Zuurberg (Eastern Cape, South Africa) revisited: human remains, dating and archaeological findings

ABSTRACT

The Later Stone Age (LSA) site of Zuurberg in the Cape Fold Mountains of South Africa was excavated in the 1920s by F.W. FitzSimons. As with his other excavations, not much is known about the exact provenance and detailed archaeological context of the remains. This paper reports the re-analysis and dating of the human skeletal remains from Zuurberg. The remains were commingled but here sorted according to previously published sources. The remains of eight individuals are present, of which only one (Zg3) is fairly complete. The assemblage includes four older adults, two younger or middle-aged adults, an adolescent and a child. Two of the three individuals who could be assessed showed signs of enamel hypoplasia, which probably attests to some hardship. One of the individuals was dated to 4795 ± 42 BP (Ua-61976), which places this site into the Wilton period of the South African LSA. However, the different levels of burials, as well as the presence of potsherds, may suggest an extended use of the site. Isotopic analysis suggests a mixed diet, which was not depleted of protein. The reassessment, sorting and dating of these remains add value to the collection, especially as they come from an open-air, non-coastal forager site. This assemblage can now be included in future studies of southern African Stone Age foragers.

RÉSUMÉ

Le site de l’âge de pierre récent (LSA) de Zuurberg, dans les montagnes du Cap Fold en Afrique du Sud, a été fouillé dans les années 1920 par F.W. FitzSimons. Comme dans les cas de ses autres fouilles, on sait peu de choses sur la provenance exacte et le contexte archéologique détaillé des vestiges. Cet article fait le rapport de la réanalyse et datation des restes squelettiques humains de Zuurberg. Les restes avaient été mélangés mais ils furent triés au cours de cette étude sur la base de sources précédemment publiées. Les restes de huit individus sont présents, dont un seul (Zg3) est assez complet. L’assemblage comprend quatre personnes âgées, deux adultes plus jeunes ou d’âge moyen, un adolescent et un enfant. Deux des trois individus qui ont pu être évalués présentaient des signes d’hypoplasie de l’émail, témoignant probablement de certaines épreuves. L’un des individus a été daté à 4795 ± 42 BP (Ua-61976), ce qui place ce site dans la période Wilton du LSA d’Afrique du Sud. Cependant, les différents niveaux des inhumations, ainsi que la présence de tessons de poterie, peuvent suggérer une utilisation prolongée du site. L’analyse isotopique indique une alimentation mixte qui ne manquait pas de protéines. La réévaluation, le tri et la datation de ces restes ajoutent de la valeur à cette collection, d’autant plus qu’ils proviennent d’un site de chasseurs-cueilleurs à ciel ouvert et non côtier. Cet assemblage peut désormais être inclus dans les futures études sur les chasseurs-cueilleurs de l’âge de pierre en Afrique australe.

KEYWORDS:

• Bioarchaeology
• Later Stone Age
• South Africa
• Tsitsikamma
• skeletal analysis
• radiocarbon dating

Introduction

The result of many excavations conducted in the first half of the twentieth century is that several museums and universities across South Africa hold fairly large, but poorly described and provenanced collections of skeletons (Rightmire 1978; Morris 1992a). This is also the case at the University of the Witwatersrand in Johannesburg, where physical anthropologists such as L.H. Wells and later R.A. Dart excavated and received (from various donors and institutions) many archaeological skeletons. These early researchers published a series of papers with detailed descriptions of skeletons (e.g. Laing 1924; Gear 1926; Wells 1929; Wells and Gear 1931), which are unfortunately of little value today in terms of their conclusions, but helpful for tracing and reassembling commingled skeletons.

During much of the previous century, museum collections mostly came about through the collection of interesting objects (Dunnell 1971), rather than focusing on context and scientific information. The collection of skeletons, and in particular skulls, was no exception, with large numbers of crania excavated to advance various aspects of ‘race’ science (Morris 2022). In their analyses of human remains from an archaeological context, most early physical anthropologists did not emphasise aspects that are of use in modern bioarchaeology, such as lifestyle, health and demographic characteristics, and their detailed anatomical descriptions have little application today. Much of the information about these skeletons and their context has been lost, yet they remain invaluable sources of information about the past. With modern radiocarbon dating, ancient DNA (aDNA) extraction and other methods of analysis, the value of these irreplaceable collections is increasingly realised and warrants further analysis (e.g. Pfeiffer et al. 2019; Steyn et al. 2019; Meyer et al. 2022).

Mr FW FitzSimons, the then director of the Port Elizabeth Museum (now Gqeberha), conducted a series of excavations in the Eastern Cape Province of South Africa during the 1920s. Sadly, these excavations were often very haphazard and poorly documented (Morris 2022). Their focus was simply on the retrieval of as many skeletons as possible, with little interest in the context of the findings. As part of the typological thinking of the time, excavations of human remains primarily focused on skulls, to be able to investigate the human ‘types’ that inhabited various regions of southern Africa. The archaeological endeavours of FitzSimons are discussed in more detail in Schauder (1963), Turner (1970), Morris (2022) and Steyn et al. (submitted). The excavations and skeletons from Zuurberg are said to be among the better-documented endeavours of FitzSimons (Schauder 1963) and some details of the excavations and skeletons have been published (FitzSimons 1923; Wells 1929). However, FitzSimons was not a trained archaeologist and, according to Morris (2022), he was somewhat of a showman who was prone to fabricating information. His interpretations should thus be treated with caution. The human skeletal remains of Zuurberg and other sites in the Tzitzikamma region were initially stored in the Port Elizabeth Museum but were transferred for study to the University of the Witwatersrand in the late 1920s, where they remain in the Raymond A. Dart Archaeological Human Remains Collection.

