Intentional tooth removal in Neolithic Italian women.

Robb, John

Cultural modification of the teeth is a widespread, often florid phenomenon, with the highly visible front teeth most commonly furnishing the canvas for dental self-expression. Native peoples of Mesoamerica and South America inlaid the incisors, sometimes with jade; various African and African-American groups filed points and geometric designs on the same teeth; the Etruscans made showy gold bridges, primarily for young aristocratic females (Becker 1995; Corruccini & Pacciani 1989), and modern Euro-Americans engage in elaborate cosmetic orthodontistry. The front teeth can also be modified or lost through craft activities, traumatic injury, and dental therapy (Scott & Turner 1988; see Milner & Larsen 1991 for a comprehensive review of cultural dental modifications). In prehistoric Europe, Jackson (1915) reports possible cases of tooth removal from the British Neolithic.

This paper describes a distinctive pattern of dental modification(1) in Neolithic Italian women. The Italian Neolithic (c. 6500-3200 bc; Skeates 1994), while not a homogeneous period, displays continuity in many aspects of culture. Social life was based upon small villages of 25-200 people, supported by unintensified agricultural economies. In spite of a rich record of art (Graziosi 1974) and burial practices (Robb 1994a), little is known about gender-related behaviour and ritual practices.

Patterns of tooth loss

Samples and methods

Neolithic Italian burials usually occur singly or in small groups, but as sufficient skeletal data accumulate, cultural patterns can be discerned statistically. This analysis is based on data from 30 adult females and 22 adult males from 27 sites in central and southern Italy; a pooled sample is both necessary for most Italian skeletal analysis prior to the Iron Age, and consistent with the social reality of the Italian Neolithic. None of the sites studied would have been demographically or culturally self-sufficient, and traditions would have been maintained and reproduced by groups living at many villages.(2)

Data were collected directly on 36 skulls, and data on 16 specimens taken from published sources (TABLE 1; for information on site location and sources, see Robb 1994a; 1994b). About half of the sample are moderately well-preserved skulls; the rest include only the upper or lower jaw, often quite fragmentary. Most were associated with postcrania. The specimens were [TABULAR DATA FOR TABLE 1 OMITTED] [TABULAR DATA FOR TABLE 2 OMITTED] sexed morphologically, using both cranial and postcranial indicators as available; juveniles and unsexed adult skeletons were excluded.

The basic method followed was to make a standard dental census of each specimen (Buikstra & Ubelaker 1994; Lukacs 1989) (TABLE 1). When a tooth is lost during life, the alveolus does not remain an empty socket but resorbs to become solid bone [ILLUSTRATION FOR FIGURE 1 OMITTED]; teeth present at death (including ones still in place and ones lost archaeologically after death) can be distinguished accurately from teeth lost during life. While assigning teeth to these categories is usually unproblematic, it can occasionally be difficult when using published data, due both to variation in observational standards and to incomplete publication. Consequently, published specimens were included in this analysis only when the status of each tooth socket in the surviving dentition was explicitly noted or when published photographs show their status. (3,4)

Results and statistical analysis

Eight women of 30 in the Neolithic Italian sample lost incisors and/or canines during life ([ILLUSTRATION FOR FIGURE 1 OMITTED]; TABLE 2). Statistically, for females, a total of 6.0% of all anterior teeth were lost before death; none of the 22 males lost incisors or canines at all [ILLUSTRATION FOR FIGURE 2 OMITTED]. The teeth most commonly affected were the central and lateral incisors, although no one tooth was the norm. It is less certain whether premolars were involved in this pattern,(5) and there is no significant difference between males and females in how many molars were lost during life.

While the sample of Neolithic Italian teeth is large, the number of front teeth lost before death is small; it is important to establish that the male-female difference observed could not be due to random fluctuations. The standard statistical methodology for testing the probability of nominal data is the [[Chi].sup.2] test of independence (Blalock 1979; Reynolds 1977).(6) When this test is carried, the male-female difference in antemortem loss of incisors and canines is highly significant; the [[Chi].sup.2] value of 9-88 has a probability value of -002, suggesting that there are only about two chances in a thousand of these results arising through random fluctuations. As a second method of assessing the observed results, random simulation was used, as this can provide a helpful context for understanding small samples (Shennan 1988: 58-9). For each simulation run, a hypothetical 'collection' was generated identical to the actual dataset in its aggregate characteristics of total size, male-female proportions and proportion of teeth lost before death. In these replica samples, the status of each tooth and the sex it was assigned to were generated randomly. In all, 100 random samples were created. To compare them, the [[Chi].sup.2] statistic was used to measure how far each sample departed from having equal rates of antemortem tooth loss for males and females. As FIGURE 3 shows, the actual data have a sex difference in tooth loss greater than do any of the 100 random 'samples'. Both statistical significance testing and the simulation study suggest the sex difference in tooth loss is genuine rather than a statistical fluke.

