Heritage Cases

EXECUTIVE SUMMARY Kruger Cave is a Later Stone Age site in eastern fringe of the Magaliesberg with exceptional organic preservation and many unique artefacts. The shelter was last excavated in 1983 and was never backfilled or stabilised. A site visit in 2017 revealed that the site is in a poor condition with the excavation sides eroding away and rich archaeological material slumping out of context. There are signs of recent occupation that will jeopardise the integrity of the archaeological deposit. I propose to stabilise the site by sandbagging and backfilling the open excavation trenches. Before this is done I will open up two small sections (maximum 2 m in length) by cutting back 50 cm from the present profile section. These sections will be excavated stratagraphically and I plan to submit sediment samples for aDNA analysis. The grass layers will be preserved in gypsum jackets (see Backwell et al. 2018) for analysis to see what plants compose these layers and whether the grass was blown in or deliberately placed. INTRODUCTION Kruger Cave is a 30 m deep rock shelter overlooking the Hex River on the farm Commissiedrift 327JQ in the Rustenburg District, Magaliesberg (Figure 1). The site preserves a Later Stone Age occupation, with two Oakhurst expressions dating from 10480-5400 BP, and a Wilton assemblage 1390-1020 BP (Mason 1988). The Oakhurst technocomplex at this site preserves unusually late in time, overlapping with expressions of the Wilton technology at nearby sites, like Jubilee Shelter (cf. Wadley 1996; Lombard et al. 2012). The Kruger Cave deposit is extremely rich in terms of organic preservation (see below). Some of the remarkable organic finds from this site include: a bladder cap with human hair preserved; a bone quiver with bone arrowheads inserted, poisoned bone and wooden points, one of which also has beautifully preserved ochre paint; finely plaited plant-fibre rope, and chewed bark (Figure 2). Two individuals of the extinct springbok (Antidorcus bondi) were also found in the deposit (Brown & Verhagen 1985). Kruger Cave is one of only seven rock art sites in the Magaliesberg. Pager (1987) recorded 57 painting on the walls of the shelter, the overwhelming majority of which are depictions of humans. Birds and animals are also recorded. Pager noted that the human depictions at Kruger Cave are on average larger than similar depictions at other sites in the area. Along with other hunter-gatherer sites in the Magaliesberg valley, Kruger Cave was abandoned between AD 300-600 when the first Bantu-speaking farmers infiltrated the landscape. Indeed, Kruger Cave preserves the earliest record (in the form of pot sherds) of Iron Age occupation of the Magaliesberg (Mason 1962), although there is no evidence for farmer occupation of the cave. As with other sites in the region, Kruger Cave was reoccupied from AD600-1300, after which the region was permanently abandoned when the farmers returned. BACKGROUND TO ORIGIONAL EXACVTIONS Kruger Cave was excavated over three field seasons. The first trial excavation was conducted by Revil Mason in 1956. A second small excavation was undertaken in 1974, after public revellers destroyed sections of the site (Mason 1974). A perimeter fence, the remains of which are still present, was erected at this time, but proved ineffective, and by 1986 had been effectively destroyed (Figure 3). Repeated vandalism of the site caused Mason to return to the site in 1982 and 1983 when Robbie Steele conducted a full-scale excavation (Mason 1987, 1988). The rationale for the ’82-’83 campaign, which removed approximately two-thirds of the deposit, was to 1) rescue data before further damage occurred, 2) ascertain the relationship between the rock art and the site’s occupants, and 3) gather more information about the plant material that people were using in the region between 10 000-1 000 BP (Steele 1987). The deposit was excavated along a NW to SE grid (Figure 4) and the deposit removed in 10 cm spits and sieved through an 8 mm mesh. Five depositional clusters were identified and three were dated. Cluster 1 calibrated to 10608-8239 BC, Cluster 2 calibrated to 6390-4053 BC, and Cluster 4 was uncalibrated at 1390-1020 BP. The sediment deposit is largely determined by the natural slope of the shelter floor. Very little sediment is present in the SE section, whereas the depth of deposit in the NW section is approximately 90 cm. natural sediment stratigraphy is horizontal to a depth of ~40 cm whereafter more complicated grass and vegetal layers occur. These vegetal layers were dated to 6600±100 BP (Mason 1988). The first millennium AD occupation was restricted to the outer section of the shelter, near the drip line, and it is likely that the paintings date to this period (Mason 1988). The artefacts from this final period of occupation resemble contemporaneous assemblages from other sites in the area, such as Munro cave and Oliebooomspoort (Mason 1988). Lithic raw material is dominated by quartzite, hornfels, jaspelite, chert and andersite. The fauna deposit is dominated by fresh