The prehistoric burial site of Zuurberg was accidentally discovered in 1922. The site is in the Zuurberg Mountains, north of Gqeberha. Unfortunately, its exact location is currently unknown, but Figure 1 shows a map with the location of the Zuurberg Mountain Village (the old Zuurberg Manor) and two roads that most closely match the descriptions provided by FitzSimons. As referenced by Wells (1929), FitzSimons reported that the site was discovered during the digging out of the soil in a road cutting. It is described as situated in:

‘a more or less wooded knoll among the many mountains which form the Zuurberg. The main road to Somerset runs through the lower portions of this minor mountain top, and it was in the higher side of the cutting that the remains were lodged. It was evident that when the road was made (about the year 1856) a prehistoric graveyard had been cut into … ’ (Wells 1929: 806).

Based on this description and the terminus post quem provided by the date of the road construction (1856), two roads running between Somerset East and Gqeberha could be identified using historic maps dating to 1878 and 1901, respectively. These maps were georeferenced and are displayed in Figure 1 in orange. The 1878 map drawn up by the Surveyor General’s Office of the then Cape Colony (https://digitalcollections.lib.uct.ac.za/collection/islandora-19613) indicated only one road connecting Somerset East and Port Elizabeth (modern-day Gqeberha) running through the town of Alicedale. In the 1901 map, produced by the Field Intelligence Department of the British Army during the Second South African War (https://digitalcollections.lib.uct.ac.za/collection/islandora-24828), an additional road running between Somerset East and Port Elizabeth, corresponding to the road associated with the old Zuurberg Manor, is depicted. However, given that the Zuurberg Manor was already built by 1861 (https://zuurbergmountainvillage.co.za/the-zuurberg-manor/) access to it would have been necessary, likely indicating that the road associated with it was already in existence, even though it was not indicated on the 1887 map. It may therefore be more likely that the Zuurberg site was located on the road running past the old Zuurberg Manor. This is confirmed to some extent by Schauder (1963), who managed to locate the site, which at that point was on a farm owned by the owners of ‘Zuurberg Manor’ approximately two miles (∼3.2 km) on the manor’s Gqeberha side. Unfortunately, no archaeological work to confirm this has been carried out since FitzSimons’ excavation.

Figure 1. Map indicating two georeferenced roads that ran between Somerset East and Gqeberha (previously known as Port Elizabeth), South Africa, with the road running past the Zuurberg Mountain Village being the most likely area where the original site may have been.

FitzSimons (1923) noted that the first group of skeletons, consisting of four individuals, was found underneath a stone circle; within this circle were four flat stones below which the four skeletons were found in a horizontally flexed position. They were poorly preserved and associated with ‘grinders or mullers of dark blue stone’ (FitzSimons 1923: 501), spearheads, knives and other stone tools. FitzSimons then excavated deeper into the bank of the road (in other words to the side of the original four skeletons) and found a stone mortar with red pigment (possibly ochre), presumed to be associated with a burial which was discovered about 18 inches (∼45 cm) below it. This individual was well preserved and was also found in a flexed, horizontal position. Three grinding stones were found on the skull and an object of hard blue slate near a scapula. Near the hands were 13 ivory implements. Photographs of the slate tool and ivory implements are in the Fitzsimons publication and are presumably curated at the museum in Gqeberha, although access to them could not be arranged to confirm this. Various chipped stone weapons and diverse stone tools were found around the skeleton, as well as a fragment of an oyster or mussel shell. According to FitzSimons (1926: 503), the area was ‘teeming with imperfect implements’, suggesting that this site may have been used for the manufacturing of stone tools. He also mentions the presence of two vertebrae, identified by Wells (1929: 809) as belonging to the painted (wild) dog (Lycaon pictus). One of these vertebrae, an atlas, is still available today.

Lower down, two more skeletons were discovered, both associated with stone implements and both of them well preserved. At a level below this, two more skeletons were found, comprising one adult and one child. These individuals, on the other hand, were poorly preserved and not associated with stone implements. This brings the total number of discovered skeletons to nine, of which eight were excavated and one could not be retrieved due to poor preservation. A drawing of the stratigraphy is shown in Wells (1929), indicating that the four individuals in the stone circle, and possibly the adult with the many grave goods, were on the same level. Deeper to this were two skeletons, and at a slightly deeper level another adult and a child. It thus seems as if human remains were found at three different depths, although Wells (1929) mentions that the remains were found on two different levels.

