9

THE POPULATION

Mounting evidence challenges the concept of ‘racial’ classification from skeletal remains and few scholars today would consider that there was any value in attempting to identify European ‘races’, as these almost certainly do not exist (Chapter 3). Nonetheless, there is still some information that can be gleaned about population affiliations from human skeletal remains.

It was generally assumed that the Pompeian population was heterogeneous, since the town was a river port with a long history. As discussed in Chapter 4, literary evidence has been invoked to support this notion, like Strabo’s description of the different groups that occupied Pompeii over time. Similarly, Pliny the Elder stated that Campania had been inhabited by Oscans, Greeks, Umbrians, Etruscans and Campanians.¹ The composition of the population would also have been affected by the colony of veteran Roman soldiers that was superimposed onto the population by Sulla as punishment for resisting Rome in the Italic War.² An explanation would have to be sought if the population were found to demonstrate some degree of homogeneity.

Various factors may have had an impact on the composition of the _AD 79 population, including partial abandonment of the settlement as a result of the _AD 62 earthquake and subsequent seismic activity in the final 17 years of occupation. It is possible that the sample of victims may not reflect theAD 79 population, depending on the time of the year of the eruption and whether it was possible for certain sections of the community to have more opportunity to escape prior to the lethal phase. The most recent evidence casts more than doubt on the generally accepted August date, which means that seasonal inhabitants would have returned to Rome after the summer (Chapter 4). The issue of Pompeian heterogeneity was examined by the collection of both metric and non-metric data, which were subjected to metric and non-metric analysis. While the contribution of DNA studies cannot be ruled out in the future, current research on the populations of the victims of the_AD 79 eruption relies only on macroscopic analysis.

Molecular biology has the potential to provide valuable insights into the population affinities of the victims of the _AD 79 eruption. DNA analysis has been attempted on samples from both Pompeii and Herculaneum, but the results to date have been disappointing. The high temperatures to which the bodies were exposed at the time of death have been invoked to explain the poor preservation of nucleic acids in samples from the Herculaneum skeletons. Human skeletal remains from Pompeii have also yielded limited information due to poor DNA preservation. Nonetheless, these preliminary studies indicate that, at least in some cases, there is sufficient endogenous DNA to enable amplification and analysis.³

Cranial evidence

The main material used for this research was the cranial collection stored in the Forum Baths, as it appeared to reflect a random sample of adults that tended to be more complete than the skulls housed in the Sarno Baths (Chapter 5). Population studies are usually confined to adult material as the results obtained from skulls where growth is not yet complete would be misleading.

Metric evidence

Population studies have traditionally been based on measurements of skulls (Chapter 3). It was appropriate to commence the study of the Pompeian sample in a similar fashion as the data collected could be compared with those published in the earlier work of Nicolucci well as Bisel’s measurements of the Herculaneum sample. Unfortunately, D’Amore et al. did not publish their raw data. Capasso only took minimal cranial measurements, enabling him to calculate a series of indices, which he considered to be most useful descriptors of the Herculaneum heads. Astonishingly, these included the now largely abandoned horizontal or cephalic index (Chapter 3). While he did present the indices for the sample, he did not include the raw data.⁴

Bisel restricted her study to adult male skulls⁵ and made 11 measurements on 50 skulls. These were compared with an early study of Howells, which was used to develop standards based on Irish monastery burials. Bisel calculated the cumulative standard deviation for her sample and found it to be greater than the norm for the Howells data. Bisel suggested that the considerable variability of the cranial metric data from the Herculaneum sample was a reflection of a heterogeneous population with the implied benefits of hybrid vigour which would have been ‘manifested in great energy and creativity’. While a large standard deviation does imply variability, recent work suggests that an isolated sample can also exhibit considerable variation. Howells, for example, demonstrated this in his later studies using Berg data, which represents a population that was geographically isolated over a number of generations.⁶ It is apparent that the standard deviation alone does not provide information about the composition of a population.

The fact that many of the Pompeian skulls were incomplete hampered the collection of cranial metric data. A series of 12 measurements were made on 117 adult male and female skulls. The analysis of these data was compared with similar analyses based on the raw craniometric data published by Nicolucci in 1882 to establish whether there was consistency between samples. The 12 cranial measurements were then compared with data collected by Howells from a variety of European and African populations, the Pompeian skeletal sample studied by Nicolucci and the data collected from the Herculaneum material by Bisel to gain some understanding of the Pompeian sample in relation to other populations.⁷

Metric evidence from the skull sample provided insuf ficient evidence to establish whether the Pompeian and Herculaneum samples reflect homogeneous or heterogeneous populations. Comparison with other samples from European and African contexts tended to confirm the European affinities of the sample. As expected, the data from the current Pompeian sample was closest to Nicolucci’s earlier sample and Bisel’s Herculaneum sample, though there were exceptions for some measurements.

It should be noted that a large proportion of the observed craniometric differences in the Pompeian sample appear to be intrapopulational rather than interpopulational and probably reflect variation between male and female skulls. Even though differences could be observed between the sexes in the Pompeian sample, the results of my study indicate that the skull, and the craniometric data in particular, do not provide a very useful sex indicator for the Pompeian skeletons (Chapter 6). By implication, these cranial measurements are of limited value for the determination of population affinities for this sample as they did not indicate any real separation into well-defined groups.

Non-metric studies

Non-metric traits are anomalous skeletal variants, which are generally nonpathological. On the whole, these present as innocuous features on the bone. It is unlikely that individuals would ever notice that they had such traits. They are mostly of interest to physical anthropologists as they are easily observed and counted. They are also known as epigenetic traits and occur with varying frequency in all populations. A study of the pattern of cranial and post cranial anomalies can provide information about population variability.

Because skeletal inheritance is multifactorial, the genetic and environmental components of non-metric traits cannot easily be distinguished. Human and mouse pedigree studies have established a genetic component for a number of traits, though a genetic basis is not essential for a non-metric trait to be a useful population descriptor. The acquisition of traits as a result of shared environmental factors, especially during the period of growth and development, can also reveal information about a population. The potential of epigenetic traits as population descriptors is supported by the consistency of the results from a large scale study of non-metric cranial traits for a number of populations with genetic and other morphological studies that have been made to establish population distance.⁸ It should be borne in mind that there is no reason to assume regional immutability over time, especially for traits that have an environmental component.

Figure 9.1 Facial view of skull, showing some of the non-metric traits that were observed in the Pompeian skeletal sample (adapted from Comas, 1960, in Krogman, 1962, 316, and Brothwell, 1981, 94)

Figure 9.2 Lateral view of skull showing some of the non-metric traits observed in the Pompeian skeletal sample (adapted from Comas, in Krogman, 1962, 317, Brothwell, 1981, 94)

Twenty-eight cranial non-metric traits were scored on 126 skulls in the Forum Bath collection.⁹ Standard definitions for scoring traits were provided by Hauser and De Stefano.¹⁰ Because of the comparatively low retention of maxillae and mandibles, and the high rate of post-mortem tooth loss as a result of poor storage techniques, only one dental non-metric trait was scored. This was the presence or absence of double-rooted canines.

