. have by geological flaking or when stones split as a result of been found in the Soan Valley. In my opinion, however, alternating cooling and heating. With flint, naturally their arguments are not conclusive. I do not wish to produced ripple marks can often be distinguished from exclude the possibility of artefacts of such great antiq- ripple marks on man-made artefacts; this is more uity, but to my mind it has not been proven that the problematical in the case of quartzite pieces. Natural Pakistan pieces are man-made. Only one piece, inter- ripples are often sharp-profiled in cross section, while preted by Dennell et al. as a core, is illustrated in their ripples on flakes produced by man mostly show smooth, report, and the photographs are extremely unclear and sinus-like cross sections (Stapert 1976). On the Soan Val- worthless as documentation of its presumed artificial ley cores, are the ripple marks similar to each other character. within all the flake scars occurring on a core, and what Some criteria that they use to identify artificial objects are their cross sections like? are relatively small remaining cortical areas, scars of Without sufficient data of these kinds, we cannot eval- several removed flakes, the presence of percussion bulbs, uate Dennell et al.'s claim. Serious problems regarding and the existence of several planes in which flakes were the identification of man-made objects are common in removed. None of these criteria is sufficient to permit Lower Palaeolithic research and often underestimated. reliable identification of man-made objects, especially In Europe, one of many examples of this problem is the for the stratigraphic context described for these finds-a material from the Belle Roche Cave in Belgium (Cordy coarse conglomerate, most probably fluviatile in origin. I98o, I98I), which can best be considered as of uncer- Archaeological research in the Netherlands at sites of tain origin (Roebroeks and Stapert I986). The Pakistan the Rhenen industry (Stapert I98I, n.d.) has made clear finds should also be considered "incerto-facts" until that great caution should be exercised in identifying ar- more convincing data on their presumed artificial char- tefacts derived from gravelly river-laid depositsp (in this acter have been produced. case deposits of the River Rhine, dated to 200,000- 300,000 years B.P.). "Flake negatives" and especially "re- touches" are produced by geological processes in enor- Reply mous quantities in gravelly riverbeds, and it is by no means easy to establish in every case whether finds from such contexts are man-made or not. The removal of R.W. DENNELL flakes in several planes does not prove human workman- Department of Archaeology and Prehistory, University ship; geological processes can do the same. Bulbs can of Sheffield, Sheffield SIO 2TN, England. io xi 88 also be produced by geological flaking. Small cortical remnants do not constitute proof of pieces' being man- Hemingway's comments display several misunderstand- made if no further details are offered. Stones present in ings of both the East African early hominid record and river gravels may have had a long geological history, and our work in Pakistan. This is shown most clearly by his "old faces" could have been produced prior to deposition criticism that we used "a priori and untested technologi- of the gravel in which they are now found. Some of these cal criteria" in judging some of the pieces from Riwat faces may not look like "remnants of cortex," but that hominid- rather than geologically flaked. The Oldowan does not mean that they originated as a result of hominid is accepted as hominid-struck because it is clearly differ- activities. In my opinion, therefore, Dennell et al. have ent in flaking and fracturing from geologically flaked not convincingly demonstrated the artificial character of material. It was on those grounds that it was first recog- their finds-which does not necessarily mean, by the nised by Leakey in I935 and defined in terms of direc- way, that these pieces are not artefacts. tionality of flaking in i95i (see Leakey I95I:34-40), long

This content downloaded from 129.125.148.19 on Mon, 29 Oct 2018 11:07:31 UTC All use subject to https://about.jstor.org/terms Volume 30, Number 3, June I989 1 319

