The South African Archaeological Bulletin 27 STRIATED GRINDING GROOVES IN CENTRAL AFRICA* ROBIN DERRICOURT 93 Albion Street, Surry Hills NSW 2010, Australia ABSTRACT The distinctive striated lenticular grinding grooves in hard and soft rocks, found primarily in Zambia, have been the subject of a range of cultural explanations. It is here suggested that the understanding of their possible causes can be advanced by a systematic analytical approach to the physical processes involved. This limits the range of possibilities and removes some earlier interpretations. One possible explanation, their creation in the preparation of clay for pottery making during the Early Iron Age, is advanced. *Manuscript received June 1985, revised December 1985. In a number of locations - primarily in the Republic of Zambia are groups of distinctive striated lenticular (canoe shaped) grooves in hard rock, customarily named in the archaeological literature 'grinding grooves'. No direct archaeological evidence for their date or purpose exists, not do local communities have knowledge of their origin. Traditional explanations do, of course, exist for such prominent phenomena. They are said to be the footprints of men who lived when the rock was still mud; to be where the elephant rubbed his penis; to be where blunt metal axe and hoe blades were once sharpened (an explanation which falls down on trial); to be where royal wives ground their grain; to be incisions of God (Derricourt 1980:9-10). David Livingstone (1874:315) was told in 1868 by people living near Lake Bangweulu that they were the footsteps of God. This lack of knowledge suggests the grooves were not created by the recent 'Late Iron Age' ancestors of the communities who have produced these stories. Modem scholars have advanced several hypotheses for their function and date. One argument (Clark 1959:191) is that they were used to grind stone axes, and date from the Later Stone Age, which was replaced by Early Iron Age farming societies early in the first millennium AD. Practical experiments to test this hypothesis, however, turned out to erode the striations in grooves (Chaplin 1961). An Early Iron Age date has also been proposed (Phillipson 1972:115-6) implying use either in metal working or in grinding food or other materials. Hammering metal blades, rather than rubbing them, has been suggested (Verboom 1977). Other grooves, without striations, have been considered used for beadmaking, or net sinkers (Fagg 1959). These explanations stem from assumed context and cultural analogies. In the course of field survey and research in the Luapula province of Zambia where many such sites occur (Derricourt 1980:9-11), and in subsequent discussion, I came to doubt the feasibility of some existing explanations for the grooves, and to support a predecessor's agnosticism (Chaplin A geologist who surveyed the sparsely populated area of the east- ern Luapula valley in Zambia has reported numerous sites on volcanic rocks (Derricourt 1980: 11). Grinding grooves of this kind are not reported from neighbouring areas. Grooves of lenticular shape are found elsewhere, but with no internal striations confirmed. Six or seven sites in South Africa, a number in Angola or western Zaire, in Equatorial Guinea, several in the southern Sudan, and a widespread tradition in Ghana (Shaw 1944) represent other grinding traditions. A different kind of groove in western Nigeria (Fagg 1959) seems to have a wider range of forms. The specific nature of the Zambian tradition encourages us to treat it as a unit for study and interpretation. Characteristics The characteristics of most of the Zambian grooves are sufficiently uniform to imply a unity of purpose and origin. The features of individual grooves are: - canoe shaped: wide in the middle, narrowing to each end; - open bowl section: deepest in the middle, curving up to each end; - deep scratches (striations): roughly parallel, rarely crossing and sometimes in groups, running the length of the groove, least visible at the end; - symmetrical latitudinally; often symmetrical longitudinally; - smooth sides. The context of some grooves and sites emphasizes other points: the significance of these features is reviewed below: - water is normally nearby (Fig. 1); - a large number may be found together, abandoned at a maximum depth of c. 120 mm (Fig. 2); - partially worn grooves at Kasembe are long and, however shallow, the parallel striations are already present, showing this was an essential part of the process, or sufficient cause for abandoning a groove (Figs 3, 4); - hard rocks are common; - angular rocks are possible. 