Percussion tools in Olduvai Beds I and II (Tanzania): Implications for early human activities

Journal of Anthropological Archaeology 24 (2005) 179–192

www.elsevier.com/locate/jaa

Percussion tools in Olduvai Beds I and II (Tanzania):Implications for early human activities

Rafael Moraa,*, Ignacio de la Torreb

a Area de Prehistoria, Facultad de Letras, Universidad Autonoma de Barcelona, 08193, Bellaterra, Barcelona, Spainb Departamento de Prehistoria, Instituto de Historia, Consejo Superior de Investigaciones Cientıficas, c/Serrano 13, 28001, Madrid, Spain

Received 21 September 2004; revision received 14 December 2004Available online 5 February 2005

Abstract

In this paper, we focus on the description of the Olduvai Bed I and II assemblages linked to percussion activities.Our revision of the sites excavated by Leakey [Excavations in Beds I and II, 1960–1963, vol. 3, 1971] has demonstratedthe dominance that lithic percussion processes—not always related to knapping activities—have in some Oldowan andEarly Acheulean assemblages. We consider the technical sequences that generated the pounded pieces, evaluating theimportance of such percussion processes in the framework of the activities performed at each site and its relevance forthe reconstruction of early human behaviour.� 2005 Elsevier Inc. All rights reserved.

Keywords: Africa; Plio-Pleistocene; Olduvai; Early hominid activities; Percussion tools

Introduction

Mary Leakey�s classification (1971) divided the Old-uvai Bed I and II lithic assemblages into four main cat-egories: tools, utilised material, debitage, andunmodified lithic items (the so-called manuports). Inthe tools group, Leakey (1971) included all the objectsthat in her opinion responded to standardised morpho-logies and were thus classifiable in discrete types ob-tained from intentional knapping, a process that wasalso linked to the debitage category. On the contrary,the utilised materials group encompassed all the itemsthat presented anthropic alterations not necessarily

0278-4165/$ - see front matter � 2005 Elsevier Inc. All rights reserve

doi:10.1016/j.jaa.2004.12.001

* Corresponding author. Fax: +34 93 4020720.E-mail addresses: [email protected] (R. Mora), itorre

@ih.csic.es (I. de la Torre).

linked to lithic knapping, but to the direct use of lithicobjects that had not undergone previous and intentionalmorphological modification.

In this paper, we focus on the description of the Old-uvai Bed I and II assemblages linked to the utilisedmaterial (Leakey, 1971) or pounded pieces category(Isaac et al., 1997). Although Leakey herself (1971) al-ready emphasised the relevance of these objects in theOlduvai assemblages, our reanalysis of some of the Old-uvai assemblages stored in the National Museum ofKenya in Nairobi (Table 1) has revealed certain con-tradictions as regards Leakey�s original classification.Therefore, and given the relevance percussion activitiesseem to have had in some points of the chrono-strati-graphic sequence, we have considered it necessary to re-view the classification system for pounded pieces,stressing the relevance of these artefacts in the activitiesof early humans at Olduvai.

d.

Table 1Stratigraphic and chronological position of the sites analysed at Olduvai

Dates from Bed I follow Walter et al. (1992) and Bed II chronology obeys Manega�s (1993) estimations. See also Blumenschine et al.(2003) for new radiometric dates.

180 R. Mora, I. de la Torre / Journal of Anthropological Archaeology 24 (2005) 179–192

The classification of percussion materials in the Early

Stone Age

Leakey�s (1971) renowned classification for Africanpercussionmaterials basically includes them in the utilisedmaterial category. As regards percussion activities, Lea-key�s utilised materials (1971, p. 7) encompass anvils,hammerstones and cobblestones, nodules and blocks.Subsequent classification systems followed Leakey�s(1971) layout, although they introduced certain varia-tions. Isaac et al. (1997) included the types Leakey had al-ready considered (anvils, hammerstones, modifiedbattered cobbles) in the pounded pieces category, andadded the spheroids and subspheroids, according to Lea-key (1971) pieces that were subjected to intentional shap-ing but which, from Isaac and his collaborators�perspective, were simple hammerstones. This same optionhas been maintained by Clark and Kleindienst (2001),including spheroids and subspheroids in the poundedgroup but not knapped material, therefore modifyingtheir own previous classifications on their role as heavy-duty tools (see Clark and Kleindienst, 1974). At MelkaKunture percussion materials comprise an extremely rel-evant percentage in the Oldowan and Early Acheuleanassemblages classified byChavaillon (1979) into twomaingroups. The former was composed by battered cobblesand hammerstones and the latter by fractured cobbles.

Despite some differences in the classification systems,all these typologies coincide in distinguishing two maingroups in the percussion material, active hammers (clas-sic hammerstones), and passive hammerstones (anvils),regardless of the subtypes and variants each authormay include when analysing the collections. Althoughin Koobi Fora anvils or spheroids are absent or merelyappear incidentally (Isaac et al., 1997), both in theOlduvai sequence (Leakey, 1971; Leakey and Roe, 1994)

and at Melka Kunture (Chavaillon, 1979; Chavaillonand Chavaillon, 1976, 1981) these percussion objectswere extremely profuse, and have in fact been used aschrono-cultural markers to distinguish the Oldowanfrom the Developed Oldowan (i.e., Leakey, 1971, 1975).

Our reanalysis of several of the Olduvai assemblageshas revealed an even greater frequency of percussion ob-jects than the amount established by Leakey (1971).Hence, it is of vital importance to further our knowledgeon these percussion materials and, most importantly, onthe technical processes used to generate them. The mor-phology of many of the objects indicates that these per-cussion materials were not always linked to knappingactivities, but to other work processes. Before verifyingthrough systematic use-wear analysis, it is risky to ven-ture which type of functional activities gave way to thematerials preserved at the sites. Nonetheless, we con-sider that on the basis of an analytical approach, it willat least be possible to discriminate the importance of thebattered items in the assemblages, and if they can beidentified as products of flaking processes or if othertechnical alternatives should be sought. Given theseparameters, and the differentiation between active per-cussion elements (hard objects that transmit a force in-tended to modify another item) and passive percussionelements (hard objects that receive the force transmittedby another item), we present the results obtained in ourstudy of various Olduvai Bed I and II assemblages.

Active percussion elements

Active hammerstones used for knapping activities

The most common active hammers in any Palaeo-lithic archaeological site are always hammerstones used

Fig. 1. Distribution by raw materials of the knapping ham-merstones in some sites of the Olduvai sequence.

Fig. 2. Detail of the battering on a ridge of a hammerstone withfracture angles from site FC West Main Floor.

