Antlers, bone pins and flint blades: the mesolithic cemeteries of Teviec and Hoedic, Brittany.
The shell middens Teviec and Hoedic are located on what are now small islands in the Bay of Quiberon in Brittany, off the Atlantic coast of northwest France. An abundant microlithic industry and a single radiocarbon estimate places them in the late Mesolithic. The sites are best known for their evidence of elaborate burial practices, with stone and red-deer antler structures, evidence for ceremonial burning and feasting, and abundant and varied grave goods. Together they constitute some two-thirds of known French Mesolithic burials. Teviec and Hoedic are critical to our understanding of the late Mesolithic and the transition to the Neolithic. The sites are part of the phenomenon of increasing 'complexity' in the late Mesolithic of northwest Europe; they also fill a geographical gap between the cemeteries of south-central Portugal and those of southern Scandinavia.
While the importance of Teviec and Hoedic has long been recognized, there has been little further analysis since their discovery and excavation in the earlier part of this century (Taborin 1974 is a notable exception). Yet opinions about the sites are frequently presented, especially in publications dealing with the Mesolithic-Neolithic transition and the origins of the megalithic tombs of Brittany in particular and of western Europe in general (e.g. Bender 1985; 1986; Boujot & Cassen 1993; Chapman 1981; Clark & Neeley 1987; Hibbs 1983; Mohen 1990; Newell 1984; Newell & Constandse-Westermann 1988; Patton 1993; Renfrew 1976; Sherratt 1990; Thomas & Tilley 1993). The intent of this paper is to present a more thorough analysis of mortuary variability at Teviec and Hoedic, examining the differences between them, as well as the similarities.
Teviec and Hoedic were excavated by M. Pequart and S.-J. Pequart in the late 1920s and early 1930s. It is due to the high quality of their excavation reports, excellent for the time, that the present re-analysis is possible. Nevertheless, some of the statements and identifications contained in the reports should be viewed with caution.
Teviec and Hoedic are roughly contemporaneous, based on similarities in tool typology and burial practices (Pequart et al. 1937; Pequart & Pequart 1954; Rozoy 1978). A single radiocarbon determination on charcoal from a hearth in the lower part of the midden at Hoedic provides an estimate of 6575+350 b.p. (Gif-227) (Delibrias et al. 1966; Patton (1993: 39) calibrates the date to 5500-5110 BC). The burials themselves remain undated. The large standard error of this date limits its usefulness; at two standard deviations it overlaps dates for Breton early Neolithic passage-graves and long mounds at c. 5700 b.p. No stratigraphic breaks were noted within the Mesolithic levels at either site, and the materials recovered were described as homogeneous throughout the 0.5 to 1.0 m of deposits. Neolithic deposits were encountered at Hoedic, but the 0.3 to 0.5 m of Mesolithic deposits were apparently entirely sealed by a layer of sterile gravel (Pequart & Pequart 1954: 10-12).
Unfortunately, the rise in sea-levels from Atlantic times means that the sites must be looked at in isolation (cf. Hibbs 1983: 274); the now-submerged coastal plain on which they were high points was undoubtedly the focus of Mesolithic settlement in the area. Additional Mesolithic sites in Brittany include both shell-midden sites (Beg-er-Vil, Point St-Gildas, Anse du Sud, La Torche) and many more non-shell-midden sites (e.g. Kerhillio, Kerjouanno, Malvant, Porz Carn, La Girardiere, Ty Lann, Ty Nancien) (Rozoy 1978: 818-20). With a 10-m drop in sea level (c. 6000 b.p.) (Ters 1973, cited in Rozoy 1978: 784; Admiralty Chart 2353, Presqu'Ile de Quiberon to Croisic 1995), Teviec becomes attached to the mainland via the Quiberon peninsula, while Hoedic joins with the island of Houat as well as a number of smaller islets [ILLUSTRATION FOR FIGURE 1 OMITTED]. This would leave Teviec and Hoedic between i and 2 km from the shore at the time of their occupation - much depends on a more precise estimate of the date of occupation. The destroyed shell-midden at Point St-Gildas appears to have been roughly the same distance from the shore during its occupation early in the Atlantic period (Rozoy 1978: 819), while Beg-er-Vil was likely somewhat closer to the coast. That such large shell-middens are found so far inland strongly suggests the presence of much larger sites closer to the coast.
Faunal remains: seasonality and domestication
The middens in which the burials were encountered contained abundant remains of shellfish, while boar and red deer dominate the mammalian fauna. Additional terrestrial mammals include roe deer, beaver, hedgehog, fox, marten, wild cat, aurochs/cow, sheep/goat, and domestic dog. Whale remains were found in small numbers at both sites; at Teviec a whalebone was incorporated into a feature associated with grave A. A few teeth of a small species of sea! were also reported for Teviec. A variety of bird species are reportedly present, the most abundant being duck (Pequart et al. 1937; Pequart & Pequart 1954; see also Rozoy 1978). Despite the lack of sieving, a large number of fish remains were found, indicating a great range in size of species. Due to the fragmentary state of the bones only the pharyngeal plates of wrasse were identified in the original report; these were apparently abundant at Teviec (Pequart et al. 1937: 102). A large incised fish mandible found in one of the graves at Teviec may belong to a meagre (Argyrosmus regium) (Rozoy 1978: 788). Plant remains were represented by carbonized pear pips and hazelnut shells.