1961). Distribution Grinding grooves are gregarious by nature; their habitat is mainly flat rocky areas near water. On suitable rock surfaces many may be concentrated. The grinding grooves proper, with parallel striations along their interior, are confirmed in central African archaeological literature at one site in Zimbabwe and at more than 28 sites in Zambia (see Appendix). They range from grooves on loose rocks, some within excavated sites, to open surfaces with 56, 197 or over 200 grooves present. Grooves on loose rocks have been recorded within two Early Iron Age sites, and one Later Stone Age site occupied contemporaneously with the Early Iron Age. Most open sites are not associated with settlement material, but are close to water in all, or almost all, cases. The grooves are found not only on the relatively soft sandstone, but on hard rocks: quartzite, granite, silicified sandstone. The distribution within Zambia is in the north, largely within the Zaire River basin and entirely within the area influenced from Zaire since the 16th century. The sites recorded must be only a sample of those in existence. Fig. 1. Grinding groove site, Kasembe by Ngona River. Explanations To help seek (or limit the range of) explanations for the grinding grooves, it is proposed to start not from archaeological hypotheses, generated from current or discarded reconstructions of past culture history. Instead, one may argue from the petrological and physical evidence, and the possibilities and limitations implied by the evidence of the grooves themselves, leading on to the archaeological content and cultural implications. The Distribution. The limited region of distribution implies alternately: (i) a geological limitation on rocks orcauses; or (ii) a human S. Afr. archaeol. Bull. 41: 27-31. 1986 This content downloaded from 149.171.67.148 on Wed, 27 Mar 2019 04:34:42 UTC All use subject to https://about.jstor.org/terms 28 The South African Archaeological Bulletin Fig. 2. Grinding grooves, Kasamba Stream, Samfya. cultural limitation on needs of causes. We have no evidence for localised geological events or deposition; therefore (ii) is more probable. Location: The waterside position of most sites implies alternately: (i) the causation of suitable rock surfaces by water; or (ii) the use of water in creating the grooves. Flat rocky areas away from water do not seem to carry grinding grooves, but there is no firm statisti- Fig. 3. Close-up of grinding grooves, showing striations, Kasamba Stream, Samfya. cal evidence for this; therefore (ii) is more possible. The group. Grouping implies alternately: (i) the requirements of many single grooves together for a single purpose; (ii) a large population of simultaneous users; or (iii) that grooves reached a 'maximum' stage and were then abandoned. Grooves are randomly positioned, which makes (i) unlikely. They can be close together, which makes simultaneous use (i) unlikely. Therefore (iii) is the probable explanation. Abandonment. If abandoned because they reached a maximum stage this could alternately be because: (i) they become too deep; (ii) they become too long; (iii) the depth/height ratio became too great; (iv) the scratches (striations) were too deep; (v) the scratches were too many; (vi) the desired form was completed; or (vii) one groove was created for one finite period, time of life. The range of sizes precludes (i) and (ii), and (iii) seems unlikely apart from a few of the grooves. It seems possible the scratches Fig. 4. Incipient grooves, showing striations, Kasembe. themselves forced abandonment (iv, v) since the flatter grooves also had scratches. If these are regarded as aborted, it is possible to see most grooves as representing a completed goal (vi); the limited number of grooves at most sites makes (vii) possible, if unlikely. The attitude. The force used to create the grooves was either primarily: (i) gravity (e.g. weights) only; (ii) pressure only; (iii) gravity aided by pressure. The grooves do not occur on vertical or overhanging rocks; they do occur on sloping surfaces; therefore neither (i) nor (ii) is an adequate explanation and elements of pressure