water mussels, and, while it is clear that gathering was the dominant economic activity, hunted bovids made up the bulk of the meat (Brown 1987). Although the sample is too small to inform on changes in hunting technique, it is apparent that larger bovids increase in the later phases of occupation. Fishing occurred throughout the deposit; unsurprisingly given the site’s proximity to perennial water. IMPORTNANT FINDS As mentioned in the introduction Kruger Cave has extraordinary organic preservation. An unusual bladder cap with human hair and wooden peg attached was recovered from the fore of the shelter in an undated deposit. Bladders and pegs were used as ornaments among the Nguni, though not the Tswana, suggesting that this artefact might date to the second quarter of the 19th century when Mzilikatzi’s Ndebele were moving through the area (Steele 1987; Mason 1988). Most of the rest of the interesting artefacts were recovered from the excavations in the rear of the shelter (Steele 1987). Sixty-three bone tools were recovered from the Oakhurst and Ceramic Final Later Stone Age deposits (Steele 1987). Some of the latter are engraved with similar motifs to those seen at Mapungubwe (Voigt 1983) and Olieboomspoort (Mason 1988). Similarly, four wooden arrowheads from the 6690-6600BP layers were recovered. One of these displays a form only seen in bone and ivory at some Eastern Cape sites (Bradfield 2014). One bone and one wooden point are covered in poison (Figure 2E), the chemical characterisation of which is in progress (Bradfield et al. in prep). These two poisoned arrowheads are the oldest unequivocal contextualised examples that I am aware of in southern Africa (but see Langley et al. 2016 for examples from Zanzibar). The shaft of a femur was found with bone points embedded into the cavity like arrows in a quiver (Mason 1988). This is the only known example of a bone quiver from southern Africa. Pieces of rope or string that have been twisted and plaited were recovered from the same layers and in the same area as the bone points. Their preservation is remarkable (Figure 2A). In the NW corner of the shelter, close to where the bone points and string were recovered, Steele found a cluster of ostrich eggshell beads, bone pendants and quartz crystals. This clutch was deliberately placed under a layer of boulders and large rocks and dates to ±6000 BP. One of the bone ‘pendants’ resembles an aerophone recently identified at Matjes River (Kumbani et al. 2019). The lithic component of the site consisted of large side-struck flakes, typical of the Oakhurst technocomplex, that were used for hide scraping and wood working (Mason 1988). The absence of cores suggests that the stone was knapped off site and the flakes brought in. PROPOSED SCHEME OF WORK I propose to cut back by 50 cm along two metres of the least eroded parts of the 1982-1983 trench (See Figure 4). I will do this in two areas in the rear of the shelter, where most of the organic artefacts were found. This will total a maximum of 1 x 2 m2 of deposit. I will apply excavation procedures and protocols following those of the recent re-investigations of Border Cave, KwaZulu-Natal (SAHRIS caseID 7645; Backwell et al. 2018). Excavated material will be sieved through a 1 mm mesh to catch micro-items that would have escaped the 8 mm sieved used by Steele. Sediment samples will be taken for micromorphology and aDNA analysis in collaboration with Prof. Marlize Lombard and Dr Riaan Rifkin. Grass layers, clearly visible in the exposed sections, could represent wind-blown deposit or deliberate bedding material. Sections will be jacketed and removed intact for further computed tomographic analysis to determine its origin. Upon completion of excavations I will place loose Perspex plates against the sides of the new excavations, remove the 36-year-old slump, and shore up the sides of the original excavations before backfilling and stabilising the site. The slump material will be sieved and any ex-situ material will be used for teaching purposes in the Palaeo-Research Institute, University of Johannesburg. Sandbags will be placed on the bedrock floor of the shelter up to 10 cm from present surface level. Thereafter, commercially available sterile coarse sand will be used to fill up the remaining cavity and ‘carpet’ the floor of the shelter. I have opted for this approach as the ‘imported’ sand is easy to recognise from the archaeological deposit, should someone wish to re-excavate in the future, and the size of the sand grains renders it is less likely to infiltrate into the archaeological deposit over time. ‘Carpeting’ the floor of the shelter is intended to impart a natural look to the site to make it less obvious to people wandering into the shelter that it is covering an archaeological excavation. I feel that this will be the best protective measure for the site, since fencing the site has clearly not worked and there is no reason to believe that erecting a new fence will have a different outcome. The site is very close to a main road and a popular fishing spot; a new fence would