R. Mora, I. de la Torre / Journal of Anthropological Archaeology 24 (2005) 179–192 181

to modify another lithic object. Typical hammerstonesare natural rounded forms, without intentional humanmodifications and with a weight and morphology thatwould have allowed them to be held. The main featurethat identifies these objects as hammerstones is the pres-ence of areas with extremely concentrated pitting. In fact,the fundamental requisite to identify hammerstones em-ployed for lithic knapping is that, regardless of the sizeand weight, the area of the piece that came into contactwith the other object maintains a compact and homoge-neous structure.Whenhammerstones present fracture an-gles, the area used for knapping is rotated or the piece isdiscarded, since in order to produce a conchoidal fractureon the object, the forcemust be transmitted from the ham-merstone uniformly; this does not occur when hammer-stones start to present fracture points.

In the Olduvai sequence, generally these classic ham-merstones are rounded blanks with ergonomic shapesthat enable their use as hand-held hammerstones. Thepredominant raw materials are always lavas, which ap-pear in greater percentages than quartzes (Fig. 1). Thus,there does seem to be an intentional selection of specificlavas when choosing hammerstones. This selection isprobably related to the nature of the blanks, since mostof the lavas present rounded shapes that denote a fluvialorigin and facilitate their use as knapping hammer-stones. In constrast, quartzes present tabular and angu-lar shapes that are not suitable for this task.

Active hammerstones with fracture angles

Leakey (1971) included in the category of classichammerstones all the pieces she considered poundingitems that did not present intentional shaping. However,in our revision of the collections, we have found a greatnumber of pieces that, although they present the batter-ing typical of active percussion activities, cannot be in-

cluded in the classic hammerstones category. Severalof these pieces can be encompassed under the name ‘‘ac-tive hammerstones with fracture angles.’’

If classic hammerstones are characterised by pittingon cortical areas, round shapes and homogenous cob-bles, hammerstones with fracture angles present batter-ing traces along orthogonal non-cortical planes. Theprocess is as follows: when hitting the object with thehammerstone, the active element is fractured, producingorthogonal planes and irregular ridges. Instead ofreplacing the hammerstone or finding an undamagedarea on the same piece—as occurs with classic hammer-stones—in the case of the active elements, the actualgenerated fracture angles are used to continue bangingthe passive object. Thus, on these active pieces the areaused for percussion becomes completely fractured. Dueto this process, the ridges that formed the fracture planesare abraded by the battering, presenting sinuous edgesalong the ridge�s silhouette (Fig. 2).

Occasionally, it is hard to distinguish these negativesgenerated naturally by percussion activities from thosecreated specifically by knapping. This led Leakey(1971) to classify some pieces as choppers and polyhe-drons which had not really been subjected to knapping,but had instead been fractured by percussion activities.Morphologically, many of the hammerstones with frac-ture angles present very similar planes to those that ap-pear in core forms like choppers. However, the similarityis exclusively morphologic. Upon analysing these piecesmeticulously, we observed that many of them presentfeatures that are not related to the principles of conchoi-dal fracture. Several of the so-called choppers and coresdo not present impact points on the negatives, nor do

182 R. Mora, I. de la Torre / Journal of Anthropological Archaeology 24 (2005) 179–192

they stem from the edge of the piece but from the centralpart of the negative. Furthermore, the scars have irregu-lar shapes without a set directionality, whilst the edgesof the ridges present rims that cannot have been gener-ated using a conventional knapping system. Moreover,they always present step and hinge scars. All these facts,linked to the battering of the ridges and the convex an-gles on the detachments, demonstrate the products weregenerated using activities other than knapping.

Consequently, we can speculate about the functional-ity of this type of hammerstone. We have already statedthat they cannot be the classic active hammerstoneslinked to knapping activities; the objects we are cur-rently describing present a support area (generally a cob-ble) that maintains the original cortical structure, whilstthe opposite area is completely covered in planes andridges generated by intense percussion activities that alsoleave the natural edges intensely battered. These anglesand irregularities on the surfaces affected by percussionshow that they could not have been used as lithic knap-ping hammerstones. There is no specific area on this ac-tive element which, upon coming into contact with thehammered element, could transmit the force uniformlyto generate a conchoidal fracture.

Quite often classic hammerstones ended up breakingafter their use and this led some to state that hammer-stones with fracture angles were simply broken knap-ping hammerstones. Yet, on the pieces describedherein the battering affects previously fractured anglesintensely, meaning that the angles generated by fractureswere employed after they were no longer effective forpercussion activities linked to lithic knapping. More-over, the location where the battering marks appear onthe ridges (see again Fig. 2) seems to indicate that itwas precisely these natural angles created by percussionwhich were used principally to perform the activity.Speculating, and until we can verify by use-wear analy-sis, the most plausible activity to have been performedwith these objects is the chopping of wood, bone orother organic elements. This technical action requiresthe combination of two factors: a force applied severelyand an obtuse dihedral angle that could resist the im-pacts on the material being processed.

Subspheroids, spheroids, and stone balls

Worked stones with spheroid shapes have been tack-led in many studies on the African Early Stone Age. Thepioneering work of Clark (1955) is the most comprehen-sive, and defines stone balls and similar objects as piecesknapped in facets until achieving a spherical shape,which presented intentional battering that reduced theirregularities of the ridges until they became completelyblunt. Later, Kleindienst (1962) established three catego-ries, missiles (practically natural pieces, with isolatedanthropic modifications), polyhedrons (objects with

many facets and negatives) and bolas (quasi-sphericalpieces with a smooth surface obtained by battering pro-cesses). Over subsequent years, successive typologicalproposals (e.g., Clark and Kleindienst, 1974; Leakey,1971) continued to classify spheroids and subspheroidsas instruments with intentional and standardised shapes.Similarly, some authors continued to suggest thesespherical objects were used as missiles (e.g., Leakey,1979), something which has been the object of muchspeculation in subsequent periods (Bingham, 2000; Cal-vin, 2002; Isaac, 1987, etc.).

Throughout the recent literature on the issue, there arecertain disagreements as regards the functionality andway of producing spheroids. Authors such as Schickand Toth (1994), Willoughby (1987) or Sahnouni et al.(1997) consider spheroids acquire their morphology afterbeing used intensely as hammerstones, without any fur-ther technical predetermination. Others such as Wynn(1989) or Texier and Roche (1995), however, conceivethese objects as the end product of an orderly and precon-ceived shaping process. Despite these opposing view-points, all these authors concur that polyhedrons,subspheroids, spheroids, and bolas are different stages ofthe same process.

However, it is not possible to put forward this sce-nario for the production of these pieces at Olduvai.Jones (1994) stresses the fact that most of the polyhe-drons in both Beds I and II were manufactured from la-vas whilst the spheroids and subspheroids were almostinvariably made of quartz. Therefore, they cannot be-long to the same reduction sequence since the raw mate-rials employed in the production of each artefactcategory do not coincide. Jones (1994, pp. 276–277) alsoprovides convincing morphometric arguments, as hedemonstrates that it is impossible for subspheroids tocome from polyhedrons. Upon analysing the size ofboth samples, subspheroids are generally larger thanpolyhedrons, therefore the spheroids could not havebeen produced during a subsequent reduction sequence.