and gravity were used. The longitudinal dimension. Creation of the groove with the regularity achieved along its length, implies longitudinal motion. This was parallel to the lines of the striations and the shallow ('un- This content downloaded from 149.171.67.148 on Wed, 27 Mar 2019 04:34:42 UTC All use subject to https://about.jstor.org/terms The South African Archaeological Bulletin 29 finished') grooves show the striations accompanied, and did not follow, groove creation. It is therefore certain that the grooves were created by repeated longitudinal motion; and the symmetry shows this was motion in both directions. One position would often be the more comfortable side to work from, with alternating directions of motion. The section. The section of the grooves is uniformly deepest where the groove is widest, shallowest where the groove is narrowest, converging to a point. This will be created by the longitudinal motion in a single or an alternating direction, of an object with a round section or round half-section, and only such objects. The grooves could not therefore have been caused by any square- ended or pointed metal, stone or bone object but only by (i) a sphere; (ii) a cylinder; (iii) a round-ended blade; (iv) a chisel form. The last is excluded by the alternating motion implied by the symmetry of the grooves. Simple experimentation in a bed of hard sand demonstrates the necessity of the round section, which creates lenticular grooves, but not striations. Possible explanations can be listed systematically: (i) Goal: to create grooves Rubbing object: same as bedrock, or more likely harder. Scratches caused by: hard inclusions in rubbing object. Against this argument: the use of hard rocks in which to create grooves; the rarity of stone harder than granite or quartzite. (ii) Goal: to crush vegetable, ochre or other matter. Rubbing object: same as bedrock, or possibly harder. Scratches caused by: hard inclusions in rubbing object. Against this argument: the force needed to create the scratches exceeds that required to crush vegetable or other matter, therefore this explanation is unlikely. (iii) Goal: to rub into a required shape the rubbing object. Rubbing object: same as bedrock, and sometimes softer. Scratches caused by: hard inclusions in rubbing object, grits harder than either rubbing object or bedrock. This seems a probable explanation. (iv) Goal: to break or crush a material itself harder, or with inclu- The striations. The striations exist on the shallowest of incipient sions harder than bedrock. grooves; they distinguish the Zimbabwian sites; and must remain Rubbing object: same as bedrock, harder, softer? Scratches caused by: hard substance being rubbed below rubthe key to testing, rejecting or accepting explanations. They are bing object. aligned with the movement which caused the groove itself, and probably form part of the same action. The grooves are deeper at We have concluded from a look at the physical elements of the lower part of the grooves, suggesting that force was acting most grinding grooves, and possible goals, and possible causes, indepenpowerfully on both. dent of other archaeological evidence. We have excluded metallurgy and the preparation of organic materials from the explanations, and have shown that hard substances were involved. Hard cannot be the effect of grit used as an abrasive between the 'grind- metals are absent from the technology of prehistoric Africa: at first appearance, stone of round section would seem therefore the ing object' and the bedrock. They must be created either by (i) The grooves, cut in hard rock, are too deep for a single scratch, even of a diamond! They can be over 1 mm deep. They therefore something pushed up and down the groove in the same position by means by which the grooves were created, The striations, how- the 'grinding object', held from lateral movement by pressure down the groove; (ii) material extruding from the surface of the ever, remain central to the analysis of the whole operation. 