Ancient History & Civilisation



Determination of health from skeletal evidence is fraught with dif ficulties, as many disorders that involve soft tissue do not present on bone. Most pathological changes to the skeleton reflect chronic ailments. Conversely, acute disorders, apart from trauma, are not likely to leave any trace on the skeleton. Further, bone can only respond to insult in a very limited number of ways; it can be lost or resorbed, new bone can be deposited or a combination of the two can occur. As a result, a number of diseases leave a similar appearance on the skeleton. It is preferable to base diagnoses on an examination of the entire skeleton as some pathology can be distinguished by the pattern of changes that can be observed on different bones.1 The disarticulated nature of the Pompeian sample limited the study to disorders that could be diagnosed with confidence from gross inspection of a single bone, like healed trauma. In addition, the lack of access to x-ray facilities and destructive techniques, like sampling for histological analysis, constrained the kinds of questions that could be asked of the evidence.

While the Pompeian skeletal record only provides a very limited view, all signs of pathology reveal clues about the general health of the sample of victims. The presence of certain pathology and other indicators, like stature, act as health markers and assist in constructing a picture, albeit a rather indistinct one, of the general well-being of individuals and the sample as a whole. Apart from elucidating oral health and diet, dental data can provide some indication of underlying health problems as bacteria associated with dental and other pathology of the oral cavity have been implicated in some soft-tissue disease, like heart-valve problems. Stature is, in part, a reflection of health and nutrition during the growing years. Similarly, bone alterations, such as flattening of the proximal shaft of the tibia and femur, have been interpreted as indicators of stress during the period of skeletal development. Healed injuries, infections and other diseases yield information about the status of the immune system, while age-related disorders are a valuable gauge of health, as they do not manifest until an individual has achieved a certain age.

The main issues to be considered were whether there was any skewing towards individuals with signs of infirmity in the Pompeian sample and if the observed pathology was likely to have impeded escape from the eruption. Another issue that was considered was whether it was possible to detect any evidence of surgical or dental intervention. Comparisons could then be made with the bones that have been excavated from Herculaneum. The Herculaneum material should provide more information as complete skeletons were available for investigation, which not only provided more reliable diagnostic opportunities, but also increased the range of pathology that could be studied. In addition, researchers in Herculaneum have had access to radiography and funding to employ various destructive techniques to facilitate diagnosis.

The skeletal record has also been invoked as a valuable guide to professional activities and other aspects of life that can be reflected in changes to bones and teeth as a result of habitual activity. Brief consideration is given to the actual potential of the Pompeian and Herculaneum skeletal record to indicate ancient lifestyles.

Oral health

The dif ficulty of attempting to gain some insight into the general health of the Pompeian victims without the benefit of soft tissue may be partly mitigated by a study of oral health. An association has been found between poor oral health, infections and various systemic conditions. Micro-organisms in the oral cavity can be responsible for infections in a number of locations in the body. These are known as focal oral infections and they result from the introduction of oral micro-organisms or toxins from oral pathology into the bloodstream or the lymphatic system. Focal oral infections are exacerbated in elderly people and individuals with compromised immune systems. The bacteria associated with the formation of dental plaque have developed special mechanisms to ensure their adhesion to hard and soft oral tissues, as well as other oral bacteria. These bacteria, when transported in the bloodstream, can reach the heart and are well suited to adhere to damaged heart valves and can cause infective endocarditis. This is a disease that involves inflammation and infection of the inner surface of the heart. Correlation has also been found, between dental pathology, particularly periodontitis, and other cardiovascular disease, pre-term low birth rate, diabetes, aspiration pneumonia and abscesses of the lung. In addition, the bacterium associated with gastric ulcers, Heliobacter pylori, has been identified in samples of saliva and dental plaque.2

An examination of the maxillae and mandibles of the Pompeian sample provides some indication of the oral health and diet of the population. The state of the teeth and the alveolar region was assessed to establish the level of oral hygiene, whether there was any evidence of dental intervention and the general health of individuals during childhood. Dietary factors, such as the impact of the milling process of flour on the teeth, were also considered.3

Ninety-seven maxillae and 80 mandibles were examined. These represent all the mandibles and maxillae from the Forum and Sarno Bath collections that were available at the time of this study. Before one could assess the dental health of the population it was necessary to establish the number of teeth that remained in situ and the degree of post mortem tooth loss. The incidence of tooth retention determined the potential value of the survey of frequency of carious or decayed teeth, alveolar bone loss, calculus or calcified plaque deposits and enamel hypoplasia in the sample. There were very few cases where the mandible and maxilla of an individual could be rearticulated. This meant that for the majority of cases it was not possible to assess occlusion or bite.

Number of teeth in situ

Only one of the 97 available maxillae retained a full complement of teeth. Most of the maxillary teeth had suffered either ante or post mortem loss. No teeth at all survived in 42.3 per cent of the maxillae. Only 14 maxillae, or 14.4 per cent of the sample, retained eight or more teeth. A higher proportion of mandibular teeth were still in situ. No full sets of teeth were preserved. Only three mandibles were completely devoid of teeth. Twenty-three, or 19.8 per cent, of the 81 mandibles contained eight or more teeth. The high incidence of post mortem tooth loss means that the cases of caries and the degree of calculus and linear enamel hypoplasia can only be interpreted as their minimum expression in the Pompeian sample.

Ante mortem loss

Ante mortem tooth loss can be distinguished from post mortem loss as the process of healing, which involves the closing of the socket hole in the mandible or maxilla, usually takes about six months. The principal function of the alveolar bone is to support and maintain teeth in position so that they can function properly. It is gradually resorbed when deprived of this function as a result of tooth loss. It is possible to determine whether ante mortem loss occurred some time before death by the degree of remodelling of the socket and subsequent alveolar bone loss. The only ambiguity that cannot be accounted for is perimortem loss as it is impossible to differentiate teeth lost just prior to or after death from the skeletal record. In addition, it is not possible to establish whether teeth that were lost ante mortem were purposely extracted.4 Ante mortem loss was scored as the number of teeth that were unequivocally lost before death, namely, in cases where some degree of bone remodelling had occurred.

At least one tooth had been lost prior to death in 69 per cent of the maxillae (n = 79) and 48 per cent of the mandibles (n = 56) in the Pompeian sample. The highest number of teeth lost from an individual mandible in the sample was 5 as compared to 14 for a single maxilla.

A strong association has been recorded between ante mortem tooth loss and advancing age for American Indian sites. The Pompeian results are not that conclusive, though a number of the mandibles and maxillae that have been interpreted as belonging to older individuals who had lost teeth prior to death.5

There are some problems in comparing the Pompeian dental data with those from the Herculaneum sample as different scholars presented their findings in various ways. Further, comparison between the results of scholars who have worked on the Herculaneum skeletons is hampered by the different methods that they have used to record their results. Bisel, for example, described her results in terms of mouths, as she had the advantage of access to articulated jaws and did not see the need for separate discussion of the upper and lower dentition. By comparison, Capasso presented his data in terms of numbers of individual teeth and related these to the quantity of affected individuals, as did Torino and Fornaciari and Petrone et al.6 These problems are exacerbated by the fact that none of the various data sets have been published in a standard format.

Bisel calculated a mean ante mortem tooth loss per mouth of 1.79 for males and 2.07 for females. As Bisel did not present her raw data it is not really possible to compare her results with those of the Pompeian sample.

Capasso presented data on the numbers and types of teeth that were lost prior to death. He noted that 37.4 per cent of a sample of 139 individuals lost, at least, one tooth prior to death. In most cases, only one tooth was lost, though there were three individuals who respectively lost 10, 12 and 30 teeth.7


The degree of attrition or tooth wear can give some indication of dietary behaviour and methods of food processing. Alternatively, it can provide clues about habitual behaviour, such as tooth grinding or bruxism, or the industrial usage of teeth as tools.8 It is often used as an indicator of age-at-death in archaeological samples (see Chapter 6). If attrition is very severe it can result in the use of the roots as an occlusal surface. The pulp cavity may then be exposed, leaving the tooth liable to bacterial infection. Resultant abscesses were potentially lethal in antiquity.

