A history in paint and stone from Rose Cottage Cave, South Africa.

Ouzman, Sven ; Wadley, Lyn

In South Africa, as in so many regions, the world of dirt archaeology in shelter. floors and of rock art on shelter walls, have also been rather separate as domains of study. In research at Rose Cottage Cave, bridges are being made to link both strands of evidence to the forager social strategies from which both derive.

Rock art and `dirt' archaeology

Southern Africa has an abundance of forager rock art(1) in the form of engravings and paintings; it is similarly rich in Later Stone Age (c. 25,000 b.p. until time of living memory) deposits containing bone, ochre, shell, stone and wood artefacts as well as archaeological features such as hearths and activity areas. For much of its history, southern African Later Stone Age research has been a descriptive discourse concerned with artefact sequences and typologies. Excavated material culture has provided information that relates principally to economic, environmental and technological aspects of forager life; only recently have social implications been considered (Wadley 1987; Mazel 1989; see also Barham 1992). Rock art research has a much longer history of using social theory to interpret imagery and has focused on issues such as religion (Vinnicombe 1976; Lewis-Williams 1981a), political strategies (Campbell 1986; Dowson 1994), forager-agriculturist contiguity (Hall 1994; Prins & Hall 1994; Jolly 1996; Ouzman 1995a) and the construction of regional histories (Loubser & Laurens 1994; Yates & Smith 1994). Though polysemic in nature, it is generally agreed that the production and content of most of the forager rock art in southern Africa centres on the hallucinatory visions and experiences of San religious adepts or shamans (Lewis-Williams 1981a; Huffman 1983; Deacon 1988; Kinahan 1991; Garlake 1995; see also Dowson & Lewis-Williams 1994). Because excavated material culture and rock art are visually disparate and because rock art was, for a significant period, considered something of a `Cinderella' body of research, southern African archaeologists tend to separate and compartmentalize studies of rock art and `dirt archaeology', with comprehensive research all the poorer for this tendency. We suggest that a more open-ended approach to the Later Stone Age, utilizing multiple strands of evidence, will throw new light on problems and may point to new research questions and directions (see, for example, David et al. 1994; Hall 1994; Tacon & Brockwell 1995). Our study attempts to understand both rock art and excavated material culture as integrated elements relating to forager social strategies. We do not, however, contend that all rock art and excavated material culture were necessarily part of similar social strategies or that they were always complementary.

We recognize that open-ended approaches are potentially fraught; for example, establishing the precise temporal and conceptual relationships between rock art and excavated cultural material is problematic (but see Humphreys 1971; Morris & Beaumont 1994; see also Campbell & Mardaga-Campbell 1993). Even a match of radiocarbon or other dates obtained from excavated deposits with dates obtained from paint samples is no guarantee of synchronous occupation.

Strands, `cables' and `tacking'

The challenge is to understand and integrate visually disparate material culture signatures such as painted images and lithics. If used wisely, each strand of evidence can act as an independent verification or limitation for an argument. In our efforts to integrate rock art and excavated material culture we use Bernstein's metaphor of `cables' and `tacking' as a heuristic device (Bernstein 1983). Wylie, citing Bernstein, argues for a practicable path between objectivism and relativism by suggesting that multiple strands of evidence can be twisted together to form a `cable' of evidence, the cumulative `weight' of which is sufficient to make an argument plausible even if no strand is sufficient in itself (Wylie 1989). We use the cable and tacking metaphor to link three strands of evidence from different yet related fields: ethnography, rock art and excavated material culture.

Strand 1: ethnography

Foragers enjoy a relationship with the land fundamentally different from that of non-foragers (e.g. Deacon 1988; Ingold 1993; Smith 1994); it is thus difficult for non-foragers, such as Western urbanites, to perceive and understand the world as foragers do. Ethnographic information is sometimes thought to bridge the gap between archaeologically observed phenomena and past social reality but this perception is highly problematic (e.g. Wylie 1985; Hammersley 1992). Cultural uniformitarianism cannot be taken for granted, nor the assumption that our western cognitive systems can assimilate and explain forager cognitive systems. Despite these problems, ethnographic analogy, as the closest known referent to the forager world, remains central to most, if not all, fields of archaeological enquiry (Wylie 1985; Testart 1988; Roth 1989; Aunger 1995), and has been used with considerable success in southern African rock art and excavation-centric research (see Lewis-Williams & Loubser 1986).

The Later Stone Age material culture produced in southern Africa during the last few thousand years attests to some activities of foraging communities ancestral to people most often referred to as `Bushman' or `San'.(2) These San communities occupied most of southern Africa until they and their social structure fell victim to black and white settlers' claims to their land and its resources (Wright 1971; Campbell 1986; Gordon 1992). There is not much information on the systemic impact of forager and agriculturist `contact' or interaction in either the colonial or pre-colonial past; ethnographers are unsure about how to recognize `pure' forager sites and material culture (e.g. Schrire 1984; 1992; Smith et al. 1991; see also Wilmsen 1989; Gulbrandsen 1991; Lee 1992). There is evidence to suggest that some San groups moved towards sedentism during the last 200 years (Kent 1992). We have no secure knowledge that ethnographically observed foragers were at all like their predecessors prior to contact with nonforagers. To complicate the issue, San communities retaining elements of a forager life-style into this century comprised numerous groups, some with mutually unintelligible, though related, languages (e.g. Marshall 1976; Lee 1979; Silberbauer 1981; Guenther 1986; Barnard 1992; Biesele 1993). These San groups, displaying great variability in social organization and material culture, yet had much in common. All groups gathered plant food and hunted, trapped or scavenged game (Marshall 1976; Lee 1979). Like foragers world-wide, these San groups practised corporate aggregation and dispersal (Carter 1970; Marshall 1976; Yellen 1984; Wadley 1987; see also Conkey 1985). Among the !Kung of Botswana, studied in the 1960s and 1970s, band size varied between family groups during the dispersal phase and up to 200 people during an aggregation phase (Lee 1979: 365). Although the specifics of aggregation vary cross-culturally, and probably varied temporally too, the aggregation phase amongst ethnographically observed San such as the /Xam of the central interior of South Africa (Bleek & Lloyd 1911), and the G/wi, !Kung and !Xo of Botswana (e.g., Heinz 1975; Silberbauer 1981; Lee 1979) is perceived as the context for initiating group gathering of plant foods, fishing, hunting, trapping, information exchange, making and mending artefacts, making and exchanging gifts, marriage brokering, socializing and staging religious and ritual activities.

Behaviour appears to have been generally more formal during aggregation than dispersal phases, and avoidance rules more strictly adhered to. The !Kung of Botswana, for example, have avoidance rules that affect speech, entering shelters and sitting near certain categories of people (Marshall 1976: 249-51). !Kung married men and women sit separately around the hearth (Marshall 1976: 88, 249); male and female tasks are often spatially separated during the aggregation phase so that women tend to work together to make beads, while men make and poison their arrows. In contrast, during the dispersal phase, family groups work closely and informally at subsistence tasks; gift manufacture is reduced to a minimum, and ritual activities, such as the medicine dance (Marshall 1969), are largely absent. Amongst !Kung the medicine dance took place most often during the aggregation phase of band life because the dance needs more than 15 adults to be effective (Lee 1979: 365).

Ethnoarchaeological studies among the !Kung and various Basarwa (San) groups in the Kutse area have identified camp site features that are good material indicators of aggregation or dispersal phase sites (Yellen 1984). Refuse dumps, for example, are said to accumulate in aggregation camps (Bartram et al. 1991) and, in the case of the Kutse Basarwa, special manufacturing areas are also restricted to the long-term camps (Kent 1992).