Intra and interobserver variation in scoring traits has been well documented.¹¹ Interobserver error was not an issue for this study as I was the sole person scoring the Pompeian skeletal sample. Intraobserver error was addressed by rescoring traits on the same bones over a period of years without reference to previous studies. An extremely high degree of concordance was found between scores taken at different times over a five-year period.

Where possible, comparison was made with the results from observations made by other scholars who have worked on the _AD 79 victims. The only available published material was produced by Nicolucci and Capasso.¹² Nicolucci scored the traits he observed from his sample of 100 skulls. It is difficult to establish the exact sample size that Capasso employed for the non-metric traits that he scored. He claimed to have based his study on the adult sub-sample of the Herculaneum skeletal collection that was available to him but, when the percentages he presents are scrutinized, it is apparent that his sample was 159, which meant that he also included sub-adult material.¹³ It is notable that Capasso used the same standard scoring scheme that was used for the Pompeian study.¹⁴

The practice of cremation as the primary method of disposal of the dead in the Roman world in the first century_AD is an impediment to obtaining appropriate comparative material. As a result, a number of the other skeletal samples used for comparison with the Pompeian material were temporally and geographically different. The cranial non-metric data collected from the excavations at the medieval monastery at San Vincenzo at Volturno provided comparative material from another Central Southern Italian site. Two groups of skeletons were unearthed at this site: the first were from the late Roman period and the second from the early medieval period. These two groups of skeletons have been interpreted as 84 workers from a large villa estate of the fifth century AD and 69 lay workers from the monastery. The latter set of burials comprised individuals from family groups of tenants who worked the monastic land. The monastery was in use from the eighth to the end of the ninth century AD.¹⁵ Further comparative Italian material was obtained from a study of frequencies of nine non-metric traits for ten skeletal series from central Italy that dated from the ninth to the fifth centuries BC. Two of these series were from Campania; one from Sala Consilina, dating between the ninth and the sixth centuries BC and one from Pontecagnano, dating between the seventh and sixth centuries BC.¹⁶ Other Italian samples included: a presumably homogeneous Iron Age sample from Alfedena in the Abruzzo, which dated from 500–400_BC, an early twentieth-century collection from the University ‘La Sapienza’ in Rome, a cranial sample from East Sicily, dating from the second and first millennia _BC, a number of Iron Age skeletal samples, dating from the first millennium _BC, from either side of the Apennine mountains, including three samples from the area to the south of Naples, a Sardinian sample of adult males and an Etruscan population from Tarquinia.¹⁷

Additional cranial material that could be used for comparison with the Pompeian sample was less satisfactory and included an assortment of European populations, which dated from the prehistoric to the modern period. In addition, where data were available, a sample of Ancient Egyptian skulls, prehistoric Africans from Mali and an historic Nubian sample were compared with the Pompeian material as an acknowledgement of the possibility of Pompeian contact with Africa.¹⁸ As there was very little published data for the population incidence of double-rooted canines and post-cranial nonmetric traits, comparisons were made with all the available material.

A major consideration in relation to comparison of traits with other published material is the lack of standardization in the presentation of trait incidence. Scholars have not always described the techniques that they used to record incidence. There are further problems when making comparisons between data sets for different populations, which include differences in sample sizes and whether they reflect random samples or are representative of the populations from which they were drawn. Apart from the acknowledgement that they exist, it is very difficult to account for these problems.¹⁹ The samples of victims of the _AD 79 eruption are particularly problematic as comparative material since they reflect a sample of disaster victims, which is unusual in the archaeological record. The fact that they share roughly the same date and cause of death means that they provide more of a snapshot than a very slow time lapse view of a population. Statistical analyses were used to identify whether there was any association between the traits, with sex, or sides for bilateral features.²⁰

The results for the majority of these traits were inconclusive. This is partly attributable to the lack of appropriate comparative material, though investing meaning in the patterns of expression of non-metric traits is problematic (see below). The non-metric traits that yielded the most interesting results in relation to the issue of heterogeneity in the Pompeian cranial sample were palatine torus and double-rooted canines. While perhaps not as significant as population indicators for the Pompeian sample, it is also worth considering some of the other Pompeian cranial traits that can be compared with the data obtained by Nicolucci and Capasso. These include the expression of the metopic suture, wormian or fontanelle bones and inca bones and frontal grooves.

Palatine torus

This is a hyperostotic trait, which means that it is associated with excessive ossification into structures that are usually made up of cartilage or dura.²¹ Palatine torii can be observed on the palate as a median, or more frequently, paramedian bony mound, varying in height, width and length. They tend to be bilateral but can be unilateral. There can be a considerable range of expression of this trait from very slight and more pronounced at either the anterior or posterior end of the bone to excessive and covering the entire length of the palate. Researchers disagree about the relationship of age and expression of palatine torus and there are huge discrepancies between reports of first appearance and when development ceases. Similarly, there is no consensus about its aetiology. Some scholars have suggested that this is a threshold trait, which means that expression is dependent on environmental factors reaching a particular level. Stress from mastication has been invoked as a contributing factor to the development of palatine torus. Higher frequencies have been observed in circumpolar populations, whose diets include frozen and dried meat, which require hard chewing. It has also been suggested that nutrition may have some impact on the appearance of this trait. The consumption of marine food, with elevated levels of omega-3 and omega-6 fatty acids as well as vitamin D, has been proposed as major contributor to the development of palatine torus. Circumpolar populations have again been cited to support this notion as they tend to rely on a marine diet. Familial studies indicate that there appears to be a genetic potential for the expression of palatine torus. Observed occurrence of this trait in monozygotic twins further supports a genetic element, though it has been suggested that its presence can be influenced by other factors, such as pathology. It is most likely that palatine torus results from the interaction between genetic and environmental components.²²

Some degree of expression for palatine torus was present on all but two cases of the 52 skulls in the Forum Bath collection where the palate had been preserved. A similar frequency was observed in the Sarno Bath collection, though the survival rate of the palate was much lower in this collection. Contrary to the literature on palatine torus,²³ this trait was not found to be associated with either sex or age in the Pompeian sample. The Pompeian incidence for palatine torus is 96.2 per cent, which is extremely high compared to that recorded for most other populations (Table 9.1). Only three cases of palatine torus were reported for the Herculaneum sample studied by Capasso.²⁴ He did not mention the sample size for his observations of this trait, but based on the frequency calculations of the other epigenetic features he documented, this apparently represents 1.9 per cent of his sample.