Data from some of the skulls have been used in modern craniometric studies such as that of Neuweger (2007), who used three crania in her undated group of Savanna Biome individuals for comparison with Matjes River material (Dreyer 1933; Sealy et al. 2006). However, no comprehensive analysis of the remains, using a modern bioarchaeological approach, has been undertaken. This study aimed to reassess the remains from Zuurberg, currently housed in the Raymond A. Dart Archaeological Human Remains Collection (hereafter referred to as the Dart Archaeological Collection) at the School of Anatomical Sciences, University of the Witwatersrand. An attempt was made to resolve the commingling resulting from both poor excavation and curation and to use primary sources to associate specific individuals with the excavation drawing of Wells (1929). The basic demographic characteristics of the remains were determined and they were assessed for signs of disease. In addition, radiocarbon dating and isotope analysis of one individual were performed.

Materials and methods

The remains are currently curated in a single box in the Dart Archaeological Collection under accession number A1110 (the ‘A’ number refers to the Dart accession numbers). Some commingling had occurred through the years and it was clear that someone had, at an unknown point in time, added new numbers onto some of the remains. Unknown, commingled remains from the Port Elizabeth Museum curated under A1108 were also assessed to establish if some of them might belong to the Zuurberg individuals (see Zg5 below; the somewhat confusing allocation of the Zg — Zuurberg — numbers is described in more detail in Wells (1929)). The remains were thus sorted, using the detailed descriptions of Wells (1929), who initially analysed the remains. Some unaccessioned teeth and bone fragments remained after our sorting, but for future purposes the remains will be curated under A1110, but separated by their Zg numbers, in accordance with the descriptions provided here.

All the skeletons were assessed using standard anthropological methods (Buikstra and Ubelaker 1994; İşcan and Steyn 2013) to first establish their sex and age. As the remains were for the most part fragmentary and incomplete, the ages of adults were only estimated to belong to either young, middle-aged or older individuals. Cranial suture closure (Acsádi and Neméskeri 1970), features of the pubic symphyses (Brooks and Suchey 1990) and general degenerative changes (Pfeiffer 2022) were assessed. For juveniles, dental development (AlQahtani et al. 2010) and skeletal maturation were employed (Schaefer et al. 2009). Sex was assessed with the use of the morphological features of the pelvis in the case of one individual, whereas in others general robusticity was assessed (Buikstra and Ubelaker 1994, İşcan and Steyn 2013).

Stature could only be reconstructed for one individual, using the femur:stature ratio of Feldesman and Fountain (1996). The maximum femur length is provided, to be available for different methods of assessing stature. All of the skeletons were morphologically assessed for signs of trauma and pathology where possible as some bones were varnished, while the teeth were investigated for evidence of dental disease and ante-mortem tooth loss. A detailed analysis of dental wear fell outside the scope of this paper, but their general wear levels were used to aid in age estimation (see Morris 1992b for the terminology used). No attempts were made to assess the ancestry/origins of the individuals, as they clearly predate the arrival of Khoekhoe pastoralists and Bantu-speaking farmers in the region and therefore would have been ancestors of San hunter-gatherers. Cranial and mandibular measurements were recorded where possible (Tables 1 and 2).

Table 1. Cranial dimensions of adults in millimetres (Buikstra and Ubelaker 1994).

Skeleton	Zg1	Zg3
Maximum length	179	194
Maximum breadth	137	139
Bizygomatic breadth	-	-
Basion-bregma height	-	-
Cranial base length	-	-
Basion-prosthion	-	-
Maximum alveolar breadth	-	62
Maximum alveolar length	-	55
Biasterionic breadth	115	-
Upper facial height	-	65
Minimum frontal breadth t	90	97
Upper facial breadth	99	103
Nasal height	-	48
Nasal breadth	-	28
Orbital breadth ec	-	37
Orbital height	-	31*
Biorbital breadth	-	96
Interorbital breadth	-	25
Frontal chord	110	117
Parietal chord	115	122
Occipital chord	-	-
Foramen magnum length	-	-
Foramen magnum breadth	-	-
Mastoid length	22	-

* indicates that measurements were recorded on the right side.

Table 2. Measurements of the mandible in millimetres (Buikstra and Ubelaker 1994)

Dimension (mm)	Zg 1	Zg 2	Zg 3	Zg 4	Zg 5	Zg 7	Zg 8
Chin height	30	32	29	28	-	24	-
Body height mental foramen	25	30	26	26	-	22	26
Body thickness mental foramen	11	11	10	13	-	9	11
Bigonial diameter	-	94	97	-	-	-	-
Bicondylar breadth	-	-	-	-	-	-	-
Minimum ramus breadth	32	35	39*	-	34	-	-
Maximum ramus breadth	41	-	44	-	38	-	-

* measured on the right side.

One of the individuals, Zg1, was dated by the sampling of a petrous bone. Material for radiocarbon measurements, dietary isotope study and ancient DNA analyses (which will feature in a future publication) was collected at the University of the Witwatersrand using a portable hand-held drill and cleaned circular diamond blades, aiming to collect at least 100 mg of bone. Bone samples were then transported to Uppsala University, Sweden, where dating and isotopic analysis were performed. A permit for destructive sampling and export of the sample was obtained from the South African Heritage Resources Agency (permit number 2789).

Radiocarbon measurements were obtained by accelerator mass spectrometry (AMS) at the Tandem Laboratory, Department of Physics and Astronomy, Uppsala University, Sweden. Stable isotopes of carbon (¹³C) and nitrogen (¹⁵N) were measured independently by isotope-ratio mass spectrometry (IRMS) at the same laboratory. Radiocarbon results were calibrated with OxCal 4.4 (Bronk Ramsey 2009) using the atmospheric curve SHCal20 (Hogg et al. 2020). Calendar ages are reported here as dates cal. BC. All calibrated ranges are given at 95.4% probability.