The frequencies of the different degrees of attrition were comparable for both mandibular and maxillary teeth, with the exception of those cases where the teeth were worn down to the roots, which had a mandibular incidence that was three times higher than for maxillae. This may just be a reflection of the greater likelihood of roots surviving in situ in mandibles, as they were generally stored with the teeth facing upwards, whereas maxillae were usually stored with the teeth facing down. However, a study of the dentition of a Roman skeletal sample from Quadrella, Molise, also revealed a higher frequency of wear on mandibular teeth.9

For the purposes of this study, attrition was generally scored to indicate relative adult age-at-death. Severe attrition was probably the result of dietary intake of stone as a result of the milling process.10 The flour the Pompeians used for baking was ground in large basalt mills (Figure 7.6), which led to small particles of stone becoming incorporated into bread. No apparent evidence was observed for industrial usage of teeth but such wear is not precluded because it was not possible to fully assess the majority of mouths due to the high incidence of post mortem tooth loss. The loss was especially evident for the anterior teeth, which are most commonly employed as tools.11 It is notable that Bisel observed what she interpreted to be industrial wear on the right maxillary incisors of one individual from Herculaneum.12 Capasso also recorded evidence of industrial wear, most notably on anterior teeth, of 18 individuals in the Herculaneum sample. Apart from attrition as a result of tool use, he noted severe tooth wear in five per cent of the 2966 teeth that he examined.13

The tooth wear patterns observed in Pompeii are consistent with those observed on other ancient populations that apparently had constantly masticated abrasive substances in their food. For example, the tooth wear patterns from a series of small Pueblo sites in north-eastern Arizona were similarly interpreted as age-related and resultant from constant mastication of abrasive substances in food, rather than other factors. Conversely, little heavy wear was observed on the Roman dental sample from Quadrella. This was interpreted either as a reflection of a diet that was low in hard fibrous foods which required vigorous chewing, or care with cleaning and food preparation to soften the final product and to minimize the presence of abrasive material.14


The presence of caries cavities gives an indication of diet and oral hygiene. This pathology is directly related to the presence of dental plaque and a dietary intake of fermentable carbohydrates. It involves the progressive decalcification of enamel or dentine. Several bacterial organisms are associated with dental caries, though the main one is Streptococcus mutans. The position of caries can provide dietary information about the afflicted individual.15 For the purposes of this study the survey of caries in the Pompeian sample was limited to the number of caries observed in each mouth and the degree of destruction. Scoring was based on the largest carious lesion in the mandible or maxilla under investigation. This study was based on direct visual inspection of the teeth with the aid of a dental probe.16 It was not possible to undertake x-ray analysis, which would have produced more reliable results, especially for the size of carious lesions.

Of the 314 surviving maxillary teeth, 50, or nearly 16 per cent, were carious. Forty-three, or 9.7 per cent, of the 444 mandibular teeth that were still in situ had carious lesions. Because it was impossible to assess cases where teeth had been lost, the sample was limited to 27 mandibles and 27 maxillae. As a result of the lack of complete dentition for virtually all the maxillae and mandibles, it is impossible to determine the number of carious lesions per mouth. The data merely provides the minimum number of caries cavities per mouth. The degree of carious involvement was substantially greater for a higher proportion of maxillary than mandibular teeth. Nearly half of the caries in the maxillary teeth were described as advanced or gross as compared to just under a quarter for the mandibular teeth.

The Pompeian data set was too incomplete to draw firm conclusions about the relationship between the degree of carious lesions and advancing age, though it is not uncommon to observe a considerable involvement of the crowns of teeth with caries cavities in the mandibles and maxillae of individuals interpreted as older on the basis of attrition.17 These results suggest that fermentable carbohydrates formed a significant part of the diet and that oral hygiene was not widely practised.

Bisel recorded an average of 0.92 caries per mouth for males and 0.68 for females in the Herculaneum sample.18 The Pompeian material, which could only be used to assess the minimum number of caries, does not give an indication of the number of caries per mouth.

Bisel attributed the apparently low number of lesions in the mouths that she examined to a lack of sugar in the Herculanean diet. Honey was known to the Romans but was not used as extensively as sugar is used in a modern Western diet. In addition, Bisel suggested that a less processed diet would have been advantageous in that caries would have been worn away before they could develop. Hillson reasoned that low sugar consumption would provide a more likely explanation than attrition for the low incidence of caries in ancient populations as attrition would have to be very rapid to remove active deep fissure caries before the pulp cavity was penetrated.19

Torino and Fornaciari undertook a preliminary study of 87 victims, 64 of whom were adults and 23 juveniles. They documented 58 carious teeth in a sample of 2,020 teeth. Four of these were deciduous, while the remainder were permanent dentition. These reflected 1.8 per cent of the deciduous sample of 222 teeth and 3 per cent of the 1,798 permanent teeth. They considered that rate of caries in the sample was extremely low by both modern and ancient standards. They partially attributed this finding to the consumption of water and food with high fluorine levels, as attested to by the presence of cases of fluorosis in the sample.20

Like Torino and Fornaciari, Capasso recorded the presence of caries in relation to the total number of teeth he studied, rather than per mouth. He recorded 135 cases in 3,236 teeth, which is a 4.17 per cent incidence of teeth with carious lesions. Nine of these teeth were deciduous and they reflected 3.33 per cent of the sub group of 270 deciduous teeth in the sample. In contrast, Petrone et al. observed that 78.6 per cent of the 56 mouths that they examined had carious teeth, with 20 per cent of the 1,358 permanent teeth displaying evidence of tooth decay. Even though there is considerable variation in size between samples, it is difficult to account for the different results obtained by different researchers from the Herculaneum sample.21 Interdental alveolar resorption

Periodontal tissues, including jaw bone, gingivae or gums, cementum and the periodontal ligament, surround and support the tooth. Interdental alveolar resorption is generally the result of periodontal disease. On a living person, periodontal disease is marked by inflammation. The first stages only involve the soft tissue, notably the gingivae. All the periodontal tissues are involved in the most advanced stage, which is known as periodontitis. The most common form in modern Western populations is mostly seen in adults over thirty years of age. It can occur as a result of a lack of dental hygiene, advanced attrition or poor diet and causes the alveolar bone to recede. As bone loss increases, the teeth become loose and, if left unchecked, can ultimately be lost. It is generally marked by a ridge-like change to the labial side of the alveolus. The degree of recession can be recorded if the teeth are still present. Though more difficult to score, alveolar loss can be observed in mouths where no teeth are in situ but teeth have been lost some time prior to death and considerable remodelling and subsequent loss of bone has occurred.22

Alveolar loss was dif ficult to assess in cases where teeth had not survived. Most of the cases that could be scored for both mandibles and maxillae showed some degree of interdental alveolar resorption. There were relatively few examples of considerable alveolar loss.

Bisel found that about 75 per cent of the Herculaneum sample exhibited some degree of alveolar resorption. Of these, 60 per cent had only slight resorption and 9 per cent displayed a moderate to marked degree of periodontal disease. Because of the post mortem loss of so many teeth it was not possible to adequately assess the degree of alveolar recession in the Pompeian sample. Nonetheless, the data that could be collected suggested the incidence and degree of alveolar resorption was comparable with that of the Herculaneum sample.23

Dental abscess

Abscesses are the result of the build-up of dental infections that can be causally related to caries, excessive, rapid attrition, trauma and periodontal disease. These can generally be detected as a clearly defined circular cavity in the alveolus near the root of the affected tooth. These cavities function as drains for the pus produced by the abscess. Bony changes to the alveolus that could be attributed to dental abscesses were recorded for each maxilla and mandible. It should be noted that deep pockets from advanced periodontal disease are virtually impossible to distinguish from abscesses.24