The same result appears if we analyse the data by individuals instead of by teeth. Among females, 8/30 (26.7%) lost at least one incisor or canine before death. For males, the corresponding figure is 0% (0/22). As before, since sample sizes are small, it is important to assess the probability of these results statistically. The [[Chi].sup.2] test of independence yields a highly significant result ([[Chi].sup.2] = 6.9, p = .009; with Yates' correction for continuity [[Chi].sup.2] = 5.0, p = .025). Fisher's Exact Test (Blalock 1978), more suitable for data with low expected frequencies, assigns these results a probability of .008, which again suggests that the sex difference in anterior tooth loss is highly unlikely to be chance.(7)

The proportion of women experiencing anterior tooth loss can be estimated in a very general way. Using the raw data, about a quarter of female skeletons display lost anterior teeth. However, these remains are fragmentary, and the rate is likely to be higher if we estimate a rate for complete remains. If 24% of all women lost one or more anterior teeth in their upper or lower dentition, around 40% of all women would have lost teeth in either the upper or the lower dentition or both.(8) Likewise, if every anterior tooth had a probability of being lost during life of .06, the chance of a woman's losing at least one of her 12 canines and incisors would have been slightly over half.(9) These rough calculations serve to make the point that probably between a quarter and a half of Neolithic women would have lost incisors or canines during life.

Age distribution

The age at which women lost their anterior teeth is both of anthropological interest and pertinent to whether anterior tooth loss could have been due to dental disease (see below). Unfortunately, disarticulation and fragmentation limit analysis. The assembled corpus, including some of the largest collections from the peninsular Italian Neolithic, represents a typical range of preservation. Of 52 specimens, 8 were disarticulated fragments with no data relevant to age; 8 were published specimens without age estimates; another 6 included published age estimates based on widely varying criteria and useful only for assignment to broad categories such as 'young adult'. Molar wear data are available for 23 specimens, and post-cranial ages estimated by the author are available for 7 specimens. The problem of age data is especially critical for the 8 women who had lost anterior teeth during life. Of these, 2 are represented by isolated fragments; broad age categorizations are available for 6, molar wear data for 4, and detailed age estimates based on non-dental criteria for only 2. Except for the molar wear data, then, age distribution analysis must remain at the level of qualitative observation.

Anterior tooth loss probably began early in adulthood. The permanent dentition does not record tooth loss before 8-12 years of age, and as Cook (1981) points out, deciduous teeth can be extracted for cultural reasons. The fact that a clear diastema is visible in most cases suggests that tooth loss did not occur before the eruption of the complete anterior dentition; otherwise erupting teeth would have partially closed the gaps in the tooth rows.

Within the adult range, one woman with anterior tooth loss (Fonteviva 1) has moderately low wear placing her in early adulthood, and three (Cala Colombo 1, Fonteviva 2 and Continenza) are placed in an undifferentiated 'adult' category, without extreme tooth wear or skeletal degeneration marking extreme age. The two women who can be aged are both mature adults (Catignano, 40+ (Robb & Mallegni 1994) and Lanciano, 50+ (Geniola & Mallegni 1975)). Anterior tooth loss is known throughout the adult age range.