Results

All the remains analysed were in a commingled condition and were sorted and reassembled as explained below. Parts of all eight excavated skeletons were present, although many were not as complete as described by Wells (1929) and it is clear that much material was lost over time. Skeletons Zg1, Zg2, Zg7 and Zg8 were identified as the four graves from the stone circle (Figure 2), while Zg3 was found to be the complete skeleton buried some distance apart. Skeletons Zg4, Zg5 and Zg6 are the remains from the deeper levels, with Zg4 and Zg5 being from the seven-foot (2.1 m) level, with Zg6 (the juvenile) somewhat apart. One additional skeleton is indicated in the original drawing as being too poorly preserved to be excavated (Figure 2) (Wells 1929).

Figure 2. Plan of excavations, redrawn from Wells (1929). No scale is available.

Zg1

The skeletal remains marked as Zg1 comprise an incomplete skull with a neurocranium lacking the cranial base. No facial bones are present. The mandible now marked as Zg8 fits the description by Wells (1929) as belonging to this individual — for example, he mentions the absence of third molars, a feature that is seen in this individual. It will thus be described with this skull and also in future curated with it. The mandible is complete except for the left gonial region and ramus. Ochre staining is evident on both parietal regions of the skull (Figure 3). This skull was dated to 4795 ± 42 BP (Ua-61976), using a sample of the already exposed petrous bone.

Figure 3. Skull of Zg1 from above. Note the ochre staining (red circles).

The few cranial and mandibular measurements that were possible are given in Table 1. All the cranial sutures are in an advanced state of closure, suggesting an older individual. This is corroborated by the heavy wear observed on the teeth. Sex estimation is very difficult when the remains are so poorly preserved and when only a skull and mandible are available. This is further complicated by the low levels of sexual dimorphism in Khoe-San populations. The remains seem fairly gracile with small mastoids, a smooth nuchal area, sharp supra-orbital margins and small supraorbital ridges. The mandible is gracile. A very tentative diagnosis of a female was therefore made. Wells (1929) did not provide an estimate of the sex of this individual.

No upper teeth are preserved, but in the mandible the right first molar, right canine, left incisors, left canine, left first premolar and left first molar are all present. The left side of the mandible beyond the first molar is absent. The right third molar was absent prior to death, most probably because it was undeveloped. The left second premolar had been lost before death, while all the other teeth were lost post-mortem. No carious lesions were noted, but abscesses are present at the roots of the right first molar and left second premolar. Heavy tooth wear is present, with exposed pulpal spaces in many of the teeth. This most probably caused the abscessing. The right first molar shows very oblique wear, which may suggest the use of the teeth as tools. The anterior teeth are heavily worn so that they appear rounded.

Zg2

The remains marked as Zg2 comprise cranial (calvaria) fragments only. However, the mandible currently marked (in pencil, probably as a later addition) as Zg7 fits Wells’ (1929: 815–816) description of this individual’s mandible exactly (complete except for left condyle, osteoarthritis of right condyle, root abscess at the first molar) and was thus reunited with them. The skull seems to have been more complete/intact before but is now broken so that it only comprises fragments of the frontal, occipital, parietal and temporal bones. The mandible is fairly well preserved and its measurements are given in Table 2. Ochre staining of the cranial vault is evident.

The remains are those of an adult. Heavy to extreme dental wear is present and the observable cranial sutures are all obliterated. Wells (1929) mentioned the presence of osteoarthritic changes on both ends of the (now lost) clavicle. These characteristics suggest that this individual survived well into adulthood and even old age. The fairly large mastoids and relatively robust mandible tentatively suggest that he was male.

As also mentioned by Wells, degenerative changes in the right condyle of the mandible are evident. No maxillary teeth are preserved, and of the mandibular teeth the right-sided first and second molars, second premolar and central incisor, as well as the left-sided central incisor and second and third molars are all present. The right-sided third molar and left-sided first molar were lost before death, while all the other teeth were absent post-mortem. No carious lesions were noted, but abscesses are present at the roots of the right-sided first and third molars and canine. The teeth are heavily worn, with a rounded appearance of the anterior teeth, and could not be reliably scored for caries or enamel hypoplasia.

Zg3

This is the most complete of all the remains and is the skeleton described by FitzSimons (1923) in some detail as having been associated with three grinding stones, an object of hard blue slate, 13 ivory implements and various other stone implements. A fragment of an oyster or mussel shell as well as two canid vertebrae were also found near or with this individual. Only one canid cervical vertebra is currently present, specifically the C1/atlas (Figure 4). This was tentatively identified by one of us (Anja Meyer) to be most similar in its morphology and size to those of the domestic dog (Canis lupus familiaris) (Figure 5).

Figure 4. Ventral and dorsal views of an atlas found in association with Zg3.

Figure 5. Ventral (A) and dorsal (B) views of the atlas associated with Zg3 (left) compared to that of a domestic dog (right).

