Masseria Candelaro

Masseria Candelaro is an extensively excavated Early–Middle Neolithic ditched village, located in a densely inhabited area of the Tavoliere Plain in northern Puglia (Figure 1) (Cassano & Manfredini, Reference Cassano and Manfredini1983, Reference Cassano and Manfredini2004). The initial early sixth-millennium bc village comprised a small 0.5 ha compound encircled by a ditch (Ditch Z) (Cassano & Muntoni, Reference Cassano, Muntoni, Cassano and Manfredini2004: 89). The village was briefly abandoned before it was reoccupied in the mid-sixth millennium bc and extended with four outer concentric ditches, enclosing an area approximately 300 × 250 m (Cassano & Muntoni, Reference Cassano, Muntoni, Cassano and Manfredini2004: 89–90). As with many villages, it was eventually transformed into a landmark for ritual and funerary activities. Burial chambers were cut into the sides of the inner ditch, and several older storage features within and surrounding an area known as ‘Structure Q’ were re-used for burial (Marconi et al., Reference Marconi, Muntoni, Cassano, Manfredini, Carboni, Curci, Cassano and Manfredini2004). Structure Q is located in the earlier compound inside Ditch Z (Marconi et al., Reference Marconi, Muntoni, Cassano, Manfredini, Carboni, Curci, Cassano and Manfredini2004: 66–81) (Figure 2). Fifteen radiocarbon dates place occupation between 5720–4903 cal bc, while eleven dates from human bone span 5630–4903 cal bc, suggesting there were several generations of occupation before burial traditions were established (Cassano & Manfredini, Reference Cassano and Manfredini1993; Skeates & Whitehouse, Reference Skeates and Whitehouse1994; Manfredini & Muntoni, Reference Cassano, Muntoni, Cassano and Manfredini2004; Conati Barbaro, Reference Conati Barbaro, Bernabò Brea, Maggi and Manfredini2014; Whitehouse, Reference Whitehouse2014) (see Table 1 and Supplementary Material 2).

Figure 1. A) Location of Masseria Candelaro and contemporary sites in southern Italy mentioned in text. B) Sites in the vicinity of Masseria Candelaro.

Figure 2. Left: outline of the village ditches with approximate location of cranial cache marked by a star (redrawn after Seager Thomas in Hamilton et al., Reference Hamilton, Whitehouse and Seager Thomas2020: 249); Right: excavated features inside Ditch Z, arrow indicating location of cranial cache in Structure Q within the circle (redrawn after Cassano et al., Reference Cassano, Manfredini, Marconi, Muntoni, Carboni, Curci, Cassano and Manfredini2004: 69). Reproduced with permission of the Accordia Institute and SABAP for the provinces of Barletta-Andria-Trani and Foggia, respectively.

Table 1. Radiocarbon dates from human bone in Structure Q, calibrated with IntCal20 (Reimer et al., Reference Reimer, Austin, Bard, Bayliss, Blackwell and Bronk Ramsey2020).

Secondary deposits of bones were found in multiple areas (Muntoni & Salvadei, Reference Muntoni, Salvadei, Cassano and Manfredini2004). These include isolated fragments within burials, disarticulated bones within the ditch, and fragments scattered across living spaces, some of which were burnt. There is little evidence for defleshing, unlike at Passo di Corvo and Grotta Scaloria (contemporary sites less than 10 km away), but there is evidence for secondary deposition, the circulation of crania, and placing the dead in open pits (Marconi et al., Reference Marconi, Muntoni, Cassano, Manfredini, Carboni, Curci, Cassano and Manfredini2004). This fits within regional practices focused on viewing, processing, and interacting with the bodies of the dead, often over considerable timespans.

Osteological remains and methods

The deposit comprised almost 400 fragments of crania and mandibles. In all cases, the calottes or calvaria were broken off from the splanchnocrania and basicrania, and maxillae and mandibles could not be re-associated with cranial fragments. The presence and completeness of cranial and mandibular zones were recorded using two methods: Knüsel & Outram (Reference Knüsel and Outram2004) for cranial elements, and Kranioti (Reference Kranioti2015) to divide the cranium and mandible into smaller zones (Supplementary Material 3, Figure S7). This allowed assessment of duplication of zones amongst the crania; combined with age estimation and assessment of refits, this produced a new minimum number of elements (MNE) and minimum number of individuals (MNI) (Supplementary Material 3, Table S4). Various MNI estimates are possible. Assuming that each unassociated fragment representing more than half of an element zone (Knüsel & Outram, Reference Knüsel and Outram2004) corresponds to a separate individual, the maximum number of individuals is fifty-three. If unassociated fragments originated in the smallest possible number of crania, the MNI is fifteen. The true number of individuals is likely toward the lower end of this range.