'grinding object' held in the same direction. The labour/force. It can be stated quite simply: a great deal of Quartz inclusions seem the obvious explanation: quartz would be quite hard enough to scratch all the rocks in which grinding Materials: the groove. There is no clear indication between the grooves have been found. To do so, however, the orientation of the rubbing/ground stone object would need to be held constant. Once a groove began its shape would limit lateral movement same substance as the bedrock would abrade both itself and the the areas of the strikingly visible grinding grooves. labour and force was expended to create one grinding groove. grooves on softer rocks (sandstone) and on harder rocks; the only(wobble). But the scratches could not come from rolling a stone ball: only from holding the grinding object steady to push it difference is in portability which allows friable rocks to bear single repeatedly up (and down) the groove. grooves found within settlement sites. Explanations therefore must be sought for the erosion of the hardest materials, and these must be tested as applicable to the softer stone. The simplest force Ground stones to explain the grooves would be a rough sphere moved forward The counterpart to such an explanation for the grinding grooves and backward in the direction of the groove by force from above. should, of course, lie with actual ground stone objects, of round Looking at alternatives: (i) any 'rubbing object' much softer than section, produced or used in the grinding. Such objects, it is true, the bedrock would go to dust without abrading the bedrock (soft would be less 'visible' to a non-archaeological public and it may stone) or blunting (metal blade); (ii) any 'rubbing object' of the therefore be unsurprising that more have not been reported from bedrock, as would most reasonably hard stones; (iii) any 'rubbing object' significantly harder than the bedrock would create a groove without itself being significantly abraded. The groove therefore cannot have been created by copper or iron; but by a substance at least as hard or almost as hard as the bedrock itself: namely, stone within a certain range. Materials: the striations. The substance which created the scratches would be external to the rubbing object. If it were the same substance as the rubbing object - e.g. a protrusion on a stone ball - it would only create deep scratches if the object were harder than the bedrock. If the same substance it would be eroded first. The alter- natives are therefore: (i) a rubbing object the same substance as the bedrock, scratches caused by a substance harder than the bed- rock; (ii) the rubbing object harder than bedrock and not ground in the process; the scratches caused by a substance the same as this; or (iii) the rubbing object as (ii), the scratching substance harder still. Reconstruction. The possible physical explanations for the grinding grooves can now be narrowed down to a small number of possibilities. Having excluded natural causes in favour of human causes, we can relate possible mechanisms to possible goals. All the physical attributes show that the grooves were caused by alternating longitudinal motion of a rounded sectioned stone object under gravity and lateral pressure, maintaining its 'horizontal' axis uniformly so as to allow either protrusions or additives to carve scratched lines. Ground stone objects do occur in the prehistory of much of subSaharan Africa: not always in a clearly defined and dated cultural framework. Stone bowls occur in East Africa, together with stone axes, in contexts dated to pre-Iron Age pastoral communities in East Africa, of the first millennium BC (Phillipson 1977:74-82). They are known also from sites in Namibia, and sparsely in the area between, including northern Zambia (Clark 1964). A pastoralist context, or a group marginal to the 'core' LSA or Iron Age, is possible for most of these bowls. Stone pipes for smoking occur sporadically from South Africa (Goodwin 1947:7, 208-10) but their design is not symmetrical enough to require grinding grooves. Bored stones - using a natural or ground rock often approximating to a flattened sphere - are drilled through from opposing sides to give a circular hole. Their use as weights to aid leverage on food- gatherers' digging sticks is the consensus explanation for these artefacts, although there have been views (e.g. Shaw 1944:50) that a use as currency was plausible. The modern unfamiliarity of bored stones is emphasized by their use near Lake Tanganyika as a charm (Livingstone 1874:224) while some bored stones were kept by Kazembe's Lunda - the people who occupy the grinding grooves area of Luapula Province in Zambia (Cunnison 1959:221). The rough regularity of