It is notable that high frequencies have been reported for this trait in populations that are very genetically distant from the Pompeian sample. For example, Pardoe reported an incidence of palatine torus of up to 74.5 per cent for Australian Aboriginal samples from the border area between New South Wales and Victoria.²⁵ Even higher frequencies have been recorded for archaeological samples of circumpolar populations (Table 9.2).

The frequency of palatine torus in the Pompeian sample is extraordinary in comparison with the majority of the available data for Italian and other Mediterranean populations, especially that from Herculaneum, and requires some investigation. It is important to note that the case of 100 per cent expression from Termoli is based on a sample size of only nine individuals. It could be argued that the high frequency is an artefact of the comparatively low survival rate of this portion of the skull. There is, however, no reason to assume that preservation is not random. Even if it were strongly skewed toward individuals with this trait, the frequency would still be higher than that for most of the comparative material. In addition, observations of palatine torus for the available Sarno Bath palate sample, which were not included in this study, were consistent with the documented frequency for the Forum Bath collection.

Consideration should also be given to scoring. Berry and Berry claimed that there was a discrepancy between observers in scoring this trait as the results obtained varied considerably from other published data. This led them to postulate the existence of two separate entities that could be scored as palatine torus.²⁶ This problem should have been solved by the publication of Hauser and De Stefano’s standard atlas in 1989, which includes photographic references to minimize ambiguity for each trait. Another source of possible over-identification of this trait in comparison to the results of other scholars is the inclusion of trace scores. Some observers score trace expression

Figure 9.3 Basilar view of skull, showing a strongly expressed palatine torus on the bony palate (cf. Fig. A3.4 where this trait is absent) (adapted from Comas, 1960, in Krogman, 1962, 318, and Brothwell, 1981, 94; Hauser and Stefano, 1989, 175–7) Table 9.1 Presence of palatine torus in various populations

Population Sample size Frequency (%)

Pompeii AD 79 52 96.2
Pompeii AD 79 (trace cases excluded) 52 73.1
Herculaneum AD 79 159 1.9
Pontecagnano (Campania) (7^th–6^th century BC) 32 21.9
Sala Consilina (Campania) (9^th–6^th century BC)9 0 Termoli (Molise) (7^th century BC) 9 100 Ardea (Latium) (8^th– 6^th century BC)190 Romans (Latium) (6^th–5^th century BC) 167 14.9
Alfedena (Abruzzo) (6^th century BC) 69 34.8
Campovalano (Abruzzo) (7^th–6^th century BC) 95 32.7
Perdasdefogu (Sardinia) (9^th century BC)17 0 Etruscans 1 (Central Etruria) (6^th–5^th century BC)56 0 Etruscans 2 (Southern Etruria) (6^th–5^th century BC)84 0 San Vincenzo al Volturno 153 23.7
Cefalu (17^th century BC) 13 7.7
Plemmyrion (16^th–14^th century BC)333 Castiglione (17^th century BC) 11 18.2
Castiglione (8^th– 6^th century BC) 6 33.3
Thapsos (16^th–14^th century BC) 41 2.4
Lentini (5^th– 4^th century BC) 6 33.3
Siracusa (8^th century BC) 29 6.9
Siracusa (3^rd century BC) 109 11.9
Piscitello (5^th – 4^th century BC)270 Carlentini (5^th– 4^th century _BC)110 Modern Roman sample 285 7.4
Undated Sardinian population 245 7.3
African sample (Mali) (1^st millennium BC) 145 0 Nubian (historic) 33 3

Sources: Adapted from Capasso, 2001, 982; Hauser and De Stefano, 1989, 178 –79; V. Higgins (University of Notre Dame, Rome) to E. Lazer, 1989–1990, personal communication; Lazer, 1995, 297; Rubini et al., 1999, 10; Rubini et al., 2007, 124.

Table 9.2 Frequency of palatine torus from Scandinavian archaeological samples Population Sample size Frequency (%)

Medieval Norway (male)
Medieval Norway (female)
Medieval Iceland (male)
Medieval Iceland (female)
Eastern Early Greenland (male)
Eastern Early Greenland (female) Eastern Middle–Late Greenland (male) Eastern Middle–Late Greenland (female) Western Greenland (male)
Western Greenland (female)
48 70.8 50 90.0 20 75.0 34 91.2 20 80.0 11 90.9 11 90.9 17 100

7 100 27 100 Source: Adapted from Halffman et al., 1992, 151.

as absent.²⁷ Trace expression only accounts for 12 cases or 23.1 per cent of the Pompeian sample. Removal of all cases with a trace score still leaves an unequivocal 73.1 per cent with palatine torus. Since Capasso used the same standard scoring system as the Pompeii study, the lower frequency reported for the Herculanuem sample cannot be attributed to differences in recording. Nonetheless, it would be valuable for other ancient and contemporary skulls from the Vesuvian region to be examined for palatine torus to establish if this is a feature that is specific to ancient Pompeii or whether there are other populations in the region where the incidence of expression is high.

It does appear that the high frequency of this trait in the Pompeian sample is not an artefact. As already mentioned, the aetiology of this trait is not well understood. It appears that both environmental and genetic components contribute to the expression of palatine torus. Whatever the mechanism for the formation of the palatine torus, it has the potential to be a useful population descriptor for ancient Pompeians. It could be argued that the almost total presence of the trait in the sample might suggest a type of homogeneity that was not necessarily based on similarity of genotype but perhaps a shared environment during the period of osseous development. Dietary factors should be considered, though it would be hard to explain why there is such a low prevalence of this trait in the nearby settlement of Herculaneum, which has been argued to have relied heavily on marine protein (Chapter 8).

Double-rooted canines

The roots of canines in the mandible or lower jaw are occasionally divided into two parts: labial (facing the lips) and lingual (facing toward the tongue). The degree of division can vary and be either partial or complete. It is most uncommon to find a bifurcated root on an upper or maxillary canine.²⁸ This characteristic can be a useful population marker. During the course of research, loose canines were routinely removed from their sockets to facilitate measurement and a number of double-rooted canines were observed. Observations were also made of the sockets of canines that had been lost postmortem. This is only useful for the identification of this trait when the roots are well divided. Scoring was limited to unequivocal cases. Only mandibular occurrence of this trait was observed by these means. Six of the 21 mandibles from which it was possible to make observations had teeth with roots that were divided. It is perhaps misleading to use percentages for such a small sample size, but for the purposes of comparison, the prevalence of doublerooted canines was about 28.6 per cent.