While molar wear does not bear a simple linear relationship to age,(10) it may be a proxy for it. On the average, women losing front teeth during life had slightly higher molar wear scores than their fellows. Median wear scores (using Scott's (1979) scale of 1-40) for women losing front teeth during life were 31.75, 29.875 and 11.0 for the first, second and third molars respectively; the corresponding values for women who had not lost anterior teeth were 23.625, 17.375 and 8.5. The difference between the two groups of women is statistically significant for the first molars and second molars but not for the third molars.(11) While this establishes the high probability that there is a relationship between tooth wear and anterior dental loss, the relationship is a weak one.(12)

This result is open to several interpretations. It may mean that anterior tooth loss increased as the individual grew older. This pattern would make sense in terms of both natural causes, such as dental disease, and cultural causes. If adult women removed a tooth to mark life-events (such as to express grief at the death of kin), the longer they lived, the greater the possibility that such an occasion would arise. Alternatively, since it is impossible to tell from a completely resorbed socket how long before death the tooth was lost, all anterior tooth loss may have taken place at a prescribed and relatively early age. If so, then women who lost anterior teeth in early adulthood may have lived longer, perhaps because of status-related factors. This possibility seems less likely, but cannot be excluded.

Why ablation? Possible causes of anterior tooth loss in Neolithic Italian women(13)

Following Hrdlicka (1940), several analysts have discussed criteria for identifying ritual tooth ablation in skeletal remains (Cook 1981; Merbs 1968) and for distinguishing intentional dental modification from attrition due to craft activity (Blakely & Beck 1984). While individual criteria may be argued, there are two key general points. First, other possible causes must be excluded; these include post-mortem damage, congenital absence, traumatic fractures, activity-related loss and dental pathology. Secondly, it is desirable to establish that tooth loss displays cultural patterns of some sort; criteria frequently cited include symmetrical tooth loss, selection of visible teeth, differential distribution for the sexes or other categories, and age patterns which distinguish tooth loss from progressive age-related loss. Ideally, one can also document the practice of tooth ablation in neighbouring societies, and in local history, myth or legend (Hrdlicka 1940), although this is impossible for most prehistoric populations.

Why did Neolithic Italian women lose their anterior teeth? Post-mortem damage can be ruled out at once; the alveolar remodelling evident in all cases could only have occurred in life. Genetic causes are also unlikely. Congenital absence of the third molars is relatively common in Italian Neolithic people; but the congenital absence of other teeth is rare, and the absence of several teeth in different positions within the same dentitions virtually unknown.(14) Anterior teeth may be broken off or knocked out by traumatic injuries due to accidents or violence. Trauma rates often do vary between the sexes, although it is not known whether this was the case in the Italian Neolithic. However, trauma seems an unlikely cause in the present case. It is hard to imagine trauma knocking out symmetrical pairs of teeth, as occurs in several cases. Moreover, only a fraction of facial traumas would have resulted in tooth loss. If Neolithic women suffered enough facial traumas to result in tooth loss in a quarter of the individuals studied, we would expect to see numerous traumatic injuries to the mandible, maxillae, nasals, zygomatics and adjacent bones. None of the female skulls studied displayed healed facial injuries.

What about tooth loss due to specialized craft activities performed only by females? Craft activities can affect the teeth through general attrition, and Merbs (1968) argues that paramasticatory use resulted in traumatic tooth loss in some Arctic groups as the front teeth were chipped and fractured during forceful gripping. The anterior teeth of Italian Neolithic women are often heavily worn, possibly from paramasticatory activities (Salvadei & Macchiarelli 1983). That wear clearly represents a phenomenon different from whatever produced anterior tooth loss: it affects all anterior teeth, wearing them down to a common occlusal plane, rather than singling out one tooth and destroying it completely. Moreover, if paramasticatory use caused anterior tooth loss in Neolithic Italian women, we would expect to see other signs of such activities, such as frequent crown chips and fractures (Merbs 1968). Except for some teeth apparently broken off at the root during life (see below), there are few such traces.

In contrast to activities involving an entire region of the dental arch, we might imagine a craft activity affecting selected teeth only, such as pulling or rotating an abrasive cord or stick across a favoured tooth. As Blakely & Beck (1984) note, attrition due to such activities should be distinguishable from intentional modification based upon its distribution in the dentition, age distribution and characteristics such as symmetry. In the Italian case, loss is sometimes symmetrical. If teeth were lost in an activity, we might imagine women switching from one tooth to the controlateral one as the first grew worn or painful or was lost, leading to symmetrical loss. Even so, women would have had to forego such an activity once the selected teeth were destroyed; adjacent teeth are rarely missing. Furthermore, missing teeth are known in at least one young adult woman, excluding progressive degeneration as the sole cause of tooth loss. Most tellingly, any such activity involving abrasive contact should have left intermediate forms - anterior teeth worn or abraded but not yet destroyed, and traces on adjacent teeth. Neither is evident in the Neolithic dentitions.