The remains of this individual are fairly complete and include a reconstructed skull (with complete face) (Figure 6), near complete mandible, both claviculae, partial scapulae, long bones of both upper limbs, partial os coxae, a sacrum, both femora, both tibiae and fragments of both fibulae. Of the ribs, only the first ribs are preserved, as is a partial vertebral column. A few hand and foot bones are also preserved. The bones are mineralised.

Figure 6. Skull of Zg 3 in anterior and lateral view.

The measurements that were possible are given in Tables 1 and 2. The individual only lived to young adulthood. All the permanent teeth had erupted and were in occlusion. All the long bone epiphyses, including the medial ends of clavicles, are fused. No sternal ends of ribs are available for age estimation, but the left pubic symphysis is in Phase 3, suggesting an age of 21–46 years. The cranial sutures are mostly open. This person was therefore most probably 25–40 years of age at the time of death.

The features of the pelvis, such as narrow sciatic notches, a narrow subpubic angle, triangular pubic bones and an absent pre-auricular sulcus, suggest a male individual. The skull is fairly robust, with a well-developed glabella and rounded supra-orbital margins (Figure 6). The chin is square. A diagnosis of a male was thus made, although he is post-cranially very gracile. The stature of this individual was calculated with the length of the left femur (maximum femur length 401 mm) and is estimated to have been about 150.1 cm. This is a short stature, but falls within the reported ranges for Khoe-San individuals (Sealy and Pfeiffer 2000; Pfeiffer and Sealy 2006).

All the upper teeth are present, except for the right-sided incisors, which were lost after death. The left-sided incisors are fractured (post-mortem). Similarly, in the mandible, all the teeth are present except for the two central incisors, which went missing after death. No caries, abscessing or enamel hypoplasia could be observed.

Although the bones are heavily varnished, it does not seem that there was any subperiosteal bone deposition. The femora are quite bowed, but no other signs of rickets are present. Squatting facets could be observed on the tibiae. The humeri are gracile, with the right being larger than the left. The right patella shows a developmental anomaly commonly referred to as a bipartite patella (Figure 7). This condition results in the incomplete fusion of the patella during adolescence. It is usually asymptomatic, occurs bilaterally or unilaterally and is more frequently observed in males (Barnes 2012; Oohashi 2015). Slight lipping is starting to appear on the lower vertebrae. An additional, partly sacralised lumbar vertebra is present. No signs of active disease could be observed.

Figure 7. Anterior and posterior views of the right bipartite patella (indicated by the arrow) associated with Zg 3.

Zg4

This individual comprises a partial skull and mandible only. Of the skull, only the cranial vault (without cranial base) and a few other fragments are present. The mandible comprises the section of bone that includes the right central incisor to the left molars, as well as part of the left ramus.

Age was difficult to assess due to the incomplete nature of the remains. All the permanent teeth had erupted except for the third molar, which may have been impacted/not developed. The sagittal suture is completely obliterated, but the coronal and lambdoid sutures are open. The teeth are moderately worn. This individual was thus possibly young to middle-aged when he died. The orbital margins seem to be rounded, and the mandible is relatively robust. A very tentative diagnosis of a male was thus made.

No upper teeth are present. Of the mandibular teeth the right-sided central incisor, as well as all the left-sided teeth except for the third molar (unerupted/undeveloped), are present. The left central incisor is fractured (post-mortem). No signs of caries or abscessing could be observed, but an enamel hypoplastic line was visible on the left lower canine.

Zg5

Zg5 comprises a heavily reconstructed skull (comprising only the parietal, temporal and occipital bones) and the left side of a mandible. After sorting through Box A1108, also curated in the Dart Archaeological Collection and containing miscellaneous bones from the Port Elizabeth Museum, the right side of a mandible and left maxilla were reunited with Zg5. Wells (1929) indicated that this individual had a maxilla and mandible and unerupted third molars, which fit the reassembled remains exactly. The right half of the mandible resembles the left half and the maxillary dentition also fits clearly with those of the mandible when articulated. The mandible remained as three separate pieces with an absent right ramus.

All the teeth present are permanent, but the third molars are still unerupted and deep in their crypts. An age estimate of 15–18 years was thus made. Sex could not be determined due to the young age of the individual and the incompleteness of the remains.

Of the maxillary teeth, only the left lateral incisor, the canine, both premolars and the first molar are present. Other than for a slightly rotated canine, no other signs of pathology could be observed on the upper teeth. In the mandible, the right-sided molars and first premolars are present, as well as the left-sided canine, both premolars and all the molars. The incisors are fractured so that only roots remain, while all other teeth were lost after death. Single enamel hypoplastic lines are observable on the left-sided premolars. No signs of caries or abscesses could be observed. No other pathological changes could be seen and no orbits are present to assess for the presence of cribra orbitalia.

Zg6

This individual is a juvenile, and the only skeletal remains present are a tibial shaft (with clearly unfused proximal epiphysis), two mandibular fragments, seven deciduous teeth and a crown of a permanent molar. The deciduous teeth include an upper left central incisor, as well as four, fully erupted lower molars and a left canine. The upper first permanent molar is also present and comprises the crown only with no roots developed. This pattern of dental development suggests a child who was about 4.5 ± 1 years old when he/she died. No other information could be gathered from this skeleton.