There is minimal comparative data for this trait in the literature. Maxillary canines of 13 Central Southern Italian Iron Age populations from either side of the Apennine Mountains and dating from the ninth to the second centuries_BC were examined for the presence of bifurcated roots. Not surprisingly, no cases were observed in the 1,114 individuals, which included three Campanian samples from the region to the south of Naples (Coppa et al. 1998: 375). The mandibular canines were not examined for this trait. Scott and Turner observed double-rooted canines in varying frequencies in a diverse sample of populations. Table 9.3 demonstrates that this trait is extremely rare in Asiatic, Oceanic and African populations. Turner found double-rooted canines to occur more frequently in European populations.²⁹

Because of the small sample size and the lack of appropriate comparative data, it is not reasonable to draw too many conclusions from the presence of double-rooted canines in the Pompeian sample. The frequency appears to be considerably higher for Pompeians than for any other recorded population. This finding does appear to be remarkable and should be investigated further. The entire Pompeian skeletal sample should be subject to more detailed examination, with x-ray analysis of the jaws which still had teeth in-situ that could not be removed for inspection, as well as the cast collection. The Herculaneum skeletal collection should also be inspected for this trait, which, as yet, has not been recorded in the sample, as well as other ancient and modern samples from Campania and the rest of Italy to establish whether this feature is unique to the site of Pompeii or whether it has a high regional occurrence.

Table 9.3 Frequencies of mandibular double-rooted canines in various regions Region Sample size Frequency (%)

Pompeii AD 79 21 28.6 Western Europe 314 0.057 Northern Europe 214 0.061 North Africa 347 0.023 West Africa 33 0.00 South Africa 192 0.00 Khoisan 14 0.00 China–Mongolia 401 0.00 Jomon 203 0.010 Recent Japan 335 0.012 Northeast Siberia 206 0.00

Sources: Adapted from Scott and Turner, 1997, 322 and Lazer, 1995, 314–15.

Metopic suture

Metopism is a hypostotic trait. Hypostotic traits involve the retention of forms that are usually visible in the embryonic or early infant state, such as metopic sutures or inca bones. They involve arrested development or incomplete ossification. At birth, the frontal bone is made up of two halves separated by a suture. This suture functions as an area where growth can occur and also enables the two frontal halves to move relative to each other during birth. If this suture persists after the first two to three years of life, it is scored as present. Studies of mice and macaques as well as x-ray examination of families suggest that there is a genetic component to metopism. It has been suggested that this trait occurs with a higher frequency in females.³⁰

Observations for metopic suture were made on 121 Pompeian skulls. Twelve cases or 9.9 per cent of the sample retained complete metopic sutures, 13.2 per cent of the sample exhibited some degree of metopism. One case displayed partial persistence at the parietal end of the suture and there were three cases of partial persistence at both ends of the suture. It is notable that Nicolucci reported a frequency of 11 per cent in the Pompeian sample he studied; a figure he considered to be high. Capasso recorded seven cases in the Herculaneum sample which reflected a frequency of 4.4 per cent in the sample he examined. Six of the cases were complete and one was an incomplete superior metopic suture. The frequencies observed in Pompeii and Herculaneum can be compared with those found in other populations (Table 9.4). Much of the comparative material was scored using the method of Berry and Berry but their definition is not clear about whether partial persistence would be scored as present.³¹ If one assumes that only total persistence was scored then the frequency of metopism in the Pompeian sample would be only 9.9 per cent.

The results of this study are relatively consistent with those of Nicolucci. The Pompeian incidence of this trait sits within the range reported for other Italian populations and is higher than the reported African samples. Possibly the most interesting observation is the fact that its frequency is more than double that of the Herculaneum sample, regardless of whether partial presence is included.

Inca bone

In the foetus, the inferior and superior squama of the occipital bone are separated by a suture that runs from asterion to asterion. This suture generally closes prior to birth, but if it persists into adult life it is classified as an inca bone. If the foetal sutures on the superior squama fail to unite, bipartite, tripartite or inca bones divided into four parts can result.³² It is important to note that inca bones were treated as a separate entity from ossicle at lambda in this study as the latter tend to be fontanelle bones and result from a different cause (see below). It is possible to distinguish inca bone variants from sutural bones by morphological features but there is considerable ambiguity in recording in the literature.³³

Six cases of inca bone variants were observed in a sample of 116 Pompeian skulls, which means that this trait occurred with a frequency of 5.2 per cent. Only one case, or 0.9 per cent of the sample, exhibited full expression of this trait. Nicolucci observed one case of an inca bone on a skull he interpreted as female in his sample of 100 Pompeian skulls.³⁴ In most populations this trait has been observed more frequently in males than females,³⁵ though no relationship between this trait and gender could be ascertained for the

Table 9.4 Frequency of metopic suture in various populations Population

Pompeii AD 79 (Lazer 1995)

Pompeii AD 79 (Lazer 1995)

Pompeii AD 79 (Nicolucci 1882)
Herculaneum AD 79
Pontecagnano (Campania) (7^th– 6^th century AD) Sala Consilina (Campania) (9^th– 6^th century BC) Termoli (Molise) (7^th century BC)5016 Ardea (Latium) (8^th– 6^th century BC)170 Romans (Latium) (6^th– 5^th century BC) 153 0.6 Alfedena (Abruzzo) (6^th century BC) 64 17.2 Campovalano (Abruzzo) (7^th– 6^th century _BC) 62 9.7 Perdasdefogu (Sardinia) (9^th century _BC) 17 29.4 Etruscans 1 (Central Etruria) (6^th– 5^th century BC) 35 5.7 Etruscans 2 (Southern Etruria) (6^th– 5^th century BC)55 0 San Vincenzo al Volturno 153 7.2 Cefalu (17^th century BC)260 Plemmyrion (16^th–14^th century BC)605 Castiglione (17^th century BC) 52 5.8 Castiglione (8^th–6^th century BC) 52 19.2 Thapsos (16^th –14^th centuryBC) 48 6.25 Lentini (5^th– 4^th century BC)70 Siracusa (8^th century BC)290 Siracusa (3^rd century BC) 121 16.5 Piscitello (5^th–4^th century BC) 28 3.6 Carlentini (5^th–4^th century BC)200 Modern Roman sample 300 10.7 Undated Sardinian population 260 8.1 African sample (Mali) (1^st millennium BC) 156 3.8 Nubian (historic) 67 0

Sample size Frequency (%)

121 13.2
121 9.9 (complete expression only)
100 11
159 4.4
38 10.5
17 23.5

Sources: Adapted from Capasso, 2001, 982; Hauser and De Stefano, 1989, 42–43; Higgins, 1989–1990; Lazer, 1995, 293; Nicolucci, 1882: 10; Rubini et al., 2007, 124.

Pompeian sample. This is in no small part due to the dif ficulty in establishing sex from the skulls in this sample. As a result of this problem, there was usually no attempt to establish sex association for cranial non-metric traits in the disarticulated Pompeian sample.