only draw unnecessary attention and potentially destructive curiosity to the site. LAND OWNERSHIP Kruger Cave is located on portion 47 of the farm Commissiesdrift 327 JQ, in the Rustenburg municipal district. According to the Deeds Office and the Department of Rural Development and Land Affairs the land was bought and registered by the Olifantsnek Development Company in 1938. As the company no longer exists and no change of title was ever registered with the Deeds Office, its land automatically reverts to the State and is administered by the Department of Public Works (see appendix). Permission to excavate Kruger Cave and access the land on which it is located must therefore be granted only by the State, which for the purposes of archaeological research and mitigation is SAHRA. I have verbal permission from Mr Lucus Oosthuisen to traverse his land immediately adjacent the site to gain access to the cave. MOTIVATION FOR RENEWED INVESTIGATION In December 2017 I visited Kruger Cave and found it in an appalling condition. The trenches from Steele’s excavations were never back-filled and have been standing open and exposed these last 36 years. The walls of most of the trench have eroded back and rich archaeological deposit is slumped out of context. Furthermore, the shelter is intermittently inhabited by vagrants (L. Oosthuisen, pers comm. 2017) as could be seen by the remains of two fire places in the rear and centre of the shelter and the numerous footprints covering the floor (Figure 3D). The sections where the deposit is still intact is likely to succumb eventually given prolonged human activity and occupation of the site. Of the approximately one third of the original deposit left after the Steele excavations, even less now survives. The paintings on the wall were extremely faded when they were recorded in 1974, but no trace of them now remains, at least based on my cursory observation of the rock face. My proposal is two-fold: 1) To undertake two small-scale excavations to document the previously undocumented stratigraphy and to tie this is to the carbon dates and archaeological finds, and 2) to stabilise the site to prevent the eventual total destruction of the archaeological deposit. 1) A re-examination of an archaeological site almost 40 years on may seem self-evident, but there are specific reasons for undertaking this work. Both Mason and Steele excavated the site in 10 cm spits, despite clearly visible stratigraphy. This means that despite the extensive collection of material from the 1982-1983 excavations we have no information on the stratigraphy and how the C14 dates relate to the depositional sequence in the shelter. It is also certain that the five occupation clusters identified by Mason (1988) are courser than reality because each 10 cm spit would have subsumed several deposition horizons into a single unit. Modern technology and techniques, not available to Mason and his team, can provide new insights in to the Kruger cave occupation. One example is D-Stretch, which has the potential to elucidate details in the paintings that were previously missed. It is uncertain whether any of the 57 paintings survive, but it would be worthwhile trying to locate them and subjecting them to a modern state-of-the-art study. Coupled to this, the refinement of ancient DNA analysis now allows us to extract microbial DNA from sediments that can tell us something about human diseases and parasites in the past (Rifkin et al. 2017; Houldcroft et al. 2017; Pimenoff et al. 2018). I will be working with Prof. Marlize Lombard and Dr Riaan Rifkin in this regard. There is good reason to believe that parasite DNA will be present, judging by the macroscopic findings of Trichuris and Ascaris eggs in a coprolite (Evens et al. 1996). The proposed re-excavations are relatively small-scale and easily managed, which means it is ideal for student training. Post-graduate students at the University of Johannesburg have few opportunities to gain field-work experience and none at Later Stone Age sites. The two excavations and rehabilitation work will be used to acquaint honours and masters students with LSA sites and stratigraphic excavation techniques. The existing material from the site is also suitable for student projects. For example, the analyst Mason hired to do the botanical study did not identify the grass and charcoal remains, nor has use-wear been conducted on more than a handful of artefacts, leaving plenty of room for new student research projects. 2) My motivation for wanting to stabilise Kruger Cave is because this site is one of the few sites that has exceptional organic preservation, making it an important site for understanding not only developments in the Magaliesberg region, but also the range of organic material culture during the 10 000- 1 000 BP period. Many unique artefacts have been recovered from the site, as mentioned above. It is not desirable to barricade the site or to remove people seeking shelter. The only way then to mitigate further destruction of the archaeological deposit is to stabilise the shelter floor.