Another possibility is that the teeth were intentionally removed to create an artificial gap to facilitate some activity - such as holding a cord or handle with the teeth, or passing a straw into the mouth while the teeth were clenched. If this were so, it might have left distinctive wear upon the surfaces of adjacent teeth where contact actually occurred, as in the case of people who habitually grip an object such as a pipe or wear a labret which rubs against their teeth (Milner & Larsen 1991). The variation in locations affected - there is no standard location for missing teeth - may also argue against tying tooth loss to a specific, and presumably standardized, habitual activity.

Anterior tooth loss and dental disease

Dental disease, the most common cause of tooth loss during life, leaves a characteristic pattern of tooth loss which may be compared against the Italian data:

1 Dental caries and ante-mortem tooth loss due to dental disease virtually always begin with the molars and proceed from the rear of the dentition to the front, reaching the anterior teeth only after all or most other teeth have been lost (Buikstra & Ubelaker 1994: 54; Frayer 1989: 258ff; Hillson 1986: 290ff; Powell 1988: 70-71, 120-21). In the Italian Neolithic sample, dental disease is common. By early middle age most individuals had lost at least a molar or two; older individuals often had several molars missing. Of the 8 women who lost anterior teeth, one (Lanciano) had already lost most other teeth, including all posterior teeth; in this case, anterior tooth loss may be completely or partially due to dental disease.(15) Excluding this case from tabulation changes the rate of loss for incisors and canines by little (5.2%); it is still significantly different from that for males. The other 7 women had most of their posterior teeth present at death; the greatest loss was two molars (Catignano), which is quite typical of tooth loss in middle-aged Neolithic people. There is little evidence for the extreme loss of the posterior teeth which almost always accompanies pathological loss of the anterior teeth.

2 Tooth loss due to disease is associated with other manifestations of dental disease, particularly carious lesions. Molar loss frequently occurred in conjunction with caries, both in the Neolithic sample as a whole and in several of the women who had lost anterior teeth in life. Yet in all the Neolithic specimens examined by the author, no carious lesions were observed on the canines and incisors, suggesting that these teeth were rarely afflicted by dental disease. Only one of the 8 women with anterior tooth loss (Lanciano, above) had any sign of apical abscesses or marked periodontal disease in the anterior arch. Several women who had lost anterior teeth had otherwise perfectly healthy dentitions with no signs of dental pathology. There is thus little which might be interpreted as intermediate stages in a disease process - anterior teeth diseased but not yet lost.

3 The distribution of teeth lost during life should mirror the distribution of dental disease among subgroups such as males and females. Dental disease, as measured by molar loss (TABLE 2) and caries rates (Robb 1995: table 3.56), occurred in Neolithic males and females to much the same extent. Both phenomena are slightly more prevalent in females; the difference is not significant, even given the larger sample size of the pooled-sex sample, and may be due to chance.(16) If dental disease caused anterior tooth loss, we would expect to find at least some anterior tooth loss in males. We would also expect to find sex differences in molar loss and caries, the most typical manifestations of dental disease, to an extent at least as marked as those in anterior tooth loss.

4 Tooth loss proceeds progressively, becoming more advanced in older people. Of the four criteria for excluding pathological causes of tooth loss, this is the most problematic, in part due to problems of age determination discussed above. Among age, dental disease and anterior tooth loss, the clearest relationship was between age and disease. While molar wear and molar loss have a complex relationship, wear scores for the second and third molars were highly correlated with rates of posterior tooth loss (Spearman's rank-order correlation coefficient -714 for M2 wear and .553 for M3 wear; p = .000 and .026 respectively).(17) There is also a weak but detectable tendency for ante-mortem tooth loss to be found in older women (see above), and between loss of anterior teeth and loss of posterior teeth: women who lost anterior teeth in life were slightly more likely to have lost at least one posterior tooth as well.(18)

The difficulty comes in establishing whether dental disease and anterior tooth loss both increase with age because the former caused the latter, or because they are parallel age-related processes (as would be dental disease and osteoarthritis, the accumulation of traumas or the accumulation of social knowledge). Using logistic regression as a predictive technique (Warren 1990), neither molar loss nor molar wear turned out to be a particularly strong predictor of anterior tooth loss. Principal components analysis of tooth loss rates for anterior teeth, premolars and molars consistently yielded two factors, the first with heavy loadings for premolar and molar loss, the second associated with anterior loss. Thus, while the relations between age, disease and anterior tooth loss are not entirely clear, there is probably at least as much ground for interpreting anterior tooth loss as an on-going process parallel to dental disease as for seeing it as manifestation of dental disease.