Capasso recorded five cases of inca bone variants in the Herculaneum crania. He is not clear about the size of the sample upon which this study was based. Based on the frequencies he documented for other cranial nonmetric traits, he apparently examined 159 individuals, which means that the incidence of inca bones was 3.3 per cent. He observed that one case was a rare bipartite inca bone variant.³⁶

It was extremely dif ficult to obtain appropriate comparative material for this trait (Table 9.5), which is testimony to its rarity. Non-metric studies of other Italian populations tended to only record ossicle at lambda, which may include inca bone variants, though the scoring system that was employed suggests that only fontanelle bones were recorded.³⁷

Comparison of this trait with other samples is complicated by inconsistencies of scoring between scholars. The comparative data presented by Hauser and De Stefano gives a range of 3.7 per cent to 18 per cent for various European populations dating from the 1st–2nd millennium_BC to the medieval period. Interpretation of these data, however, requires some consideration of the definitions for inca bone for each sample. For example, a medieval French sample was recorded as having an incidence of 11.6 per cent based on the definition of Berry and Berry, which is very vague and does not distinguish between the ossicle at lambda, the preinterparietal bone, an inca bone variant that should be scored separately, and the inca bone. Similarly, the definitions for the medieval Bohemians and Alamannes also pooled these three variants. If, for comparative purposes, the preinterparietal bones were pooled with the inca bones in the Pompeian sample, the incidence would increase to 6.9 per cent. Inclusion of the ossicle at lambda as well would raise the Pompeian incidence to 25 per cent. Only inca bone variants were scored for prehistoric Ukrainians and first to second millennium_BC Lithuanians. The European and North African male and female samples and the Italian sample were all recorded in a consistent manner with the Pompeian sample.³⁸ The Pompeian and Herculaneum incidence is higher than the Italian and European frequencies recorded for these samples.

Table 9.5 Frequency of inca bones in various populations Population Sample size Frequency(%)

Pompeii AD 79 (Lazer 1995) 116 5.2 Pompeii _AD 79 (Nicolucci 1882) 100 1 Herculaneum AD 79 159 (?) 3.1 (?) Italy (Frosinone, Rome, Sicily, Otranto, 202 1.5

Abruzzo, recent soldiers)
European male 651 1.8
European female 176 1.1
North African male 537 3.2
North African female 345 2.0
Medieval French 69 11.6
Bohemian (8^th–10^th century AD) 555 18 Alamannes (6^th– 8^th century AD) 265 12.1
Lithuanian (1^st– 2^nd century BC) 2292 3.7
Prehistoric Ukrainian 153 4.6
Nigerian undated 40 15

Sources: Adapted from Capasso, 2001, 982; Higgins, 1989–1990; Hanihara and Ishida, 2001a, 141–43; Hauser and De Stefano, 1989, 102–103; Lazer, 1995, 305; Nicolucci, 1882, 10.

Wormian and fontanelle bones

Wormian bones are sutural bones or ossicles in the cranial vault. An ossicle is defined as any bone completely surrounded by a suture. They can be found, for example, in the coronal, sagittal and lambdoid sutures. Sutures permit very slight movement of the skull bones during birth to facilitate delivery. They also function as areas where postnatal bone growth can occur and contribute to the final shape and size of the skull. The embryology associated with the development of ossicles and fontanelle bones was discussed by Hauser and De Stefano. There is no satisfactory explanation for the function of ossicles, though there has been considerable discussion as to whether stress or pathology contributes to their presence.³⁹

The practice of cranial deformation is apparently correlated with an increased frequency of lambdoid ossicles, as well as frequency changes in a range of other non-metric traits. It has been argued that this has little impact on the determination of biological distance between populations. El-Najjar and Dawson examined a sample of American Indian skulls with particular reference to the number of ossicles per side. They observed wormian bones on skulls that had not been subjected to cranial deformation but noted that in the cases of asymmetrical deformation, the number of wormian bones was higher on the side of deformation. There also was a positive correlation between an increased number of lambdoid wormian bones and the pressure associated with deformation. No significant side differences were observed in undeformed crania. It was concluded that cranial deformation could influence the development of lambdoid sutural bones but was not the sole factor that determined their presence. Ossicles are often found in undeformed skulls. In addition, skulls that have been intentionally deformed do not always contain wormian bones.⁴⁰

There is an undoubted association between ossicles and the presence of some pathologies; for example they are almost invariably present and numerous in skulls of hydrocephalous individuals. It has been suggested that there is a link between the size, number and configuration of wormian bones and the presence of specific disorders. Similarly, this can be the case where there is delayed closure of sutures or fontanelle bones. A radiographic study of 81 skulls of individuals with osteogenesis imperfecta yielded a strong correlation between large numbers of wormian bones of a so-called ‘significant’ size, which are defined as more than ten bones of at least 6 mm by 4 mm which were arranged in a mosaic pattern.⁴¹

‘ Significant’ numbers of wormian bones have also been found in relation to a host of other disorders including cretinism, familial osteoarthropathy, kinky-hair syndrome and hypogonadism. Infantile-type osteoporosis, Down syndrome and rickets have been found in association with large sutural bones that do not necessarily occur in a particular pattern or number. It has been claimed that the presence of wormian bones is the result of developmental malfunctions, possibly with a genetic component. Cremin et al. suggested that the occurrence of more than ten ‘significant’ wormian bones may reflect an underlying environmental or genetic problem that has affected skull growth in the early developmental stages. Pedigree studies in humans suggest that wormian bone expression has a genetic component. Mouse studies appear to confirm that this trait is subjected to normal population variability.⁴²

Fontanelle bones can be found at the bregma, lambda and the asterion. There can also be fontanelle bones in the anterior lateral fontanelles. It has been suggested that the purpose of ossicles at bregma is to protect the brain in late foetal and early natal life. No similar explanation has been advanced to explain the presence of other fontanelle bones.⁴³

Ossicles are age-related in that sutures tend to be obliterated with advancing years.⁴⁴ When scoring skulls with partially fused sutures in the Pompeian sample, ossicles were only counted as present when it was certain that they could not be artefacts of other phenomena, such as complex suture patterns.

Pardoe found a positive correlation between the six wormian bone traits he used in his study of Australian Aboriginal skeletal populations, from which he concluded that the presence of one sutural bone on a skull would increase the probability of there being further sutural bones on that skull. This correlation has also been reported by other scholars. A correlation has also been observed between the presence of lambdoid ossicles and inca bones. Sutural bone intercorrelation has also been reported by other scholars.⁴⁵

The choice of which sutural bones would be studied was based on which bones could be identified without ambiguity and which were more likely to have survived, in order to provide the largest possible sample size. Squamo-parietal wormian bones were excluded as they were difficult to differentiate with certainty from post mortem damage. Additionally, the squamous portion of the temporal bone often had ‘sprung’, thus diminishing the possible number of observations.⁴⁶

OSSICLE AT LAMBDA

The ossicle at lambda was present in 20.2 per cent of the 116 Pompeian skulls that were scored for this trait. More than half of the cases, about 11.4 per cent, involved single or multiple large ossicles. Two of these cases were classified as interparietal bones. Capasso did not report any cases of ossicle at lambda in his sample.⁴⁷ The observed frequency of this trait in the Pompeian crania is within the upper end of the range of the Italian and African populations that have been recorded and lower than some of the second millennium Sicilian populations (Table 9.6). Most notable is the absence of the trait in the Herculaneum sample.