However desirable and necessary occasional aggregation phases may have been, they were not without their problems, particularly interpersonal tensions. Story-telling and the medicine dance helped to relieve tensions, and shamans used their healing powers to cure people of illness and sickness (Marshall 1961; 1969). The medicine dance typically provided the medium through which the shaman would enter an altered state of consciousness, perceived as the spirit world in which s/he experienced religious revelations (Lewis-Williams 1981a; 1992). The shaman would then relate visions and experiences of the spirit world to the rest of the group in the form of oral accounts. The San communities of the southeastern mountains of southern Africa seem to have used rock art to this same end, indicating a certain continuity in San religious thought and social organization (Lewis-Williams 1984; Deacon 1988; Dowson & Lewis-Williams 1994) which may justify the cautious application of certain San ethnographies to the parts of the archaeological record. Indeed, the consonances between the comprehensive ethnographies of the /Xam who inhabited the Northern Cape Province of South Africa (e.g. Bleek & Lloyd 1911; Hewitt 1986) and the San communities of what is today Botswana and Namibia (e.g. Barnard 1992; see also Lewis-Williams & Biesele 1978) far outweigh the dissonances. These two bodies of ethnography in turn provide general support for the scanty ethnographic information relating to the San of the southeastern mountains of southern Africa (Casalis 1861; Orpen 1874; Ellenberger 1953). Here Qing, a Maluti San, made one of the very rare direct comments on forager rock art (Orpen 1874); comments that appear to confirm the hypothesis that San rock art is chiefly shamanistic (Lewis-Williams 1980: 473-6; but see Jolly 1996).

Introducing the site

We draw on these ethnographies to inform our discussion of a pre-eminent southern African forager site, Rose Cottage Cave. We now turn to the geographical, environmental and cultural context of Rose Cottage Cave and its former inhabitants.

Rose Cottage Cave is situated in the Caledon River Valley at 29 [degrees] 13'S, 27 [degrees] 02'E, near the town of Ladybrand in the Free State Province of South Africa (FIGURE 1). The cave, surrounded by a grassland and scrub thicket ecotone, is unusual in that it was formed by a huge boulder falling out of a cliff face leaving a 125 sq. m cavern hidden behind it (FIGURE 2). The cave faces north; it is sun-filled in winter and cool in summer. A perennial stream is within 40 m of the cave, and rocks such as opaline, used for stone-knapping, are located in the Caledon River 7 km to the northeast. The rock ledge adjacent to the cave affords a panoramic view of grassy plains that teemed with game until about 100 years ago (Dornan 1909). Missionary accounts indicate that the Ladybrand area was still inhabited by San foragers in the 1830s (Casalis 1861; Dornan 1909). Excavations of agriculturist settlements in the area suggest that agriculturists first settled near Ladybrand as recently as the last 200 years; further north, the earliest agriculturist occupation in the Free State Province is determined at 500 [+ or -] 95 b.p. (Maggs 1976: 146). On a terrace below the entrance to Rose Cottage Cave is a defensive agriculturist stone-walled settlement, probably dating to a period of unrest and war which took place about AD 1790-1820 and was especially pronounced in the eastern Free State (Cobbing 1988: 517; see also Hamilton 1995). Research more specific to foragers also hints at the profound effect agriculturists had on the world of the eastern Free State foragers (Thorp 1996).

[FIGURES 1 AND 2 ILLUSTRATION OMITTED]

Sixty-one painted shelters, all with Later Stone Age archaeological deposit as well as numerous unpainted LSA sites have been recorded in the Ladybrand Magisterial District, suggesting a high density of forager occupation during the late Holocene. The excavation of 10 shelters with Later Stone Age deposit in the eastern Free State (Wadley 1995) indicates that Rose Cottage Cave contains the richest and most varied material culture: it was a pre-eminent site on the forager landscape (Ouzman in press). This pre-eminent status has material expression in an exceptionally deep archaeological deposit of 6 m relating to the last 100,000 years which, together with the approximately 25 sq. m of rock paintings, indicates that the cave and its immediate surroundings provided a long-term locus for foraging people.

The long-term and even cross-cultural attraction of Rose Cottage Cave may be an example of `topophilia' (Tuan 1974; Deacon 1988; Tacon 1990), the often indefinable but very real attraction that certain places hold.

To explore and understand the allure of Rose Cottage Cave we carefully `tack' (cf. Wylie 1989) between certain San ethnographies and the Rose Cottage Cave material culture record.

Strand 2: the rock art

In February 1994 we spent approximately 96 hours recording the 186 complete and 462 partial painted images remaining in Rose Cottage Cave. There were surely many more paintings: Rose Cottage Cave must have been a magnificent gallery in forager times. Fifty years ago the French archaeologist Abbe Henri Breuil (1945: 353) described as the Rose Cottage Cave rock art as `rather faded', but exceptionally complex. Approximately 6x2 m of the east wall and 6[cdot]5x2 m of the south wall (Figure 2) were painted, with the denser concentration of paintings on the south wall. The paintings comprise a wide range of subjects such as antelopes, felines, fish, a snake and human figures in a variety of postures and activities.

Absolute dating of the Rose Cottage Cave paintings is not possible at present. However, there are at least four strands of dating evidence that may be used to situate the imagery in a relative chronological context.

* First, oxides such as ochre and haematite, including some ground specimens, have been found in all the excavated levels of Rose Cottage Cave; the use of pigment, not necessarily for rock art only, but perhaps also for body painting and hide processing, has a deep antiquity.

* Secondly, eight painted finger dots are visible on the wall of Rose Cottage Cave that used to be 2-8 cm below the depositional surface. Erosion of deposit has made these finger dots visible and they are contemporaneous with or slightly older than the uppermost occupation level.

* Thirdly, Loubser & Laurens (1994) have convincingly suggested that in the Caledon River Valley paintings executed in a bright, blocked, poster-paint-like pigment are the most recent images, often comprising images of domesticates, shields, spears and such like and which are consistently superimposed on top of shaded images executed in deeper hues. It also appears as if this change in representational manner just pre-dates the physical arrival of agriculturists on the eastern Free State landscape. With the exception of two cows painted in a now faded black pigment, the bulk of the Rose Cottage Cave imagery is shaded and thus appears to pre-date the arrival of agriculturists and, given their sheltered location, may be of considerable antiquity. Work by Sampson (1972) and Brooker (1980) in the Wepener District, 120 km south of Rose Cottage Cave, corroborate the findings of Loubser & Laurens (1994); even including a secure date on a blocked image of a cow of less than 300 years (Sampson 1972).

* Fourth, almost every painted shelter in the Caledon River Valley has some measure of occupational deposit in it is improbable that none of the occupation episodes relate to the painting episodes.

The majority of the present Rose Cottage Cave rock paintings appear to pre-date a physical agriculturist presence on the landscape and could have survived for a considerable time, even more than 1000 years (for more on the dating of southern African rock art see van Riet Lowe 1933; Holm 1958; Goodwin 1960; Singer & Wymer 1969; Humphreys 1971; Sampson 1972: 309-10; Wendt 1976; Butzer et al. 1979; Thackeray 1983; van der Merwe et al. 1987; Breunig 1989: 33-5; Sampson & Vogel 1989; Binneman & Hall 1993; Mazel 1994; 1995; Morris & Beaumont 1994; Whitley & Annegarn 1995; Yates & Jerardino 1996).

Most of the Rose Cottage Cave paintings appear, on visual inspection, to be of some antiquity, though the depiction of two cattle in black pigment at the cave indicates that at least some of the art should be considered in the context of a landscape in which forager residents and more recently arrived pastoralists and/or agriculturists affected each other. Though the Rose Cottage Cave paintings do not include abundant depictions of domesticates, pastoralists or agriculturists such as those found in other Free State art (Loubser & Laurens 1994), archaeological excavation indicates that both domesticates and pottery pre-date, by up to 400 years, the colonization of the Free State by Bantu-speaking agriculturists in the last 500 years (Klatzow 1994). This suggests exchange networks, and possibly other types of forager-agriculturist interaction established many hundreds of years before Bantu-speaking farmers permanently settled in the Free State.