According to three of four criteria, anterior tooth loss is quite distinct from tooth loss due to dental disease; the fourth criterion is inconclusive. A reasonable interpretation is to resolve tooth loss in the Italian Neolithic sample into two patterns. The first is the pathological pattern of loss proceeding from back to front, accompanied by dental disease and found in both sexes. The second is the idiosyncratic loss of incisors and canines in females.

Discussion and conclusions

Anterior tooth loss in Neolithic Italian women is unlikely to have resulted from taphonomic, genetic, traumatic, activity-related or pathological causes; its primary cultural pattern is a marked sex distribution in favour of females. Ethnographically, many groups around the world are known to have practiced intentional dental ablation for cultural reasons (Milner & Larsen 1991), and this seems the most likely possibility for Neolithic Italian women.

Ante-mortem anterior tooth loss is known in samples from cultures spanning the entire Neolithic, including the Early Neolithic Impressed Ware, Middle Neolithic groups with various forms of painted wares and the Late and Final Neolithic Ripoli, Serra d'Alto and Diana-Bellavista groups. It shows equally little regional distinction: while the excavated remains, and known cases of ante-mortem tooth loss, are biased heavily towards Puglia, cases are known from both coastal and highland Abruzzo as well. As a cultural behaviour, it appears to have occurred in all the Neolithic groups for whom skeletal data are available.

Breakage patterns may provide some insight into how teeth were removed. Some tooth breakage in Neolithic samples is undoubtedly due to post-mortem destruction. But in several skulls, broken edges of the anterior tooth roots are rounded by wear, implying that the subject lived and chewed for some time after the tooth was broken. In one case (Madonna di Grottole, [ILLUSTRATION FOR FIGURE 4 OMITTED]; see Scattarella & De Lucia (1988) for archaeological context), broken anterior teeth with worn roots are known from a female skeleton which was excavated in an undisturbed context, further precluding the possibility that wear could be due to taphonomic causes. In female skulls, incisors and second premolars are often broken off, in a number of cases symmetrically. As it is difficult to distinguish pre-mortem from post-mortem breakage, no quantitative analysis was carried out. The similarity to the pattern of ante-mortem loss may imply that breakage and ante-mortem loss represent stages of a single process. If teeth were intentionally removed via direct force, torsion or pulling, the manoeuvre may have broken the crown off at the neck in some cases, followed in some cases by the loss of the root and resorption of the socket.

The immediate circumstances of the intentional, and presumably voluntary, extraction of front teeth are unclear. Initiation into adult status is ethnographically a common occasion for tooth removal. In Neolithic Italy, women both with and without tooth removal are found at several sites (e.g. Continenza, Passe di Corvo). This suggests that tooth ablation was not universally prescribed, as we might expect for a normal rite of passage, but was instead optional or contingent, possibly via a combination of social distinctions such as adult female status and initiation into a particular group or status.(19) Non-initiatory rites or ceremonies could have been involved; the ultimate intent could have been cosmetic, with the dental diastema associated with ideas of beauty. These possibilities are not mutually exclusive; a cosmetic operation could have formed part of a transition to adulthood, with its visible effect furnishing an aesthetic symbol of its completion. Alternatively, tooth removal could have been a ritualized gesture, for instance of mourning dead kin. This view may be supported by the fact that women with lost teeth were slightly older at death. It is also possible that teeth were removed early in adulthood and that women losing teeth lived longer for some other reason.

All societies modify and ornament the human body for symbolic reasons. Anterior tooth removal is only one of three kinds of human body modification known in Italian prehistory. Tattoos are known from the 'Ice Man' found on the Austrian border (Hopfel et al. 1992; Barfield 1994). Trepanations known from the Neolithic onward are much commoner in males (Germana & Fornaciari 1992). As with tattoos and trepanation, the overall significance of tooth removal clearly depended on its relation to symbolisms of the body as yet poorly understood, such as the tendency in Italian Neolithic groups to use the female body as a symbol. Whatever the symbolic meaning, the identity or relationship dental ablation was intended to mark must have been a lifelong one, to judge from the fact that an irreversible bodily operation was the chosen symbol.