The abundant duck (Anas sp.) remains and the presence of fur-bearers such as marten (Mustela mattes) suggest winter occupation of Teviec (cf. Kayser 1986: 71). (Hoedic bird remains are unidentified.) Other species, such as guillemot (Uria aalge), razorbill (Alca torda) and stork (Cinocia sp.), suggest a spring presence, the season in which most intensive use of shellfish could also be expected; pear and hazelnut suggest autumn presence. Thus there are hints of occupation over most of the year at Teviec, and the possibility of permanent or semi-permanent occupation presents itself. More formal burial rites would be expected at such locations rather than at temporary camps.
There is some confusion over the presence of domestic fauna, other than dog, at the sites. The excavators note the presence of small Bovidae remains at Hoedic (Pequart & Pequart 1954: 14), and a single sheep/goat tooth is reported for Teviec (Pequart et al. 1937: 101).
Remains of small cattle are also reported for the more recent (c. 5000 b.p.) site of Beg-an-Dorchenn (Hibbs 1983). In this slight and uncertain documentation, Bender (1985: 23) sees evidence for early domestication, while Hibbs (1983: 279) deduces a significant pastoral element in the economy. There seems to be little basis for either statement. It is difficult to distinguish female aurochs from domestic cattle based on limited fragmentary remains, and remains of indigenous sheep have been suggested for Mesolithic contexts in the west Mediterranean (Geddes 1985; although see Zilhao 1995). The purported sheep/goat tooth itself has since been lost (Rozoy 1978). The few bones of domesticates, if such they are, may be evidence of interaction with contemporary Neolithic stock-keeping communities (cf. Kayser 1986; Thomas & Tilley 1993); alternatively, they may represent the hunting of strays.
Description of the cemeteries and burials
Teviec and Hoedic each occupy an area of only some 200 sq. m; both, actively eroding prior to excavation, were certainly once larger (Pequart et al. 1937: 5; Pequart & Pequart 1954: 10) and additional graves may have been present. At Teviec, the graves were all excavated partly into the sterile beach deposits on which the site lay, and partly through overlying midden material. Many were associated with substantial stone structures, both lining the grave and rising above it to a height of 0.6-0.8 m. At Hoedic, the graves lay within bedrock depressions at the base of the midden. No graves at either site were solely within the midden deposits, arguing for some degree of contemporaneity in their use for burial, followed by a period in which they were used for non-burial activities. Possibly these later activities were still associated with funerary practices, although it is difficult to envisage such continuity at Hoedic since the burial features would not have been visible for at least part of this time. On the other hand, wooden structures may have marked the graves - post-holes can be difficult to identify in shell-middens.
The 10 graves found at Teviec held the remains of some 23 individuals. What has been identified as a cenotaph (a grave without a body) (Pequart et al. 1937: 59) was also present. Seven of the ten graves contained multiple individuals. A total of nine graves were recovered from Hoedic, containing 14 individuals (a tenth grave, M, was assumed to have contained a child based on its size, although no skeletal remains were preserved). Three graves contained multiple individuals, while the others each contained one individual. One or two graves held the remains of a child only, and the others the remains of adults interred with or without children. The legs of the skeletons remaining in primary position were generally tightly flexed, and may have been bound in position (Pequart & Pequart 1954: 49).
In addition to the graves themselves, other features at Teviec include a series of stone-lined hearths showing varying degrees of burning [ILLUSTRATION FOR FIGURE 2 OMITTED]. Pequart et al. (1937: 19-23) classify these into three types - domestic, feasting, and ritual - on the basis of size, placement in relation to the graves, and degree of burning. At least the feasting and ritual hearths appear to have been involved in activities associated with funerary behaviour. The feasting hearths are generally associated with the larger graves, the best example being the close association of Feature 7 with grave D (Pequart et al. 1937: [ILLUSTRATION FOR FIGURE 3 OMITTED]). These relatively large (c. 0.8-1-0 m diameter) hearths contain considerable quantities of calcined bone fragments and charcoal. The ritual hearths also have signs of burning, but less intense and less prolonged. These 'hearths', which appear to represent light fires made directly on the covering slabs of the graves, are found over the majority of the graves. With one exception, the underlying skeletons show no evidence of burning, again indicating that the fires were light. Pequart et al. suggest that they were used for the burning of funerary offerings (1937: 32). Invariably they include one or two mandibles of red deer and/or wild boar - token offerings of the animals killed and consumed during funerary feasts? The validity of the distinction between these different types of hearths is certainly open to question, particularly that between domestic and feasting hearths. At least the ritual hearths appear clearly defined by the criteria of light burning directly over a grave and the association with a red deer and/or boar mandible.
These features are less common and less well defined at Hoedic (Pequart & Pequart 1954: 623). The site plan published for Hoedic (Pequart & Pequart 1954: [ILLUSTRATION FOR FIGURE 42 OMITTED]) does not identify which grave is which, and it is difficult to reconcile with the written descriptions of the grave locations in relation to one another; FIGURE 3 represents Newell & Constandse-Westermann's (in preparation) efforts at a best approximation.