Zg7

Zg7 comprises a partial mandible only. It is marked as Zg9 (probably by later curators) but fits the description of Wells as being Zg7. It consists of a body of a mandible without rami. The shape of the chin and very gracile appearance tentatively suggest a female individual. The teeth are heavily to extremely worn (Figure 8). The third molar (right side) is undeveloped/lost ante-mortem, whereas the corresponding part of the mandible on the left side is missing. Based on the tooth wear, the individual was judged as having been an older adult. Of the mandibular teeth, the right first and second molars, second premolar and both incisors are present. So, too, are the left-sided premolars and first molar. Other absent teeth were lost post-mortem, none of them before death. The teeth could not be assessed for caries or enamel hypoplasia as they are too heavily worn.

Figure 8. Mandible of Zg7, showing the advanced dental wear.

Zg8

The remains of this individual comprise a partial mandible marked as Zg10. However, the description fits that by Wells (1929) as belonging to Zg8 and the number Zg10 was probably added by later curators. Only the left and right halves of the mandibular body are preserved.

The individual was judged to have been an older adult, based on the heavy dental wear (Figure 9). The mandible is more robust than that of Zg7, which may tentatively suggest a male individual, but this is difficult to substantiate. The right-sided second and third molars, first premolar, canine and both incisors, as well as all the left-sided teeth, are present. The right-sided second premolar and first molar were lost before death. A dental abscess could be observed at the tooth socket of the ante-mortem lost right first molar. The teeth could not be scored for enamel hypoplasia or dental caries due to their heavy wear.

Figure 9. Mandible of Zg8. Note the advanced dental wear.

Radiocarbon dating and isotopic analysis

Radiocarbon dating of Zg1 indicated a date of 4795 ± 42 BP (Ua-61976) with a calibrated age of between 3660 to 3380 cal. BC (at a 95.4% probability). This date corresponds to the age of the Later Stone Age Wilton technocomplex of southern Africa (Lombard et al. 2022). The isotopic values obtained suggested a δ¹³C value of -14.7‰ and a δ¹⁵N value of 10.5 ‰.

Discussion

The remains from the Zuurberg site comprise eight individuals, with a ninth that could not be excavated. Poor initial preservation, coupled with poor curation, led to the absence of many skeletal elements and later also a degree of commingling. As is the case with many other older (often commingled) remains kept in various repositories around the world (Fox and Marklein 2013; Lambacher et al. 2016; Osterholtz 2019; Landsman et al. 2022), we reassessed these remains using a modern, bioarchaeological approach. Although some commingling occurred during transport from the Port Elizabeth Museum and the subsequent storage, there is no evidence to suggest that the remains were commingled in situ.

Through the detailed descriptions of Wells (1929), who also identified eight individuals, we were able to reassemble the remains where possible and they will be curated with the numbers indicated above in future, under the collection number A1110. Of the eight excavated individuals, four were older adults (Zg1, Zg2, Zg7, Zg8), one was a child (Zg6), one an adolescent (Zg5) and two younger or middle-aged adults (Zg3, Zg4). Unfortunately, the incompleteness of the remains prevented more detailed analyses and closer estimates of age. The four older adults were reportedly buried in the stone circle, whereas the adult individual Zg3 was buried on the same level (or somewhat deeper) but some distance apart. Individuals Zg4, Zg5 and Zg6 were buried at a deeper level(s) according to Wells (1929). From this, it seems that the individuals were interred at probably at least three levels, suggesting different dates/ages and thus a longer use of the site. Wells (1929) mentions the presence of potsherds in the superficial layers, which would also suggest a more recent use of the site than the nearly 5000 BP date of Zg1 in the superficial layer. The possibility that this site was used over an extended period should thus be considered, even though this is unusual for an open-air site in southern Africa. Both Hall (2000) and Lazarides (2015) mention that high-density, concentrated burials are usually found in caves/rock shelters, whereas open-air LSA burials are usually single. This is not the case here and raises some interesting questions. Although the datable material from Zuurberg is limited due to poor preservation, the dating of the other individuals (as well as of the canid vertebra) should be considered to gain a clearer picture of the time depth of this site.

Further investigation into the canid vertebra is important given the early radiocarbon date obtained for Zg1, and therefore the site. Canis lupus familiaris is not indigenous to Africa, having been introduced into North Africa from the Middle East (Van Sittert and Swart 2003; Mitchell 2014, 2015). The spread of domestic dogs into southern Africa likely occurred at a staggered rate and appears to have broadly coincided with the spread of farming populations in the first millennium AD (Mitchell 2014). However, archaeological evidence remains scarce and is often inconclusive when identification is based solely on osteological metrics and morphology, given the similarities between domestic dog breeds and other canids such as the black-backed jackal (Lupulella mesomelas) and African wild dogs (Lycaon pictus) and the general lack of comparative studies (Mitchell 2014). The oldest confirmed presence of domestic dog remains in South Africa comes from the Diamant site in Limpopo Province and dates to the sixth century AD (Plug and Robertson 1989), while remains from Kasteelberg B, a likely pastoralist site in the Western Cape, date to between the mid-seventh and late thirteenth centuries AD (Mitchell 2014). Given the general scarcity of evidence of domestic dogs at LSA sites in South Africa, further investigation of the possible domestic dog vertebra found with Zg3 is warranted and may shed light on the introduction of domesticated dogs to southern Africa and their potential use by early San ancestral populations. Recalling that Wells (1929: 809) thought that the vertebrae found might be those of an African wild dog, such analyses should include absolute dating and potential DNA sequencing or ZooMS analysis to confirm identification to species.