There was a higher frequency of abscesses in the maxillae than the mandibles in the Pompeian sample, with at least one abscess in almost 43 per cent of the maxillae and just under 19 per cent of the mandibles. This is consistent with observations from other archaeological sites. A strong association was observed between advancing age and the presence and degree of abscesses in the skeletal series from Pueblo sites in Northeastern Arizona that were studied by Martin et al.25 The Pompeian data was difficult to assess in this respect as it was not possible to estimate age-at-death in a number of cases where teeth and other age indicators had not survived. In the cases where assessment was possible, there did appear to be some correlation between age and the presence of dental abscesses, especially severe abscesses, which were almost always associated with older individuals (for example, Figure 8.1).26

Bisel recorded a mean of 0.73 abscesses per male mouth and 0.66 per female mouth in the Herculanean sample. This is a higher frequency than that for the Pompeian sample. Capasso observed a total of 52 abscesses in 23 individuals in his sample, 51 of which were periapical. Nine of these individuals had only one abscess, eight had two, one had three, two individuals presented with four and two had five, while one individual had a total of eight oral abscesses. Capasso suggested that these abscesses were associated with the development of carious lesions and severe tooth attrition, which involved exposure of the pulp cavity.27


Calculus is mineralized plaque. It is attached to the surface of the tooth. The mineral, though deposited from plaque fluid, derives from saliva. As a result the greatest amount of calculus formation occurs at sites on teeth that are nearest to the salivary glands. The reasons for plaque mineralization are not fully known, though it is thought that bacteria are involved. The presence and degree of calculus observed in the Pompeian sample was recorded.28

Figure 8.1 Large sinus for abscess drainage (TF NS 86: 1). This abscess formed as a result of severe attrition (7.4), where tooth wear was so great that the pulp cavity was exposed to the air, making it susceptible to bacterial infection

Most of the surviving teeth of the Pompeian sample had, at least, a slight calculus deposit. It is important to note that it was only possible to score minimal expression as the storage conditions were not conducive to the preservation of larger deposits. The degree of calculus deposited on approximately half of the teeth for both mandibles and maxillae was recorded as slight. Only 19.5 per cent of the available maxillary and 11.1 per cent of the mandibular teeth displayed no evidence of calculus. This suggests that oral hygiene, as it is known to modern Western communities, was not a high priority.

Bisel did not present data on the presence of calculus on the Herculaneum teeth, whereas Capasso commented that there was a particularly high frequency of calculus deposits in the sample he studied. Forty-two, or 39 per cent, of the 139 mouths he studied showed some degree of calculus deposition. Twenty-eight individuals displayed slight; ten, medium; and four had considerable deposits of calculus.29

Linear enamel hypoplasia

The presence of linear enamel hypoplasia is the direct result of the failure of the enamel to properly form in the developing tooth. It can result from a variety of factors, such as nutritional stress, infection, poisoning or trauma to the tooth or pulp of a deciduous tooth lying over a growing permanent tooth. Enamel hypoplasia often presents as transverse line of indented enamel on the sides of the tooth crown, though it can also present as pits and grooves. They can only form during the period of crown development, which means that if all the dentition is examined, presence of hypoplasia can potentially reflect the health and nutrition of an individual between approximately the second and fourteenth year of life. The position of the line gives some indication of the age of the individual when such periods of stress occurred. It should be noted, however, that absence of hypoplasia does not necessarily infer that there were no periods of stress in the years of crown formation. While the presence of enamel disruption suggests insult, its absence cannot automatically be interpreted as evidence of a healthy and well-nourished person. Modern individuals with a history of lengthy and major illness in early childhood do not always display enamel hypoplasia.30

Goodman and Armelagos discovered that anterior teeth were more susceptible to hypoplasias than posterior teeth because their developmental timing can be easily disrupted. Individuals exposed to the same environmental stressors may exhibit varying degrees of enamel hypoplasia. Martin et al. suggested that it was only really useful to record linear enamel hypoplasia in the case of anterior teeth.31

Equipment was not available for making accurate readings for the distance between lines in the Pompeian sample, so no attempt was made to determine the period in juvenile tooth growth when the disruption occurred. Instead, the presence or absence of these lines was noted and a score was allocated with respect to the number of lines and degree of disruption to the enamel surface. Some degree of linear enamel hypoplasia was observed on 19 of the 33 maxillary teeth and 36 of the 45 mandibular teeth.

The Pompeian sample exhibited both a higher incidence and a greater degree of linear enamel hypoplasia on mandibular teeth. Nonetheless, most of the recorded cases of hypoplasia for the sample were slight. It is not reasonable to extrapolate interpretations from the small Pompeian sample onto the population at large, beyond the observation that a number of individuals experienced episodes of stress, possibly as a result of illness, during the period of crown formation.

In the posthumous publication of Bisel ’s work, the frequency of linear enamel hypoplasia is recorded at about 50 per cent. Her earlier publications indicate lower levels, with observations of some degree of linear enamel hypoplasia on about 25 per cent of the male and 16 per cent of the female Herculanean sample. Capasso observed some degree of enamel hypoplasia in the teeth of 71 individuals in the sample he studied. Of these, 42 were male and 29 were identified as female, which meant a male to female ratio of approximately 3 : 2. Petrone et al. examined the mouths of 51 individuals and recorded a frequency of enamel hypoplasia in 96.1 per cent, with some degree of expression in 36 males and 15 females.32

Evidence of dental intervention

No unequivocal evidence of dental intervention was observed in the available Pompeian sample. It is possible that teeth were extracted ante mortem, though it is impossible to distinguish intentional from unintentional ante mortem tooth loss from the skeletal evidence. There is some indication that tooth extraction was not practised even when removal might have alleviated a problem. A good example of this can be seen in the maxilla of a skull in the Forum Baths collection (see Figures 8.2 and 8.3).33

Only two teeth are still in situ, the right second molar and the left first molar. The former tooth is marked by a very thick deposit of calculus over the entire tooth, with the exception of a small polished facet on the mesial side of the crown. This wear facet is not on the normal occlusal surface. Considerable alveolar loss is apparent, with the exposure of more than half the roots of the remaining molar. The left first molar displays no sign of calculus. The degree of alveolar loss is comparable to that on the right side of the maxilla.

The excessive build-up of calculus on the right molar suggested that this individual was avoiding the use of this side of the mouth for mastication. The apparent reason for this was a large abscess associated with the palatal root of the adjacent first molar. This involved most of the lingual alveolar surface as evidenced by the marked resorption of the palatal process of the alveolar region. There is also a small sinus from an abscess associated with the distal buccal root of this tooth (Figure 8.2). It appears that this molar

Figure 8.2 Lateral view of skull showing excessive build-up of calculus on tooth adjacent to an abscessed tooth (TF 6)

Figure 8.3 View of maxilla from below showing excessive build-up of calculus on tooth adjacent to an abscessed tooth (TF 6)

was lost post mortem. Tooth extraction would have provided an effective means of draining pus from chronic abscesses.34
Non-dental evidence for dental practice

It is likely that, at least, some of the observed ante mortem tooth loss was the result of deliberate extraction, even though no tooth forceps have been found amongst the repertoire of medical instruments discovered in Pompeii. General bone forceps may have been substituted for this purpose. Other examples of tooth forceps found in Roman archaeological contexts were made of iron and it is remotely possible that this material did not survive burial in Pompeii.35

There were specialists in extraction in the Roman world as can be seen in the case of a shop built into the podium of the Temple of Castor and Pollux in the Roman Forum, where dental procedures were undertaken, along with the sale of pharmaceutical and cosmetic goods. It was suggested that a combined dental and pharmaceutical practice would have been valuable as related problems could have been treated simultaneously. A total of 86 decayed teeth, including two molars from children, were found during excavation. There is evidence that they were removed by skilled practitioners as great care had been taken to ensure that the entire tooth was excised and that no portion of the roots remained in the sockets. In one case, a portion of alveolar bone was extracted with the tooth so that the extremely decayed crown did not break from the root. The dangers of leaving dental fragments in the jaw were discussed by Celsus in the sections on dentistry in his treatise, De Medicina, which he wrote in Italy in the early part of the first century AD. It is notable that he considered extraction to be a last resort, after the failure of other measures, like treatment of the gums. He also recommended the use of gold wire to bind teeth that had been loosened, by an accident or a blow, to firm teeth in order to stabilize them.36