In the case of the multiple burials at both Teviec and Hoedic, it appears that the last additions were always found in a largely articulated state. The skeletal elements of previous interments were in some cases clearly pushed aside to make room for later interments (in particular, graves H and K at Teviec). Grave K contained the remains of a total of six individuals [ILLUSTRATION FOR FIGURE 4 OMITTED]. Apparently the ritual hearth associated with this grave was used repeatedly, as it shows evidence of a greater degree of burning than the others. The cranium of skeleton 4, underlying the ritual hearth, showed light traces of burning (Pequart et al. 1937: 23). Presumably it was subjected to repeated light burning as additional bodies were added to the grave. More severe burning is evident on the remains of two children last interred in grave C at Hoedic, where the interpretation is the same (Pequart & Pequart 1954: 34).
Skeleton 6 in grave K is somewhat unusual. The first interment in this grave, it remained distinct and separate from the others - it was not moved aside to make room for later interments. The skeleton, a young adult male, lay in a stone-lined depression, its skull within a 'box' constructed of stone slabs. The positioning of the hands crossed on the abdomen is also unique at the site (Pequart et al. 1937: 4951). Fragments of two microlithic projectile points were found embedded in the 6th and 11th thoracic vertebrae of this individual, with no signs of healing (Pequart et al. 1937: 52-4, [ILLUSTRATION FOR FIGURES 23 & 24 OMITTED]). The same individual exhibits a healed fracture of the mandible. Evidence of violence was also found on a number of other skeletons at Teviec (Newell et al. 1979: 137).
Alexander Marshack (1972: 348) observed a series of up to 110 intentional markings on two faces and one edge of a child's rib at Teviec [ILLUSTRATION FOR FIGURE 5 OMITTED]. The marks, clearly not related to defieshing, emphasize the accessibility and use of human remains after interment. Grave H contained three individuals (two adults and one child), of which the child was the second interment; some period of time appears to have elapsed between each successive addition. Skeleton 1 was the most disturbed, being piled into one corner of the grave; the remains of the child were also disturbed, but to a lesser extent. If any articulation remained, it seems strange that an element would be incised and then replaced with the same individual. Is it possible that the rib is an isolated element from yet another child?
Grave goods found with the burials at Teviec and Hoedic include flint implements, utilized quartzite and schist cobbles, bone pins, 'daggers', bi-points (fishing gorges?), and awls, an antler 'baton', antler picks and/or clubs, worked boar tusks, perforated red-deer teeth and an abundance of perforated marine shells of various species. The placement of shells around the skeleton in many cases suggests necklaces, bracelets and headgear (Pequart et al. 1937; Taborin 1974). There is one example of an engraved mandible from a large species of fish at Teviec (Pequart et al. 1937: plate XI, no. 7). Red ochre, found in many graves, tends to be concentrated in the chest area of the skeletons (Pequart & Pequart 1954: 77).
From the very beginning, Pequart et al. (1937: 62-70) discuss the possibility of social differentiation, placing the individuals of graves A and D, together with skeleton 5 in grave K, as elevated in rank above other individuals at Teviec, and recognizing the adult of grave J as the richest individual at Hoedic (Pequart & Pequart 1954: 53). Interestingly, one of the three infants of grave C is also suspected to be of higher status than the more ordinary children at Teviec; taken at face value, this would suggest some degree of hereditary status. Clark & Neeley (1987: 125) seem to concur (see also Patton 1993: 37):
The French sites [referring to Teviec and Hoedic] contain the only relatively clear-cut indicators of an ascription-based differential in status that we have found so far.
Thomas & Tilley (1993: 228) see little indication of social hierarchy. Newell & Constandse-Westermann (1988: 171) agree, contrasting a group of late Mesolithic cemeteries with recent finds at Skateholm:
Whereas the grave accoutrement profiles of Moita do Sebastiao, Hoedic, Teviec and probably Vedbeek-Bogebakken indicate societies based on earned status, closely following the status dimensions of age and sex, Skateholm I and Skateholm II largely ignore the former source of patterning, while the frequencies of the latter are indicative of a system of ascribed status, possibly based upon social ranking.
Yet, in an earlier publication, Newell (1984: 75) interpreted Teviec and Hoedic as showing the beginnings of ascribed status. Rozoy (1978: 1167) sees in the French Mesolithic no evidence 'for any structures [burials or dwellings] richer than others'. Bender (1985: 23; 1986: 24-5), making a distinction between the two sites, sees a slight degree of social differentiation, more pronounced at Teviec. Given the range of opinion presented, it is worthwhile examining the evidence from these two sites in greater detail.
The mortuary ritual presents evidence both for homogeneity and for variability at the intra-and inter-site levels. At Teviec, the lower limbs of the skeletons were always flexed, most on their right sides. A few individuals are described as 'seated'. Where observation was possible, the hands, with two exceptions, were positioned at the hips (Pequart et al. 1937: 60-61). Orientation was more variable and seems to exhibit little correlation with other aspects of mortuary behaviour. At Hoedic, burials lay on either side, and there was one example of a seated individual. Greater variability is seen in the placement of the upper limbs. Orientation was somewhat more patterned at Hoedic; while their general north-south orientation is conditioned by the same tendency of the bedrock depressions into which the burials were placed, the placement of the head where it could be determined was invariably to the north half of the compass (Pequart & Pequart 1954: 66) which cannot be reduced to the orientation of the depressions. The use of red ochre is common to both sites, although it appears far less frequently at Hoedic. Evidence for burning associated with the mortuary ceremony is seen at both sites; traces of burning on a few skeletons appear to have resulted not from intentional burning of the body but from overlying ceremonial fires. Finally, the graves at Teviec appear to cluster more as a unit, while at Hoedic there are two more isolated graves.