After reanalysing the Olduvai assemblages, our con-clusions are similar to those of Jones (1994). The issueof polyhedrons and their contribution to the differentsites in the Olduvai sequence is complex, since a greatpart of the items Leakey (1971) classified as polyhedronsare nothing of the kind. As occurred with the choppers,Leakey (1971) often used purely morphological criteriato classify polyhedrons. This resulted in a multitude ofchunks being assigned to this category that—albeit pre-senting multiple angles and ridges—had marks createdby natural fractures not by knapping surfaces. As a con-sequence, in many of the so-called polyhedrons the sup-posed extractions do not present negative bulbs, or theseelements are located on the central part of the scar, andthey present impossible angles, natural ridges, etc. (Fig.3). In all, it can be concluded that a high number of theso-called polyhedrons are merely natural irregular

Fig. 3. Examples of chunks devoid of anthropic modificationfrom DK, classified originally as polyhedrons.

R. Mora, I. de la Torre / Journal of Anthropological Archaeology 24 (2005) 179–192 183

chunks. In other cases, some of the objects consideredpolyhedrons can be reclassified as belonging to otherknapping systems with bifacial structures. In any case,most of the polyhedrons we have identified are quitesmall, made of lava and do not present traces of batter-ing. Thus, they do not seem to be related to percussionactivities but with knapping processes. Overall, we arein agreement with Jones (1994) in that we consider thequartz subspheroids and spheroids at Olduvai are froma sequence different to that of the polyhedrons, andshould therefore be described individually.

The first problem encountered upon studying spher-oid forms at Olduvai is the actual distinction betweenpieces modified anthropically and pieces with non-artifi-cial rounded forms. As pointed out by Willoughby(1987), natural spheroids are not rare, generated by dif-ferent processes such as fluvial abrasion, volcanic lapilli

and even spheroid weathering. Therefore some of the

objects classified previously as spheroid artefacts are,in fact, naturally rounded pieces. Conversely, some ofthe objects classified as subspheroids or spheroids are,according to our study, irregular chunks presentingtraces of battering, and not objects that have been useddirectly for percussion activities. These pieces are actu-ally fragments that have been detached by battering.Thus, classifying the small fragments that have comefrom genuine objects used during percussion activitiesas spheroids or subspheroids demonstrates that the fre-quencies of these categories were elevated artificially(Leakey, 1971). It seems that this problem does not ap-pear exclusively in the counts performed in Beds I andII, since Jones (1994) points out that many of the so-called subspheroids in Bed III, Bed IV, and the MasekBeds were merely simple chunks or broken artefacts.

Focussing on the group of quartz objects that wereclassified as polyhedrons, subspheroids, and spheroids,displaying traces of use from activities linked to percus-sion, two different situations are presented. This dichot-omy can be established on the basis of the sedimentaryorigin of the quartzes employed. Although the quartzesused in the Olduvai sites are usually tabular, there arealso (especially in Bed II) quartz cobbles from streams.It can be observed that many of the so-called spheroidsare quartz cobbles with natural rounded shapes. Thesepieces present traces of battering that indicate their func-tion as hammerstones, and it is most probable that theintensity of much of this pitting led to their original clas-sification (Leakey, 1971) as spheroids. Even though theycould ultimately be used for the same tasks as otherspheroids, the morphological genesis process is radicallydifferent to that of tabular quartzes, since quartz cobbleshave a naturally rounded shape.

The processes involved in the production of spheroidshapes from tabular quartz blanks are different nonethe-less. Here the pounding process generated natural facetson the quartz blocks. These facets usually present in-tense battering on the ridges, quite often with simulta-neous extractions on both sides of the edge. Theseobjects represent the first stage (stage 1) of the use ofthe quartz blocks that present multiple facets given per-cussion activities, and actually include the pieces alreadydescribed in the previous section referring to the differ-ent types of hammerstones with fracture angles. Withcontinuous use, battering spread all over the piece andthe ridges collapsed, starting to blur the original shapeof the quartz block. This could be designated stage 2of the reduction process, although it is actually the sameprocess as that of the previous stage, but entailing agreater level of intensity of the percussion (see Fig. 4).Finally, quartz pieces totally rounded by battering havebeen found at Bed II in Olduvai, which could be consid-ered genuine spheroids and compose stage 3 or the finalstage of modification once they have lost their originalshape completely.

Fig. 4. Examples of so-called subspheroids and spheroids fromFLK North (Sandy Conglomerate Level), and diagrams repre-senting the ideal stages of the reduction of the quartz blocks,from an initial battering stage on the ridges with angularfragments, to the final stage envisaging the rounding of thespheroids.

184 R. Mora, I. de la Torre / Journal of Anthropological Archaeology 24 (2005) 179–192

In all, the process described here is the same as theone proposed by Schick and Toth (1994), in which thequartz blocks, after being used as hammerstones, endup taking on a totally round shape. At least as regardsOlduvai, Texier and Roche�s (1995) hypothesis on theseartefacts� shaping is not fulfilled; in Beds I and II therounded shapes of the quartz blocks are obtained viaan extremely intense battering of the artefacts.

It is a different matter to attempt to clarify if theseartefacts have a casual or intentional morphology. Anintermediate solution between Schick and Toth�s(1994) conclusions and Texier and Roche�s (1995)hypotheses could be the one presented by Jones (1994),who—albeit considering that the swiftest manner to ob-tain spheroids is by using them as hammerstones—con-siders it must have been a deliberate option used by the

artisan, in an attempt to produce round shapes suitablefor specific purposes. As Clark (1955) pointed out, thespheroid phenomenon appears throughout the Africancontinent and ranges over a long period of time thatstarts at the Olduvai sites and continues throughoutthe whole of the Acheulean and the Middle Stone Age.This morphological standardisation seems to be indica-tive of a certain interest in attaining perfectly roundedshapes. The fact that the blocks were used for percussionactivities during a certain stage (our stage 1) in which,due to the irregularity of tabular shapes, they couldnot have been used as classic hammerstones, makes ithard to believe that the intense battering processes thatled to the creation of completely spheroid shapes (ourstage 3 with percussion) are linked to lithic knapping.Consequently, we may have to pursue other functionalalternatives to explain these active hammerstones, eventhough they have not been verified by use-wear and sys-tematic experimental analyses.