Acknowledgements. I am grateful to F. Bartoli (Pisa), M. Bellati (Cambridge), S. Borgognini Tarli (Pisa), A. Canci (Pisa), F. Facchini (Bologna), R. Foley (Cambridge), R. Macchiarelli (Museo Pigorini, Rome), F. Mallegni (Pisa), G. Manzi (Rome), M. Mazzei (Manfredonia), S. Sublimi Saponetti (Bari) and V. Scattarella (Bari) for the opportunity to study skeletal collections in their care. I am also grateful to Debra Gold and to two anonymous reviewers for critical comments on the manuscript. Financial support was provided by the Wenner-Gren Foundation for Anthropological Research, the University of Michigan, and the Cotton Foundation for Mediterranean Archaeology.

1 Many physical anthropologists term intentional tooth modification 'dental mutilation', an unnecessarily negative term for a procedure which in contemporary societies has been voluntary and culturally significant.

2 In sometimes small samples, different sites do not appear to differ noticeably in their proportions of males and females, in frequencies of dental disease, or in other characteristics which might affect sex-related patterns in tooth loss.

3 Specimens were inspected from several collections for which detailed published descriptions exist (Fonteviva, Passo di Corvo, Cala Colombo and Lama dei Peligni). No discrepancies were found between the author's census of ante- and post-mortem tooth loss and published descriptions; this suggests that explicit published descriptions can furnish this data reliably.

4 Danby (1987: 207) also notes a case of symmetrical antemortem loss of the upper lateral incisors in an unsexed fragment from Passo di Corvo. This specimen, not found in re-examining the collection, was not included in the tabulated data.

5 Females lost 9.3% of all premolars before death, while males lost only 1.0%. As for anterior teeth, the difference is highly significant (p = .013). Moreover, in several cases, females with otherwise healthy teeth lost a symmetrical pair of second premolars. However, when the three women in the sample who display severe dental disease and premolar loss (Grotta di S. Nicola, Villa Badessa 1 and Lanciano) are excluded, the rate of premolar loss falls to 2.9% and the sex difference is no longer significant. Given this, premolars are excluded from this discussion of ante-mortem loss, though in some cases they may have formed part of the same pattern as the anterior teeth.

6 The data tested conform to the assumptions of the [[Chi].sup.2] test; in particular, the expected cell frequencies are well above the standard threshold of 5 in all cases.

7 Though less immediately obvious, it is probably slightly more accurate to compare only complete upper or lower dentitions rather than all specimens, since the remains vary widely in preservation. When this is done, the results are essentially identical (7/29 complete female dentitions and 0/25 complete male dentitions have lost anterior teeth; [[Chi].sup.2] = 6.9, p = .009; with Yates' correction for continuity [[Chi].sup.2] = 4.9, p = .026; Fisher's Exact Test p = -009).

8 Given a probability of 7/29 or .241 of missing an anterior tooth in either the upper or the lower law, the probability of an individual's missing at least one tooth in at least one jaw would have been 1-[(1-.241).sup.2] or 42.24%.

9 Given a probability of .0595 of a tooth's being lost, the probability that it was not lost would be .9405; the probability that none of 12 teeth would be lost would be (.9405)(12) or -479; and the probability that at least one tooth out of the complete dentition would be lost would be 1-.479 or .521. This estimate is likely to be slightly high, as it assumes that tooth loss is randomly distributed, while in some cases women lost multiple teeth.

10 For instance, skeletally aged specimens suggest that in Italian prehistoric: specimens, molar wear could reach its maximum possible extent - exposure of dentine on the entire occlusal surface - by the age of 40-50; survival beyond this age is not measurable through wear.

11 Statistical significance was tested using the Mann-Whitney U test comparing the average rank for each group, as is appropriate for ordinal categories of non-normally distributed data.

12 Parallel analyses using the Kolmogorov-Smirnov Z test, also appropriate for ordinal wear stage data, did not give significant results for any molar.

13 This section is based on specimens examined by the author, with occasional reference to Neolithic dentitions not included in this analysis (e.g. unsexed specimens}; published data generally do not include sufficient detail to be used here.