Age and sex differences
The burial sample from both sites includes adult males and females, and children of all ages [ILLUSTRATION FOR FIGURE 6 AND 7 OMITTED]. There is little indication that any age or sex group is under-represented relative to a living population (Weiss 1973).
Such distinctions as are visible in mortuary behaviour at both sites seem to be related partly, although not entirely, to the subordinate dimensions of age and sex. Two shell species - Trivia europea (cowrie) and Littorina obtusata (periwinkle) [ILLUSTRATION FOR FIGURE 8 OMITTED] - dominate the burial assemblage, despite the much greater diversity of locally available species seen in the shell midden and inferred through a reconstruction of the local environment (Taborin 1974: 158). This in itself suggests symbolic importance for these particular species - they are certainly too small to have been of any dietary value. Taborin (1974) found that pierced T. europea shells were associated with males, and L. obtusata shells with females. The association is not one of presence/absence, since nearly equal numbers of males and females are found with both species. Rather, it is seen in the quantity with which they are represented, with the distinction seeming to hold for the various kinds of objects - bracelets, necklaces, and headgear - made of shell (Taborin 1974: 171). Thus a bracelet on a male would be comprised of both species, but predominantly T. europea. Taborin also identifies a tentative trend between age and quantity of shell ornaments; young and middle-aged adults seem to have been interred with more ornamentation, while both younger and older individuals exhibited less (1974: 174-7).
Ritual hearths containing deer and/or boar mandibles are found with all graves at Teviec except for B, C, and L (grave J is too disturbed to determine whether or not a ritual hearth was in association). Interestingly, these contained the remains of infants and children (C and L) and an adolescent (B). By contrast, all the adult graves have ritual hearths (Pequart et al. 1937: 62). This suggests that meat feasts, involving red deer and/or boar, were not held for sub-adults at death, and argues for the absence of hereditary social rank. Yet some sub-adult burials are found with grave inclusions, and are otherwise accorded a similar mortuary treatment, one perhaps not available to all members of the society.
In the following discussion, richness refers simply to the number of artefact types present (TABLE 1), without taking into account their abundance (Leonard & Jones 1989). There are both theoretical and practical reasons for this approach. A number of studies have suggested that presence/absence is a more sensitive indicator of status than quantity (Goldstein 1980; Macdonald 1990; O'Shea 1984; Ravesloot 1988; Rothschild 1990; Schulting 1995a). In some cases it was not possible to associate all the artefacts present in a grave with specific individuals; the shell beads are sometimes found strewn throughout the grave, mitigating against the use of absolute quantity of items in the analysis. But in many cases objects are clearly associated with individuals, often placed in the hands or on the chest.
Sub-adults show lower artefact richness than the adult group at Teviec [ILLUSTRATION FOR FIGURE 9 OMITTED]; among them are the only three individuals at the site lacking non-perishable grave inclusions entirely (with the exception of red ochre). Significance tests (TABLE 2) quantify and provide more detail on these differences. Adults at Teviec show significantly greater number of artefact types overall. Breaking down types into utilitarian and sociotechnic groups (TABLE 1) (Binford 1962; Winters 1968; Schulting 1995b) shows that adults dominate in both, although in the socio-technic group the difference does not quite reach the standard 0.05 level of significance. Males and females do not differ significantly in their overall artefact richness. However, males on average are interred with over twice as many utilitarian artefact types as females, and this difference is statistically significant. The distribution of sociotechnic types, on the other hand, is almost equal between the sexes. Tests of association show that neither sex is associated with particular artefact classes, although this may be a function of small sample size. Within the adult age-group itself, no significant differences based on age are seen at either site.
In general, the structures at Hoedic appear simpler, lacking the stone linings and superstructures of Teviec; also in contrast to Teviec, there appears to be little differentiation of treatment along the dimensions of age and sex. None of the differences in artefact richness between age and sex classes attain statistical significance ([ILLUSTRATION FOR FIGURE 10 OMITTED] & TABLE 3). Contrary to what might be expected, sub-adults actually have a somewhat higher average incidence - although remaining statistically insignificant - of utilitarian artefact types. This contrasts not only with Teviec, but also with other studies of mortuary data from hunter-gatherer societies (see O'Shea 1984; Schulting 1995a; 1995b). Females on average tend to have slightly higher artefact richness, due largely to the influences of two very 'rich' female burials from graves H and J. Only one male, from grave K, is comparable. No relationship was noted between grave location within the cemetery and sex, age or artefact richness at either site.