Ante mortem tooth loss was clearly not uncommon in ancient Italy as dental prostheses have been found, dating from Etruscan times, and dentures are mentioned in some of the epigrams of Martial. In addition, one of the laws of the Twelve Tables at Rome expressly forbade the burial of gold with a corpse unless it was fastened to the teeth of the deceased, which indicates its use for ligatures and dentures.37

The dental evidence from the Pompeian sample provides no indication that oral hygiene was practised, though other sources from antiquity suggest that some of the associated health and social benefits were appreciated. Writers like Pliny the Elder and Scribonius Largus provide recipes for dentrifices. Some of these contained substances like rabbit’s skull mixed with mouse head ashes, while others just prescribed the use of salt water. The charcoal from burnt bones in these recipes would have acted as a cleaning agent. Pumice was also employed for the same purpose. The use of toothpicks was also mentioned in the ancient literature as were ‘chew sticks’, which were used in place of toothbrushes, along with bare fingers.38

It would be reasonable to conclude from the number and degree of carious lesions, abscesses, ante mortem tooth loss and the degree of alveolar resorption and calculus deposition, that oral hygiene, as it is known in modern Western societies, was not practised. In addition, there appeared to be no evidence of dental intervention, with the possible exception of tooth extraction. While not enough evidence exists to draw firm conclusions from the results of an examination of the oral health of the Pompeian sample of victims, it still provides some insight into their general health. The high frequency of calculus and other dental problems indicates the potential presence of a number of systemic conditions, such as cardiovascular disease, diabetes, and gastric and respiratory disorders. Further, the presence of some degree of linear enamel hypoplasia on a number of the anterior teeth suggests that a proportion of the sample was exposed to stress, such as serious or chronic illness, during the period of crown development.

Due to the high frequency of post mortem tooth loss, the data set is too incomplete to determine with certainty whether the degree of dental pathology is correlated with age. The available evidence is consistent with such a conclusion.


Bone inheritance is multifactorial, which means that the ultimate height an individual attains is a compromise between their genetic potential and environmental factors, especially health and nutrition, during the years of bone growth. Information about the stature of earlier populations can provide insight into the health status of individuals in a sample.

The height of a human body is directly correlated with limb bone length. Regression formulae have been developed over time by a number of scholars to estimate the stature of individuals from long bone measurements.39 Trotter and Gleser’s formulae for white males and females have generally been favoured by physical anthropologists for height reconstruction for skeletal material from ancient European populations.40 It should be noted that reconstructions from these formulae only provide a rough guide to the actual height of individuals from an unknown population as they were primarily devised for modern American forensic use. Even when applied to known modern populations, they only produce an estimate of stature with a margin of error of several centimetres. A brief consideration of the background of stature estimation and the establishment and application of techniques, particularly for archaeological material from Central and Southern Italy, highlights the problems associated with the interpretation of data that has been collected to date. The Pompeian material requires detailed attention as the disarticulation of this skeletal sample creates further difficulties for stature estimation.

Trotter and Gleser developed their regression formulae for stature reconstruction from the bones of European and African American soldiers who died in World War II and the Korean war, along with data from the Terry skeletal collection. Separate regression formulae were calculated for females.41 It should also be recognized that the proportions of long bones relative to body length vary between populations; for example, Japanese people generally have proportionately shorter lower limb bones than Europeans, whilst those of black Africans are proportionately longer.42 The proportions of long bone length to height can also vary within a single population. It has been recognized that as people age, their height diminishes and modifications have been made to the formulae for their application to the bones of older individuals to accommodate these changes.43Unfortunately, they could not be applied to the Pompeian sample as it was not possible to estimate age from individual adult femora.

Given the dif ficulties associated with height reconstruction, it has been suggested that long bone length alone would provide a more reliable indicator of the general health of archaeological populations. This would dispense with the problem of attempting to compare results between scholars who have used different techniques to estimate height.44 As already stated, stature reconstruction, particularly using formulae that weren’t designed to accommodate unknown populations is, at best, only an approximation. Nonetheless, until standardized data sets of long bone length are available for comparison, it is necessary to compare with the published material, which is based on stature estimation.

Measurements of skeletons from archaeological contexts in situ, prior to excavation, provide an indication of actual height and can be used as a basis for choosing the most appropriate regression formulae for stature estimation. Scholars working on Central and South Italian skeletal samples reported that the most reliable estimates were obtained from the application of the 1952 Trotter and Gleser formulae for the group they describe as ‘American Negroes’.45

Unfortunately, the majority of scholars who estimated stature from Campanian skeletal samples have not specified which formulae were applied, though it is likely that they used those devised for ‘American Whites’ as these have been traditionally thought to be most appropriate. Both ‘American Negro’ and ‘American White’ sets of formulae were applied to the data collected from the Pompeian sample, the former as an acknowledgement of the results of recent research on South Italian material and the latter to facilitate comparison with other skeletal studies of victims of the AD 79 eruption.

The average heights of both males and females in the Pompeian sample were calculated from the maximum length of 148 left femora. Ideally, height reconstruction should be based on a combination of bones.46 As it was not possible to do this for the disarticulated Pompeian sample, the femur was chosen because it is considered to be the most useful single bone for the reconstruction of height. Further, it has been observed that lower limbs are apparently more sensitive to environmental stressors than upper limbs, which means that they are better health indicators of a population.47

Apart from the problems associated with the application of formulae derived from modern populations onto an ancient unknown sample, one of the major difficulties in obtaining height estimates from the Pompeian femora was the need to separate them by sex (see Chapter 6). It is generally assumed that an individual whose stature is to be determined is of known sex so that the appropriate height reconstruction formula can be used. This was not possible for the ancient Pompeians as the sample of femora, though apparently dimorphic, exhibited considerable overlap. The femur measurement which showed the most bimodality was maximum length. As a result, it was considered to be reasonable to use the frequency distribution of maximum length to obtain height guesstimates for the Pompeian sample.

The use of the frequency histogram of maximum length measurements (Figure 6.5) for this purpose was based on several assumptions: first, that the histogram is, in fact, bimodal and that the separation is by sex, and second, some degree of normal distribution would be expected for each component curve for a random normal biological sample. The appearance of the histogram is consistent with what would be expected, based on these assumptions.

The degree of overlap between male and female curves differs between populations. This was illustrated by Pheasant, who presented curves of the range for male and female stature for a number of different human groups.48 Populations that are very dimorphic with minimal overlap between the sexes have well-separated curves with two distinct peaks. The Dinka Nilotes from the southern Sudan provide an example of such a population. By contrast, Efe and Basua pygmies display very little dimorphism; with so much overlap it is difficult to separate the peaks of the two curves. It is notable that the Dinka are amongst the tallest people in the world and the Efe and Basua among the shortest. The frequency histogram for the Pompeians most closely resembles that of the British, the only European population presented by Pheasant.49 The degree of overlap for this population is between that observed for the two extremes of the Dinka and the Efe and Basua.

The closeness of the mean of the curve to the sex mean of the sample is dependent on the degree of separation between the peaks of the curves. The further apart the peaks, the nearer the estimate to the mean of the peak, because overlap decreases with distance. A very rough guide to the mean stature of the ‘male’ and‘female’ peaks could be obtained by applying the female formula to the mean of the left peak and the male for that of the right. This provides a rough guide to the Pompeian male and female mean stature.

It would be preferable if the numbers of males and females were roughly equivalent. The results of sex separation for the femora were inconclusive but suggest that the sample may have been skewed towards the more gracile bones (see Chapters 5 and 6).

The estimated mean for the right or ‘male’ peak was about 44 cm and approximately 40.75 cm for the left or ‘female’ peak. The results for height estimation using the various sets of regression formulae are presented in Table 8.1. The ‘American Negro’ mean heights are a few centimetres less than those calculated using the formulae devised for ‘American Whites’, though the margin of error means that there is considerable overlap.

It cannot be stressed too highly that the stature means for the Pompeian sample are rough estimates and need to be confirmed by results from complete skeletons, which can be more reliably sexed. These data may either come from skeletons revealed by new excavations at Pompeii or from x-rays of the sample of casts.