The Rose Cottage Cave images are painted in monochrome, bichrome, polychrome and shaded polychrome. The colours used include shades of red, ochre, white and, more rarely, black. On some of the images, particularly the eland (Taurotragus oryx) on the east wall, individual brush strokes may be discerned. The general impression of the paintings is of complexity, and of great care and precision in execution. Some of the images are over 2 m above the present ground surface level, thus necessitating some form of scaffolding. This elevated positioning of imagery suggests that the paintings were intended to provide a dominant visual focus in the cave space.

We select three painted panels from Rose Cottage Cave for exploring forager beliefs and practices.

Panel 1: The shaman's fish

There are five paintings of fish in Rose Cottage Cave. Two of these fish (Figure 3), following Breuil's 1945 identification, were considered to depict whales, dolphins or even sharks (Breuil 1945) on the basis of body size, the appearance and position of the anal and ventral fins, and the presence of a pronounced proboscis. Breuil's identification, endorsed by other researchers (Battiss 1945; Wells 1946; van Riet Lowe 1947), has persisted (Editorial, South African Archaeological Bulletin 1992). A more adequate identification of the Rose Cottage fish is that it represents a freshwater mormyrid (Willcox 1962: 6; Skelton 1996 pers. comm.): the fish's diagnostic features are incomplete, but the image contains a singular diagnostic mormyrid element in the form of a `beak'. The beak is somewhat exaggerated in the Rose Cottage image as though the artist was ensuring that the diagnostic feature was not missed. The beak is absent from the local Clarias, Barbus or Labeo species, which do not seem to be represented.

[FIGURE 3 ILLUSTRATION OMITTED]

The closest occurrence of mormyrids is in the northern KwaZulu-Natal Province of South Africa (Skelton 1993: 92), some 200 km away from Ladybrand. Other rock paintings of mormyrid fish cluster in and around the southeastern mountains of southern Africa (Figure 4). The discrepancy between the natural and painted mormyrid distributions is puzzling, unless the mormyrid paintings are taken to be deliberate statements about places to or from which the shaman artists travelled.

[FIGURE 4 ILLUSTRATION OMITTED]

Mormyrids are not the only `exotic' creatures depicted in the local rock art. First, three elephant are painted at the site of Tandjesberg (Loubser 1993: 359-60), 16 km northeast of Rose Cottage Cave. With the exception of one tooth found in a riverbed, which is probably a lost hunting trophy, there is no evidence of elephant skeletal remains or sightings in the Free State (Skead 1980; Skinner & Smithers 1990: 549; Lynch 1991: 13). Though elephants may occasionally have wandered into the eastern Free State the topography is unsuited to elephants, even during more favourable climatic regimes. Though local San may have been aware of elephant, their depiction at Tandjesberg would have been perceived as out of place. Secondly, there are two rock paintings of giraffe at Roma and Thaba Bosiu in Lesotho, some 25 km east of Rose Cottage Cave (Loubser & Zietsman 1994). The closest known reported occurrence of giraffe is almost 350 km east of Thaba Bosiu (Skead 1980; Skinner & Smithers 1990: 605; Lynch 1993: 32-3). The mountainous, arid terrain of Thaba Bosiu and, indeed, much of the Free State and Lesotho, is unsuited to savannah-loving giraffe. A giraffe depicted in the Bethlehem District 220 km to the northeast is also outside of the known distribution range of elephant.

Ethnographic evidence of San visiting and exchange networks (e.g. Bleek & Lloyd 1911: 303, 381-5; Cashdan 1983; Wiessner 1983) and the presence of exotic material culture in Later Stone Age sites in the southeastern mountains (Vinnicombe 1976: 86; Mazel 1989: 143-4; Mitchell 1995: 35; 1996) indicate that distances of up to 200 km were not unusual for forager exchange networks. Similarly, the presence of domesticates and pottery in forager contexts substantially prior to agriculturist colonization (Klatzow 1994) suggests extensive exchange networks in the eastern Free State during the last few hundred years. We see foragers in the general area of Rose Cottage Cave as involved in long-distance exchange networks with other forager, and even agriculturist communities: rock paintings of exotic species demonstrating access to the resources of distant exchange partners (see also Johnson 1989). Resources, social as well as economic, would have included shamanic power. There are not many instances of exotic paintings or exotic material culture. Long-distance exchange networks may not have been a significant element contributing to rock art production. Alternatively, long-distance exchange networks may have been jealously guarded and material culture relating to these networks used carefully in order to make the presence of exotics and the people who obtained them all the more telling.

Mormyrid depictions are ideal metaphors for the supernatural nature and activities of San shamans (for a fuller discussion see Ouzman 1995b). Mormyrids are electrogenic and electroreceptive and give off an electric shock when handled. The tingling sensation experienced when touching a freshly caught mormyrid is similar to that reported by contemporary Botswana San shamans who describe supernatural potency travelling up their spine immediately prior to entering an altered state of consciousness (Katz 1982). The mormyrids' electrical discharge may have been perceived as evidence that these fish were supernaturally potent creatures, much like shamans themselves. Fish, shaman and peregrination became conflated in people's perceptions.

Panel 2: The enemy within

The second polysemic painted panel comprises a magnificent ochre and white lioness with teeth, an off-set ear and between three and seven claws on each paw (FIGURE 5). The lioness has a characteristically bent but exaggeratedly long tail, an extremely powerful and pronounced nuchal hump, and an odd ochre-coloured body line. The lioness is depicted walking amongst a herd of 17 eland (eight shown) and almost touches the back of one of the eland. Interestingly, Breuil mentions `a number of lions' in the cave (Breuil 1945: 353). Today Figure 5 is the only visible example. There are definitely no other lions associated with the eland herd.

[FIGURE 5 ILLUSTRATION OMITTED]

The association between solitary felines and herds of eland is a numerically small but persistent feature of southern African rock art (e.g. Pager 1971; see also Lewis-Williams 1985: 54). !Kung and /Xam foragers believed that felines represented a powerful and sometimes threatening, unstable force (Bleek & Lloyd 1911: 182; Marshall 1969: 352). It was believed that especially powerful and sometimes even malevolent shamans could sometimes assume the form and potency of felines to harm people (Bleek & Lloyd 1911: 189; Biesele 1993: 111). Felines, like shamans, were believed to possess supernatural knowledge (Bleek & Lloyd 1911: 187). The !Kung word for out-of-body travel, one of the shaman's activities, is jum, also the word used for `pawed' creatures (Lewis-Williams 1981a: 97; 1985: 55). A common appellation for shaman is `beast of prey'; felines are obvious candidates for the visual representation of a shaman and their potentially dangerous qualities.

Eland, on the other hand, were said by /Xam informants to be the favourite animal of /Kaggen, the trickster deity of the southern San; they were believed to possess quantities of supernatural potency (Lewis-Williams 1981a: 84). Eland, a natural analogue for ideal human behaviour, are docile, gregarious animals that personify social cohesion and harmony (Vinnicombe 1976: 163-4). The juxtapositioning of feline with eland sets together the distinctive negative and positive associations each animal represents (see Lewis-Williams 1985:.54-5).

The panel is highly ambiguous. We suggest this ambiguity may have been conscious, intended as a metaphor expressing both the desirability and the danger of staging large social gatherings. The large number of eland depicted at Rose Cottage Cave may have referred to the way that eland aggregate after the rains (Vinnicombe 1976:163-4; Skinner & Smithers 1990: 701-2). During aggregation, herds of eland must have represented a localized but immensely potent phenomenon. San groups may have considered human aggregation to be similarly potent when, for example, a number of shamans were brought together. Vinnicombe has thought that the San of the KwaZulu-Natal Drakensberg mimicked the aggregative and dispersive behaviour of the eland, hence the title of her book People of the eland (1976: 163).