TABLE 1. Division of artefacts into utilitarian and sociotechnic categories.
utilitarian sociotechnic artefact types artefact types
truncated flint blade incised fish bone microlith antler baton piercer antler structures scraper boar/deer mandible flint debitage boar tusk misc. flint implement perforated deer canine schist pebble tool perforated ray ornaments quartz pebble tool schist bead bone bipoint Trivia bone awl Littorina misc. bone implement Patella antler implement Nassa bone pin(*) Cardium Dentalium Murex Pecten Pectunculus Purpura Haliotis red ochre
* probably belongs in the sociotechnic category TABLE 2. Age and sex differences in richness of grave inclusions at Teviec.
adult subadult male female
total n umber of types
n 15 8 7 8 [Mathematical Expression Omitted] 7.33 3.38 8.43 6.38 t 2.40 0.97 p 0.025 0.348
number of utilitarian types
[Mathematical Expression Omitted] 2.73 0.75 4.00 1.63 t 2.18 2.21 p 0.041 0.045
number of sociotechnic types
[Mathematical Expression Omitted] 4.60 2.63 4.43 4.75 t 1.89 -0.24 p 0.072 0.816
[TABULAR DATA FOR TABLE 3 OMITTED]
Red-deer antlers, bone pins and flint blades
Structures of red-deer antler are associated with two adults (one male and one female, graves A and D) at Teviec, and with four adults (two males and two females, graves F, H, J and K) at Hoedic; these appear to have formed small tent-like arrangements over the heads of these individuals [ILLUSTRATION FOR FIGURE 11 OMITTED]. The hypothesis that these represent special treatment is an obvious one to make. The six individuals with antler structures have over twice the average artefact richness of individuals lacking these structures; this difference is statistically significant (TABLE 4). If these structures are indeed associated with hunting, either in reality or symbolically, it suggests that this was perceived as a high-status activity (agreeing with the findings of O'Shea & Zvelebil (1984) at the site of Oleneostrovski Mogilnik, but contra Jacobs (1995)). The stone superstructures at Teviec appear to be more related to the number of inviduals in the grave, and show no clear trend in terms of richness when adjusted for this.
Apart from the nearly ubiquitous shell beads, truncated flint blades and bone pins form the two most common classes of grave inclusions. Pequart & Pequart (1954: 68) remark on an interesting difference between Teviec and Hoedic in the size of the truncated blades found with the burials. At Hoedic, these blades, which have been interpreted as knives (Rozoy 1978), are similar in size to specimens found in the general midden; at Teviec those placed with burials were reportedly larger than those found in the midden. This suggests that perhaps long flint blades were being manufactured specifically as grave offerings, which appears to fit in with an emerging pattern of slightly more elaborate mortuary practices at Teviec. However, in the published photographs and drawings of flint blades from Teviec (Pequart et al. 1937), the blades found with the graves do not seem significantly longer. But the idea that flint blades are associated with high status does receive [TABULAR DATA FOR TABLE 4 OMITTED] support from a comparison of the artefact richness in those graves with blades and those without (TABLE 4). Furthermore, the frequency of truncated blades in the graves, approximately equally distributed among adults of both sexes as well as sub-adults, can be contrasted with their scarcity in the midden (Rozoy 1978: 783).
The distinctive bone pins, or stylets, also show differences; with one or two exceptions, those from Teviec are made of wild boar fibulae while those from Hoedic are made of deer bone (Pequart & Pequart 1954: 69). Three examples, two from Teviec and one from Hoedic, are decorated with similar patterns of fine incised lines. In many cases the pins were found placed on the chest in a manner suggesting garment fasteners [ILLUSTRATION FOR FIGURE 12 OMITTED] (Pequart et al. 1937: 67; Pequart & Pequart 1954: 69). Pequart & Pequart (1954: 69; echoed by Bender 1985: 22) suggest that they may have been made specifically as grave offerings, although Pequart et al. (1937: 91) note that some of the worked bone fragments from the midden deposits at Teviec may be broken stylets. Again, graves with pins show a strong tendency to be associated with greater-than-average artefact richness (TABLE 4). The pins, while not restricted to either sex or age group, are more common with adults.
Tests of association show a relationship between antler structures, flint blades and bone pins. Moderate but significant correlations occur between antler structures and bone pins ([Phi] = 0.45, Fisher's p = 0.009; [Phi] is the equivalent of Pearson's r using only presence/absence data), and between bone pins and flint blades ([Phi] = 0.56, Fisher's p = 0.003). Only one individual of the six with an antler structure lacks a bone pin. All this suggests bone garment-pins were strongly associated with high status at Teviec and Hoedic. It is likely that the garments worn by these individuals were of superior quality; clothing is one of the earliest-appearing and most effective means of communicating differences in status. While flint blades appear to also have functioned partly as status markers, they show no similar association with the elaborate antler structures ([Phi] = 0.27, Fisher's p = 0.184). All of this indicates a number of complex status roles that cannot simply be reduced to age and sex, or to random variation. Nor does it seem likely - although this remains a possibility - that the observed patterning is based solely on achieved forms of status.
Mortuary behaviour at Teviec and Hoedic presents a number of parallels to that of the Early Neolithic. Thomas & Tilley (1993: 228) remark on the association between feasting, ritual, death, and burning at the sites and the continuation of these practices into the Neolithic (see also Hibbs 1983: 310-13). Bender (1986: 25) remarks on the resemblance of the mortuary structures at Teviec and Hoedic to the paved area and small cairn at the early Neolithic site of Les Fouaillages, Guernsey, where Mesolithic flints were recovered from the surface underlying the cairn (Kinnes 1982). The re-use of graves seen at Teviec and Hoedic, with earlier interments moved aside to make way for the new additions, is common to the early Neolithic of both Brittany and Britain. Bone garment-pins, possibly a special feature of mortuary rites, are also seen in both Mesolithic and Neolithic contexts. Others have found parallels between the arrangements of cists, hearths, and pits in the shell middens of Teviec and Hoedic and the features found underlying the Neolithic Manio long mound at Carnac (Kirk 1993: 212; Thomas & Tilley 1993: 239).