Table 8.1 Stature estimates based on Pompeian femora

Trotter and Gleser ‘American Whites’ male height (1958) 167.60 ± 3.94 cm Trotter and Gleser ‘American Whites’ female height (1958) 154.75 ± 3.72 cm Trotter ‘American Whites’ male height (1970) 166.13 ± 3.27 cm Trotter ‘American Whites’ female height (1970) 154.75 ± 3.72 cm Trotter and Gleser ‘American Negroes’ male height (1952, 1977) 163.19 ± 3.94 cm Trotter and Gleser ‘American Negroes’ female height (1952, 1977) 152.67 ± 3.41 cm

Subsequent stature estimates have been made, based on the maximum length of the right femur and humerus from the disarticulated Pompeian sample. A number of regression formulae and estimates of proportional relationships between limb and body length have been applied to these data, though it is not clear whether those established for European or other populations were applied. The results that were obtained are consistent with those obtained from this study from the application of ‘American White’formulae to the left femora.50

Possibly of greater signi ficance than the calculated mean heights is the difference between them. Depending on the formulae that are applied, the difference between male and female stature varies from almost 11cm to just under 13 cm, which is comparable to the difference between the means of modern North American males and females from a civilian sample.51 If the observed difference in the Pompeian sample is not an artefact of the disarticulated sample, it indicates that the population was not exposed to many environmental stressors during the period of growth. It has been argued that males are affected more by stress, such as inadequate nutrition or disease, than females. This is manifested in diminished stature as a result of failure to achieve genetic potential for height. If this theory is correct it would follow that the greater the observed sexual dimorphism for stature in a population, the more likely its members were healthy and well nourished during the years of skeletal development.52

Stature reconstructions for the Herculaneum sample would be expected to be more reliable because the sexing of individuals was based on entire skeletons. It is notable that Capasso obtained slightly lower values than Bisel for mean height from what was essentially the same sample (see Table 8.2). He found it difficult to account for the difference between his results and Bisel’s as he claimed that she did not provide details of the methodology for her calculations, though Bisel recorded that she employed Trotter and Gleser’s 1958 height reconstruction formulae. These are the same formulae that Capasso used for his later study.53 Capasso was in turn criticized for failing to state whether he applied Trotter and Gleser’s formulae based on American ‘Whites’ or ‘Negroes’.54 As a result of the lack of methodological documentation, it is difficult to compare these results with those obtained from the Pompeian femora. Those obtained using the ‘American White’ are

Table 8.2 Stature estimates for the Herculaneum skeletal sample and a modern Neapolitan sample

Herculaneum males 169.1 cm (Bisel and Bisel 2002 – using Trotter and Gleser 1958)
Herculaneum females 155.2 cm (Bisel and Bisel 2002 – using Trotter and Gleser 1958)
Herculaneum males 163.8 cm (Capasso 2001 – using Trotter and Gleser 1958)
Herculaneum females 151.7 cm (Capasso 2001 – using Trotter and Gleser 1958)
Modern Neapolitan males 164 cm (D’Amore et al. 1964)
Modern Neapolitan females 152.6 cm (D’Amore et al. 1964)

comparable to those obtained by Bisel for the Herculaneum skeletal sample as they fall within the same range when the margin of error is considered. Similarly, the results based on the ‘American Black’ formulae are more comparable to those recorded by Capasso.

The reconstruction of height from long bones in the Pompeian sample enables comparison with modern Neapolitan average stature for both sexes, based on a study of living people that was published in 1964.55 If the mean statures were observed to be comparable with those of the modern population, an argument could be constructed for regional continuity. Smaller average heights than those of the modern Neapolitans could imply that the ancient population had a lower standard of health and nutrition or constituted a different population. Higher means in the ancient sample could be interpreted as either a higher standard of health and diet or as a population that was not related to that of modern Naples.

It is noteworthy that when the ‘American White’ formulae are applied, both the Pompeian and Bisel’s Herculaneum samples have comparable but slightly higher mean stature than a modern Neapolitan sample (see Tables 8.1 and 8.2).56 The height estimates obtained from the Pompeian sample using the ‘American Negro’ formulae are closer to those of the modern Neapolitans. Regardless of the formulae that are applied, the heights mostly do not diverge by more than a few centimetres and the range of errors show considerable overlap. It is therefore possible to mount an argument for regional continuity, which may reflect a relationship between the ancient populations in the region and the modern Campanians.

While it is clear that stature is a useful health indicator, some scholars push the evidence to relate stature to social class and have incorporated data from height studies into social and economic histories.57 There needs to be a word of caution about the extent to which stature reconstruction data can be interpreted. It has been claimed that as taller populations or tall subsets of a population are apparently the result of better health and nutrition, they reflect individuals or groups of higher status or social classes. It should be recognized that higher status is no guarantee of a more nourishing diet during the growing years, as can be demonstrated by the consumption of more processed food by wealthier English people between the Industrial Revolution and World War II (see Chapter 1).

As a result of the disarticulation of the sample, it was only possible to roughly estimate the mean Pompeian male and female stature. The results are consistent with the potentially more reliable mean heights that have been obtained from the recently excavated Herculaneum sample. There is a need, however, to apply the formulae for stature estimation that have been found to be most appropriate for Central and South Italian archaeological skeletal remains to all the AD 79 victims. The lack of clear documentation of techniques employed impedes comparison between skeletal samples and makes it difficult to draw conclusions. Despite these problems, the data that have been collected suggest regional continuity for height and that the ancient Campanians had adequate diets and were in relatively good health during the period of bone growth.

Platymeria and platycnemia

Flattening of the proximal end of the shaft of the femur is known as platymeria and platycnemia when it occurs in the tibia. Like stature, the presence of platymeria and platycnemia have been interpreted as indicators of general health as they are considered to be stress related, either biomechanical, or as a result of nutritional deficiencies. Platymeria and platycnemia are expressed as an index using the antero-posterior and transverse diameters of the proximal part of the shaft.58

A sample of 156 left femora from the Forum Bath collection were examined for platymeria. The mean index was 80.9 ± 0.65, which is well within the platymeric range. The Pompeian sample is apparently slightly more platymeric than that of the Herculaneum sample, which according to Bisel had a platymeric index of 83.1 for females and 81.9 for males. Capasso calculated combined indices for left and right femora of 84.2 for males and 85.1 for females. Bisel calculated an average Herculanean platymeric index of 82.4 as compared to Capasso’s combined left and right index of 84.6. There is no apparent reason for the differences in the results of Bisel and Capasso, though they may just be a reflection of the larger sample available for the latter study. It was not possible to sex segregate the Pompeian sample for platymeria but it should be noted that the greater degree of platymeria in the Herculaneans was observed in femora of individuals interpreted as male which is not consistent with the view that this trait is more commonly observed in females.59
Fifty left and 51 right tibiae were examined for the presence of platycnemia. The mean index for the sample of left and right bones was 70.3 ± 0.61 which is beyond the range for platycnemia and is described as eurycnemia. Only 2.3 per cent of the left and 12.8 per cent of the right tibiae were platycnemic. No bimodality was observed, which implies that there is no sex separation for these measurements. There was no appreciable difference between the left and right samples. This means that both the left and right tibiae probably reflect the same population and can reasonably be compared.

The averaged Herculanean sample, like the Pompeian sample, is eurycnemic. This implies no significant flattening and suggests a sample more consistent with a modern population, such as that of the modern French, which has a cnemic index range of 71 to 74.60

It should be noted that there is no universal agreement as to the cause of proximal shaft flattening of the lower limbs. This is partly because of the difficulty in obtaining information on the causes of bony changes for human bones due to the ethical problems involved in human experimentation. It has been suggested that certain pathologies, such as osteoarthritis and osteoperiostosis may be associated with platymeria. Another possible cause for platymeria could be excessive strain on femora during childhood. The causes of platycnemia are also debated. Pathological factors have been cited. It has also been suggested that it is caused by constant squatting. Various pathological factors, including treponemal diseases and rickets, have been implicated for the production of long bone bowing.61 It is reasonable to assume that no single factor is fully responsible for these changes and that interpretations must be based on an examination of whole skeletons in the context of their population, rather than of individual bones.