14 Congenital absence of teeth is also not known to be linked to sex.

15 Even so, this individual retained a greater proportion of her anterior teeth (with 4/6 still in place) than of her premolars (0/4 remaining) or molars (0/6 remaining).

16 This difference between the sexes in disease may be related to a slight difference in ages at death; males had slightly lower amounts of molar wear. The difference in wear scores is not marked and is not statistically significant.

17 First molar wear scores bore little correlation to dental disease, probably because the first molars wore down sooner and were often the first teeth to be lost.

18 The statistical details of this vary according to how the data is treated (e.g. with pathology rate treated as an interval, ordinal or categorical data), and varied significance tests never yield results more significant than .05. In the simplest summary, 6 of 21 women without anterior tooth loss had lost at least one molar, while 5 of 5 women with anterior tooth loss had also lost one or more molars; using Fisher's Exact Test, p = .055. These figures change slightly if we exclude the one woman whose anterior tooth loss probably is actually due to very advanced dental disease (see text and note 15).

19 Burial goods provide little clue to possible statuses. While one or two women with missing front teeth were deposited in special ways (notably Lanciano, deposited as an isolated skull beneath a house floor (Geniola & Mallegni 1975)), others such as Fonteviva I and II and Catignano were completely typical depositions in villages with no or few goods (Robb 1994a).

References

BARFIELD, L. H. 1994. The Iceman reviewed, Antiquity 68: 10-26.

BECKER, M. 1995. Etruscan gold dental appliances: female status loss in Etruria during the period of Romanization. Paper presented at the Society for American Archaeology meetings, Minneapolis (MN).

BLAKELY, R.L. & L.A. BECK. 1984. Tooth-tool use versus dental mutilation: a case study from the prehistoric South, Midcontinental Journal of Archaeology 9: 269-77.

BLALOCK, H. M. 1979. Social statistics. Revised 2nd edition. New York (NY): McGraw-Hill.

BUIKSTRA, J. & D. UBELAKER. 1994. Standards for data collection from human skeletal remains. Fayetteville (AR): Arkansas Archaeological Survey. Research series 44.

CAPITANIO, M. 1978. Resti scheletrici umani d'epoca neolitica rinvenuti nelle contrade Guadone e S. Rocco di San Severo (Foggia), Rassegna di Studi Dauni 1: 89-104.

COOK, D.C. 1981. Koniag Eskimo tooth ablation: was Hrdlicka right after all?, Current Anthropology 22: 159-63.

CORRAIN, C. & M. CAPITANIO. 1968. I resti scheletrici dei depositi neolitici di Maddalena di Muccia e di Ripabianca di Monterado nelle Marche, Rivista di Scienze Preistoriche 23: 223-44.

CORRUCCINI, R. & E. PACCIANI. 1989. 'Orthodontistry' and dental occlusion in Etruscans, Angle Orthodontist 59: 61-64.

DANBY, G. 1987. The skeletal material from Passo di Corvo, in G. Jones (ed.), Apulia: 206-7. London: Society of Antiquaries. Research report 44.

FRAYER, D. 1989. Oral pathologies in the European Upper Paleolithic and Mesolithic. in I. Herskovits (ed.). People and culture in change: 255-81. Oxford: British Archaeological Reports. International series S508.

GENIOLA, A. & F. MALLEGNI. 1975. Il calvario neolitico di Lanciano (Chieti): note paletnologiche e studio antropologico, Atti, Societa Toscana di Scienze Naturali 82: 237-54.

GERMANA, F. & G. FORNACIARI. 1992. Trapanazioni, craniotomie e traume cranici in Italia dalla preistoria all'eta moderna. Pisa: Giardini.

GERMANA, F., F. MALLEGNI, C. DE POMPEIS & D. RONCO. 1990. Il villaggio neolitico di Villa Badessa (Pescara): aspetti paletnologici, antropologici e paleopatologici, Atti, Societa Toscana di Scienze Naturali 97: 271-310.

GRAZIOSI, P. 1974. L'arte preistorica in Italia. Florence: Sansoni.

HILLSON, S. 1986. Teeth. Cambridge: Cambridge University Press.

HOPFEL, F., W. PLATZER & K. SPINDLER 1992. Der Mann im Eis. Innsbruck: Eigenverlag der Universitat Innsbruck.