LAMBDOID OSSICLES

Lambdoid ossicles were observed more frequently on the right than the left side in the 112 observations that could be made for each side of the Pompeian cranial sample. They were scored as present to some degree in 34.8 per cent of cases for the left and 39.3 per cent for the right.⁴⁸ Capasso recorded 22 cases or 13.8 per cent cranial incidence in the Herculaneum sample, though he did not score them in any further detail than presence or absence. He reported an even division between the sexes. Nicolucci only observed eight cases of lambdoid ossicles in his sample, involving four males and four females.⁴⁹

It is important to distinguish between cranial and side index for this bilaterally expressed trait. Tables 9.7 and 9.8 show the side and cranial frequencies that have been recorded for this trait for different populations. To calculate side incidence for bilateral traits, some scholars⁵⁰ pool the number of observations for both sides, which means that these can represent a figure greater than the number of crania that were examined. This is important to bear in mind when examining Table 9.7 as the column labelled sample size,

Table 9.6 Presence of ossicle at lambda in various populations Population Sample size Frequency (%)

Pompeii AD 79 116 20.2 Herculaneum AD 79 159 0 Pontecagnano (Campania) (7^th– 6^th century BC)42 19 Sala Consilina (Campania) (9^th– 6^th century BC)16 25 Termoli (Molise) (7^th century BC)5010 Ardea (Latium) (8^th– 6^th century BC) 17 5.9 Romans (Latium) (6^th–5^th century BC) 153 22.2 Alfedena (Abruzzo) (6^th century BC) 83 19.2 Campovalano (Abruzzo) (7^th– 6^th century BC) 40 22.5 Perdasdefogu (Sardinia) (9^th century BC) 17 11.8 Etruscans 1 (Central Etruria) (6^th–5^th century BC) 35 22.8 Etruscans 2 (Southern Etruria) (6^th–5^th century BC) 55 18.1 San Vincenzo al Volturno 153 18.6 Cefalu (17^th century BC) 12 25 Plemmyrion (16^th–14^th century BC)5034 Castiglione (17^thcentury BC) 42 33.3 Castiglione (8^th– 6^th century BC)4030 Thapsos (16^th –14^th century BC)502 Lentini (5^th– 4^th century BC)70 Siracusa (8^th century BC) 29 13.8 Siracusa (3^rd century BC) 131 13 Piscitello (5^th– 4^th century BC) 29 3.4 Carlentini (5^th– 4^th century BC)220 Modern Roman sample 255 25.1 Undated Sardinian population 238 14.3 African sample (Mali) (1^st millennium BC) 154 20 Nubian (historic) 67 10.4

Sources: Adapted from Capasso, 2001, 982–83; Hauser and De Stefano, 1989, 178–79; Higgins, 1989–1990; Lazer, 1995, 294; Rubini et al., 1999, 10; Rubini et al., 2007, 124.

as for all the other non-metric data presented in this chapter, refers to the total number of observations upon which scoring was based.

Lambdoid ossicles, like other sutural bones, have been found to be correlated with the presence of other wormian bones. The only statistically significant correlation for the Pompeian sample was between the left lambdoid ossicle and the ossicle at lambda.

Comparison between the different samples is complicated by the way that the material is presented with some scholars only calculating cranial frequency and others just side incidence. The majority of Italian and other populations presented in Table 9.7 have a higher side incidence of lambdoid ossicles than the Pompeian sample, with only three other populations displaying a similar frequency, and two a lower incidence. The cranial incidence for this trait in the Pompeian sample is lower than that for other Italian populations documented in Table 9.8, though it is substantially higher than that recorded for the Herculaneum sample. It is difficult to account for the huge discrepancy with the very low incidence recorded by Nicolucci.

Table 9.7 Side incidence of lambdoid ossicles in various populations Population

Pompeii AD 79 (Lazer 1995)
Pontecagnano (Campania) (7^th – 6^th century BC) Sala Consilina (Campania) (9^th– 6^th century BC) Termoli (Molise) (7^th century BC)
Ardea (Latium) (8^th – 6^th century BC)
Romans (Latium) (6^th–5^th century BC)
Alfedena (Abruzzo) (6^th century BC)
Campovalano (Abruzzo) (7^th– 6^th century BC) Perdasdefogu (Sardinia) (9^th century BC)
Etruscans 1 (Central Etruria) (6^th–5^th century BC) Etruscans 2 (Southern Etruria) (6^th–5^th century BC) San Vincenzo al Volturno
Modern Roman sample
Undated Sardinian population
African sample (Mali) (1^st millennium BC) Nubian (historic)

Number of Frequency (%) observations

224 37.1
78 56.4
29 72.3
100 24
34 35.3
306 49.3
169 32
72 37.5
32 56.2
70 51.4
110 52.7
153 55.8
516 58.3
220 67.3
296 46.3
134 25.4

Sources: Adapted from Hauser and De Stefano, 1989, 178–79; Higgins, 1991; Lazer, 1995, 294; Rubini et al., 1999, 10; Rubini, et al., 2007, 124.

CORONAL OSSICLES

Of the 111 left and 117 right side observations for coronal ossicles, there was only one medium ossicle observed on the left side of a skull. This means that there is a cranial incidence of 0.9 per cent and a side incidence of 0.4 per cent. Nicolucci also recorded one case of a coronal ossicle in the sample that he studied, which translates into a cranial index of 1 per cent. Similarly, Capasso recorded one case in his sample, which is a cranial index of 0.63 per cent.⁵¹

ThePompeiansideincidenceofcoronalossicles is considerably lower than that recorded for the majority of the populations shown in Table 9.9. Similarly, the cranial incidence (Table 9.10) is much lower than that of the other recorded Italian populations, with the exception of the Herculaneum sample. The cranial incidence for this trait in the Pompeian sample is consistent with that obtained by Nicolucci and only minimally higher than that of the Herculaneum sample.

OSSICLE AT BREGMA

Only one strongly expressed ossicle at bregma was observed in the entire sample of 116 skulls, which meant that there was a cranial index of just under 0.9 per cent. Capasso did not record any cases of this trait. Inspection of Table 9.11 demonstrates that the frequency of this trait in the Pompeian sample is within the range observed for other Italian samples, though it should be noted that the sample sizes for a number of these populations are rather small.