There is no signi ficant difference between the Pompeian and Herculanean samples for these features. As it is difficult to isolate the cause or causes for the flattening of the proximal shafts of long bones, it is not really possible to interpret these results beyond stating that both the Pompeian and Herculanean populations may have had lifestyles that involved the application of greater stresses to the femora than the tibiae.

Pelvic brim index

Bisel argued that bone softening due to poor nutrition would result in a certain degree of flattening of the pelvis as it bears a great deal of the weight of the body above it. This could be expressed as an index of the pelvic brim. The Herculaneum sample displayed a mean figure of 83.9, which she compared to a mean of 93.3 for modern Americans. She interpreted this as a reflection of the better level of nourishment in the latter population.62 This is not a commonly used skeletal marker and it is, therefore, difficult to assess its usefulness.

Pathological change: trauma

Trauma includes various bone injuries, caused by cutting or piercing of the bone by sharp implements or crushing by blunt objects. It includes fracture. Trauma also includes certain types of surgical intervention, such as amputation, trepanation or trephination. Trauma is the second most common cause of pathological change to bones after degenerative changes.63

Observations were made on all the bones stored in the Forum and Sarno Baths as well as the skeletal material in the Casa del Fabbro (I, x, 7) and the Casa del Menandro (I, x, 4). No obvious signs of trauma were discerned on bones other than skulls and long bones in the available sample. At least 350 skulls, 500 right and left femora, 400 right and left tibiae, 150 left and right fibulae, 400 right and left humeri, 200 right and left radii and 200 right and left ulnae were inspected for signs of trauma or surgical intervention.64

Only one case presented with an injury that was consistent with having occurred at or around the time of death. The fracture pattern suggested a perimortem blow to a skull, which may reflect a tephra-related injury (see Chapter 4).65 All the other cases of trauma that were observed in the sample were either healed or healing.

No evidence of trauma was apparent on the bones from the Houses of the Fabbro or the Menander. Gross inspection of the 1,800 or so bones that were stored in the Forum and Sarno Baths yielded a total of seven unequivocal fractures and one unequivocal case of surgical intervention. There was also one case of a healed injury on a skull. Six of these fractures involved long bones. Three of these had healed with no bone displacement and three had healed with malalignment. The seventh was a depressed fracture of a skull. All the bones were identified as adult and all of these fractures had healed some time prior to death.

Fractures naturally commence the healing process shortly after the trauma occurs. The fracture causes blood vessels in the bone to rupture. Blood then flows into the area of the fracture and forms a bloody mass or haematoma, which then stimulates new bone formation, ultimately leading to the development of a hard callus. The callus acts as a natural splint and will remodel as the fracture heals. If the fracture ends are in line and the bone is immobilized during the period of healing, it can be difficult to detect evidence of the fracture in the remodelled bone, except in radiographs.66

The three bones that had healed with no bone displacement were identifiable by callus formation around the fracture site. They were a right ulna, a left radius and a right tibia.67

The right ulna and the left radius exhibited callus formation on the proximal third of the shaft, whilst the right tibia displayed similar pathological alteration to the distal third of the shaft. No signs of secondary infection were apparent. There is no reason to assume that there had been any medical intervention in these three cases as these all occurred in paired bones. If only one bone had been broken the other could have acted as a splint.

The three bones, which displayed compound fractures with some malalignment were a right femur and a right tibia and fibula from the same individual.68

The fracture of the right femur involved the proximal third of the bone (Figure 8.4). This was a compound fracture with associated secondary infection. Compound fractures are open to the external environment and therefore at risk of bacterial infection, which is what appears to have occurred in this case. Bone infection is described as osteomyelitis, which involves bone destruction and the formation of pus. In this case there are osteomyelitic lesions in the form of sinuses or cloacae, which would have initially formed inside the bone interior as abscesses containing pus. These abscesses eventually penetrated the compact bone wall to enable the pus to drain from the bone to the surface of the leg.69 The broken sections of bone were malaligned, which meant that the femur healed with some angulation. This resulted in the shortening of the bone.

The right tibia and fibula70 (Figure 8.5) are among the few cases in the stored Pompeian skeletal collection in the Sarno Bath complex where an individual is clearly represented by more than one bone. The remains were stored in a basket. They are fragmentary, but include a mandible and a maxilla,71 a right and left tibia and two portions of the right fibula. The features of the maxilla are consistent with a male sex attribution.

Figure 8.4 Healed fracture of a right femur (TdS R 11) with associated osteomyelitic lesions in the form of sinuses on the bone surface

Figure 8.5 Right tibia and fibula from one individual (TdS #28.1) with healed compound fractures exhibiting pronounced displacement and override at the midshaft of the bone

The tibia and fibula display healed compound fractures. The tibia exhibited an oblique fracture with pronounced displacement and override at the midshaft of the bone. This injury has resulted in considerable shortening of the leg. The maximum length of the right tibia is 264 mm as compared to the left tibia, which has a maximum length of 323 mm. There is no sign of secondary infection on the tibia or the remains of the fibula. It is difficult to assess the damage to the fibula as it is fragmentary but it is clear that it suffered a compound fracture with apparent displacement as a result of the same event that caused the fracture of the tibia.

A healed depressed fracture was observed on the remains of a skull in the Sarno Baths (Figures 8.6 and 8.7).72 The occipital bone and parietals are essentially all that has survived of this skull. As a result, it lacks most of the diagnostic features for sex determination. What remains appears rather robust. An age-at-death estimate based on an examination of the ectocranial suture closure on the remaining sutures suggested an age consistent with an individual in, at least, the fifth decade at the time of death. The injury is located on the left parietal bone superior to the anterior articulation for the squamous portion of the temporal bone and just posterior to the lateral portion of the coronal suture. It is roughly circular in shape and covers an area of 34 x 30 mm. Both the inner and outer tables are involved. The wound presents on the inner table as a hemisphere about 23 x 30 mm in area and protrudes about 10 mm from the normal surface of the bone. It was interpreted as a healed depressed fracture without comminution, or splintering of the bone, consistent with a wound made by a blunt implement. The fracture appeared to be fully healed, as evidenced by the rounded and remodelled edges of the site of injury. There was no sign of infection.

Figure 8.6 Healed depressed fracture of a skull (TdS 199), displaying involvement of both the inner and outer tables. View of the outer table

Figure 8.7 Inner table of a skull (TdS 199) with a healed depressed fracture

A healed injury was also observed on the right parietal bone of a skull stored in the Forum Baths.73 It is approximately oval in shape and covers an area of 30 x 10 mm. The affected area presents as an oval region of granulated healing bone, which is ringed by a scar-like bony ridge. It is about 50 mm to the right of the sagittal suture. The outer table and inner tables are both involved though it is only just visible on the inner table. The injury has healed completely with no sign of secondary infection. It has been interpreted as being consistent with a wound resulting from a blow to the head.74

Apart from these cases there were two left femora that displayed pathological changes that were consistent with trauma, but interpretation required confirmation from histological and/or x-ray analysis. Two75 appeared to exhibit healed fractures, one76 of which was consistent with a bending fracture on the proximal third of the shaft.77

Another left femur78 exhibited major changes to the femoral head and a shortening of the neck, which could variously be interpreted as a slipped femoral capital epiphysis as a result of a stress fracture, though the most likely interpretation is hip displacement, Perthes’ disease or osteoarthritis to the head with extensive remodelling and shortening of the neck. Osteoarthritic changes could have been the result of displacement or dislocation as the articular cartilage would have been compromised and would not have been nourished by the synovial fluid, which in turn would have caused the cartilage to break down and the commencement of osteophytic change to the bone.79

A number of bones show evidence of osteophytic change (see below). Interpretation of such changes usually requires an examination of the entire skeleton to establish whether they were the result of arthropathy or trauma. It is possible that some of the osteophytic change, such as the presence of lipping and eburnation on the distal condyles of 11 of the sample of about 320 left and right femora in the Forum Bath collection, may reflect trauma. This interpretation may be supported by the fact that most of these cases display no signs of osteophytic change at the femoral head. It is possible that the local roads, with their deep gutters and unevenly worn stone blocks, ridged with ruts from carriage wheels, could have contributed to such injuries.