Aggregations were also times of social tension with many opportunities for conflict (e.g. Marshall 1961). Depicting a solitary feline in the midst of an aggregation of eland may have been an acknowledgement of the risk inherent in staging large gatherings. There is neither obvious threat in the posture of the lioness, nor indication of fear among the eland. The danger from an anti-social agent may remain latent if social norms are respected. Depicting a lioness, not a lion, makes a gender statement: the lioness, a more active hunter, travels more widely and is more dangerous than the lion. Depicting a lioness may acknowledge the ritual power of women. Significantly, Qing, a Maluti San, recounted how Maluti San women were a destructive influence (Orpen 1874: 3), particularly in the sphere of shaman's activities (1874: 7). Similarly, certain G/wi of the central Kalahari regard women as the carriers of evil and misfortune (Silberbauer 1981: 174). The being who brings misfortune, G//awama, is said to throw arrows containing evil to earth; they lodge in the bodies of certain women who harm the rest of the group, the spreading evil destroys its harmonious atmosphere (Silberbauer 1981:175). This panel (Figure 5) may present the threat of evil in order to maintain group harmony.

Panel 3: The medicine dance

In all San communities group harmony, or professed desire for harmony, is expressed in the medicine dance (Marshall 1969) at which people dance, clap and sing to create a resonant, spiritually charged atmosphere. The shamans are enabled to enter the spirit world (hallucinate) and to control the movements of game herds, to heal and to make rain (Lewis-Williams 1981a). A medicine dance is depicted on the south wall of Rose Cottage Cave.

A group of 11 humans and two eland Figure 6) is depicted in red. The heads of the humans and eland were painted in fugitive white pigment. That this is a depiction of a medicine dance is indicated by two features.

[FIGURE 6 ILLUSTRATION OMITTED]

First, among the lowermost four human figures is a faded human figure in an arms-back posture, unmistakably characteristic of a San shaman desirous of acquiring sufficient super-natural potency to perform shamanistic tasks (Lewis-Williams 1980). The acquisition of potency and execution of shamanistic tasks is almost always conducted in the context of a medicine dance, a context indicated by the presence of 10 other human figures (Figure 6). The division of this panel into an active uppermost and a seated lowermost group visually corresponds to known paintings of medicine dances (e.g. Vinnicombe 1976: figure 227; Lewis-Williams 1981a: figure 18; Loubser 1993: figure 11).

Secondly, the thin red line connecting both uppermost and lower human figures, a feature of the rock art of the southeastern mountains, has been explained as representing a line of supernatural potency or even a transformed shaman (Lewis-Williams 1981b). Both the line of potency and transformation are generated at a medicine dance. The painted line joins the uppermost six human figures to the painted remains of two eland, one of which is upside-down. This is unlikely to be a representation of a `real' event and the upside-down eland may represent a dead animal. Death, in !Kung society a metaphor for trance (Katz 1982), is a spiritually potent state; the upside-down human figure, in mimicking the upside-down eland, imbues this panel with additional potency.

Depicting a medicine dance at Rose Cottage Cave may have constituted an attempt to mould people's perceptions of harmonious social order, the image functioning as a model for a sanctioned code of appropriate behaviour. We cannot infer the context of this behaviour; although medicine dances are associated with aggregation, paintings of the dances cannot be used to infer aggregation at the painted site because an artist can depict real or imaginary scenes at any venue at any time. One cannot read from the art a list of features that correlate with San aggregation and dispersal phases. As rock art is often willfully ambiguous in its provision of insights, it is necessary to integrate the imagery into a broader archaeological context, establishing relations of relevance between rock art and excavated material culture.

Strand 3: the evidence from the earth

We may be reasonably certain that the rock art of Rose Cottage Cave is broadly contemporaneous with the three uppermost occupation levels; Mn, A, and A2. Level Mn is determined at 500 [+ or -] 50 b.p. (Pta-6788; charcoal), Level A is determined at 680 [+ or -] 50 b.p. (Pta-5622; grit-tempered pottery) and A2 2240 [+ or -] 60 b.p. (Pta 7117; charcoal) (Wadley 1995: 576). We discuss levels Mn and A in detail. Level Pt, immediately below level A2, is determined at 5970 [+ or -] 70 b.p. (Pta 59341; charcoal). The lack of excavated material culture dating 2200-6000 b.p. indicates a significant hiatus in the late Holocene occupation of the cave. This hiatus, peculiar to Rose Cottage and some Lesotho sites (Mitchell et al. 1994), is not general in the eastern Free State (Wadley 1995).

Large-scale excavations of 38 sq.m have mapped features and activity areas in the cave (Figure 2). The arrangement of features provides crucial spatial information because artefacts and food debris are likely to cluster around features (e.g., Clarke 1977). As previous excavations in the 1940s and 1960s (Malan 1958; Beaumont 1963; 1978) have left only half of the cave's deposit intact, it is not possible to plot all the site's features, though the remaining deposit is immensely informative.

The artefacts from the surface floor of the cave reflect an occupation by foragers and black and white settlers in the last few hundred years. Some pottery types from Level Mn suggest that foragers may not have occupied the cave after about 500 b.p. (Thorp pers. comm.). The 19th-century European artefacts -- bullets, buttons, European pottery and glass (Behrens 1992) -- possibly indicative of the first white settlers in the area in the 1830s (Casalis 1861; Loubser & Laurens 1994), may have been left by the settlers themselves or they may be objects curated by foragers or Bantu-speaking agriculturists. Some pottery and bone from domestic animals may have been introduced into the cave during the late 18th- and early 19th-century unrest in the area (e.g. Cobbing 1988). The defensive agriculturist settlement just below Rose Cottage Cave may date to an occupation in the Ladybrand area by this time period.

Level 1: Mn Level

Mn contains a predominance of artefacts and fauna typical of Later Stone Age campsites; also pottery and some domesticates (Wadley 1991; 1992). At 500 b.p no agriculturist settlements appear to have been established in the area and the pottery and animals may have been brought from forager or agriculturist settlements at some distance. The Stone Age material culture in Level Mn comprises many finely made end-scrapers, some worked bone and a few ostrich-eggshell beads. The fauna includes a wide range of large and small bovids, for example, red hartebeest (Alcelaphus buselaphus), eland (Taurotragus oryx) and springbok (Antidorcas marsupialis), and many small collectable creatures such as rock hyrax (Procavia capensis), freshwater mussels (Cafferia caffer) and fish; Barbus spp. (yellowfish), Clarias gariepinus (catfish), Labeo spp. (mudfish).

Two inconsistencies occur between the rock art and the excavated material. Exotic mormy-rids, absent from the excavated material, are present in the rock art, and agriculturist pottery was excavated but is not present or alluded to in the rock art. These are useful inconsistencies. The excavated fish remains are of Labeo spp., Barbus spp. and Clarias gariepinus (Hall pers. comm.). There are no mormyrid bones in the deposit, and palaeoclimatic evidence indicates that mormyrids probably never occurred in the environs (Butzer 1984; Butzer & Vogel 1979; Wadley 1991:127; Tyson & Lindesay 1992: 275 -- 6; Skelton 1993; Mitchell 1994: 85). The mormyrid images indicate that the forager exchange networks of the southeastern mountains involved movement of personnel, objects and information over considerable distances.

Since pastoralists and agriculturists did not settle in the Ladybrand area prior to 200 b.p. (Maggs 1976), the domesticates and pottery may have arrived at the cave via forager exchange networks or via direct exchange at farming settlements when foragers visited areas outside the Free State or farmers made exploratory forays into it (see also Mazel 1992; Loubser & Laurens 1994). Thorp (pers. comm.) considers:

Neither scenario is likely to have involved the foragers in the type of intense relationship with agriculturists (such as Campbell 1987 inferred in the Drakensberg) that would have encouraged them to incorporate agriculturist images into their shamanistic experiences. Clearly this situation changed after agriculturists settled in the area, but Rose Cottage seems not to have been occupied by foragers at this stage.