Thomas & Tilley (1993: 228) address the oft-remarked dichotomy between the Neolithic as ritual phenomenon and the Mesolithic as economic phenomenon. They would have us view the middens of Teviec and Hoedic as reflecting a conscious decision to 'accumulate rather than scatter debris over generations' [emphasis in original], and that they are therefore more than 'simply rubbish heaps'. They go further (1993: 228):
These shell mounds were obviously foci for large social gatherings and feasting. They also provided an arena for ritual and the burial of the dead [emphasis added].
I would argue that the shell middens first and foremost are indeed rubbish heaps, and not obviously foci for large social gatherings and feasting - unless by this is meant a habitation site at which such activities, among others, may have occurred. Any other interpretation must be supported with additional evidence. The presence of burials does not itself transform a location into ritual space; this is an ethnocentric view. And, while we may concede that the act of burial itself, particularly given the elaborate nature of the Teviec and Hoedic graves, might dominate how the site was viewed at the time, there is no reason to believe that this remained its main emphasis at other times. The duration of occupation of Teviec and Hoedic is unknown (based on its greater depth of deposits Teviec may have been in use for a longer period of time). The graves are partly in the lower midden deposits and partly dug into sterile beach gravels; the sites were used both before and after burial for what appear to be residential purposes. Recall also that both sites were eroded when excavated; their original extent is unknown. The burials may have been at the back of much larger settlements, as they are in the roughly contemporaneous cemeteries of Vedbaek-Bogebakken and Skateholm in southern Scandinavia. While it is possible that the sites did indeed have ritual significance, and that this was the major focus of activity, this remains to be shown. Every activity may indeed have a ritual aspect to it, but to call every activity 'ritual' on this basis is, I believe, to erode the usefulness of the term. The formation processes behind shell middens have been observed in many parts of the world - the Northwest Coast of North America and northern Australia to provide two examples - where they are indeed habitation sites full stop. And the occurrence of burials in shell middens is the rule rather than the exception throughout most of Northwest Coast prehistory.
The burial regimes at Teviec and Hoedic exhibit both similarities and significant differences (TABLE 5). There is evidence for social differentiation based on age and sex, and hints of ascribed status differentiation. The differences between males and females appear to reflect a horizontal rather than a vertical status dimension. In general the graves at Teviec appear more elaborate, with a more structured distribution of grave inclusions. If the admittedly small sample sizes allow meaningful comparison, then it may be suggested that Teviec shows more formal patterning and more 'complexity' than Hoedic in some respects, though not all. The relationship between the two sites then becomes of interest. There are a number of possibilities. The sites may both have been larger and contained additional burials, so that different parts of cemeteries may be represented. Or they may represent two contemporaneous communities with slightly differing socio-economic organization. Finally, the difference may reflect diachronic variability. A programme of directly dating human bone from both sites would be helpful.
[TABULAR DATA FOR TABLE 5 OMITTED]
The Breton shell-midden sites are at the fringe of the late Mesolithic coastal settlement system; larger, more complex sites would be expected on the coast itself. In a study of northwest Zealand's Ertebolle site distribution, Paludan-Muller (1978) found that the zone 1-5 km from the coast displayed relatively little settlement, it being more efficient to exploit this zone from coastal base camps, with specialized interior sites further than 5 km inland.
The red-deer antler structures at Teviec and Hoedic are not entirely unique. There are three adult graves with red-deer antlers at the late Mesolithic cemetery of Vedbaek-Bogebakken, Denmark (Albrethson & Brinch Petersen 1976: 26), but with no indication of significant association with higher than average richness ([Mathematical Expression Omitted] vs 2.05, t = 0.49, p = 0.63). Comparisons may also be made with three adult graves at Skateholm II, Scania (Larsson 1989; Persson & Persson 1988). Finally, a small test excavation at the recently discovered and roughly contemporaneous site of Beg-er-Vil, located between Teviec and Hoedic [ILLUSTRATION FOR FIGURE 1 OMITTED], revealed an intriguing feature: a pit, surmounted by three antlers and a schist manuport, contained - in addition to burned bone fragments and stone tools - an incised shell fragment and two incised bone artefacts (Kayser & Bernier 1988; Poissonnier & Kayser 1988). One of these, made on a deer tibia and described as a poignard, bears considerable resemblance to the bone pins of Teviec and Hoedic [ILLUSTRATION FOR FIGURE 13 OMITTED]. It is tempting to interpret this feature as a 'structured deposit' of the type seen in the Neolithic. Comparable features have been found in association with the Mesolithic cemteries of Skateholm (Larsson 1989).
The use of bone from different species for the pins at the two sites recalls a similar distinction in tooth ornamentation noted by O'Shea & Zvelebil (1984: 10) at the large Mesolithic cemetery site of Oleneostrovski Mogilnik in Karelia, where the distinction was interpreted as identifying a binary division in society, i.e. along moiety lines. The boar bone pins from Teviec are easily visually distinguished from those made of red deer at Hoedic [ILLUSTRATION FOR FIGURE 13 OMITTED]. While I would certainly not argue on this basis for a moiety division of society (particularly given the distance between the sites), the use of different bone garment pins may have been one means of group identification (cf. Wiessner 1983). Alternatively, it may reflect availability of raw materials, i.e. boar and red deer. In terms of habitat, there is no reason to expect such a difference between the hunting catchments of the two sites; boar is placed first in the species lists at both sites, which appear to be given in order of abundance (Pequart et al. 1937: 101; Pequart & Pequart 1954: 19).