shape of a bored stone is important: the ground hollow core more so. In South Africa careful attention to the outer surface is pre-Iron Age, as argued in the classic study by Goodwin (1947:8) but many bored stones in Zimbabwe and Zambia have ground exteriors. In Zambia bored stones have been This content downloaded from 149.171.67.148 on Wed, 27 Mar 2019 04:34:42 UTC All use subject to https://about.jstor.org/terms 30 The South African Archaeological Bulletin found as surface chance finds at over 40 sites (Derricourt 1976:119) but within excavated assemblages they typify phase I, Ila-b and much more rarely, phase III of the Nachikufan Industry of the Later Stone Age in northern Zambia (Miller 1968, 1972; Derricourt 1980:4). Regular external finish is not an essential of these stones and they therefore provide an unlikely cause of the grinding grooves. Of related technique are ground stone rings and discs, known in South Africa (Goodwin 1943) but these are rare, occurring within a few selected areas in the Later Stone Age (LSA). Finally, ground stone axes and celts form a class of artefact not adequately studied in central and southern African prehistory. Sampson (1974) notes their occurrence in the Wilton-related Pfupian industry of Zimbabwe, the Zambian Wilton, and from South African coastal sites (e.g. fig. 158). Ground stone axes also occur in proportionately small numbers in the Nachikufan Ila, Ilb and III phases, the dominant LSA industry of northern Zambia (Miller 1972; Phillipson 1977:40). In eastern Zambia ground stone axes are characteristic of the LSA sequence (Phillipson 1976: 56, 91 ff, 151 ff.) but these do not have a regular section of a kind which would require the use of the grinding grooves described. The extension of polished axes into Zaire has long been known (Clark 1959:191), and may relate to the finely ground stone 'celts' of Fernando Po. In lower Zaire such celts are associated with pottery in the last two centuries BC and thus immediately precede the assumed Iron Age settlement (De Maret 1976; Phillipson 1977: 221-2) while later polished stone tools exist within the Iron Age (De Maret et al. 1977:493-4). It may be that such polished axes in Zaire represent different uses in different periods (Van Noten 1982:65) but they are very rare in the east of Zaire. The stone axes are not found further east in the area of the Zambian grinding Fig. 5. Woman grinding clay for pottery, Geji, near Bauchi, grooves. Polished stone axes, most of hard fine grained rocks, are northern Nigeria (photo: Hamo Sassoon). of course found further away from coastal west Africa (Davies 1967:190-203). There is therefore no clear evidence for ground or rubbed stone objects being created with the grinding grooves: of explanation we have defined. The grooves had a human agency; either in direct association or in similar regional distribution. Therewater is needed in pottery as it is not in stoneworking or metal::::~~~~~~- ' 4 '0; 4 - is no significant coincidence of the 'grinding groove' area and that lurgy. Grooves would reach a point where the depth and striations would have an unwelcome effect on the smooth working of clay, and require their abandonment; pressure and gravity were used; Pottery repeated motion is used in working clay; a hard rubbing stone on There does remain, however, one other explanation which is the the clay would create the section of the grooves; quartz extrusions only contemporary example noted of striated grooves actually from clay could create the striations with adequate force. being created. In Bauchi, in northern Nigeria, Hamo Sassoon We may not conclude with certainty about the origins of the (1962:145) observed: striated grinding grooves of Zambia. We have narrowed the field of ground stone finds. . . . amongst the Gyezawa people of Geji, near Bauchi, in of possible explanation and defined the criteria against which northern Nigeria, the women prepare their clay for pottingexplanations by must be tested, but at present the most probable cause is grinding it to a fine powder with a quartz rubbing stone on suitthe only one 'tested' by observation (Sassoon in Nigeria). This able flat rock surfaces. Owing to the grains of quartz sand in the suggests the grooves were most probably created in pottery makclay, this activity almost certainly wears deep grooves in the ing by Iron Age people. The commonness of quartz