Passive percussion elements

Passive hammerstones or anvils, i.e., the elementsthat receive the force transmitted by another item, areanother important category in the Olduvai sequence.The description of the Olduvai anvils Leakey offers(1971, p. 7) is still valid, as she considered these piecesas ‘‘cuboid blocks or broken cobblestones with edgesof approximately 90� on which there is battered utilisa-tion, usually including plunging scars.’’ In an attempt tocomplete this description, we could add that in these an-vils, the natural tabular planes of the quartz blocks actas platforms that receive the impacts from the activehammerstones. This is due to the regular surface visibleon these tabular shapes, which allow one of the flat sidesto be used as a percussion platform (A) whilst the oppo-site side (B) is positioned on stable ground. Given thepercussion processes, platform A is full of impact marks,especially by the edges of the block and cause the abrad-ing of the whole periphery of the platform. Platform B,although it does not receive direct impacts, also experi-ences ecailles and fractures given the force transmittedto the block and being in contact with the ground, espe-cially on the edges of the piece. Furthermore, and as aresult of this whole process, the surface of the block ex-posed between the two natural planes (plane C) is alsomodified by percussion, generating numerous scars withhinged and stepped morphologies throughout the wholeperiphery of the block.

The dynamics involved in the modification of theblocks and the generation of anvils is surprisingly similarthroughout the whole of the Bed I and II sequences,being—according to our study—particularly relevantin sites such as FLK North, TK, and FC West. Inall these assemblages, the dominant raw material in

Fig. 5. Examples of quartz anvils in sites TK (A) and FLK North (B).

Fig. 6. Different modalities of the products generated duringthe activities in which anvils are used.

R. Mora, I. de la Torre / Journal of Anthropological Archaeology 24 (2005) 179–192 185

the anvil category is quartz, probably due to the afore-mentioned tabular morphology, which ensures the stabil-ity of the passive element during the percussion process.These tabular quartz anvils (see Fig. 5) vary as regardssize, ranging between 85 (e.g., FLK North Levels 6-1)and 90 mm (e.g., TK Lower and Upper Floors) in lengthand 555 and 733 g, respectively. Thus, they are not espe-cially large pieces and could be handled easily. Conse-quently, although we have observed the presence ofnecessarily static anvils such as MNK (samples weighingover 10 kg) or SHK (with an anvil weighing over 20 kg),in the examples from FLK North, TK or FC West theirsize should not be considered the criterion to distinguishthese objects as passive hammerstones.

The correct identification of the formation process ofthe anvil is important, since the appearance of involun-tary detachments on the surface of the blocks obviouslyimplies the generation of positives detached from theanvils. Our reanalysis has shown that at sites such asFLK North or TK, a high number of the objects classi-fied as flakes or flake fragments are actually positivesspontaneously detached from the anvils due to percus-sion activities and not intentional products from flaking,as thought initially (Leakey, 1971). All these fragmentspresent a series of common characteristics: the firstand most relevant are the traces of battering on theexternal surfaces. Furthermore, the majority of thesepositives do not present a butt or any other attributethat could indicate the direction of the force appliedto obtain the product. Likewise, the dorsal surfaces ofthese positive do not present defined ridges or tracesof previous extractions. Conversely, battering and therepetition of concrete morphologies allow us to link ahigh number of these quartz fragments to percussionactivities and, more specifically, to processes using an-vils and the modifications generated involuntarily onthese pieces. Given the morphological patterns ob-served, we have distinguished several types of what wehave designated ‘‘positives detached from the anvil’’(see Fig. 6).

This reclassification is of paramount importance,since a great number of the small items can be assignedto processes that resulted in the involuntary fragmenta-

186 R. Mora, I. de la Torre / Journal of Anthropological Archaeology 24 (2005) 179–192

tion of the anvils and not to activities involving the pro-duction of flakes. This leads us to reflect on the role ofpercussion processes in the activities performed by hom-inids and the functionality of the studied sites. In thissense, it is interesting to point out that most of the worksthat have attempted to offer an explanation that is eithertypological (i.e., Chavaillon, 1979; Isaac et al., 1997;Leakey, 1971, etc.) or technological (i.e., Sahnouniet al., 1997; Schick and Toth, 1994; Texier and Roche,1995, etc.) regarding percussion artefacts have focussedon the resulting objects (hammerstones, spheroids, an-vils, etc.), but not on the products generated duringthese activities (an exception could be found in Jones,1994). Thus, when the positive fragments are revisedmeticulously, as we have done in sites such as FLKNorth or TK, it has become apparent that many of themstem from the use of anvils and are not related to flakingprocesses.

Finally, we must question the functionality of theseanvils. In the Upper Beds, Leakey and Roe (1994) linkedthe existence of pitted anvils to bipolar knapping and theproduction of outils ecailles (sensu Leakey, 1971). Alongthis same line, Jones (1994) performed replication exper-iments and proposed that both the outils ecailles and thepunches and pitted anvils from Beds III and IV were cre-ated by striking small quartz/quartzite flakes between ananvil and a hammerstone. However, this does not seemto be the case for the sites we have analysed at Beds Iand II: the so-called outils ecailles seem rather like pos-itives with battering detached from the anvils and notactual flakes obtained from a bipolar technique. Fur-thermore, the severe fractures and battering on the Old-uvai anvils do not seem to respond to isolatedmodifications generated by the positioning of a coreon the surface, as required by the bipolar technique. Itmust also be considered that these passive hammers werepart of the reduction sequence linked to the block-on-block or anvil-chipping technique, consisting of strikinga core held in both hands on a fixed anvil on the floor(see, i.e., Kleindienst and Keller, 1976; Shen and Wang,2000). However, we do not believe this to be the case forthe Olduvai anvils either, since in Oldowan sites such asFLK North the flakes obtained are smaller, and inAcheulean assemblages such as TK, the few large flakesseem to have been obtained by direct percussion with ahard hammerstone.

Another alternative could be that the Olduvai anvilshad been used to process small nuts, as documented atother archaeological sites (i.e., Chavaillon and Chavail-lon, 1976; Goren-Inbar et al., 2002) and is widely re-corded in ethological contexts (Boesch and Boesch,1983, 1993; Boesch and Boesch-Achermann, 2000; Mer-cader et al., 2002; etc.). However, it is difficult to assessthis hypothesis for sites such as FLK North or TK, sincethe anvils do not present the typical pits described atMelka Kunture (Chavaillon and Chavaillon, 1976) or

Gesher Benot Ya0aqov (Goren-Inbar et al., 2002). Nev-ertheless, activities related with nut processing shouldnot be discarded, and require further investigation bycomparisons between anvils used by chimpanzees andthe archaeological samples.

As occurs in the examples described in the Sahara(Alimen, 1963) and Ubeidiya (Bar-Yosef and Goren-In-bar, 1993), most of the battering on the anvils studied atOlduvai appears on the contact area between the hori-zontal (platforms A and B) and transversal planes (planeC), where the ridges are completely disfigured by percus-sion. In view of these statements, and given the majoralteration by battering noticeable on many of these an-vils, these fractures must have been generated by muchstronger processes. Therefore, and presenting the Bar-Yosef and Goren-Inbar (1993, p. 110) proposal for theexamples in Ubeidiya, we have established—as ahypothesis—that the majority of the Olduvai anvilscould have been used for interposing elongated elementssuch as bone diaphyses or wood branches between theedge of the anvil and the ground. In doing so the batter-ing would primarily affect the ridge of the anvil andwould unintentionally generate a large number of lithicpositives from the fracture of the passive hammerstone.