Of the four individuals reportedly buried in the stone circle, two were male and two female. Overall, of the individuals that could tentatively be sexed, four were male (Zg2, 3, 4 and 8), while two (Zg1 and 7) were assessed to be female This small assemblage thus includes the remains of adults of both sexes as well as children. In some other LSA sites (e.g. Dewar 2010; Parkington and Dlamini 2015; Pfeiffer et al. 2020) group burials are associated with possible episodes of violence. However, this seems not the case here as we saw no evidence of trauma on the remains and the fact that the burials occurred at different levels (with the possible exception of the four individuals reportedly buried in a stone circle) suggests that they most probably date from different periods and were not communal or buried at the same time. Lazarides (2015) also mentions that communal burials do not usually include the remains of male individuals, thus also arguing against the four burials (Zg1, 2, 7 and 8) being a communal grave as two of the individuals in it were probably male.

FitzSimons reported the presence of many associated artefacts at the Zuurberg site, which is unusual (Lazarides 2015). Their association with the graves, as well as the presence of a stone circle with burials facing the centre (as described by FitzSimons 1923), can unfortunately not be verified. Two of the skulls did show signs of ochre staining, possibly attesting to some ritualistic behaviour, and many other LSA burials have been associated with ochre (Lazarides 2015). The Zuurberg remains were in a horizontally flexed position, like the majority of LSA burials (Inskeep 1986; Hall 2000; Lazarides 2013; Loftus and Pfeiffer 2023), including the much more recent Kouga mummy, which was bundled up and tied with a rope to keep it in a tightly flexed position (Steyn et al. 2007). Graves from the more superficial layers, presumed to be more recent, seem to have had more objects associated with them than those from the deeper layers, although it cannot be determined what their exact relationship with the remains was.

Assessment for disease and pathology is difficult due to the small sample size and poor preservation. However, two of the three individuals that could be assessed had enamel hypoplasia. These lesions are caused by hardship during the years in which the teeth develop (Goodman and Rose 1990), possibly suggesting some stress during early life. None of the individuals had cribra orbitalia, which is associated with non-specific anaemia. These skeletons come from a period before the intensification of populations in this general region, which is thought to have given rise to heightened stress, particularly in relation to the availability of resources (Hall 2000; Sealy and Pfeiffer 2000; Pfeiffer and Sealy 2006; Pfeiffer and Harrington 2011).

Of the five individuals with scorable permanent teeth, none had caries. However, three out of the seven adults had dental abscesses, and three also had ante-mortem tooth losses. Due to the small sample size, it is difficult to compare this pattern statistically to other populations. The abscesses and ante-mortem tooth losses all occurred in older adults. This suggests a diet, as expected, that was low in refined carbohydrates, but very abrasive. The coarse food gave rise to rapid dental wear, leading to pulpal exposure, abscessing and subsequent tooth loss. In this, the Zuurberg sample seems similar to the people from Whitchers’ Cave, also in the Eastern Cape Province, (Steyn et al. submitted), although the relatively high incidence of enamel hypoplasia may suggest higher stress levels than was the case there.

One unusual finding is the many individuals with absent third molars. Of the seven individuals with permanent teeth, three (Zg1, Zg4, Zg7) had no third molars, presumably due to agenesis. This is an unusual and interesting finding, which may possibly be related to the gracility of individuals with limited space for third molars. It could also suggest close genetic relationships with a limited genetic pool, as has been suggested for the Whitcher’s Cave group (Steyn et al. submitted). However, currently, nothing is known about the genetic relationships between these individuals, although future analysis may provide pertinent information.

The date of 4795 ± 42 BP obtained from individual Zg1 is a new addition to our knowledge of the Zuurberg site, with other individuals probably interred there before that. In a recent study, Loftus et al. (in press) reported just over 80 dated individuals in total from the Wilton period in South Africa, making this a valuable contribution to this relatively small pool. Interestingly, Wells (1929) estimated that the site had been occupied after 6000 BP, but before 1800 BP, based on what was known about the climate at this time. Another Zuurberg skeleton (ALB 206), dated to 4610 ± 50 BP (Pta-8713), has been reported (Morris 1992a; Ginter 2011), but, although this date is very similar to that of Zg1, this was a rock-shelter burial and thus most probably comes from a different site in the region. Other skeletal material from this area (Silberbauer 1979; Morris 1992a) include that from the Vygeboom farm (Middelkop Kloof Cave and Mookrantz Cave) which has dates similar to that of Zg1, while Melkhoutboom (H. Deacon 1976) is more recent. All these burials, however, come from rock-shelter sites and there is relatively little information available from non-coastal, open-air sites during this period.