before it was found associated with hominid remains, Hemingway is unduly concerned about our collecting probable structures, and so-called living floors. What the strategy at Riwat in I983. The claimed artefacts in ques- discoveries of the early I96os did was (a) to enable Mary tion were "initially extracted from an unknown sample" Leakey to suggest that the Oldowan also included items in the sense that we did not count the total number of that would otherwise have been indistinguishable from clasts exposed in the surface of the conglomerate that we geological material-hammerstones, manuports, anvils, examined. However, I estimate that at least 2,ooo were modified blocks, etc. (Leakey 1971:3-8) and (b) to open inspected. a As stated, only 23 showed any indications of window on early hominid behaviour in general. Even flaking or fracturing: of these, as indicated, only 5 (at though the trend of research over the last i 5 years most has 0.25%) are considered to be extremely convincing been to doubt the strength of association at many of the as instances of hominid flaking. Olduvai and Koobi Fora sites between hominids, butch- He is correct in saying that we did not demonstrate ered bone, and probable structures, the Oldowan is still differences between the claimed artefacts from Riwat recognised as hominid-struck on primarily technological and a sample from a high-energy fluvial environment. and stratigraphic criteria. Conversely, the main reason We have, however, inspected conglomerates at Rohtas, the Kafuan material from the MN horizon at Nsonsezi Dina, and Jalapur, one in the Pabbi Hills, and many in in Uganda (once regarded as arguably second only in the im- Soan Valley since I 98I. These conglomerates can be portance to Olduvai as evidence for early tool making up to 30 m thick and entirely clast-supported. In I983 [see Cole 1954:125]) was discredited was that Clark Ind I985 we inspected as many accessible conglomerate (i95 8) was able to demonstrate on technological grounds ections as we could, stone by stone, in the search for that it could not be distinguished from naturally flaked unambiguous examples of Palaeolithic artefacts (see material. Rendell and Dennell I985). Of the several thousands of In other words, we have simply followed long-estab- clasts that we have examined in those conglomerates, lished practice in attempting to demonstrate that pieces the overwhelming majority are unflaked and typically of flaked stone were modified by hominids and not spherical, sub-spherical, or ovoid; some are split; very by geological agencies. For example, Acheulean bi- occasionally one finds clasts with one or two shallow faces were once regarded as thunderbolts: their accep- flakes detached; signs of battering are also common. tance as artefacts came about not by finding them asso- Clasts with more than three flake scars are extremely ciated with hominids, butchered animals, or structures rare. Whilst this work has no statistical basis, it but with the realisation that they could not have been confirms our view that the pieces under discussion from produced naturally. Using the same procedure, we iden- the Riwat lower conglomerate are highly anomalous. tified handaxes in conglomerate horizons in I983 (see We have also documented in detail the flaking and Rendell and Dennell I985, Dennell I986); no one has fracturing of clasts in the lower conglomerate in Riwat. ever suggested that they were natural because they were In I988, an additional piece of flaked quartzite that we not found in association with hominids, structures, consider to be hominid-struck (fig. I) was found in sec- floors, and butchered bone. tion 40 m from where the piece of which a cast is avail- Hemingway additionally fails to understand our sys- able was found in i983. To demonstrate that these came tem of grading the Riwat artefacts on a scale of con- from the same horizon, this 40-m section was drawn at fidence from o to 5. A major aim of our survey work in a scale of 1:20, and every clast more than 2 cm in size the Pabbi Hills since I986 has been to gather stone tools. was plotted and noted (Rendell and Dennell n.d.). In To date, some 500 pieces of flaked quartzite have been found whilst collecting vertebrate fossil assemblages from erosional surfaces between o.9 and 2.5 million years old. Some pieces-as expected on surface sur- veys-are more convincing than others as examples of hominid-struck stone; the most convincing include hemispherical disc cores, small flakes with no cortex but with positive and negative bulbs of percussion, and cores with multiple flake scars and little cortex. There is no question that some of this material is hominid-struck; what needs clarification at present is its stratigraphic context. The purpose of grading this material after re- cording in detail its flaking characteristics was to isolate N cm localities that looked especially promising for future re- search. The reason for grading the Riwat assemblage the same way in I987 was simply to place it in the wider O 5 CM context of what we had collected since its discovery. Far from introducing an element of circularity, the point of FIG. I. Quartzite flake (R88/i) from lower that ranking was simply to compare the Riwat material conglomerate horizon at Riwat. (For comparison with a with a larger set of data that we had already recorded in utilised light-duty flake from DK, Lower Bed I, detail. Olduvai, see Leakey I97I:fig. i8, no. 3.)