Patterns of percussion processes in the Olduvai sequence

Throughout the previous pages we have attempted topresent a meticulous description of the technologicalpatterns involved in percussion activities in Olduvai. Itis now necessary to portray a quantitative assessmentof the percussion items, with a view to evaluating the rel-evance of the specific percussion processes in the frame-work of the activities performed at each site. First, itseems clear that the Olduvai hominids always used lavasand quartzes simultaneously as raw materials for theirpercussion activities. Hence, although the percussionmaterials denote a gradual increase of the relevance ofquartz (Fig. 7), the increase of metamorphic rocks withthe development of the sequence seems proven in all sitesand lithic categories in Bed II, and is therefore not exclu-sive to the items linked to percussion (i.e., Kyara, 1999).

As regards the distribution of raw materials, the jointanalysis of all the percussion categories indicates a gen-eral preference for lavas as knapping hammerstones(Fig. 8). With reference to anvils, they do seem to be clo-sely linked to the availability of quartzes, and this isprobably due to the tabular nature of the piece as statedearlier, which allows their stable positioning on thefloor. The representation of the categories of objectsthroughout the sequence offers interesting patterns.Thus, we have found (Fig. 9) that knapping hammer-stones are always the most abundant pounding artefacts;in EF-HR 100% of the percussion artefacts are classichammerstones, and in DK these objects compose

Fig. 8. Total number and raw materials of active and passivepercussion objects in the analysed Olduvai sequence (DK, FLKZinj, FLK North all levels, FC West, EF-HR, and TK (bothlevels)), excluding the products (chips and fragments) generatedspontaneously during work procedures.

Fig. 9. Absolute frequencies of the different pounded piecescategories in each of the analysed sites.

Fig. 10. Relative frequencies of the different percussion cate-gories in the sites with the greatest variety of pounded pieces.

Fig. 7. Weight in kilograms of the raw materials representedfor the percussion items (including active and passive objects aswell as generated products) from each of the analysed sites.

R. Mora, I. de la Torre / Journal of Anthropological Archaeology 24 (2005) 179–192 187

97.1% of the total, with a very similar pattern to FLKZinj (90%). We found a slightly lower percentage atFC West (72.1%) and FLK North I (Levels 6-1)(63.2%), whilst at TK (both levels) it drops to 54.4%and at FLK North II (Deinotherium Level and SandyConglomerate Level) the rate of classic hammerstonesdecreases to 40.4%. Fig. 9 also shows that with the devel-opment of the sequence, except for EF-HR, differentmodalities of pounded pieces start to accompany classichammerstones at sites, which could be linked to a great-er variety of activities performed at each site.

On the basis of the relative frequencies of the objectswith the greatest variety of categories of pounded piecesat the sites, we have tried to discern a pattern linked to

their distribution (Fig. 10). Willoughby (1987) proposeda functional association between spheroids and anvils, asuggestion collected subsequently (i.e., Texier andRoche, 1995). Unfortunately, our results are not veryenlightening in this respect: although at FC West andTK there is a co-variation in both categories of items(Fig. 10), at FLK North II (Deinotherium and SandyConglomerate Levels), where spheroids are the mostabundant category (47.5%), anvils are very scarce(9.1%).

One aspect that does become very obvious whenstudying the relative frequencies is the strong negativecorrelation between classic hammerstones and spheroids(see Fig. 10). Thus, the percentage of spheroids at FLKNorth I (Levels 6-1) is practically nonexistent (0.5%),whilst there is 63.2% of classic hammerstones. The samepattern occurs at FC West, with 72.1% knapping ham-merstones but not a single spheroid. The opposite occursat FLK North II and TK, where classic hammerstones

Fig. 11. (A) Mean weight of the different categories in severalof the sites. EF-HR excluded given the low number of items. (B)Mean length of the different categories in several of theanalysed sites.

188 R. Mora, I. de la Torre / Journal of Anthropological Archaeology 24 (2005) 179–192

attain their lowest frequencies and spheroids appear inthe highest percentages (47.5 and 28.7%, respectively).

Figs. 11A and, especially, 11B are also very illustra-tive as regards the real nature of the hammerstones withfracture angles; the first of them shows that hammer-stones with fracture angles also have a similar weightto classic hammerstones and spheroids. Fig. 11B denotesan identical co-variation in the mean sizes of the classichammerstones and the hammerstones with fracture an-gles. This variation is most probably due to the size ofthe cobbles available in the environment of each site,and not to the selection performed by the hominidsthemselves. In sum, as occurred with the spheroids, weare dealing with a type of object that is very similar toclassic hammerstones. Actually, they are probably thecontinuity of a process that began with the use of cob-bles as knapping hammerstones which, when theystarted to break, were still used for complementaryactivities. We can observe that in this case the quantita-tive analyses do not provide new information, sincethese types of hammerstones with fracture angles weremade on the same types of blanks as classic hammer-stones. Therefore, the only characteristic that would

differentiate both types of objects would be—if weaccept the arguments set out above—that the batteredridges observed on the hammerstones with fractureangles are not suitable for knapping. Consequently wemust look for another alternative use.

Discussion: Percussion activities in Olduvai

The Oldowan and African Early Acheulean definedin Olduvai have always been considered a paradigm toassess the technical capacity of Plio-Pleistocene homi-nids. Nonetheless, these capacities have been linkedexclusively to the knapping activities described in eachsite. In fact, only some authors (i.e., Chavaillon, 1979)have performed a specific analysis of the percussionproducts in the oldest African archaeological sequences.Remarkably, ethological studies (i.e., Boesch andBoesch-Achermann, 2000; Mercader et al., 2002, etc.)have underlined the significance of percussion processesamongst chimpanzees and the more than probable sim-ilarities with the archaeological record.

Alongside these works on chimpanzees, zoo-archae-ologists have also insisted on the relevance of some per-cussion processes carried out in the oldest sites (i.e.,Binford, 1984; Blumenschine and Selvaggio, 1991;Bunn, 1989; Capaldo and Blumenschine, 1994; Madrigaland Blumenschine, 2000). Thus, bone marrow extractionactivities carried out in Olduvai using percussion pro-cesses are well documented (Blumenschine, 1995; Bunn,1989; Shipman, 1989). In fact, even the existence of boneanvils probably related to this type of bone marrow pro-cessing have been identified (Leakey, 1971; Shipman,1989). However, both experimental studies on the ham-mer-on-anvil technique (Blumenschine and Selvaggio,1991; Blumenschine et al., 1996; Bunn, 1989; Capaldoand Blumenschine, 1994) and the analyses of archaeo-logical materials from the Olduvai fauna (Blumenschine,1995; Bunn, 1982, 1986, 1989; Shipman, 1989) have fo-cussed on marks produced on the bones, but not onthe modifications generated on the lithic materials.