Sex was estimated using cranial and mandibular traits (Ferembach et al., Reference Ferembach, Schwidetzky and Stloukal1980; Buikstra & Ubelaker, Reference Buikstra and Ubelaker1994; Brickley, Reference Brickley, Brickley and McKinley2004). At least five individuals were male, six were probably male, and sex could not be determined for the rest. Sex estimation from cranial and mandibular traits is between 70.6 and 96.9 per cent accurate in populations with known sex (Inskip et al., Reference Inskip, Scheib, Wilder Wohns, Ge, Kivisild and Robb2019). As crania of both sexes are quite robust in this population, we have assessed sex conservatively. Dental eruption and attrition were the most informative indicators of age-at-death (Smith, Reference Smith1984; AlQahtani et al., Reference AlQahtani, Hector and Liversidge2010). Using these, the crania include a wide age range: at least two adolescents (twelve to eighteen years old), one young adult (eighteen to twenty-five), one adult between twenty-five and thirty-five, two adults between thirty-five and forty-five, and one adult older than forty-five years.

Where did the crania come from?

The two most obvious possibilities for the origins of the crania are that they were collected from enemies as trophies or curated from within the group as ancestral remains. Trophy skulls and ancestral remains may be treated in indistinguishable ways; both may show selective age and sex profiles. The key differentiating factors between these scenarios are: peri-mortem trauma suggesting violent acquisition; peri-mortem removal; decapitation; or morphological, genetic, or biochemical evidence of differences between where the crania were found and where they originated.

As elsewhere in Neolithic Europe, violence and raiding are documented in southern Italy. Village ditches may have been defensive. Healed and unhealed traumatic injuries are occasionally observed, including peri-mortem cranial fractures. Careful inspection of the Candelaro cranial cache revealed no peri-mortem or healed trauma on the extant fragments, although they are highly fragmentary. The absence of peri-mortem trauma on the remaining elements may suggest that they were not forcibly ‘collected’. There are at least eight examples of skeletons lacking crania or complete skulls in southern Italy, probably through ritual activity, although no examples of decapitation per se are known. For instance, the mandible and cervical vertebrae were present in a cranium-less child burial at Grotta Scaloria, implying that the cranium was removed after soft tissue decomposition (Knüsel et al., Reference Knüsel, Robb, Tafuri, Elster, Isetti, Robb and Traverso2016). At least two burials at Cala Colombo (Bari) lacked crania; in one, it could be seen where the grave had been cut into to retrieve the cranium (Pesce Delfino et al., Reference Pesce Delfino, Scattarella, de Lucia, Ferri, Giove and de Lucia1977: 150–51). Therefore, local comparanda suggest that the Candelaro crania probably originated from within the group.

To test whether the cache represents an endogenous or exogenous group, δ³⁴S (‰) isotopic concentrations in bone were measured from thirteen crania and mandibles representing separate individuals. Because almost none of the remains had associated dentitions, and the extant dentitions usually displayed taphonomic abrasion of the enamel, strontium isotope analysis of dental enamel was not applicable. Faunal remains from the site were used to measure the local baseline. The results are reported in Supplementary Material 4. Given this baseline, sulphur isotopes can be used to observe mobility to or from coastal and inland areas (Richards et al., Reference Richards, Fuller and Hedges2001). Masseria Candelaro is located inland, at least 12 km from the Adriatic Sea but close to the Candelaro river. The δ³⁴S values from the cranial cache fall within the distribution of δ³⁴S values obtained from the fauna assumed to be local (Figure S8, Supplementary Material 4), suggesting that the sampled individuals lived in an environment similar to that of the fauna. While we cannot exclude that some of these crania came from individuals who lived at nearby sites a similar distance from the sea, the results suggest that these crania did not form part of a deposit of trophies obtained from individuals from an isotopically distinct locality (see Table 2).

Table 2. Characteristics of crania retained for specific practices (data from Bonogofsky, Reference Bonogofsky2006, Reference Bonogofsky2011; Chacon and Dye, Reference Chacon, Dye, Chacon and Dye2007b; Bello et al., Reference Bello, Saladié, Cáceres, Rodríguez-Hidalgo and Parfitt2015; Tiesler & Lozada, Reference Tiesler and Lozada2018; Santana et al., Reference Santana, Rodríguez-Santos, Camalich-Massieu, Mart ín-Socas and Fregel2019).