Although agriculturist images were not incorporated into the rock art of Rose Cottage, the excavated material shows that the foragers and farmers were already in contact. The early traveller Stow mentions that the San occupying what is today the Ladybrand and adjacent Magisterial Districts were called the Boroa ba Makhoma Khotu or `Bushmen of many cattle' by the Basotho (1905: 183). The rock art of Rose Cottage Cave may not be unaffected by contiguity with Bantu-speaking farmers.

Features in level Mn include three hearths and a few small pits (Figure 7). Chips and chunks (the waste from stone-tool knapping) and unretouched flakes (the primary products of knapping) are clustered near the northwestern cave entrance, where the lighting is good. The flake cluster is of low density and the flake distribution diffuse, perhaps because flakes were removed from the knapping area for tasks in other activity areas. Adzes, ostrich-eggshell fragments, pottery and waste bone fragments have their highest densities in the northeastern part of the cave. Just south of this concentration, a dense cluster of bone shafts and bone points that were probably used as arrowheads (Figure 7) may show where hunters made and maintained their hunting equipment.

[FIGURE 7 ILLUSTRATION OMITTED]

Six upper grindstones form a cluster near the back of the cave in an area where there are no other artefact clusters. Residue analysis on the grindstones shows starch grains (Wallace pers. comm.) indicating the probable processing of gathered plant foods. A separate cluster of ochre and haematite associated with a single grindstone and hearth is located in square Q1, several metres away. Heating ochre causes it to change colour to a richer hue, making it easier to crush and prepare as well as providing it with additional supernatural potency (How 1970:35). The preparation of pigment, according to Mapote, the Basotho man with San stepbrothers and How's informant, was performed by a woman (How 1970: 35). Like the lioness painted in Rose Cottage Cave which provided an enduring visual metaphor of certain forager gender relations, the role of a feminine ideology in the preparation and production of rock art appears to have been fundamental (see also Solomon 1994; Ouzman 1996:46-51; Stevenson 1996). Thus hearth, ochre and grindstone may represent a discrete, even sacred, ochre preparing area. The separate plant- and ochreprocessing features illustrate the different uses to which otherwise identical material culture items may be put. A communal arrangement of the plant-processing grindstones would be an anomaly in terms of !Kung ethnography which indicates that each !Kung woman usually prepares food around her own family hearth. Such anomalies indicate that Later Stone Age gender and social relations need not be the same as those of modern San, anomalies to be an expected part of the historical process though consonances between the two sometimes exist.

Interestingly, the spaces devoted to plant-food processing and to the manufacture and mending of hunting kits are separate, a situation due to gender segregation in contemporary San ethnography. We cannot be sure that similar gender segregation occurred at Rose Cottage Cave, but the material culture signatures are tantalising.

Level 2: 1

Level A contains pottery and some domesticated animal bones in addition to a richer Later Stone Age material culture than that in level Mn. The material culture includes engraved bone and ostrich-eggshell, delicate bone `fish hooks', finely made lithics, ostrich-eggshell beads and a perforated marine shell (Nerita spp.) which probably came from the east coast of South Africa some 300 km distant (Wadley 1991; see also Mazel 1989: 148; Mitchell 1995: 35). This find provides independent verification of extensive exchange networks hinted at by the painting of the `exotic' mormyrid fish. Obtaining the marine shell would have necessitated negotiating the social and physical boundaries of other forager or even agriculturist communities, a negotiation likely in the context of corporate aggregation (e.g. Wiessner 1984) and its concomitant opportunities for conspicuous display and prestige. Rose Cottage Cave, as a pre-eminent site, would have been an appropriate locus for an extensive exchange network.

Level A features include three hearths, each with their own large ash accumulations, which cluster in the northeast of the excavated grid. Several individual hearths form another cluster some distance away (Figure 8). Widely broadcast chips, chunks and flakes have their main concentration around and in three hearths and their associated ash scatters, which must have been the focus of many camp activities. Around hearths the boundaries between activity areas may become blurred and artefacts become associated by chance (Yellen 1977). Bone points, bone shafts, waste bone, colouring material, grindstones, knapping debris and pottery cluster as part of a `smudging' of artefacts and debris around the three hearths in level A.

[FIGURE 8 ILLUSTRATION OMITTED]

By the model of !Kung ethnography, the material culture of level A indicates an aggregation camp; these camps generate so much activity that household borders soon become confused (Brooks 1984). As aggregation sites are often reoccupied but are never used in precisely the same way, scatters of bone, charcoal and tools become increasingly diffuse through time (Brooks 1984). In another possible interpretation, the three ash scatters represent refuse dumps into which all activity debris from around the hearths was swept. In !Kung camps such clusters represent secondary disposal areas which, in turn, suggest long-term camps (Bartram et al. 1991: 142). The two interpretations of the hearth and ash accumulations in Rose Cottage are not mutually exclusive; both suggest an aggregation camp.

Further support for aggregation is provided by the enormous quantity (33.6 kg) of waste bone from a wide range of species (Plug & Engela 1992); large and small game suggests that bow-and-arrow hunting and snaring took place. Amongst the !Kung, bow-and-arrow hunting is usually associated with the aggregation phase when men conduct group hunts (Biesele 1993).

More evidence of aggregation is provided by a discrete ostrich-eggshell bead-manufacturing area. Incomplete beads cluster in the centre of the grid, well behind the cluster of three hearths (Figure 8). Two grooved beadstones for polishing the rough-hewn beads were recovered from the same square. Since ostrich-eggshell beads were important items in foragers' trading ensembles (e.g. Wiessner 1984), the dedicated bead-manufacturing area within Rose Cottage Cave provides another strand of evidence for participation in exchange networks. Ostrich-eggshell bead manufacture was the preserve of San women in the ethnographic past. If comparable gendered activities extend back 700 years then we suggest that women controlled at least one of the main trading commodities (see also Gero 1991).

Towards synthesis

Rose Cottage Cave provides an opportunity to integrate rock art and excavated material culture, each field complementing the other. The rock art of Rose Cottage Cave provides glimpses of Later Stone Age religious and social beliefs, and of their articulation with economic practices: paintings of `exotic' fish both as metaphors for the supernatural and as signs of exchange networks; beliefs relating to large social gatherings in the image of a lioness and 17 eland; images of an idealized San social order in the painting of a medicine dance. These valuable aspects cannot be accessed from excavated data, while excavated material culture can provide contextual information and insights not derived from the rock art.

In some instances the rock art and material culture corroborate: the long-distance exchange suggested by the mormyrid painting is confirmed by the pierced marine shell from the east coast. The shell could have passed along exchange networks without involving movement of the Rose Cottage Cave people whereas the depiction of the mormyrid and of other animals `exotic' to the area, show that foragers travelled distances.

The prominence of Rose Cottage Cave as a social and ritual centre may have raised certain expectations as to what was appropriate at the site: visually striking rock paintings and worked bone, shell and stone represented a considerable investment in time and effort. The cave floor space was organized using social mores that may have included gender segregation of the type practised in !Kung aggregation camps, with discrete special purpose activities -- ostrich-eggshell bead and bone point manufacture, plant food and pigment processing, and stone-tool knapping -- alongside broad `smudging' of activities around some hearths. All these attributes are consonant with group aggregation, and so is the wide range of large and small animals represented among the faunal remains.