Although a great number of authors have performedstudies on the lithic industries of the Olduvai sequence(i.e., Kimura, 1999, 2002; Ludwig, 1999; Leakey, 1971;Potts, 1988, 1991; etc.), none of them (except perhapsfor Leakey with her typological descriptions) haveunderlined the relevance of percussion processes on thesites. Researchers such as Potts (1988) have insisted onthe scarce incidence of battered artefacts in OlduvaiBed I, where according to this author the poundingpieces would only compose 1–12% of the total, andtherefore, consider bone marrow processing activitiesirrelevant (Potts, 1988, p. 238; contra Binford, 1984).

However, Fig. 12 demonstrates the absolutely funda-mental relevance of percussion activities at the differentsites. Without considering the great amount of kilo-

Fig. 12. Weight in kilograms of the general categories repre-sented at each of the analysed sites. We have chosen torepresent the weight of the general categories and not theabsolute frequencies given the enormous amount of chips insome sites, which at a quantitative level can distort the realimportance of each activity. In fact the complexity of assigningpart of the products to knapping activities or to percussionactivities has led us to present maximum and minimumestimates for objects linked to percussion for several sites(FLK North I, FC West, and TK).

R. Mora, I. de la Torre / Journal of Anthropological Archaeology 24 (2005) 179–192 189

grams of unmodified material—which could in fact notbe linked to human activities in many of the assemblages(de la Torre and Mora, 2005)—we find that the volumeof raw material linked to percussion processes in somesites like TK, FC West or FLK North (all levels) exceedsknapping activities. This enormous abundance of per-cussion processes over knapping activities has neverbeen mentioned in previous studies (i.e., Kimura, 1999,2002; Leakey, 1971; Ludwig, 1999; Potts, 1988, 1991),and leads us to consider both the activities performedby the hominids at these locations and the actual func-tionality of the sites from a radically different perspec-tive. Consequently, in opposition to the ideas proposedby Potts (1991) based on a technology focussing essen-tially on detaching cutting flakes, the production of tools(i.e., from knapping processes) actually had a secondaryimportance in some of the Olduvai sites, which in realityspecialised in the intensive use of artefacts linked topercussion.

We are also aware of the problems raised with thecategorisation of active percussion elements. Fig. 11Ais a perfect example of how these active percussion ele-ments compose a homogenous group that is very distinctfrom anvils from a morphometric perspective. As men-tioned previously, it is relevant to stress that we assumethat the classic hammerstones, hammerstones with frac-ture angles, with battered ridges, spheroids, and even an-vils do not compose discrete morphotypes, and can beelements of the same reduction sequence. Yet this does

not refer solely to these objects; we could find (and thishas actually been documented) cores used previously asanvils. We also have anvils that present typical batteringdenoting their use as active hammerstones. Obviously,there are also objects with uniform pitting linked toknapping activities with planes that present completelyabraded ridges due to a complementary but differentuse.

In summary, the Olduvai artefacts compose a dy-namic sequence in which objects had a polyfunctionaluse and in which the morphotypes identified by archae-ologists were interrelated with one another. Despitethese considerations, we believe a distinction can bemade between different categories based on the stageof use in which the items were abandoned in order todiscriminate the activities performed. We have workedwith this goal in mind when creating this contribution,in this case emphasising the enormous relevance of theprocesses linked to percussion. It is important to empha-sise that at sites such as TK or FLK North I over 100 kgof raw material was used for percussion activities (seeFig. 12), activities which in these assemblages and inothers such as FC West or FLK North II were the mostsignificant documented procedures. Thus, we hope thiswork will help us to clarify the variability of the activi-ties performed by the Plio-Pleistocene artisans: the Old-uvai hominids did not only use lithic material forknapping, they also invested a great amount of the stockof raw material in activities linked to the percussion ofother elements.

Conclusions: Percussion processes and early human

activities

The classic conception of the first human technolo-gies assumes that stone working in the Oldowan andEarly Acheulean focussed on obtaining cutting edges,mainly through knapping flakes (i.e., Toth, 1982). How-ever, this analysis has highlighted the fundamental roleof percussion processes during early human technologi-cal activities. These findings are supported by the recenthypothesis proposed by de Beaune (2004), who stressesthe importance of percussion activities on the historyof human technological evolution. In fact, in her opinionsuch percussion activities could be the most importantones in the most ancient phases of stone working.

The Olduvai example studied in this paper generatesreflections on several levels, such as technological, cul-tural, chronological, and functional concerns. Regard-ing technology, it is fundamental to stress the implicitvariability which characterises early technologies. Thus,it is demonstrated that stone working was not restrictedto obtaining flakes or simply cutting edges but includes awider spectrum of activities as the abundant percussionitems suggest.

190 R. Mora, I. de la Torre / Journal of Anthropological Archaeology 24 (2005) 179–192

In this sense, it is important to ask if the percussionprocesses documented at Olduvai can be extrapolatedto similar chronological and geographical contexts. Itis known that these activities were relevant at MelkaKunture, where Chavaillon and Chavaillon (1976,1981) underlined the importance of active and passivehammerstones on the Oldowan sites. However, in otherareas such as Koobi Fora the same pattern does not oc-cur, and Isaac et al. (1997) were surprised about the ab-sence of percussion items contrasting with thosedocumented at Olduvai. The difference cannot be under-stood through chronological explanations, since theOldowan at Koobi Fora and Olduvai share the sametemporal range. Furthermore, percussion activities atOlduvai are not restricted to the Oldowan in Bed Iand Lower Bed II, as they have been widely documentedin later Acheulean contexts. Thus, any cultural associa-tions linking percussion processes with one or othertechnology are also excluded.

If chronological or cultural differences cannot explainthis pattern, alternative reasons must be sought to ex-plain the abundance of percussion processes at Olduvai.In this paper, we have speculated on a hypothesis linkingpercussion activities with carcass processing. However,this possibility also presents problems, because in someof the sites with more percussion objects such as TK,the bone evidence is very scarce. For this reason, it isimportant to return to the discussion of chimpanzeetechnological activities related with nut-cracking. Thus,vegetal processing could be included as another hypoth-esis in order to explain percussion activities at Olduvai.In addition to the ethological analogy from the nut-cracking behaviour of chimpanzees, the archaeologicalexample from Gesher Benot Ya0aqov—where there arenut seeds associated to anvils (Goren-Inbar et al.,2002)—are also available.