How long were the crania used?

Structure Q's use began in the second quarter of the sixth millennium bc, and its duration is modelled as lasting between 90–225 years (at 68.3 per cent probability) or 40–345 years (at 95.4 per cent) (Figure 4; Supplementary Material 2). The ninety-five per cent probability intervals for Cranium 6 (5618–5482 cal bc) and Cranium 1 (5476–5335 cal bc) do not overlap. Therefore, the crania almost certainly do not represent a collection from a single episode or a narrow timeframe. Instead, they represent people who died over a period of about two centuries, perhaps six to eight generations.

Figure 4. Bayesian model of the duration of Structure Q's use (produced in OxCal version 4.4.4: Bronk Ramsey, Reference Bronk Ramsey2021).

The timeframe of their deposition is equally important. Integrating stratigraphic information, we carried out Bayesian modelling (Bronk Ramsey, Reference Bronk Ramsey2009) on the six available radiocarbon dates from Structure Q. The cranial group in layer 2 seems to have been placed in a single event around the late fifty-sixth or early fifty-fifth century bc. The final cranium in layer 1 was deposited between the late fifty-fifth to early fifty-fourth century bc (Figure 5). Cranium 6 may have come from an individual who lived during an earlier settlement phase, while Crania 1 and 2 are more consistent with the later phase. The cranial cache represents a collection that was constantly changing, but its long duration suggests an enduring tradition of use, alteration, and augmentation. It contains crania gathered over a maximum of two centuries, kept above ground and used, and eventually deposited in a group.

Figure 5. OxCal plot (version 4.4.4: Bronk Ramsey, Reference Bronk Ramsey2021) of probability distributions for radiocarbon dates from each of the layers in Structure Q, modelled stratigraphically.

How were they used?

Taphonomic study provides insights into how post-mortem ritual and use modified the remains of the dead (Knüsel & Robb, Reference Knüsel and Robb2016). We highlight three common patterns of use for cranial collections and their expected taphonomic features (Table 2).

• • Preservation and display, for example Papua New Guinean tribal ancestral or trophy crania, Pre-Pottery Neolithic A plastered skulls, or medieval saints’ relics. This can involve reassembling the cranium or skull, over-modelling it with features resembling flesh and eyes, or perhaps displaying it in a reliquary, i.e. enhancements intended to give it greater visual and spiritual effect.

• • Subjecting the element to patterned reduction, such as ‘skull’ cups (e.g. from Upper Palaeolithic levels in Gough's Cave, Cheddar Gorge, England: Bello et al., Reference Bello, Parfitt and Stringer2011) or calottes of consistent form (e.g. at Early Neolithic Herxheim, Rhineland-Palatinate, Germany: Orschiedt & Haidle, Reference Orschiedt and Haidle2006). This may be observed alongside evidence for cannibalism and peri-mortem trauma.

• • Using the cranium actively in ways that modify the bone. For example, a repeatedly-used teaching collection undergoes damage, with fragile elements lost and bones reduced (Caffell et al., Reference Caffell, Roberts, Janaway, Wilson and Williams2001); another instance might be if bone is understood to have inherent power within magical actions (Chacon & Dye, Reference Chacon and Dye2007a: 623).

With such variability, there is little use in defining typological categories, whether for how remains were modified or for their contextual meaning. For example, enemy trophies and venerated ancestral skulls may show similar kinds of modification for display (Bonogofsky & Graham, Reference Bonogofsky, Graham and Bonogofsky2011; Hastorf, Reference Hastorf, Tiesler and Lozada2018).

Taphonomic analysis of the Candelaro crania included examination of fragment zonation, element completeness, surface modifications, and bone breakage characteristics (Knüsel & Outram, Reference Knüsel and Outram2004; McKinley, Reference McKinley, Brickley and McKinley2004; Bello, Reference Bello, Zakrzewski and Clegg2005; Fernández-Jalvo & Andrews, Reference Fernández-Jalvo and Andrews2016). These revealed several salient features:

• • Concreted sediment was observed on most bones, typical of the calcareous geology of the Tavoliere plain. The surfaces of many fragments were finely abraded and root-etched, and occasionally perforated by roots, reflecting the post-depositional environment.

• • No evidence was found of burning or animal intervention such as gnawing or insect activity. The absence of the latter may suggest that the bones were not exposed in an open environment.