The spatial arrangement of features and material culture sets Rose Cottage Cave apart from other sites, as does the scale and complexity of the art and other material culture, even the appearance and position of the site itself. We suggest that it was the `power of the place', an intangible topophilia (cf. Tuan 1974; Deacon 1988; Tacon 1990; Ouzman 1995a), that repeatedly attracted people there. Even today Rose Cottage Cave attracts pilgrims: many come to view the large cave or what remains of its rock art, while archaeologists are lured by the material culture that the site provides. We, the archaeologists, not unlike our Later Stone Age predecessors, aggregate seasonally in Rose Cottage where we work and rest within carefully demarcated areas. The power of the site continues; its place in the landscape is important now as it was to its Later Stone Age occupants who could never have dreamed of the changes that would and will be wrought in this remarkable cave.

Acknowledgements. We thank Simon Hall for identifying the Rose Cottage Cave fish remains, Paul Skelton for identifying the rock-paintings of mormyrids and for detailed discussion of the problems associated with the identification, James Brink and Nico Avenant for identifying the cattle depictions, Callie Lynch for providing information relating to the distribution of mammals in the Free State, Peter Mitchell for alerting us to the Roma giraffe painting, Carol Wallace for residue analyses of the grindstones, Carla Botha for preparing Figures 1 & 4, Kim Sales for preparing Figures 2, 7 & 8, and the Ladybrand Municipality for its interest in the site. Geoff Blundell, John Clegg, Janette Deacon, Simon Hall, Jannie Loubser, Peter Mitchell, Claire Smith and Paul Tacon made useful comments on a draft of this paper. The Caledon River Rock art Recording Project and the Rose Cottage Cave project are funded by the Human Sciences Research Council. Fieldwork and research expenses have also been met by the National Museum, Bloemfontein and the University of the Witwatersrand Archaeology Department is thanked for its constant support and interest. Opinions expressed in this publication and conclusions arrived at are those of the authors and do not necessarily represent the views of the Centre for Science Development of the Human Sciences Research Council.

(1) Not all rock-art in southern Africa was produced by foragers. Some rock-art is the product of Bantu-speaking agriculturists (e.g. Maggs 1995). We restrict our discussion to forager or `San' rock-art.

(2) The ethnicity of these foragers is an extremely contentious issue (Wilmsen 1989). Both `Bushman' and `San' are unsatisfactory terms for the foraging groups that populated most of southern Africa until recent times. We reject any pejorative associations that the terms `Bushman' or `San' may have. In addition, the `San' or `Bushmen' consisted of many communities, often separated spatially and temporally. We use `San' and `forager', rejecting any negative meanings these words may have. Similarly, Bantu-speaking farmers or Late Iron Age communities are referred to as `farmers' or `agriculturists' (Maggs 1992).

References

Aunger, R. 1995. Ethnography: storytelling or science?, Current Anthropology 36: 97-130.

Barham, L.S. 1992. Let's walk before we can run: an appraisal of historical materialist approaches to the Later Stone Age, South African Archaeological Bulletin 47: 44-51.

Barnard, A. 1992. Hunters and herders of Southern Africa: a comparative ethnography of the Khoisan peoples. Cambridge: Cambridge University Press.

Bartram, L.E., E.M. Kroll, & H.T. Bunn. 1991. Variability in camp structure and bone food refuse patterning at Kua San hunter-gatherer camps, in E.M. Kroll & T.D. Price (ed.), The interpretation of archaeological spatial patterning: 77-148. New York (NY): Plenum Press.

Battiss, W.W. 1945. Addendum on `Prehistoric fishing scenes', South African Journal of Science 51: 356-60.

Beaumont, P.B. 1963. Two interesting artefacts from Rose Cottage Cave, South African Archaeological Bulletin 18: 60-61. 1978. Border Cave. Unpublished MA thesis: University of Cape Town.

Behrens, J. 1992. European artefacts from Rose Cottage Cave, South African Archaeological Bulletin 47: 13-15.

Bernstein, R.J. 1983. Beyond objectivism and relativism: science, hermeneutics and praxis. Oxford: Blackwell.

Biesele, M. 1993. Women like meat: the folklore and foraging ideology of the Kalahari Ju/'hoan. Johannesburg: Witwatersrand University Press.

Binneman, J. & S.L. Hall. 1993. The context of four painted stones from the eastern Cape, Southern African Journal of Field Archaeology 2: 89-95.

Bleek, W.H.I. & L.C. Lloyd. 1911. Specimens of Bushman folklore. London: George Allen.

Breuil, H. 1945. Sea animals amongst the prehistoric rock paintings of Ladybrand, South African Journal of Science 51: 353-60.

Breunig, P. 1989. Archaeological investigations into the settlement history of the Brandberg, in H. Pager, The rock paintings of the Upper Brandberg. Part I: Amis Gorge: 17-47. Cologne: Heinrich Barth Institute.

Brooker, M. 1980. Rescue excavations at the Welbedacht dam site, Humanitas 6: 35-47.

Brooks, A. 1984. San land-use patterns, past and present: implications for southern African prehistory, in Hall et al. (ed.): 40-52.

Butzer, K.W. 1984. Archaeology and Quaternary environment in the interior of Southern Africa, in R.G. Klein (ed.), Southern African prehistory and palaeoclimates: 1-64. Rotterdam: Balkema.

Butzer, K.W. & J.C. Vogel. 1979. Archaeosedimentological sequences from the sub-montane interior of South Africa: Rose Cottage Cave, Heuningneskrans and Bushman Rock Shelter. Paper presented at the 6th Biennial Conference of the Southern African Association of Archaeologists, Stellenbosch.

Butzer, K.W., G.J. Fock, L. Scott & R. Stuckenrath. 1979. Dating and context of rock engravings in southern Africa, Science 203: 1201-14.

Campbell, C. 1986. Images of war: a problem in San rock art research, World Archaeology 18: 255-68. 1987. Art in crisis: contact period art in the south-eastern mountains of southern Africa. Unpublished MA thesis, University of the Witwatersrand.

Campbell, J.B. & M. Mardaga-Campbell. 1993. From micro- to nano-stratigraphy: linking vertical and horizontal dating of archaeological deposits with the direct dating of rock art at `The Walkunders', Chillagoe (north Queensland, Australia), in J. Steinbring, A. Watchman, P. Faulstich & P.S.C. Tacon (ed.), Time and space: dating and spatial considerations in rock art research: 57-63. Melbourne: Australian Rock Art Research Association. Occasional publication 8.

Carter, P. 1970. Late Stone Age exploitation patterns in southern Natal, South African Archaeological Bulletin 25: 55-8.

Casalis, E. 1861. The Basutos or, twenty-three years in South Africa. London: James Nisbet & Company.

Cashdan, E. 1983. Territoriality among human foragers: ecological models and an application to four Bushman groups, Current Anthropology 24: 47-66.

Clarke, D. 1977. Spatial information in archaeology, in D. Clarke (ed.), Spatial archaeology. 1-32. London: Academic Press.

Cobbing, J. 1988. The Mfecane as alibi: thoughts on Dithakong and Mbomolompo, Journal of African History 29:487-519.

Conkey, M.W. 1985. Ritual, communication, social elaboration and variable trajectories of Palaeolithic archaeology, in T.D. Price & J.A. Brown (ed.), Prehistoric hunter-gatherers: the emergence of cultural complexity: 299-323. Orlando (FL): Academic Press.

David, B., I. Mcniven, V. Attenbrow, J. Flood, & J. Collins. 1994. Of Lightning Brothers and white cockatoos: dating the antiquity of signifying systems in the Northern Territory, Australia, Antiquity 68: 241-51.

Deacon, J. 1988. The power of a place in understanding southern San rock engravings, World Archaeology 20: 129-40.

Dornan, S.S. 1909. Notes on the Bushmen of Basutoland, Transactions of the South African Philosophical Society 18: 437-50.

Dowson, T.A. 1994. Reading art, writing history: rock art and social change in Southern Africa, World Archaeology 25: 332-45.