This hypothesis also presents some problems, sinceanvils at Olduvai Beds I and II do not usually presentpits documented in anvils related to nut processing.However, at this stage we cannot rule out the possibilitythat the Olduvai anvils were used in the processing of adifferent kind of vegetal, or even other organic materi-als. . .Until comparative studies are carried out, it is pos-sible to speculate that hominids at Olduvai were usingthe hammer-anvil technique for accessing fruits that, fol-lowing some works (i.e., Peters, 1987), could be a veryrelevant resource in the Olduvai basin during certaintimes of the year (see also Blumenschine and Peters,1998).

On balance, a realistic explanation could be a combi-nation of all these options, with hominids at Olduvaiusing the lithic material for breaking bones, smashingnuts and processing diverse organic elements. New com-parative, experimental, residual analysis and use-wearstudies are needed to give additional information aboutthese processes. For the moment it is relevant to under-

line the absolute importance of percussion processes inmany of the Olduvai sites. These findings make a signif-icant contribution to the discussions on site functionalityand emphasise the need to seek alternative explanationsfor understanding early technological strategies.

Acknowledgments

We would like to thank the Governments of Kenyaand Tanzania for authorising us to research the Olduvaicollections. We are indebted to Jane Hallos for her sty-listic corrections. This work has been financed partiallyby the Estades per a la Recerca a Fora de Catalunya

2001 (BEAI400198) of the Generalitat de Catalunyaand Accion Especial (BHA2002-12145-E) from theSpanish Ministry of Science and Technology.

References

Alimen, M.-H., 1963. Enclumes (percuteurs dormants) asso-ciees a l� Acheuleen superieur de l� Ougartien. Bulletin de laSociete Prehistorique Francaise LX, 43–47.

Bar-Yosef, O., Goren-Inbar, N., 1993. The Lithic Assemblagesof Ubeidiya. A Lower Palaeolithic Site in the Jordan Valley.Quedem, Monographs of the Institute of Archaeology no.34, Jerusalem.

de Beaune, S., 2004. The invention of technology: Prehistoryand cognition. Current Anthropology 45, 139–162.

Binford, L.R., 1984. Faunal Remains from Klasies RiverMouth. Academic Press, Orlando.

Bingham, P.M., 2000. Human evolution and human history: Acomplete theory. Evolutionary Anthropology 9, 248–257.

Blumenschine, R.J., 1995. Percussion marks, tooth marks, andexperimental determinations of the timing of hominid andcarnivore access to long bones at FLK Zinjanthropus,Olduvai Gorge, Tanzania. Journal of Human Evolution 29,21–51.

Blumenschine, R.J., Selvaggio, M.M., 1991. On the marks ofmarrow bone processing by hammerstones and hyenas:Their anatomical patterning and archaeological implica-tions. In: Clark, J.D. (Ed.), Cultural Beginnings:Approaches to Understanding Early Hominid Lifeways inthe African Savannah. Romisch-Germanisches Zentralmu-seum, Mainz, pp. 17–32.

Blumenschine, R.J., Marean, C.W., Capaldo, S.D., 1996. Blindtests of inter-analyst correspondence and accuracy in theidentification of cut marks, percussion marks, an carnivoretooth marks on bone surfaces. Journal of ArchaeologicalScience 23, 493–507.

Blumenschine, R.J., Peters, C.R., 1998. Archaeological predic-tions for hominid land use in the paleo-Olduvai Basin,Tanzania, during lowermost Bed II times. Journal ofHuman Evolution 340, 565–607.

Blumenschine, R.J., Peters, C.R., Masao, F.T., Clarke, R.J.,Deino, A.L., Hay, R.L., Swisher, C.C., Stanistreet, I.G.,Ashley, G.M., McHenry, L.J., Sikes, N., van der Merwe,N.J., Tactikos, J.C., Cushing, A.E., Deocampo, D.M.,

R. Mora, I. de la Torre / Journal of Anthropological Archaeology 24 (2005) 179–192 191

Njau, J., Ebert, J., 2003. Late Pliocene Homo and Hominidland use from Western Olduvai Gorge, Tanzania. Science299, 1217–1221.

Boesch, C., Boesch, H., 1983. Optimization of nut-crackingwith natural hammers by wild chimpanzees. Behaviour 83,265–286.

Boesch, C., Boesch, H., 1993. Diversity of tool use and tool-making in wild chimpanzees. In: Berthelet, A., Chavaillon,J. (Eds.), The Use of Tools by Human and Non-humanPrimates. Clarendon Press, Oxford, pp. 158–168.

Boesch, C., Boesch-Achermann, H., 2000. The Chimpanzees ofthe Taı Forest. Behavioural Ecology and Evolution. OxfordUniversity Press, Oxford.

Bunn, H.T., 1982. Meat-eating and human evolution: Studieson the diet and subsistence patterns of Plio-PleistoceneHominids in East Africa. Unpublished PhD, University ofCalifornia, Berkeley.

Bunn, H.T., 1986. Patterns of skeletal representation andhominid subsistence activities at Olduvai Gorge, Tanzania,and Koobi Fora, Kenya. Journal of Human Evolution 15,673–690.

Bunn, H.T., 1989. Diagnosing Plio-Pleistocene Hominid activ-ity with bone fracture evidence. In: Bonnichsen, R., Sorg,M.H. (Eds.), Bone Modification. Center for the Study of theFirst Americans, Orono, pp. 299–315.

Calvin, W.H., 2002. A Brain for All Seasons: Human Evolution& Abrupt Climate Change. University of Chicago Press,Chicago.

Capaldo, S.D., Blumenschine, R.J., 1994. A quantitativediagnosis of notches made by hammerstone percussionand carnivore gnawing on bovid long bones. AmericanAntiquity 59, 724–748.

Chavaillon, J., 1979. Essai pour une typologie du materiel depercussion. Bulletin de la Societe Prehistorique Francaise76, 230–233.

Chavaillon, J., Chavaillon, N., 1976. Le Paleolithique ancien enEthiopie. Caracteres techniques de l� Oldowayen de Gom-bore I a Melka-Konture. In: Clark, J.D., Isaac, G.L. (Eds.),Colloque V: Les Plus Anciennes Industries En Afrique.UISPP IX Congres, Nice, pp. 43–69.

Chavaillon, J., Chavaillon, N., 1981. Galets amenages etnucleus du Paleolithique Inferieur. In: Roubet, C., Hugot,H.-J., Souville, G. (Eds.), Prehistoire Africaine, MelangesAu Doyen Lionel Balout. ADPF, Paris, pp. 283–292.

Clark, J.D., 1955. The stone ball: Its associations and use byprehistoric man in Africa. In: Balout, L. (Ed.), II CongresPanafricain De Prehistoire, Alger. Actes De La IIe Session.Arts et Metieres graphiqies, Paris, pp. 403–417.

Clark, J.D., Kleindienst, M.R., 1974. The Stone Age culturalsequence: Terminology, typology, and raw material. In:Clark, J.D. (Ed.), Kalambo Falls Prehistoric Site, vol. II.Cambridge University Press, Cambridge, pp. 71–106.