This content downloaded from 129.125.148.19 on Mon, 29 Oct 2018 11:07:31 UTC All use subject to https://about.jstor.org/terms 320 CURRENT ANTHROPOLOGY

all, i,277 were inspected. Of those, io were gritstone visible and (b) that all the flakes must have been de- nodules, 2 were sand casts of bone, and I was a mud ball. tached before incorporation into the gritstone. In I988, Only I had been fractured, and none of the other I,263 we traced the artefact-bearing horizon at Riwat over 5 clasts exhibited any evidence of flaking. Thus the over- kilometres and can confirm that pieces that we regard as whelming majority of clasts in the type of moderate- to hominid-struck and clearly embedded within that con- high-energy conglomerates represented at Riwat are very glomerate are localised to within ioo m of the pieces unlike the pieces under discussion, and pieces regarded found in I983. Hemingway is probably incorrect in as- as hominid-struck form a minute proportion of the total. serting that the conglomerate is "substantially older" In addition, I have recorded cobbles in colluvial con- than 2 million years. As indicated by Rendell, Hailwood, glomerates eroding from the base of Upper Pleistocene and Dennell (i987), the normal event under the con- loess nearby at Riwat. These were near scatters of glomerate is likely to be Reunion rather than Gauss: freshly struck flakes and blades and were recorded to see further palaeomagnetic analyses by Rendell are in hand whether or not these were localised: in other words, one to confirm this. would expect to find evidence of human modification. I am totally at a loss to understand what Hemingway Of a total of 3IO cobbles, 249 were unflaked and 24 were means by-"The sites occur in a conglomerate in which, split; 24 had one flake scar, 2 had two, I had three, and according to Dennell et al., they are not out of sedimen- io were flakes. Almost all clasts are complete, and none tary context," but I assume that he is trying to make could be described as a core or like the piece under dis- some point about the contextual evidence. We would cussion here. certainly not claim that we have found "sites" in the I have also recorded i,2i8 clasts in a conglomerate normal sense of the term. As indicated in Rendell's sec- surface at Site 5 5, Riwat, that was covered by loess tion ca. (Rendell, Hailwood, and Dennell I987: fig. 2), the 45,000 years ago. This particular conglomerate con- conglomerate/gritstone is sedimentary. It is clearly a tained on its surface a fresh quartzite blade, flake, and fluvial deposit associated with cross-bedding, pebble core assemblage. The interest in this particular con- stringers, and pebble orientation and varies laterally over glomerate was that some of the cobbles in its surface a short distance, from a largely clast-free gritstone near might either have been naturally flaked when first incor- where the main piece was found to a predominantly porated into the conglomerate or used as a source of raw clast-supported one 20 m downstream. Whilst the material either when or before the structure was used. claimed artefacts are clearly not in their original context Consequently, cobbles were recorded in detail in respect of discard, they are unlikely to have been transported far, to numbers of flakes removed. There were 700 complete as is evidenced by the freshness of the flake scars. clasts, 20I split ones, and I2 quartered ones. Seventy- He inadvertently makes one useful point with the ref- seven clasts had had one flake removed, 83 two flakes, erences he cites in his first paragraph. The overwhelm- 88 three flakes, 45 four flakes, i i five flakes, and i six ing amount of early hominid research in the last few flakes. The I04 cores and 65 fragments from this site were years has taken place in East Africa. It is worth bearing usually flaked considerably more than six times. In gen- in mind that absence of evidence is not evidence of ab- eral, they were usually flaked in several directions, dis- sence: whereas East African fieldwork can demonstrate played clear bulbs of percussion, and had most of their a marked degree of continuity from I 93 I to today, virtu- cortex removed. In I I cases, they could also be joined to ally all the major field programmes in China, India, and flakes and/or blades. We therefore consider the