Dowson, T.A. & J.D. Lewis-Williams (ed.). 1994. Contested images: diversity in southern African rock art research. Johannesburg: Witwatersrand University Press.

Editorial. 1992. South African Archaeological Bulletin 47: 2.

Ellenberger, V. 1953. La fin tragique des Bushmen. Paris: Amiot Dumont.

Garlake, P. 1995. The hunter's vision: the prehistoric art of Zimbabwe. London: Zimbabwe Publishing House.

Gero, J.M. 1991. Genderlithics: women's roles in stone tool production, in J.M. Gero & M.W. Conkey (ed.), Engendering archaeology. women and prehistory: 163-93. Oxford: Basil Blackwell.

Goodwin, A.J. 1960. Chemical alteration (patination) of stone, South African Archaeological Bulletin 15: 67-76.

Gordon, R. 1992. The Bushman myth: the making of a Namibian underclass. Boulder (CO): Westview Press.

Guenther, M.G. 1986. The Nharo Bushmen of Botswana: tradition and change. Hamburg: Helmut Buske Verlag. Quellen zur Khoisan-Forschung 3.

Gulbrandsen, O. 1991. On the problem of egalitarianism: the Kalahari San in transition, in R. Gronhaug. R, G. Haaland & G. Henriksen (ed), The ecology of choice and symbol: essays in honour of Frederik Barth: 81-110. Bergen: Alma Mater Forlag.

Hall, M, G. Avery, D.M. Avery, M.L. Wilson & A.J.B. Humphreys (ed.). 1984. Frontiers: southern African archaeology today. Oxford: British Archaeological Reports. International series 207.

Hall, S.L. 1994. Images of interaction: rock art and sequence in the eastern Cape, in Dowson & Lewis-Williams: 61-82.

Hamilton, C. 1995. (ed.). The `Mfecane' aftermath: reconstructive debates in southern African history. Johannesburg: Witwatersrand University Press.

Hammersley, M. 1992. What's wrong with ethnography?: methodological explorations. London: Routledge.

Heinz, H-J. 1975. Elements of !Ko religious belief, Anthropos 70: 17-41.

Hewitt, R.L. 1986. Structure, meaning and ritual in the narratives of the southern San. Hamburg: Helmut Buske Verlag. Quellen zur Khoisan-forschung 2.

Holm, E. 1958. Dating our petroglyphs, South African Archaeological Bulletin 13:34-35.

How, M.W. 1970. The mountain Bushmen of Basutoland. Pretoria: J.L. van Schaik.

Humphreys, A.J.B. 1971. Age determination of the rock art of Southern Africa, in M. Schoonraad (ed.), Rock paintings of Southern Africa, South African Journal of Science Special Issue 2: 86-90.

Huffman, T.N. 1983. The trance hypothesis and the rock art of Zimbabwe, South African Archaeological Society Goodwin Series 4: 49-53.

Ingold, T. 1993. The temporality of the landscape, World Archaeology 25: 152-72.

Johnson, M.H. 1989. Conceptions of agency in archaeological interpretation, Journal of Anthropological Archaeology 8: 189-211.

Jolly, P. 1996. Symbiotic interaction between black farmers and southeastern San: implications for southern African rock art studies, ethnographic analogy, and hunter-gatherer cultural identity, Current Anthropology 37: 277-305.

Katz, R. 1982. Boiling energy: community healing among the Kalahari !Kung. Cambridge (MA): Harvard University Press.

Kent, S. 1992. The current forager controversy: real versus ideal views of hunter-gatherers, Man (NS) 27: 45-70.

Kinahan, J. 1991. Pastoral nomads of the Central Namib Desert. Windhoek: New Namibia.

Klatzow, S. 1994. Roosfontein, a contact site in the eastern Orange Free State, South African Archaeological Bulletin 49: 9-15.

Lee, R.B. 1979. The !Kung: men, women and work in a foraging society. Cambridge: Cambridge University Press. 1992. Art, science or politics?: the crisis in hunter-gatherer studies, American Antiquity 94: 31-54.

Lewis-Williams, J.D. 1980. Ethnography and iconography: aspects of southern San thought and art, Man 15: 467-82. 1981a. Believing and seeing: symbolic meanings in southern San rock paintings. London: Academic Press. 1981b. The thin red line: Southern San notions and rock paintings of supernatural potency, South African Archaeological Bulletin 36: 5-11. 1984. Ideological continuities in prehistoric southern Africa, in Schrire: 225-52. . 1985. Testing the trance explanation of southern African rock art: depictions of felines, Bolletino del Centro Camuno di Studi Preistotici 22: 47-62. 1992. Ethnographic evidence relating to `trance' and `shamans' among Northern and Southern Bushmen, South African Archaeological Bulletin 42: 56-60.

Lewis-Williams, J.D. & M. Biesele. 1978. Eland hunting rituals among northern and southern San groups: striking similarities, Africa 48: 117-31.

Lewis-Williams, J.D. & J.H.N. Loubser. 1986. Deceptive appearances: a critique of southern African rock art studies, Advances in Archaeological Method and Theory 5: 253-89.

Loubser, J.H.N. 1993. A guide to the rock paintings of Tandjesberg, Navorsinge van die Nasionale Museum Bloemfontein 9: 345-84.

Loubser, J.H.N. & G. Laurens. 1994. Depictions of domestic ungulates and shields: hunter-gatherers and agropastoralists in the Caledon river valley area, in Dowson & Lewis-Williams: 83-118.

Loubser, J.H.N. & P.C. Zietsman. 1994. Rock painting of postulated Brunsvigia sp. (Amaryllidaceae) at Thaba Bosiu, western Lesotho, South African Journal of Science 90: 611-12.

Lynch, C. 1991. Elephants in the Orange Free State?, Culna 40: 13. 1993. Giraffes in the Orange Free State? Culna 45: 32-3.

Maggs, T.M.O'C. 1976. Iron Age communities of the Southern Highveld. Pietermaritzburg: University of Natal Press. 1992. Name calling in the Iron Age, South African Archaeological bulletin 47:131. 1995. Neglected rock art: the rock engravings of agriculturalist communities in South Africa, South African Archaeological Bulletin 50: 132-42.

Malan, B.D. 1958. The final phase of the Middle Stone Age in South Africa, Proceedings of the Pan African Congress on Prehistory (1947): 188-94. New York (NY): Philosophical Library.

Marshall, L. 1961. Sharing, talking, and giving: relief of social tensions among !Kung Bushmen. Africa 31: 231-49. 1969. The medicine dance of the !Kung Bushmen, Africa 39: 347-81. 1976. The !Kung of Nyae Nyae. Cambridge (MA): Harvard University Press.

Mazel, A.D. 1989. People making history: the last ten thousand years of hunter-gatherer communities in the Thukela Basin, Natal Museum Journal of humanities 1: 1-168. 1992. Early pottery from the eastern part of southern Africa, South African Archaeological Bulletin 47: 3-7. 1994. Dating the Collingham Shelter rock paintings, Pictogram 6: 33-5. 1995. Drakensberg rock paintings directly dated, Natal Witness, 8 July 1995.

Mitchell, P.J. 1995. Revisiting the Robberg: new results and a revision of old ideas at Sehonghong rock shelter, Lesotho, South African Archaeological Bulletin 50: 28-38. 1996. Prehistoric exchange and interaction in southeastern southern Africa: marine shells and ostrich eggshell, African Archaeological Review 13: 35-76.

Mitchell, P.J., J.E. Parkington & R. Yates. 1994. Recent Holocene archaeology in western and southern Lesotho, South African Archaeological Bulletin 49: 33-52.

Morris, D. & P. Beaumont. 1994. Portable rock engravings at Springbokoog and the archaeological contexts of rock art of the Upper Karoo, in Dowson & Lewis-Williams: 11-28.

Orpen, J.M. 1874. A glimpse into the mythology of the Maluti Bushmen, Cape Monthly Magazine (NS) 9: 1-13.