grit in Early rock more quickly than when corn is being ground .. . it takes Iron Age pottery of Zambia, the lack of cultural associations with may years of intermittent grinding before a groove has to be the Late Iron Age, and the absence of adequate explanation of the abandoned because it is too deep to be serviceable. The grains grooves among the historic descendants of the Late Iron Age of quartz sand which are included in the clay make deep longigroups makes an Early Iron Age date the most probable. tudinal striations in the bottom of the grinding grooves. (Emphasis added) Acknowledgements This operation (Fig. 5) provided a specific demonstration of the I am grateful to Mr John Tether, Geological Survey Departgrinding object, which erodes the grooves; the 'inclusions' which ment, Lusaka; Dr David Phillipson, Cambridge; Dr J. Vise, Unicreate the striation, and the existence of clay as an additional item versity of Zambia and Dr G. A. Chinner, Cambridge for discusin the operation. In the cultural context of northern Zambia this sions on aspects of this study; Mr N. Walker, Harare, for informaexplanation for the striated grinding grooves seems the most simtion from the Zimbabwe survey list; and Mrs S. Christiansen for ple and probable. secretarial assistance. The field research in Zambia was underPottery with a gritty ware has been recovered from Isokwe taken while I was director of the Zambia National Monuments Island in Lake Bangweulu (Derricourt 1980:28, 37-8) but not from Commission. Samfya Early Iron Age site. At Kasaba Bay on Lake Tanganyika (ibid: 92, 94, 98) is gritty ware, as in the Kalambo Falls Early Iron References Age pottery and the later local Lungu ware (Fagan in Clark CHAPLIN, J. H. 1961. A note on rock grooves in Northern Rho1974:30-1). desia. S. Afr. archaeol. Bull. 16:149. In the Copperbelt, Early Iron Age (EIA) pottery is characterCLARK, J. D. 1950a. The newly discovered Nachikufu culture of ised by gritty or sandy paste, while the Late Iron Age pottery has finer paste (Phillipson 1972:94), and indeed through many parts of central and southern Africa EIA pottery is marked by more gritty inclusions in the temper than Late Iron Age ware. Production of the grooves during pottery making fits the limits Northern Rhodesia. S. Afr. archaeol. Bull. 5:86-98. CLARK, J. D. 1950b. The Stone Age Cultures of Northern Rhodesia. Claremont: South African Archaeological Society. CLARK, J. D. 1959. The Prehistory of southern Africa. Harmondsworth: Penguin. This content downloaded from 149.171.67.148 on Wed, 27 Mar 2019 04:34:42 UTC All use subject to https://about.jstor.org/terms The South African Archaeological CLARK, J. D. 1964. Stone vessels from northern Rhodesia. Man 64:69-73. CLARK, J. D. 1974. Kalambo Falls Prehistoric Site II. Cambridge: Cambridge University Press. COOKE, C. K. 1961. Letter to the Editor. S. Afr. archaeol. Bull. 31 VAN NOTEN, F. L. 1982. The archaeology of Central Africa. Graz: Akademische Druck. VERBOOM, W. C. 1977. Lines of weakness in the forests, and early human settlements. ITC Journal (3):531-544. 16:116. CUNNISON, I. 1959. The Luapula peoples of Northern Rhodesia Appendix Manchester: Manchester University Press. DAVIES, 0. 1967. West Africa before the Europeans. London: *near water Methuen. DE MARET, P. 1976. Datation des haches polies. Actes du 9e Bulletin Central African striated grinding grooves (and references), Zimbabwe: Peak Mine Selukwe, 65+, granite (Guy 1965). Eastern Province Zambia: Mfuwe Luangwa, unconfirmed, Msoro congres de l'Union des Sciences Prehistoriques et ProtohistoriMission, granite (both Derricourt 1976:143). ques, Nice. Nice: UISPP. Northern Province Zambia: Malole, Kasama, 3 grooves; Mbala DE MARET, P., VAN NOTEN, F. & CAHEN, D. 1977. RadioRoad, Kasama, 2 grooves (both Geological Survey department carbon dates from West Central Africa: a synthesis. J. Afr. Hist. collection); Kasanshi*, numerous on sandstone (Chaplin 1961, 4:481-505. plate la, and National Monuments Commission files); Milima*, DERRICOURT, R. M. 1976. A classified index of archaeological loose rocks (NMC files, Chaplin 1961, Verboom 1977:534); and other sites in Zambia. Livingstone: National Monuments Mungwe Mashitu, loos rock with slag and stone anvil; Misamfwu, Commission Research Publication 3. Kasama, loose rock (both Chaplin 1961, Verboom 1977:534); DERRICOURT, R. M. 1980. People of the Lakes: archaeological Chibolya (Derricourt 1980:144); Nsalu, loose rocks with Nachikustudies in northern Zambia. Lusaka: Institute of African Studies, University of Zambia, Zambia Papers 13. DIXEY, F. 1953. Stone-age tool grinding sites on Lake Bangweulu. Northern Rhod. J. 2:72-4. FAGG, W. 1959. Grooved rocks at Apoje near Ijebu-Igbo, West- ern Nigeria. Man. 59:205. GOODWIN, A. J. H. 1943. Edged discs and armrings. S. Afr. J. Sci. 40:296-302. GOODWIN, A. J. H. 1947. The bored stones of South Africa. Ann. S. Afr. Mus. 37:1-210. GUY, G. L. 1965. Grooves and stone mortars - Selukwe, Rhode- sia. Arnoldia (Rhod.) 8(2):1-3. HILLMAN, J. & S. 1984. Archaeological observations in Bangangai Game Reserve, south-western Sudan. Azania 19:115-120. KENNEDY, R. A. 1962. 'Grinding benches' and mortars on Fernando Po. Man 62:129-30. fan LSA (Clark 1950a:92-3, 1950b: 118). Luapula Province Zambia: Kasamba Stream*, 197 grooves, quartzite and silicified sandstone, near unoccupied cave, Early Iron Age sites in the area (Dixey 1953; Clark 1959: plate 8b; Dermcourt 1980:10); Kasembe*, 200-300 grooves, some shallow; Ndongola Farm*, 56 grooves; Kasumpa*, unconfirmed; Ngona hydrological gauge*, 18 grooves, one in Livingstone Museum; Muchelenge, 2 groups on granite; Shoma (Yombwe)*, several groups; Muchinga escarpment*, numerous sites, most volcanic rocks; one sandstone (all Derricourt 1980:11); Lifunge Island; Chambeshi (both Livingstone 1874:315). Copperbelt Province, Zambia: Kamfinsa, 8 grooves on quartzite; (Phillipson 1966: plate II, 1972:115), Mishila*, loose boulder with daga and slag nearby; Mutenda* (both Phillipson 1972:115-6), Chondwe*, loose fragment, sandstone block in Early Iron Age set- LIVINGSTONE, D. 1874. Last Journals. London: John Murray. MILLER, S. F. 1968. The Nachikufu industries of the Later Stone Age in Zambia. Unpublished Ph.D. thesis: University of California, Berkeley. MILLER, S. F. 1972. Archaeological sequence of the Zambian Later Stone Age. In Hugot, H. J. ed. Actes du 6eme Congres Panafricain de Prehistoire, Dakar 1967: 565-572. Chamberry. MILLS, E. A. C. & FILMER, N. T. 1972. Chondwe Iron Age site, Ndola, Zambia. Azania 7:129-146. PHILLIPSON, D. W. 1966. The Late Stone Age and Zambia's first artists. In Fagan, B. M. ed. A Short History of Zambia: 56-80. Nairobi: Oxford University Press. PHILLIPSON, D. W. 1972. Early Iron Age sites on the Zambian Copperbelt. Azania 7:93-128. PHILLIPSON, D. W. 1976. The Prehistory of Eastern Zambia. Nairobi: British Institute in Eastern Africa Memoir 6. PHILLIPSON, D. W. 1977. The later prehistory of eastern and southern Africa. London: Heinemann. PHILLIPSON, D. W. 1981. A preliminary archaeological reconnaissance of the southern Sudan 1977-8. Azania 16:1-6. SAMPSON, C. G. 1974. The Stone Age archaeology of Southern Africa. New York: Academic Press. SASSOON, H. 1962. Grinding grooves and pits in northern Nigeria. Man 62:145. SHAW, C. T. 1944. Report on excavations carried out in the cave known as "Bosumpra". Proc. prehist. Soc. 20:1-67. SINGER, R. 1961. Incised boulders. S. Afr. archaeol. Bull. 16:27. tlement (Mills & Filmer 1972:136); Musiyakupalwa*, loose fragment, Early Iron Age open site (Phillipson 1972:106). Northwestern Province, Zambia: Chanda Falls; Lumanwa Stream, unconfirmed; Mushima, unconfirmed (all Derricourt 1973:143). Other African grooves South Africa: Zoutkop, OFS, sandstone; Kasteelberg, SW Cape, sandstone, with LSA artefacts; Kameelrock, SW Cape, LSA shelter; Witkoppies, SW Cape, LSA shelter; Clanwilliam, Cape; Pomeroy*, Natal; Mfongesi, Natal (all Singer 1961:27). Zimbabwe: Tengwai River (Cooke 1961); Mazoe area; Lake McIlwaine area; Kwe Kwe (all National Museums and Monuments index). Zaire: Kasai and Shaba regions (Van Noten 1982:65); KandaKanda and Kasenga regions (Shaw 1944); Lower Congo (Singer 1961:27; see also Clark 1959:192). Angola: see Clark 1959:192. Equatorial Guinea: Rio Muni, (stone rocks, slabs and polished grooves (Perramon, cited by Van Noten 1982:65); Fernando Po (Bioko Island) (Kennedy 1962). Ghana: Mutrakai Point* (Davies 1967:45); forest and southern savannah (Davies 1967:45; see also Shaw 1944). Senegal: Jebel Riwa and Tsimbia, with stone flaking debris (Phillip- son 1981: 3 & plate 1); Bangangai (Hillman 1984:115-6), striations in grooves. This content downloaded from 149.171.67.148 on Wed, 27 Mar 2019 04:34:42 UTC All use subject to https://about.jstor.org/terms