An interesting finding is that those graves with antler structures - all adults - have markedly greater artefact richness than those without such structures. Bone garment pins and flint blades, both of which may have been made specifically as grave inclusions, are also associated with significantly higher than average artefact richness. There is a moderate positive correlation between antler structures and bone pins, and between bone pins and flint blades. What do these differences mean? There do not appear to be any exotic materials in the artefact assemblage, nor is there evidence for great input of labour into the manufacture of some items (as there is for the finely polished stone axes of the early Neolithic). Nevertheless, there is patterning in the distribution of these items. Pierced marine shells, despite their local abundance, were clearly distinguished along the dimension of sex (Taborin 1974). And the more elaborate graves with antler structures also display significantly greater artefact richness - such patterns cannot be fortuitous. They are meaningful, even if the full depth of their meaning remains elusive.
A number of similarities between late Mesolithic mortuary behaviour at Teviec and Hoedic and early Neolithic mortuary practices of Brittany and Britain have been noted, but closer comparisons are needed to move beyond simple statements of 'continuity' or 'discontinuity'. Mortuary analysis undoubtedly has an important role to play in understanding the Mesolithic-Neolithic transition, one that is currently under-utilized (cf. Solberg 1989: 284). But it is essential to first come to terms with the range of variation in mortuary behaviour within the late Mesolithic; this has been the goal of this paper.
Acknowledgements. A shorter version of this paper was presented at the From the Jomon to Star Carr conference held at Cambridge/Durham, 4-8 September 1995. I would like to thank the conference organizers, and especially Simon Kaner, for providing this opportunity. I would also like to thank Richard Bradley, Robert Chapman, Trinette Constandse-Westermann, Nyree Finlay, Steve Mithen and Raymond Newell for their comments on earlier drafts of this paper. I am grateful for Newell & Constandse-Westermann's permission to reproduce FIGURE 3. Comments by two anonymous ANTIQUITY reviewers are much appreciated. Any remaining faults are of course my own. Lastly, I would like to acknowledge the support of the British Office of the Commonwealth Association and the Social Sciences and Humanities Research Council of Canada.
ALBRETHSEN, S.E. & E. BRINCH PETERSEN. 1976. Excavation of a Mesolithic cemetery at Vedbaek, Denmark, Acta Archaeologica 47: 1-28.
BENDER, B. 1985. Prehistoric developments in the American Midcontinent and in Brittany, in T.D. Price & J.A. Brown (ed.), Prehistoric hunter-gatherers: the emergence of cultural complexity: 21-57. Madison (WI): Prehistory Press.
1986. The archaeology of Brittany, Normandy and the Channel Islands. London: Faber & Faber.
BINFORD, L.R. 1962. Archaeology as anthropology, American Antiquity 28: 217-25.
BOUJOT, C. & S. CASSEN. 1993. A pattern of evolution for the Neolithic funerary structures of the west of France, Antiquity 67: 477-91.
CHAPMAN, R. 1981. The emergence of formal disposal areas and the 'problem' of megalithic tombs in prehistoric Europe, in I. Kinnes et al. (ed.), The archaeology of death: 71-81. Cambridge: Cambridge University Press.
CLARK, G. & M. NEELEY. 1987. Social differentiation in European Mesolithic burial data, in P. Rowley-Conwy et al. (ed.), Mesolithic northwest Europe: recent trends: 12-17. Sheffield: Department of Archaeology and Prehistory, Sheffield University.
DELIBRIAS, G., M.-T. GUILLIER & J. LABEYRIE. 1966. Gif natural radiocarbon measurements II, Radiocarbon 8: 74-95.
GEDDES, D. 1985. Mesolithic domesticated sheep in west Mediterranean Europe, Journal of Archaeological Science 12: 25-48.
GOLDSTEIN, L.G. 1980. Mississippian mortuary practices: a case study of two cemeteries in the Lower Illinois Valley. Evanston (IL): Northwestern Archaeological Program.
HIBBS, J. 1983. The Neolithic of Brittany and Normandy, in C. Scarre (ed.), Ancient France: Neolithic societies and their landscapes 6000-2000 BC: 271-323. Edinburgh: Edinburgh University Press.
JACOBS, K. 1995. Returning to Oleni' astray: social, economic, and skeletal dimensions of a boreal forest Mesolithic cemetery, Journal of Anthropological Archaeology 14:359-403.
KAYSER, O. 1986. Les amas coquilliers d'Armorique, Archeologia 218: 68-74.
KAYSER, O. & P. BERNIER. 1988. Nouveaux objets decores du Mesolithique Armoricain, Bulletin de ia Societe Pre-historique Francaise 85: 45-7.
KINNES, I. 1982. Les Fouaillages and megalithic origins, Antiquity 56: 24-30.
KIRK,T. 1993. Space, subjectivity, power and hegemony: megaliths and long mounds in earlier Neolithic Brittany, in Tilley (ed.): 181-223.
LARSSON, L. (ed.). 1988. The Skateholm Project I: Man and environment. Stockholm: Almqvist & Wiksell International. Acta Regiae Societatis Humaniorum Litterarum Ludensis 79.