Effect of these injuries on survival potential

The perimortem skull fracture80 is likely to have been the cause of death of this individual. The radius, ulna and tibia that had healed without displacement81 would probably not have had an appreciable effect on the individuals associated with them and certainly would not have prevented their departure from Pompeii. While having a shorter leg would not have been an asset, it would probably not have impeded the escape of these individuals associated with the healed compound leg fractures82 in the considerable period of time83 associated with the first phase of the eruption.

The impact of the depressed fracture84 requires further consideration. A fracture of this magnitude, with its involvement of both the tables, is likely to have resulted in some damage to the brain. The area involved corresponds with Broca’s area of the brain, which is concerned with the production of speech. It is notable that for most humans the location of the main language centre is in the left hemisphere of the brain. This is almost independent of handedness, which means that language was probably the function that was affected in this individual regardless of which hand they favoured. The person who sustained these injuries would have had a problem with syntax, especially in relation to the generation of sentences. It is possible that another portion of the brain may have taken over this function over time. It is unlikely that this injury would have hampered the escape of an individual from the erupting volcano.85

Surgical intervention

Only one unequivocal case of surgical intervention was observed in the sample. This took the form of a trephination on a skull,86 which from the degree of healing, had been performed a considerable time prior to death (Figure 8.8). The perforation associated with the surgery is surrounded by an almost circular region of granulated healing bone which, in turn, is ringed by a scar-like bony ridge. Both tables are involved, though the surface of the inner table is not raised. This suggests that the injury is not a depressed

Figure 8.8 Healing trauma, interpreted as a trephination on the skull of individual (TF 74) with hyperostosis frontalis interna

fracture. The bone defect is about 4 x 1 mm. It is in the middle of a depression about 2.5 mm below the normal surface of the surrounding parietal bone. This depression is two stepped and oval in shape. The outer ring measures about 35 x 25 mm. The inner ring is more circular in shape and covers an area of 20 x 20 mm. Elongated growths of healing bone do not quite bridge the defect. There is no sign of secondary infection. The changes observed on this bone are all consistent with a healing trephination.

Trephination, or trepanation, is a surgical procedure that involves incision of the scalp, followed by cutting the skull, often into a circular shape to make a hole and the excision of bone. While most scholars use the terms interchangeably, the distinction between trepanation and trephination is that the former is more general and refers to an opening of the cranium made by any instrument or procedure, whilst the latter specifically relates to the surgical removal of a disk of bone. There is evidence in the skeletal record that this operation has been performed, often successfully, over a wide geographical area for many millennia. It is a very dangerous procedure as it ultimately involves exposure of the cranial cavity to the external environment, though before the problem of potential infection is faced, the surgeon has to avoid damage to the dura mater, which, if perforated, would result in fatal leakage of the cerebrospinal fluid in which the brain floats. The meninges, brain and blood vessels also have to be avoided for the patient to survive the operation.

From the degree of healing, it can be established that the trephination was performed a considerable time before death. The long-term survival of this individual with no sign of osteomyelitis is not uncommon for archaeological cases of trephination. It has been suggested that one of the reasons for the high success rate was the high vascularity of the soft tissue that overlies the cranial vault, which would have decreased the likelihood of secondary infection. The lack of observed infection also indicates the use of clean instruments for this procedure. It is notable that the operation was performed on the left parietal bone. No satisfactory explanation has been established for the observation that most archaeological cases of trephination involve the left frontal or parietal bones.87

Non-skeletal evidence for medical intervention

While there is minimal evidence of surgical intervention in the Pompeian skeletal sample, discoveries of other classes of archaeological material, like medical implements, indicate that it was practised more widely in the Campanian settlements. In Pompeii alone, medical instruments have been identified with certainty at 21 sites and equivocal examples at a further 6. In addition, finds of small collections of about ten tools have been interpreted as home medical kits used for the treatment of minor injuries and ailments.

The range of instruments discovered is impressive and includes scalpels, specula, catheters, tweezers, probes and implements for cauterizing. Identifi- cation of these instruments has largely been based on the considerable documentary evidence for the use of surgical procedures in the Roman Empire in the form of medical texts. The most useful for the Pompeian material is the De Medicina of Celsus, which was written in Italy during the reign of Tiberius. In addition, much medical equipment remained virtually unchanged in form until the first part of the twentieth century and function was easily established by medical practitioners in the late nineteenth and early twentieth centuries.88

Some of the instruments that have been found in excavations of the Vesuvian sites would have specifically been used to treat bones. For example, two bone elevators were identified in the collection that is now housed in the Naples museum, and records suggest that another one or two were discovered in Pompeii prior to 1826 but have subsequently been lost. Like most of the medical implements found in Pompeii, these were made of copper alloy and had a handle with an arched rectangular plate at each end. The plates were smooth on the outside and serrated on the inner surface to assist with grip. A bone elevator, or vectis, was used to lever fractured bones into position so that they could heal without distortion. Bone elevators could also be used to raise depressed bone and it has been suggested that they may have been used to facilitate tooth extraction. As there is no clear evidence that bone files have been discovered at the site, it has been argued that the serrated edges of the bone levers could also have been used for this function.89

About ten implements found in the Campanian excavations have been interpreted as chisels. The exact provenance is not known for a number of these pieces and interpretation is hampered by the fact that they tend to be represented by the bronze handles, rather than the blades, which would have been made of iron or steel, and have not survived. The function of the bone chisel or scalper is largely known through ancient medical writers, like Celsus,90 who describes their use for the removal of diseased or fractured bones, especially from the cranium. He also mentions their application for smoothing out projecting bone from compound fractures after it was filed or cut.91 Apparently, this tool was also employed for the removal of additional digits. It is notable that the bone chisel was generally smaller than those used for carpentry but it was possible that carpenter’s chisels could also be employed for surgical use.92

According to Bliquez,93 the discovery of such tools suggests that surgeons in the Campanian region were equipped to treat severe fractures. In view of this, it is interesting to note that a number of instruments that might have been expected as part of the bone surgeon’s tool kit, such as gouges, bone files, drills and saws, have, so far, not been discovered at Pompeii. It is possible that these tools did not survive as they were made of materials that were not durable in an archaeological context, such as steel, though Bliquez is doubtful that this provides a sufficient explanation. It is possible that the full repertoire of tools has just not yet been excavated.94

Though drills, trephines and saws have not been discovered in Pompeii, Roman methods of trephination deserve a brief consideration in relation to the interpretation of the healing surgical procedure on the skull described above.95 Bow-drills were often used for treatment of diseases or injuries to the skull. They were made of a straight steel or bronze rod with a rotating handle at the upper end and a steel augur or circular saw at the lower end. A piece of cord was attached and passed tightly around the drill. The augur or saw was placed against the bone and was rapidly rotated by the movement of the bow drill, which enabled rapid perforation of the bone. For the removal of small circular pieces of bone, the augur was replaced by a crown trephine or modiolus, which was a short steel tube with teeth at the lower end. After a disc of bone was excised, the edges of the hole were cut back with the chisel until healthy bone was revealed. Celsus96 writes about the use of an instrument to protect the membrane that surrounds the brain during this type of surgery. It is possible that this was a version of a double blunt hook. The instrument was pushed under the bone that was to be removed as a guard to prevent the chisel penetrating the membrane.97

In view of the documentary and archaeological evidence for medical knowledge and practice in the Roman era, it perhaps seems remarkable that there was no apparent attempt to set the compound fractures of the femur and tibia and fibula.98 There are a number of possible explanations for this. One is that these injuries might have been sustained in a geographical region where medical attention was not available. Another is that these individuals may not have had access to medical practitioners for financial or social reasons.