Ouzman, S. 1995a. Spiritual and political uses of a rock engraving site and its imagery by San and Tswana-speakers, South African Archaeological Bulletin 50: 55-67. 1995b. The fish, the shaman and the peregrination relation: rock paintings of mormyrid fish as religious and political metaphors, Southern African Field Archaeology 4: 3-17. 1996. Thaba Sione: place of rhinoceroses and rock art, African Studies 55(1): 31-59. In press. Hidden in the common gaze: journey and metaphor at Rose Cottage Cave, South African Journal of science.

Pager, H. 1971. Ndedema. Graz: Akademische Druk.

Plug, I. & R. Engela. 1992. The macrofaunal remains from recent excavations at Rose Cottage Cave, Orange Free State, South African Archaeological Bulletin 47: 16-25.

Prins, F.E. & S. Hall. 1994. Expressions of fertility in the rock art of Bantu-speaking agriculturalists, African Archaeological Review 12: 171-203.

Roth, P.A. 1989. Ethnography without tears, Current Anthropology 30: 555-70.

Sampson, C.G. 1972. The Stone Age industries of the Orange River Scheme and South Africa, Memoirs of the Notional Museum, Bloemfontein 6:1-288.

Sampson, C.G. & J.C. Vogel. 1989. A painted pebble and associated C-14 date from the Upper Karoo, Pictogram 2: 1-3.

Schrire, C. (ed). 1984. Past and present in hunter-gatherer studies. Orlando(FL): Academic Press. 1992. The archaeological identity of hunters and herders at the Cape over the last 2000 years: a critique, South African Archaeological Bulletin 47: 62-4.

Silberbauer, G.B. 1981. Hunter and habitat in the central Kalahari desert. Cambridge: Cambridge University Press.

Singer, R. & J. Wymer. 1969. Radiocarbon date for two painted stones from a coastal cave in South Africa, Nature 224: 508-10.

Skead, C.J. 1980. Historical mammal incidence in the Cape Province, Vol 1. Cape Town: Department of Nature and Environmental Conservation of the Provincial Administration of the Cape of Good Hope.

Skelton, P. 1993. A complete guide to the freshwater fishes of Southern Africa. Johannesburg: Southern Book Publishers.

Skinner, J.P. & R.H.N. Smithers. 1990. The mammals of the Southern African subregion. Pretoria: University of Pretoria Press.

Smith, A.B. 1994. Metaphors of space: rock art and territoriality in Southern Africa, in Dowson & Lewis-Williams: 373-84.

Smith, A.B., K. Sadr, J. Gribble & R. Yates. 1991. Excavations in the south-western Cape, South Africa, and the archaeological identity of prehistoric hunter-gatherers within the last 2000 years, South African Archaeological Bulletin 46: 71-91,

Solomon, A. 1994. `Mythic women': a study in variability in San rock art and narrative, in Dowson & Lewis-Williams: 331-71.

Stevenson, J. 1996. Man-the-shaman: is it the whole story? A feminist perspective on the San rock art of southern Africa. MA thesis: University of the Witwatersrand.

Stow. G.W. 1905. The native races of South Africa. London; Swan Sonneschein.

Tacon, P.S.C. 1990. The power of place: cross-cultural responses to natural and cultural landscapes of stone and earth, in J.M. Vastokas, J. Paper & P.S.C. Tacon (ed.), Perspectives of Canadian landscape: native traditions: 11-43. York (ONT): York University, Robarts Centre for Canadian Studies.

Tacon, P.S.C. & S. Brockwell. 1995. Arnhem Land prehistory in landscape, stone and paint, Antiquity 69: 676-95.

Testart, A. 1988. Some major problems in the social anthropology of hunter-gatherers, Current Anthropology 29: 1-31.

Thackeray, A.I. 1983. Dating the rock art of southern Africa, South African Archaeological Society Goodwin Series 4: 21-6.

Thorp, C.R. 1996. A preliminary report on evidence of interaction between hunter-gatherers and farmers along a hypothesised frontier in the eastern Free State, South African Archaeological Bulletin 51: 57-63.

Tuan, Y-F. 1974. Topophilia: a study of environmental perception, attitudes and values. Englewood Cliffs (NY): Prentice-hall.

Tyson, P.D. & J.A. Lindesay. 1992. The climate of the last 2000 years in southern Africa, Holocene 2: 271-8.

Van Der Merwe, N.J., J. Sealey & R.J. Yates. 1987. First accelerator carbon-14 date for pigment from a rock painting, South African Journal of science 83: 56-7.

Van Riet Lowe, C. 1933. The age of rock engravings in South Africa, South African Journal of Science 30: 525-6. 1947. Rock paintings of marine animals in the interior of South Africa, South African Archaeological Bulletin 2: 41-5.

Vinnicombe, P. 1976. People of the eland: rock paintings of the Drakensberg bushmen as a reflection of their life and thought. Pietermaritzburg: Natal University Press.

Wadley, L. 1987. Later Stone Age hunters and gatherers of the southern Transvaal. Oxford: British Archaeological Reports. International series 380. 1991. Rose Cottage Cave: background and a preliminary report on the recent excavations, South African Archaeological Bulletin 46: 125-30 1992. Rose Cottage Cave: the Later Stone Age levels with European and Iron Age artefacts, South African Archaeological Bulletin 47: 8-12. 1995. Review of dated Stone Age sites recently excavated in the eastern Free State, South Africa, South African Journal of Science 91: 574-9.

Wells, L.H. 1946. Marine animals in a rock painting near Fouriesburg, O.F.S, South African Journal of Science 52: 236-9.

Wendt, E.W 1976. `Art mobilier' from the Apollo 11 cave, South West Africa: Africa's oldest dated works of art, South African Archaeological Bulletin 31: 5-11.

Whitley, D.S. & H.J. Annegarn. 1994. Cation-ratio dating of rock engravings from Klipfontein, Northern Cape, in Dowson & Lewis-Williams: 189-97.

Wiessner, P. 1983. Style and social information in Kalahari San projectile points, American Antiquity 48: 253-76. 1984. Considering the behavioural basis for style: a case study among the Kalahari San, Journal of anthropological Archaeology 3: 190-234.

Willcox, A.R. 1962. Marine animals in rock paintings, South African Journal of science 5: 6-7.

Wilmsen, E.N. 1989. Land filled with flies: a political economy of the Kalahari. Chicago (IL): University of chicago Press.

Wright, J.B. 1971. Bushman raiders of the Drakensberg 1840-1870. Pietermaritzburg: University of Natal Press.

Wylie, A. 1985. The reaction against analogy, Advances in Archaeological Method and Theory 8: 63-111. 1989. Archaeological cables and tacking: the implications of practice for Bernstein's `Options beyond objectivism and relativism', Philosophy of the Social Sciences 19: 1-18.

Yates, R.J. & A. Jerardino. 1996. A fortuitous fall: early rock paintings from the west coast of South Africa, South African Journal of science 92: 110.

Yates, R.J. & A.B. Smith. 1993. A reevaluation of the chronology at Oudepost: a reply to Schrire, South African Archaeological Bulletin 48: 52-3.

Yellen, J.E. 1977. Archaeological approaches to the present. New York: Academic Press.

Yellen, J.E. 1984. The integration of herding into prehistoric hunting and gathering economies, in Hall et al. (ed.): 53-64.

Sven Ouzman & Kyn Wadley, Ouzman Department of Rock Art, National Museum, PO Box 266, Bloemfontein, 9300, SOuth Africa. E-mail: rockart@internext.co.za.

Wadley, Department of Archaeology, University of the Witwatersrand, PO Wits, 2050, South Africa. E-mail: 107lym@cosmos.wits.ac.za

Received 27 August 1996, accepted 11 January 1997, revised 17 March 1997.