1989. Late Mesolithic settlements and cemeteries at Skateholm, southern Sweden, in C. Bonsall (ed.), The Mesolithic in Europe: 367-78. Edinburgh: John Donald.
LEONARD, R.D. & C.T. JONES (ed.). 1989. Quantifying diversity in archaeology. Cambridge: Cambridge University Press.
MACDONALD, W.K. 1990. Investigating style: an exploratory analysis of some Plains burials, in M. Conkey & C. Hastorf (ed.), The uses of style in archaeology. 52-60. Cambridge: Cambridge University Press.
MARSHACK, A. 1972. The roots of civilization. New York (NY): McGraw-Hill.
MOHEN, J.-P. 1990. The world of megaliths. New York (NY): Facts On File.
NEWELL, R.R. 1984. On the Mesolithic contribution to the social evolution of western European society, in J. Bintliff (ed.), European social evolution: archaeological perspectives: 69-82. Bradford: Bradford University Press.
NEWELL, R.R. & T.S. CONSTANDSE-WESTERMANN. 1988. The significance of Skateholm I and Skateholm II to the Mesolithic of western Europe, in Larsson (ed.): 164-74.
In preparation. The mortuary practices of the Mesolithic societies in western Europe I: The nine later Mesolithic cemeteries.
NEWELL, R.R., T.S. CONSTANDSE-WESTERMANN & C. MEIKLEJOHN. 1979. The skeletal remains of Mesolithic man in western Europe: an evaluative catalogue, Journal of Human Evolution 8: 1-228.
O'SHEA, J.M. 1984. Mortuary variability: an archaeological investigation. Orlando (FL): Academic Press.
O'SHEA, J.M. & M. ZVELEBIL. 1984. Oleneostrovski mogilnik: reconstructing the social and economic organisation of prehistoric foragers in northern Russia, Journal of Anthropological Archaeology 3: 1-40.
PALUDAN-MULLER, C. 1978. High Atlantic food gathering in northwestern Zealand: ecological conditions and spatial representation, Studies in Scandinavian prehistory and early history 1: 120-57.
PATTON, M. 1993. Statements in stone: monuments and society in Neolithic Brittany. London: Routledge.
Pequart, M. & S.J. PEQUART. 1954. Hoedic, deuxieme station-necropole du Mesolithique cotier Armoricain. Anvers: De Sikkel.
PEQUART, M., S.J. PEQUART, M. BOULE & H. VALLOIS. 1937. Teviec, station-necropole du Mesolithique du Morbihan. Paris: Archives de L'Institut de Paleontologie Humaine XVIII.
PERSSON, O. & E. PERSSON. 1988. Anthropological report concerning the interred Mesolithic populations from Skateholm, southern Sweden: excavation seasons 1983-1984, in Larsson (ed.): 89-105.
POISSONNIER, B. & O. KAYSER. 1988. Les bois de cerfs Mesolithiques de Beg-er-Vil a Quiberon (Morbihan), Revue Archeologique de l'Ouest 5: 35-43.
RAVESLOOT, J.C. 1988. Mortuary practices and social differentiation at Casa Grandes, Chihuahua, Mexico. Tucson (AZ): University of Arizona Press.
RENFREW, C. 1976. Megaliths, territories and populations, in S.J. De Laet (ed.), Acculturation and continuity in Atlantic Europe: 198-220. Bruges: De Tempel.
ROTHSCHILD, N.A. 1990. Prehistoric dimensions of status: gender and age in eastern North America. New York (NY): Garland.
ROZOY, J.-G. 1978. Les derniers chasseurs: l'Epipaleolithique en France at en Belgique. Essai de synthese. Charleville: Bulletin de la Societe Archeologique Champenoise, J.-G. Ruzoy.
SCHULTING, R.J. 1995a. Mortuary variability and status differentiation on the Columbia-Fraser Plateau. Burnaby: Archaeology Press, Simon Fraser University.
1995b. Age and sex differences in burial inclusions on the Plateau. Paper presented at the 28th Meeting of the Canadian Archaeological Association, Kelowna, British Columbia.
SHERRATT, A. 1990. The genesis of megaliths: monumentality, ethnicity and social complexity in Neolithic northwest Europe, World Archaeology 22: 147-67.
SOLBERG, B. 1989. The Neolithic transition in southern Scandinavia: internal development or migration?, Oxford Journal of Archaeology 8: 261-96.
TABORIN, Y. 1974. La parure en coquillage, Gallia Prehistoire 17:101-79.
TMOMAS; J. & C. TILLEY. 1993. The axe and the torso: symbolic structures in the Neolithic of Brittany, in Tilley (ed.): 225-324.
TILLEY, C. (ed.). 1993. Interpretative archaeology. Oxford: Berg.
WEISS, K.M. 1973. Demographic models for anthropology. Washington (DC): Society for American Archaeology. Memoir 27.
WIESSNER, P. 1983. Style and social information in Kalahari San projectile points, American Antiquity 48: 253-73.
WINTERS, H.D. 1968. Value systems and trade cycles of the Late Archaic in the Midwest, in S.R. Binford & L.R. Binford (ed.), New perspectives in archaeology: 175-221. Chicago (IL): Aldine.
ZILHAO, J. 1993. The spread of agra-pastoral economies across Mediterranean Europe: a view from the far west, Journal of Mediterranean Archaeology 6: 5-63.