piece Java were terminated by the outbreak of the war with from the lower conglomerate at Riwat far more like the Japan, and little significant research into Asian hominid archaeological material from Site 5 5 than like the clasts origins occurred until the late I970S. Merely that there from conglomerates that we have examined but not doc- are few additional data yet from Asia is no guarantee umented in detail, from the i,277 clasts in the 40-m that they are not there; at present, it may well reflect the section at Riwat, from the 3IO clasts recorded in collu- lack of fieldwork since I939 . vial deposits eroding from the base of Upper Pleistocene Stapert's comments are useful and informed. I agree loess, and from the I,2i8 recorded in the conglomerate that the photographs are not very satisfactory, but he is at Site 5 5. free to inspect the cast of the one illustrated. His main We feel, therefore, that we can demonstrate significant point concerns the crucial question of differentiating differences between geological flaking and the key items hominid from geological flaking. He does not specify the from the Riwat lower conglomerate assemblage. In addi- type or size of stone(s) making up the gravel deposits he tion, we have shown a cast of this specimen (which mentions from the Rhine, but I assume that they are Hemingway could easily have seen for himself) to any- largely flint/chert, which fractures more easily than one expressing interest. To date, no one who has seen it quartzite, and perhaps smaller on average than the clasts has regarded it as natural; instead, discussion has con- from the Riwat horizon. The Rhine deposits appear to cerned its context and dating. differ substantially from the sections we have inspected. Two points can be raised here on context and dating. Far from finding "enormous quantities" of "flake nega- In I988, a latex-and-plaster mould was made of the socket left after chiseling the main piece (RooI) from the gritstone five years previously. This shows clearly (a) i. I thank Linda Hurcombe for critical comments and P. Smith for that the impression of the artefact in the socket is still some of the raw-data estimates.

This content downloaded from 129.125.148.19 on Mon, 29 Oct 2018 11:07:31 UTC All use subject to https://about.jstor.org/terms Volume 30, Number 3, June I989 1 32I

tives" or "retouches," we have found these in only negli- type ancien dans la region de Carmona (Province de Seville, Es- gible frequencies. As indicated above, flaking of any pagne). Comptes Rendus de l'Academie des Sciences, Paris kind is extremely rare in the fluvial horizons that we 269:I946-47. BROWN, F., J. HARRIS, R. LEAKEY, AND A. WALKER. i985. Early have examined. None of the pieces that we consider ar- Homo erectus skeleton from west Lake Turkana, Kenya. Nature tefacts show signs of thermal fracturing, and we do not 3I6:788-92. feel that this factor can account for the multiple flaking CHAVAILLON, J., J-L. BOISAUBERT, M. FAURE, C. GUERIN, J-L. seen here. MA, B. NICKEL, G. POUPEAU, P. REY, AND S. A. WARSAMA. i987. Le site de d6pecage pl6istocene a Elephas recki de Barogali I thank him for emphasising the importance of regu- (R6publique de Djibouti): Nouveaux r6sultats et datation. larity of flake size and proportion, obliqueness of strik- Comptes Rendus de l'Academie des Sciences, Paris ing angle, and presence of sinuous ripple scars as criteria 305: I259-66. of hominid flaking that we could have included. Details CLARK, J. D. I958. The natural fracture of pebbles from the of piece Rooi (taken from the cast) are as follows: Flake Bakota Gorge, Northern Rhodesia, and its bearing on the Kafuan industries of Africa. Proceedings of the Prehistoric Society sizes (mm) are (Face I) IO7 X IO3*, 32 x 4o, and > 46 x 24:64-77. io5*;(Face2)60 x 78,66 x 64,and35 x 40; (Face3) 52 COLE, S. I954. The prehistory of East Africa. London: Pelican. X 26 and 70 x 74*. Those asterisked have clear bulbs of C*O P P E N S, Y. I 977. Hominid remains from the Plio/Pleistocene percussion and striking angles of I 120, IIO, and 960 re- formations of the Omo Basin, Ethiopia. Journal of Human Evo- lution 6:i69-73. spectively; flake angles on the other five could not be C*O RD Y, J - M. I 980. Le pal6okarst de la Belle Roche (Sprimont, measured as precisely but are more than goo. Length/ Liege): Premier gisement pal6ontologique et arch6ologique de breadth ratios are I.04, o.8o, > 0.44, 0.77, I.03, o.88, 2.0, Pl6istocene moyen en Belgique. Comptes Rendus de and o.95. Ripple marks on Rooi are sinuous and not l'Academie des Sciences, Paris 29i D:749-52. sharp-profiled in cross section; they are also consistent . I98I. D6couverte d'un gisement karstique du Pal6oli- thique inf6rieur a la carriere de la Belle Roche, commune de with one another where they occur. Five of the flakes Sprimont. Activites 8o S.O.S. Fouilles 2:92-98. from Roo I are similar in size and proportion to some of DENNELL, R. W. i986. Reply [to D. J. Ganjool. CURRENT AN- the larger ones in Oldowan assemblages from Bed I and THROPOLOGY 27:I86. Lower Bed II at Olduvai. Maximum lengths of flakes at GOWLETT, J. A. J., J. W. K. HARRIS, D. WALTON, AND B. A. W OOD. I 98 I. Early archaeological sites, hominid remains, and DK, FLK, and FLK North were 64, 65, and 57 mm and traces of fire from Chesowanja, Kenya. Nature 294:I25-29. maximum breadths 36, 8o, and 54 mm; average length/ HARRIS, J. M., F. H. BROWN, M. G. LEAKEY, A. C. WALKER, AND breadth ratios for quartzite flakes from these localities R. E. LEAK E Y. I988. Pliocene and Pleistocene hominid-bearing were I.30, I.I7, and I.3I if end-struck and o.87, 0.77, and sites from west of Lake Turkana, Kenya. Science 239:27-33. o.96 if side-struck (Leakey I97I: 37, 55, 83). The other HARRIS, J. W. K., P. G. WILLIAMSON, J. VERNIERS, M. J. TAP- PEN, K. STEWART, D. HELGREN, J. DE HEINZELIN, N. T. three flakes on Roo I are within the size range of the BOAZ, AND R. V. BELLOMO. I987. Late Pliocene hominid occu- longest (i IO mm, MNK occupation floor) and the widest pation in Central Africa: The setting, context, and character of (II 3 mm, EF-HR) flakes in Upper Middle Bed II (Leakey the Senga sA site, Zaire. Journal of Human Evolution I6:70I- I97I:I55, I36). Striking angles are also similar in that 28. ISAAC, G. LL., AND G. H. CURTIS. I974. Age of earlyAcheulian the Olduvai ones are mostly oblique; Stapert should industries from the Peninj group, Tanzania. Nature 249:624-27. note that striking angles of hominid-produced flakes are JOHANSON, D. C., AND M. A. EDEY. i98i. Lucy: The beginnings not always > goo, as up to io% of the flakes from FLK of humankind. New York: Simon and Schuster. and HWK East at Olduvai had angles of

This content downloaded from 129.125.148.19 on Mon, 29 Oct 2018 11:07:31 UTC All use subject to https://about.jstor.org/terms 3221 CURRENT ANTHROPOLOGY

RENDELL, H. M., E. HAILWOOD, AND R. W. DENNELL. I987. result of more critical analysis. I am especially thinking Palaeomagnetic dating of a two-million-year-old artefact-bear- of multiple burials and burials at various levels in a sin- ing horizon at Riwat, northern Pakistan. Earth and Planetary gle cave. Adornment is no criterion, since this can be- Science Letters 85:488-96. ROEBROEKS, W., AND D. STAPERT. I986. On the "Lower come fossilized by accident as well. If some of the Nean- Palaeolithic" site La Belle Roche: An alternative interpretation. dertal "graves" were to withstand the present critique CURRENT ANTHROPOLOGY 27:369-71. and some of the early Upper Palaeolithic "graves" were SALVATORI, N. 1984. Un Italiano di 700 000 anni fa. Airone to be explained away by referring to taphonomic pro- 40:78-IOI. cesses, then the Middle/Late Palaeolithic transition STAPERT, D. I976. Some natural surface modifications on flint in the Netherlands. Palaeohistoria I8:7-4I. would be more gradual. A major difference is not so . I98I. Archaeological research in the Kwintelooijen Pit, much the symbolic behaviour reflected in burial itself municipality of Rhenen, the Netherlands. Mededelingen Rijks but the presence of adornment, absent with Neandertal Geologische Dienst, n.s., 24: I 39-5 6. man and present in Late Palaeolithic burials from the . n.d. A progress report on the Rhenen industry (central Netherlands) and its stratigraphical context. Palaeohistoria. In Aurignacian onward (for instance, Grotte des Enfants, press. Menton) concurrent with the appearance of rock art as STEKELIS, M. I966. Archaeological excavations at 'Ubeidiya another expression of symbolic thinking. I960-63. Jerusalem: Israel Academy of Sciences and Chase and Dibble (i987) state that "it appears that the Humanities. actual number