SAGITTAL OSSICLES

Sagittal ossicles were observed in the Pompeian sample with a frequency of 7.8 per cent in the sample of 116 skulls that could be scored. Of these, 6.9 per cent were of medium expression or greater. Nicolucci only recorded one instance of this case in his sample of 100 skulls. Capasso recorded two cases of this trait, which means that there was a cranial incidence of 1.3 per cent. There was minimal appropriate comparative material for this trait (Table 9.12).⁵²

Given the paucity of comparative material, interpretation of the incidence of sagittal ossicles is difficult, though it is notable that the Pompeian frequency is significantly lower than that recorded for the Sardinian sample but does seem to be at the upper end of the range of frequencies for a disparate

Table 9.8 Cranial incidence of lambdoid ossicles in various populations
Population Sample size Frequency (%)

Pompeii AD 79 (Lazer 1995) 112 50 Pompeii AD 79 (Nicolucci 1882) 100 8 Herculaneum AD 79 159 13.8 San Vincenzo al Volturno 153 55.8 Undated Sardinian population 220 67.3

Sources: Adapted from Hauser and De Stefano, 1989, 178–79; Higgins, 1989–1991; Lazer, 1995, 294; Nicolucci, 1882, 11; Rubini et al., 1999, 10; Rubini et al., 2007, 124.

Table 9.9 Side incidence of coronal ossicles in various populations Population

Pompeii AD 79
Pontecagnano (Campania) (7^th – 6^th century BC) Sala Consilina (Campania) (9^th– 6^th century BC) Termoli (Molise) (7^th century BC)911 Ardea (Latium) (8^th– 6^th century BC) 34 11.8 Romans (Latium) (6^th–5^th century BC) 306 12 Alfedena (Abruzzo) (6^th century BC) 135 3 Campovalano (Abruzzo) (7^th– 6^th century BC)50 10 Perdasdefogu (Sardinia) (9^th century BC)340 Etruscans 1 (Central Etruria) (6^th–5^th century BC) 70 18.5 Etruscans 2 (Southern Etruria) (6^th–5^th century BC) 110 9 African sample (Mali) (1^st millennium _BC) 287 1.5 Nubian (historic) 134 13.5

Sample size Frequency (%)

228 0.4
64 3.1
21 4.8

Sources: Adapted from Hauser and De Stefano, 1989, 178–79; Higgins, 1989–1990; Rubini et al., 1999, 10; Lazer, 1995, 294; Rubini et al., 2007, 124.

Table 9.10 Cranial incidence of coronal ossicles in various populations Population

Pompeii _AD 79 (Lazer 1995) Pompeii AD 79 (Nicolucci 1882) Herculaneum AD 79
San Vincenzo al Volturno
Modern Roman sample
Undated Sardinian population

Sample size Frequency (%)

117 0.9
100 1
159 0.63
153 26.8
300 41.7
260 3.5

Sources: Adapted from Capasso, 2001, 982; Hauser and De Stefano, 1989, 178 –79; Higgins,
1989–1990; Lazer, 1995, 294; Nicolucci, 1882, 11; Rubini et al., 1999, 10; Rubini et al.,
2007, 124.

collection of European populations. Again, it is dif ficult to account for the much lower incidence of this trait in Nicolucci’s sample. It is possible that he did not take too much interest in these traits, except when they were strongly expressed, as they were only mentioned as a side issue in his article. Perhaps the most significant finding is the much higher frequency for the Pompeian sample than for that from Herculaneum.

OSSICLE AT THE ASTERION

Ossicles at the asterion were slightly more frequent on the right side than the left in the 101 observations that were possible for each side. There was an 8.9 per cent presence on the left side and a presence of 10.9 per cent on the right. Only one case involved a large ossicle, the rest were classified as small or medium. Capasso recorded just one case of ossicle at the asterion,

Table 9.11 Frequency of ossicle at bregma in various populations Population Number of Frequency (%) observations

Pompeii _AD 79 (Lazer 1995) Herculaneum _AD 79
116 0.9 159 0

Pontecagnano (Campania) (7^th–6^th century _BC)37 0
Sala Consilina (Campania) (9^th–6^th century _BC)18 0
Termoli (Molise) (7^th century _BC)482
Ardea (Latium) (8^th–6^th century _BC)17 0
Romans (Latium) (6^th–5^th century _BC) 153 2.6 Alfedena (Abruzzo) (6^th century _BC) 73 1.4 Campovalano (Abruzzo) (7^th–6^th century _BC) 51 3.9 Perdasdefogu (Sardinia) (9^th century _BC)17 0
Etruscans 1 (Central Etruria) (6^th–5^th century BC)35 0
Etruscans 2 (Southern Etruria) (6^th–5^th century BC) 55 1.8 San Vincenzo al Volturno 153 0
Modern Roman sample 296 1
Undated Sardinian population 243 2.5 African sample (Mali) (1^st millennium _BC) 153 0
Nubian (historic) 67 0

Sources: adapted from Capasso, 2001, 982–83; Hauser and De Stefano, 1989, 178–79; Higgins, 1990, Table 4; Lazer, 1995, 294; Rubini et al., 1999, 10, Rubini et al., 2007, 124.

Table 9.12 Presence of sagittal ossicles in various populations Population Sample size Frequency (%)

Pompeii AD 79 (Lazer 1995)
Pompeii AD 79 (Nicolucci 1882) Herculaneum AD 79
Undated Sardinian population Lithuanians (1^st–2^nd millennium BC) Bohemians (8^th–10^th century _AD) Alamannes (6^th– 8^th century _AD) Prehistoric Ukrainians
116 7.8 100 1 159 1.3 174 23.6

2681 5.4
557 3.8
243 1.6
112 7.8

Sources: Adapted from Capasso, 2001, 982–83; Hauser and De Stefano, 1989, 94–95; Lazer, 1995, 303, Nicolucci, 1882, 110.

which reflects a cranial incidence of 0.63 per cent. This individual also displayed lambdoid ossicles.⁵³

Table 9.13 indicates that the Pompeian side incidence for this feature was in the mid range for Italian populations, though it again should be noted that a number of these populations are represented by very small sample sizes. It was considerably lower than the frequencies observed for a modern Roman sample and two African populations. The cranial incidence documented for San Vincenzo al Volturno (Table 9.14) is about double that of Pompeii. Again, it is notable that the cranial frequency for this trait is much higher for the Pompeian than the Herculaneum sample.