HRDLICKA, A. 1940. Ritual ablation of front teeth in Siberia and America, Smithsonian Miscellaneous Collections 99: 1-37.

JACKSON, J. W. 1915. Dental mutilations in Neolithic human remains, Journal of Anatomy and Physiology 49: 72-9.

LUKACS, J. 1989. Dental paleopathology, in K. Kennedy & M. Iscan (ed.), Reconstruction of life from the skeleton: 261-86. New York (NY): Alan R. Liss.

MACCHIARELLI, R. & L. SALVADEI. 1987. Resti scheletri umani, in S. Cassano et al. (ed.), Coppa Nevigata e il suo territorio: testimonianze archeologiche dal VII al H millennia a. C.: 78-81. Rome: Quasar.

MALLEGNI, F. 1979. Su alcuni reperti scheletrici umani del villaggio neolitico del Gaudone di S. Severo (Foggia), Archivio per l'Antropologia e la Etnologia 109: 449-59.

MALLEGNI, F. & G. FORNACIARI. 1980. I resti scheletrici umani del villaggio neolitico del centro di Foggia, Archivio per l'Antropologia e la Etnologia 110: 461-96.

MANGILI, G. 1954. I reperti ossei della Grotta Patrizi (Sasso Furbara): il cranio trapanato. Rivista di Antropologia 41: 52-67.

MERBS, C. F. 1968. Anterior tooth loss in arctic populations, Southwestern Journal of Anthropology 24: 20-32.

MILNER, G. & C. LARSEN. 1991. Teeth as artifacts of human behavior: intentional mutilation and accidental modification, in M. Kelley & C. Larsen (ed.), Advances in dental anthropology: 357-78. New York (NY): Wiley-Liss.

MOSSO, A. 1910. La necropoli neolitica di Molfetta, Monumenti Antichi dei Lincei 19: 305-96.

PARENTI, R. 1957. Lo scheletro umano della 'Tomba del Cane' a Ripoli, Archivia per l'Antropologia e la Etnologia 87: 5-94.

POWELL, M.L. 1988. Status and health in prehistory: a case study of the Moundville chiefdom. Washington (DC): Smithsonian Institution Press.

REYNOLDS, H.T. 1977. Analysis of nominal data. Beverley Hills (CA): Sage.

ROBB, J. 1994a. Burial and social reproduction in the Peninsular ItaLian Neolithic, Journal of Mediterranean Archaeology 7: 29-75.

1994b. The Neolithic of peninsular Italy: anthropological synthesis and critique, Bullettino di Paletnologia Italiana 85: 189-214.

ROBB, J. 1995. From gender to class: inequalities in prehistoric Italy. Ann Arbor (MI): University Microfilms. Ph.D dissertation, University of Michigan, Ann Arbor.

ROBB. J. & F. MALLEGNI 1994. Anthropology and paleopathology of human remains from Catignano (Pescara, Italy), Rivista di Antropologia 72:197-224

SABATINI, A. 1932. Un crania eneolitico della Calabria, Rivista di Antropologia 29: 577-82.

SALVADEI, L. & R. MACCHIARELLI. 1983. Studi antropologici, in S. Cassano & A. Manfredini (ed.), Studi sul Neolitico del Tavoliere della Puglia: 253-64. Oxford: British Archaeological Reports. International series S160.

SCATTARELLA, V. & A. DE LUCIA. 1988. Su una duplice sepoltura neolitica a Polignano a Mare (Bari), Taras 8: 39-53.

SCOTT, E. 1979. Dental wear scoring technique, American Journal of Physical Anthropology 51: 213-18.

SCOTT, G. R. & C. TURNER. 1988. Dental anthropology, Annual Review of Anthropology 17: 99-126.

SHENNAN, S. 1988. Quantifying archaeology. Edinburgh: Edinburgh University Press.

SKEATES, R. 1994. A radiocarbon date-list for prehistoric Italy (c. 46,400 BP-2450 BP/400 cal. BC), in R. Skeates & R. Whitehouse (ed.). Radiocarbon dating and Italian prehistory: 147-288. London: Accordia Research Centre. Accordia Specialist Studies on Italy 3.

WARREN, R.E. 1990. Predictive modelling in archaeology: a primer, in K. Allen et al. (ed.), Interpreting space: GIS and archaeology: 90-111. London: Taylor & Francis.