of cases where intentional burial is more STEKELIS, M., 0. BAR-YOSEF, AND T. SCHICK. I969. Ar- chaeological excavations at 'Ubeidiya I964-66. Jerusalem: or less certain is small compared to the total number of Israel Academy of Science and Humanities. burials found to date" (p. 272). Perhaps the number is even zero and it is all pure taphonomy, just like "can- nibalism" and the "cave bear cult." After a period of upgrading, Neandertal man is now being downgraded on On the Evidence for the modernity scale. In respect to burial it would, how- ever, be interesting to investigate whether other animals Neandertal Burial are also accidentally fossilized more or less complete and articulated and in what numbers. The cave bear (Ur- sus spelaeus) might rank first because of its use of deep L. P. LOUWE KOOIJMANS caves as winter sleeping dens. Are there any other ani- Institut voor Prehistorie, Rijksuniversiteit Leiden, mals on this list, and, if not, can this be explained by Postbus 9515, 2300 RA Leiden, The Netherlands. man's living and/or sleeping in caves, in contrast to ani- 24 x 88 mals? Perhaps some arguments for the purposeful cover- ing of bodies could be found along this line. Gargett (CA 30:I57-77) gives plausible alternative Since "the evidence from Middle Palaeolithic burials taphonomic explanations for the major cases tradi- does not demonstrate the presence of symbolism or cul- tionally presented as proof of-be it incidental- turally defined values" (Chase and Dibble i987:276), we purposeful burial of Neandertal man by his fellow hu- should look to other fields, and I would like to point to man beings. We were prepared for his conclusion by the the sometimes very careful and sophisticated shaping of paper of Chase and Dibble (i 987 ) on the totality of Nean- handaxes, far beyond the practical requirements for a dertal symbolic behaviour, and it is this broader view ready-made, transportable cutting edge and raw-material that one misses here, especially in this anthropological supply. There is no contextual evidence pointing in this journal. direction, but one should not exclude the possibility that For one thing, we should try to discriminate between handaxes as such or especially carefully made specimens the burial or covering of corpses for very practical, for had special meanings in Middle Palaeolithic society. example, hygienic, reasons and burial with ideational There are artifacts showing a beginning of regionalisa- motivation. Burial in itself does not say very much about tion of material culture in the later Middle Palaeo- Palaeolithic man's ideas or level of abstract thinking, lithic-the differentiation of a Central European province and it is precisely here that some of the basic and most with Blattspitzen and a Western European one lacking challenging questions lie. One should first distinguish them. This is a process that becomes more marked intentional burials and then try to identify burials with in the Upper Palaeolithic-again, a gradual rather than "symbolic meaning," and, while it is difficult to identify an abrupt transition. sound archaeological correlates for the latter that work I think that we are confronted with Neandertals as in the Palaeolithic, one should try. Intentionally placed essentially non-modern and that this is of great impor- grave goods (other than adornment) seem to be the best; tance for our study of their cultural remains. This view considerable quantities of red ochre rank second, while is corroborated by the fact that Mousterian interassem- specific, traditional body postures could be a clue, too. blage variation cannot be well understood in terms of Gargett's critical approach to Neandertal burials tends ethnoarchaeological analogy. The organisational and be- to overstress the Middle/Late Palaeolithic contrast when havioural system of Neandertals must have been quite Late Palaeolithic burials are not given similar attention. different from those of modern hunter-gatherers. The so- Quite a number of Late Palaeolithic "burials" might need lution to the Mousterian problem calls for "indepen- to be rejected as such or treated more cautiously as a dent" prehistoric reasoning. Perhaps Neandertals did not

This content downloaded from 129.125.148.19 on Mon, 29 Oct 2018 11:07:31 UTC All use subject to https://about.jstor.org/terms