Table 9.13 Side incidence of ossicle at asterion in various populations Population
Pompeii AD 79

Sample size Frequency (%)

202 9.9
Pontecagnano (Campania) (7^th – 6^th century BC)48 6 Sala Consilina (Campania) (9^th– 6^th century BC) 16 6.2
Termoli (Molise) (7^th century BC)430 Ardea (Latium) (8^th– 6^th century BC) 34 8.8
Romans (Latium) (6^th–5^th century BC) 306 7.8
Alfedena (Abruzzo) (6^th century BC) 142 8.5
Campovalano (Abruzzo) (7^th–6^th century BC) 73 9.6
Perdasdefogu (Sardinia) (9^th century BC) 32 15.6
Etruscans 1 (Central Etruria) (6^th–5^th century BC) 70 17.4
Etruscans 2 (Southern Etruria) (6^th–5^th century BC) 110 15 Modern Roman sample 553 20.3
Undated Sardinian population 500 6 African sample (Mali) (1^st millennium BC) 315 12.7
Nubian (historic) 134 18.6

Sources: Adapted from Hauser and De Stefano, 1989, 198–99; Lazer, 1995, 305; Rubini et al., 2007, 124.

Table 9.14 Cranial incidence of ossicle at asterion in various populations Population Sample size Frequency (%)

Pompeii AD 79 101 13.9
Herculaneum AD 79 159 0.63 San Vincenzo al Volturno 153 20.6

Sources: Adapted from Capasso, 2001, 982; Higgins, 1989–1990; Lazer, 1995, 305.

Post-cranial evidence

Ten post-cranial non-metric traits were scored for the Pompeian sample.⁵⁴ In addition to recording many of these traits, Capasso was able to examine the Herculaneum sample for numerous post-cranial non-metric features that did not survive in the incomplete, disarticulated Pompeian collection, such as anomalies of the vertebrae.⁵⁵ He also had the advantage of being able to associate bilaterally expressed traits and different non-metric features on the same individuals. Comparison between the results obtained from the two sites was hampered by the limitations associated with the Pompeian material. While there are some apparent differences in the observed frequencies for two of the femoral non-metric traits,⁵⁶ for the purposes of this study, only one post-cranial non-metric trait – lateral squatting facets on the tibia – will be presented in detail. This trait was singled out as it appears to be most useful as a population marker for the Pompeian sample of victims.
Lateral squatting facets

This trait is scored present when the inferior articular surface of the tibia extends into the lateral fossa of the transverse depression on the lower anterior surface (Figure 9.4). This latter forms the attachment for the articular capsule of the ankle joint. It has been argued that tibial squatting facets result from biomechanical stress, caused by a particular type of squatting or kneeling posture. They have been observed in fruit pickers, whose occupation included regular flexion of the foot while standing or climbing ladders. A correlation has been observed between changes in habitual posture as a result of the introduction of stoves and chairs and a diminution in the presence of this trait in French and American archaeological samples. The observation of this trait on European foetuses suggests that biomechanical stress may not be the only cause of such facets, and a genetic origin should also be considered.⁵⁷

A sample of 127 left and 124 right tibiae in the Pompeian collection were examined for the presence of medial and lateral squatting facets. There was a much higher prevalence of lateral than medial squatting facets. This trait was observed to occur more frequently on the right side, with a prevalence of 87 per cent, as compared to an incidence of 78.7 per cent on the left side. It is notable that the majority of cases for both sides exhibited strong expression of the trait. Capasso documented only five cases of squatting facets, which represents about 3.8 per cent of the Herculaneum tibiae that he studied. A modern American sample of white males yielded a left side frequency of 26 per cent and a right side frequency of 23 per cent. The side incidences of American white females are 24 per cent for the left and 33 per cent for the right.⁵⁸

Figure 9.4 Anterior view of distal portions of two tibiae with medial and lateral squatting facets on the left and an absence of facets on the right (after Finnegan, 1978, 23–37).

Lateral squatting facets are so common in the Pompeian sample that one could hazard a guess that the inhabitants shared certain habitual behavioural traits. Shared behaviour that results in similar skeletal changes in a population can indicate a type of homogeneity. Further comparison with other contemporary skeletal samples from the region and other sites from the Roman Empire would be necessary to determine whether this trait was specific to Pompeii, or was common for other Roman communities.

Conclusion

While the cranial metric results were inconclusive for the Pompeian sample, the relatively high frequency of certain non-metric traits in relation to other populations, like palatine torus, lateral squatting facets of the tibia and double-rooted canines, may indicate homogeneity, either as a result of common genes or a shared environment during the period of growth and development. Since it has commonly been assumed that the Pompeian population was heterogeneous, an explanation is required for these unexpected results.

It could possibly be argued that the apparent tendency towards homogeneity in the skeletal record is the result of sample bias produced by selective collection in the nineteenth century. Nicolucci (1882), for example, intentionally selected what he considered to be the more unusual skulls for examination. It appears that at least some of these skulls are now in collections in Naples and are no longer accessible (D’Amore et al. 1979; p.3 01; also see Chapter 3). It is impossible to assess the impact of this practice on the sample. The skulls that were excavated and stored on site at Pompeii over the last hundred years have not apparently been subjected to this treatment and so should be random.

The evidence presented in Chapters 6, 7 and 8 for sex, age-at-death and frequency of pathology, like HFI, indicates that the sample is random and representative of a normally distributed population. It should be remembered that the Pompeian skeletal sample is a reflection of the victims of the eruption and, while it might constitute a statistically representative sample, it may exclude portions of the original population. It is possible that the Pompeian population was never as heterogeneous as suggested by the literary evidence. The various populations mentioned by Strabo and Pliny the Elder, that were said to have inhabited Pompeii and Herculaneum, may have been distinguished by culture and language rather than variations in genetic material. However, consideration should also be given to the notion that there may have been some alteration to the composition of the population by some sections of the community leaving Pompeii, either as a direct result of the AD 62 earthquake or because of continuing seismic activity in the last 17 years of occupation (see Chapter 3). The population could also have changed for other reasons, as would be expected in a dynamic community with a long occupation history.
While Nicolucci reported similar frequencies for metopic suture and coronal ossicles, he recorded significantly fewer cases of other ossicles in the sample he studied. The discrepancies between the results obtained for the current study and that of Nicolucci for the incidence of inca bones, lambdoid ossicles and sagittal ossicles can possibly be explained by the fact that the observation of non-metric traits was more of passing interest than a major research consideration. Nicolucci’s primary research objective involved a craniometric analysis of the Pompeian skulls and he may not have either noticed or recorded cases that were not strongly expressed.

Perhaps the most remarkable results of this study are the huge differences in frequency between Pompeii and Herculaneum for a number of cranial and post-cranial non-metric traits. It is difficult to establish exactly what the non-metric results mean as their aetiology is not well understood. These results may imply that there were either significant genetic differences between the two samples of victims or that they experienced substantially different environments during the growing years. It would be instructive to reassess all the available Pompeian and Herculaneum material for as many non-metric traits as possible to establish whether the differences are real and not the result of interobserver error. It would also be extremely valuable to score these traits on other Italian skeletal material, especially in the Campanian region. Calculation of the frequency of these traits over time and space should aid in the determination of whether they are regional features or if they are specific to Pompeii.