Clark, J.D., Kleindienst, M.R., 2001. The Stone Age culturalsequence: Terminology, typology and raw material. In:Clark, J.D. (Ed.), Kalambo Falls Prehistoric Site, Vol. III:The Earlier Cultures: Middle and Earlier Stone Age.Cambridge University Press, Cambridge, pp. 34–65.

Goren-Inbar, N., Sharon, G., Melamed, Y., Kislev, M.E., 2002.Nuts, nut cracking, and pitted stones at Gesher BenotYa0aqov, Israel. Proceedings of the National Academy ofSciences of the United States of America 99, 2455–2460.

Isaac, B., 1987. Throwing and human evolution. The AfricanArchaeological Review 5, 3–17.

Isaac, G.L., Harris, J.W.K., Kroll, E.M., 1997. The stoneartefact assemblages: A comparative study. In: Isaac, G.L.(Ed.), Koobi Fora Research Project. Volume 5: Plio-Pleistocene Archaeology. Oxford University Press, Oxford,pp. 262–362.

Jones, P.R., 1994. Results of experimental work in relation tothe stone industries of Olduvai Gorge. In: Leakey, M.D.,Roe, D.A. (Eds.), Olduvai Gorge. Volume 5. Excavations inBeds III, IV and the Masek Beds, 1968–1971. CambridgeUniversity Press, Cambridge, pp. 254–298.

Kimura, Y., 1999. Tool-using strategies by early hominids atBed II, Olduvai Gorge, Tanzania. Journal of HumanEvolution 37, 807–831.

Kimura, Y., 2002. Examining time trends in the Oldowantechnology at Beds I and II, Olduvai Gorge. Journal ofHuman Evolution 43, 291–321.

Kleindienst, M.R., 1962. Component of the East Africanacheulian assemblage: An analytic approach. In: Mortel-mans, G., Nenquin, J. (Eds.), Actes Du IV CongresPanafricain De Prehistoire Et De L� Etude Du Quaternaire,Leopoldville, 1959. Belgie Annalen, Musee Royal de l�Afrique Centrale, Tervuren, pp. 81–108.

Kleindienst, M.R., Keller, C.M., 1976. Towards a functionalanalysis of handaxes and cleavers: The evidence fromEastern Africa. Man 11, 176–187.

Kyara, O.A., 1999. Lithic Raw Materials and Their Implica-tions on Assemblage Variation and Hominid BehaviorDuring Bed II, Olduvai Gorge, Tanzania. UnpublishedPh.D. University of Rutgers, New Brunswick.

Leakey,M.D., 1971. Olduvai Gorge, Vol. 3. Excavations in BedsI and II, 1960–1963.CambridgeUniversityPress,Cambridge.

Leakey, M.D., 1975. Cultural patterns in the Olduvai sequence.In: Butzer, K.W., Isaac, G.L. (Eds.), After the Australop-ithecines. Stratigraphy, Ecology, and Cultural Change in theMiddle Pleistocene. Mouton, Chicago, pp. 477–493.

Leakey, M.D., 1979. Olduvai Gorge: My Search for Early Man.Williams Collins, London.

Leakey, M.D., Roe, D.A. (Eds.), 1994. Olduvai Gorge. Volume5. Excavations in Beds III, IV and the Masek Beds, 1968–1971. Cambridge University Press, Cambridge.

Ludwig, B.V., 1999. A technological reassessment of EastAfrican Plio-Pleistocene lithic artifact assemblages. Unpub-lished PhD, University of Rutgers, New Brunswick.

Madrigal, T.C., Blumenschine, R.J., 2000. Preferential process-ing of high return rate marrow bones by Oldowan Hom-inids: A comment on Lupo. Journal of ArchaeologicalScience 27, 739–741.

Manega, P.C., 1993. Geochronology, geochemistry and isotopicstudy of the Plio-Pleistocene Hominid sites and theNgorongoro Volcanic Highland in Northern Tanzania.Unpublished PhD, University of Chicago, Boulder.

Mercader, J., Panger, M.A., Boesch, C., 2002. Excavation of achimpanzee stone tool site in the African rainforest. Science296, 1452–1455.

Peters, C.R., 1987. Nut-Like Oil Seeds: Food for Monkeys,Chimpanzees, Humans, and Probably Ape-men. AmericanJournal of Physical Anthropology 73, 333–363.

Potts, R., 1988. Early Hominid Activities at Olduvai. Aldineand Gruyter, New York.

192 R. Mora, I. de la Torre / Journal of Anthropological Archaeology 24 (2005) 179–192

Potts, R., 1991. Why the Oldowan? Plio-Pleistocene toolmakingand the transport of resources. Journal of AnthropologicalResearch 47, 153–176.

Sahnouni, M., Schick, K., Toth, N., 1997. An experimentalinvestigation into the nature of faceted limestone spheroidsin the early Palaeolithic. Journal of Archaeological Science24, 701–713.

Schick, K.D., Toth, N., 1994. Early stone age technologyin Africa: A review and case study into the natureand function of spheroids and subspheroids. In:Corruchini, R.S., Ciochon, R.L. (Eds.), IntegrativePaths to the Past. Paleoanthropological Advances inHonor of F. Clark Howell. Prentice Hall, New Jersey,pp. 429–449.

Shen, C., Wang, S., 2000. A preliminary study of the anvil-chipping technique: Experiments and evaluations. LithicTechnology 25, 81–100.

Shipman, P., 1989. Altered bones from Olduvai Gorge,Tanzania: Techniques, problems, and implications of theirrecognition. In: Bonnichsen, R., Sorg, M.H. (Eds.), Bone

Modification. Center for the Study of the First Americans,Orono, pp. 317–334.

Texier, P.-J., Roche, H., 1995. Polyedre, sub-spheroıde,spheroıde et bola: des segments plus ou moins longs d�une meme chaıne operatoire. Cahier Noir 7, 31–40.

de la Torre, I., Mora, R., 2005. Unmodified Lithic Material atOlduvai Bed I: Manuports or Ecofacts? Journal of Archae-ological Science 32, 273–285.

Toth, N., 1982. The stone technologies of early hominids atKoobi Fora, Kenya; an experimental approach. Unpub-lished PhD, University of California, Berkeley.

Walter, R.C., Manega, P.C., Hay, R.L., 1992. Tephrochronol-ogy of Bed I, Olduvai Gorge: An application of laser fusion40AR/39AR dating to calibrating biological and climaticchange. Quaternary International 13/14, 37–46.

Willoughby, P.R., 1987. Spheroids and Battered Stones in theAfrican Early Stone Age. BAR International Series 321,Oxford.

Wynn, T., 1989. The Evolution of Spatial Competence.University of Illinois Press, Illinois.