• • There was no evidence of peri-mortem trauma, injury, or ‘fresh’ bone breakage. Most breakage margins were consistent with the way dry bone breaks.

• • There was no consistent pattern as to which elements of the crania were preserved.

The latter point is especially significant. Previous work comparing archaeological crania with known examples of decorated crania and ‘skull’ cups has shown that crania subjected to post-depositional taphonomic processes present lower element completeness and poorer preservation of the facial elements; in contrast, decorated crania often retain most or all elements, including facial bones (Bello et al., Reference Bello, Crete and Parfitt2022). In the Candelaro cranial cache, several crania retain parts of more than one facial element, but maxillae and nasals were often broken or absent. This was also observed in primary burials from Masseria Candelaro; it likely represents typical post-depositional taphonomic processes. For the cranial cache, the approximate anatomical position of each fragment was digitally transcribed, superimposing the fragments to create a composite diagram of bone representation (Figure 6). Each cranium preserves a different selection of elements, with the frontal and occipital bones predominating (Supplementary Material 3). Breakage patterns and element representation do not fit the expected profile of either trophy heads or of intentional, patterned reduction of the cranium (Table 2).

Figure 6. Approximate position of all fragments from the cranial cache superimposed. The colour gradient on the greyscale indicates the representation of overlapping fragments with that location preserved (MNE). The maximum (100 per cent) was set at seventeen as at least seventeen separate groups of elements were identified. The highest MNE derived from this illustration is 9.

The Candelaro crania may have become fragmented through several pathways. Some breakage probably occurred in situ in primary burials through compaction and crushing. The under-representation of robust regions such as the temporal bones, the extensive reduction of all elements, and concretion across broken margins show that they were broken before deposition in Structure Q. The lack of cutmarks, chopmarks, and percussion pits sets the Candelaro cache apart from other cranial collections such as Herxheim and Gough's Cave, where there is convincing evidence for cannibalism alongside processing of bone and production of ‘skull’ cups or calottes (Boulestin et al., Reference Boulestin, Zeeb-Lanz, Jeunesse, Haack, Arbogast and Denaire2009; Bello et al., Reference Bello, Saladié, Cáceres, Rodríguez-Hidalgo and Parfitt2015). In contrast, long-term storage and regular handling of bone leads to increased fragmentation, often beginning at the margins of bones (Caffell et al., Reference Caffell, Roberts, Janaway, Wilson and Williams2001). It seems most likely that these crania and mandibles were retrieved from burials post-mortem and actively used over a century or more. Their circulation and handling caused further breakage, perhaps explaining why some crania are represented by multiple, morphologically similar but non-refitting, fragments. Secondary burial within Structure Q may have caused further breakage.

Ex-ancestors: the meaning of the deposit?

After up to two centuries of use, around 5450 bc or later, this group of ancestral crania and mandibles were deposited in Structure Q, in a discrete area on the surface of its upper levels. The crania were not buried, but they may have been lightly covered with soil or placed in some organic container. Later, around 5400 bc, the final cranium was placed on top. Depositing the cranial cache was not necessarily an important ritual act; it may have been an unelaborate, final disposal of a heavily used collection of crania at the end of their use-lives. In some respects, it parallels the decommissioning of disarticulated human bone at Grotta Scaloria (Robb et al., Reference Robb, Elster, Isetti, Knüsel, Tafuri and Traverso2015).

Why these crania were taken out of circulation and transformed into ex-ancestors is an open question. It is possible that this was how the collection was normally stored, and simply abandoned one day. Alternatively, it may represent disposal of well-worn ritual material that had fallen out of use, while the practice of cranial curation continued in ways that have not been archaeologically detected. It is also possible that this signified the end of cranial collection during a time of social change. In later centuries, crania continue to be important, bones continue to be circulated, and villages remain places of ancestral memory, but this is accomplished in other ways. For instance, Late Neolithic groups often returned to the site of earlier villages to bury their dead. The placement of the final cranium in Structure Q could signify a commemoration of the older ancestors and final closure of this ritual cycle. If this is the case, it fits well with ethnographic studies of ritual practices which emphasize the temporary nature of ritual cycles, particularly in small-scale groups (Barth, Reference Barth1995). The post-mortem lives of the Masseria Candelaro ancestral crania extended over four to eight generations; this is likely to have surpassed the duration of living memory and may well be at the limit of most oral histories in tribal societies (Wiessner & Tumu, Reference Wiessner and Tumu1998).