Hilary Deacon’s (1976) study of Melkhoutboom (which is also in the Cape Fold Mountains) and other sites led him to propose that, although these groups were probably ancestral to modern populations, they lived in environments and conditions for which there is no modern analogue. They therefore showed important differences in their adaptations. The Cape Fold Mountains of the Eastern Cape do not support many large animals so only smaller game was available for hunting, predominantly including steenbok (Raphicerus campestris), grysbok (Raphicerus melanotis), bushbuck (Tragelaphus scriptus) and — at the larger end of the spectrum — kudu (Tragelaphus strepsiceros). The protein base was thus limited and the reliance on gathering edible plants from the local fynbos flora correspondingly high. This can be confirmed to some extent by the stable isotopic results obtained for Zg1, which suggest a mixed terrestrial diet with a slightly higher reliance on C₃/C₄ plants compared to animal-derived protein. This is in contrast to typical coastal forager diets, which have higher nitrogen delta values (Lee-Thorp et al. 1993; Lewis and Sealy 2018), suggesting that Zg1 most probably relied mainly on a terrestrial diet obtained from hunting and foraging. Similar isotopic values were observed in the other Zuurberg individual (ALB 206) mentioned above, as well as in a further individual from Whitcher’s Cave (Sealy and Pfeiffer, 2000). ALB 206 has a δ¹³C value of 12.0‰ and a δ¹⁵N value of 11.9‰ (laboratory number 113; Judith Sealy, pers. comm.), whereas the individual from Whitcher’s Cave (A118 VIII) has a δ¹³C value of -14.9‰ and a δ¹⁵N value of 10.5‰ (laboratory number: Pta-6973). It has been mentioned that the interaction between populations is relatively unknown for the earlier Holocene, but isotopic and archaeological data seem to suggest a shift in population movements and dietary habits just after 4500 BP (Pfeiffer and Sealy 2006). There are many debates in the literature as to the use and availability of marine resources by LSA people, with the possibility also mentioned that there was a degree of territoriality when it came to access to these resources (Sealy and Pfeiffer 2000; Pfeiffer and Sealy 2006; Sealy 2006; Pfeiffer 2013), something that would account for the lower δ¹⁵N values seen in some of these individuals. However, it is interesting to note that both the Zuurberg and the Whitcher’s Cave individuals’ isotopic values also reflected lower δ¹⁵N, even though they predate the 4500 BP shift, perhaps suggesting earlier restrictions on mobility and a higher reliance on local food resources even when population densities are still assumed to have been low (J. Deacon 1974).

A large portion of the Zuurberg site apparently remains unexcavated (Wells 1929; Schauder 1963), but unfortunately its exact provenance is not recorded. It may nevertheless be possible to relocate the site based on Schauder’s descriptions, in which case a further investigation would certainly be warranted. Dating and isotope analysis of the other individuals from the site should also be considered, as the seemingly extended use of an open-air site is unusual and raises interesting questions.

Acknowledgements

We thank the curators of the Raymond Dart Archaeological Collection as well as the Eastern Cape Provincial Heritage Resources Agency (2/2/APM-BG 18/10/001) for giving us the opportunity to study the human remains that form the basis of this paper. The isotope analyses were funded by a Knut and Alice Wallenberg Foundation grant awarded to Carina Schlebusch. We should also like to acknowledge Renier van der Merwe for assisting with the development of the map and Rita Peyroteo-Stjerna for her help during the sampling. SAHRA permit 2789 provided us with the permission to sample and export the remains. The paper was jointly written by all four of us with primary responsibilities as follows: Maryna Steyn — skeletal analysis, writing of the first draft and revision of later drafts, co-ordination; Anja Meyer — faunal analysis, images, assistance with drafts, data collection and interpretation of stable isotope data, editing; Cecile Jolly — sampling and preparation for isotope analyses, curation and co-ordination of data; and Carina Schlebusch — co-ordination of the Uppsala University-based analyses and contributions, funding for sample collection and isotope analysis, contribution to drafts

Additional information

Notes on contributors

Maryna Steyn

Maryna Steyn is a biological anthropologist who qualified as a medical doctor in 1983 and obtained a PhD from the University of the Witwatersrand in 1994. She consults for the South African Police Service and forensic pathologists on skeletonised human remains and also conducts research on human remains from archaeological contexts, with a special interest in palaeopathology. She has published about 170 scientific papers and several book chapters.

Cecile Jolly

Cecile Jolly is a trained physiologist and molecular biologist. She received a PhD degree at the University of Bergen. She has several years of experience applying methods used for Illumina sequencing for both modern and ancient DNA and has a strong personal interest in history and archaeology. She currently works as laboratory manager for the Schlebusch group, assisting students and researchers in their projects, both in the laboratory and during fieldwork.

Carina M. Schlebusch

Carina Schlebusch is Associate Professor in Human Evolution at Uppsala University and heads a research group investigating human evolutionary history with a focus on Africa. Born in South Africa, she completed her PhD at the University of the Witwatersrand on the genetic history of southern African population groups. After moving to Uppsala for a Wenner-Gren Foundation funded post-doctoral research fellowship, she started her own research group there funded by the European Research Council, the Swedish Research Council and the Knut and Alice Wallenberg Foundation.

Anja Meyer

Anja Meyer is a bioarchaeologist currently employed as a senior lecturer in the School of Anatomical Sciences, University of the Witwatersrand. She completed her PhD in Anatomical Sciences at the same institution. She is also the curator of the Raymond A. Dart Archaeological Human Remains Collection, housed in its School of Anatomical Sciences. Her research focuses on southern African bioarchaeology and forensic anthropology.