Frequency of trauma in the Pompeian and other samples

The frequency of trauma in the entire bone sample that could be identi fied from gross inspection was about 0.6 per cent. If the 11 cases of osteophytic change to the distal articular surface of the femur could be demonstrated to have resulted from trauma, the percentage would rise to 1.2 per cent. The rate of observed trauma in the radius and ulna was 0.5 per cent. For the tibia and fibula it was also 0.5 per cent and 0.8 per cent for the femur. If the additional eleven equivocal cases were added the frequency for trauma to the femur would rise to 3 per cent.

Bisel documented the cases of trauma that she observed in the Herculaneum sample that was available to her, though she did not distinguish between fractures, dislocations and inflammatory responses. This makes it difficult to compare her results with the Pompeian data. She reported that 32 per cent of the male sample and 11.4 per cent of the female sample displayed evidence of trauma, with a population average of 22.7 per cent. She did not consider that these figures indicated bias towards injury in the Herculaneum sample.99

Capasso recorded cases of fractures separately from evidence of other types of trauma. He reported 31 fractures in 17, or about 10.5 per cent, of the 162 individuals that he studied. Like Bisel, he found the majority of cases involved males, with a ratio of 4.7: 1. He attributed the higher incidence of fractures in males to gender-related division of labour. Four of the fractures involved the frontal bone of the skull, one the temporal bone, one the nasal region and one mandible. There were three fractures of the radius, four of the ulna, one of the humerus and one of the femur. It is notable that Capasso had access to x-ray technology for his study.100

It could be misleading to compare the Pompeian figures for the frequency of fractures with those obtained from other archaeological sites and modern populations as many of these studies would have utilized x-ray technology, and in the case of modern living populations, there is the advantage that all the skeleton is available for investigation. Bearing this in mind, it is instructive to make a brief survey of fracture rates in archaeological samples for comparison with the Pompeian results as they seem rather low compared to those from Herculaneum.

A fracture rate of between 1.0 and 3.6 per cent was reported for Greek and Turkish skeletal samples, dating from the seventh millennium to the second centuryAD. A 1.8 per cent rate of bone fracture was observed for a prehistoric central Californian population, the skeletons from the Libben site in Ohio had a 3 per cent fracture frequency and skeletal samples from various medieval British sites showed a range of between 0.3 to 6 per cent. Similarly, a visual survey of 6000 Egyptian skeletons revealed that about 3 per cent of the sample had sustained fractures. The bones most commonly affected are those of the arm, especially the forearm. Fractures to the femur tend to be fairly uncommon in most archaeological skeletal samples, as are fractures to the tibia and fibula.101

The Pompeian figures are a little lower than most of the above-mentioned populations, but still within the range that has been observed for numerous archaeological sites, whilst the figures from Herculaneum are somewhat higher. It is possible that the survival of complete skeletons and the use of radiography has contributed more to the disparity in observed frequency than possible skewing towards disability. Fifteen of the fractures observed in the Herculaneum sample involve ribs, carpals, metacarpals, tarsals and metatarsals, bones which generally did not survive in the disarticulated Pompeian sample. The possibility of sample bias must also be considered. Ciprotti, for example, mentions the removal to the museum in Naples in the nineteenth century of a femur that demonstrated pathological changes and while it is likely that there were other cases, this activity probably would not have had a significant effect on the post-cranial sample.102 Attention was usually focused on crania in the nineteenth century and they were more likely to be collected. The use of femora for hinge manufacture (see Chapter 5) is unlikely to have contributed to the loss of bones affected by trauma as pathological cases would not have been selected for this purpose since the hinges needed to be as close to perfect cylinders as possible.

Even though the available Pompeian sample does not re flect the entire corpus of excavated human skeletal material, it appears that it is random, with the exception of femora. Though some bones with pathology may have been selectively removed in the nineteenth century, there is no evidence for the removal of specific bones in the major part of the twentieth century. The frequency of observed trauma in the Pompeian skeletal collection is low, which indicates that the sample was not biased towards the infirm. There was no sign of any infirmity in the sample that would have absolutely prevented an unaided escape from the eruption. In fact, it appears unlikely that most of the cases of healed trauma would have had a significant impact on the survival prospects of these people. It is possible, however, that the decision of a few individuals to either flee or remain in Pompeii may have been influenced by the effects of pathology that had resulted in impaired mobility, as in the case of the individual with the malaligned healed compound fractures of the tibia and fibula.103 The cases of healed trauma in the Pompeian sample are testimony to the robust immune systems of these individuals.


While signs of skeletal change that were consistent with infective lesions were observed on some of the disarticulated skeletal elements of the Pompeians, they were nearly impossible to interpret, unless they were related to a specific cause, like a dental abscess or osteomyelitis resulting from a compound fracture.104 The Herculaneum material is much better suited to such studies as the entire skeleton is usually available and patterns of bony change across the body can be interpreted to establish a diagnosis.

Capasso and his colleagues documented a number of bone reactions in response to inflammation on the skeletons of the Herculaneum sample. Capasso and Di Tota noted a small flattened oval area on the outer table of the skull in 37.6 per cent of the adult sample, usually about 20 mm in diameter, with superficial bone remodelling. It was claimed that these reflected healing irritation to the periosteum as a result of scratching to relieve discomfort from infestations of head lice. This interpretation was based on the discovery of a single louse egg found preserved on a hair on the head of one of the individuals who presented with this pathology. This may well be stretching the evidence as it would take quite a bit of scalp scratching to involve the underlying bone. Capasso argued that inflammatory bone reaction observed on the dorsal surface of the first and fifth metatarsals and the superior surface of a tarsal on 25.8 per cent of the adults, was due to wearing Roman sandals and shoes. He also argued that local inflammatory reactions observed on the pleural surface of the ribs of 20.4 per cent of the adult sample were a reflection of lung infections from long-term exposure to particle pollution due to cooking and heating in poorly ventilated living areas. It should be noted that these conclusions are based on very limited evidence. Further, he observed local inflammatory reactions in the maxillary sinus as a result of the spread of dental infection.105

Pathological skeletal changes to the ribs and vertebrae that were indicative of tuberculosis were documented in two adults in the Herculaneum sample. Tuberculosis is caused by a bacterium that belongs to the genus Mycobacterium. Skeletal involvement generally only occurs after long-term illness, which means that most cases are not detected in the archaeological record. Capasso suggested that it was possible that the disease may have been contracted by the consumption of undercooked infected oxen viscera after ritual sacrifice, though there were probably other sources of infection.106

Capasso observed skeletal changes, which he considered to be a re flection of brucellosis, in the lumbar vertebrae in 16 adults of the 162 individuals he examined in the Herculaneum sample. While this disease can occur in juveniles, it does not tend to affect the vertebrae and is unlikely to be detected archaeologically. The human form of this infectious disease is chronic infection of the lungs or other organs, with recurrent episodes of fever. It only presents on bones in about 10 per cent of cases, which means that it is difficult to establish the incidence of the disorder solely from the skeletal record. The vertebrae, especially those from the lumbar region, are most commonly involved, followed by long bones. Infection usually occurs as a result of ingesting dairy products made from infected animals, though it can also be contracted from meat and cuts that have come into contact with diseased animals. It is not easily transmitted between humans. Capasso invoked the discovery of bacteria in the one portion of carbonized cheese that has been discovered in Herculaneum as supporting evidence for his diagnosis. He considered that one of the two types of bacteria that he observed was consistent in dimensions and morphology with Brucella. The heat of the eruption, coupled with the subsequent burial environment, made it impossible to use molecular techniques to confirm identification.107

Subsequent research into the presence of bacteria in other carbonized food remains from Herculaneum led Capasso to suggest that the comparatively low rate of non-specific bone infection that he observed in the skeletal sample he studied could be attributed to the consumption ofStreptomyces,a bacterium that produces a natural tetracycline with antibiotic properties.108

Porotic hyperostosis

Porotic hyperostosis refers to porosity of the cranial vault and/or orbits. It is usually related to the resorption of the underlying cortical bone. This results in thinning and can lead to the complete destruction of the outer table of the cranial vault. Though the exact cause for these bony changes has not been isolated, they have been associated with nutritional deficiencies. It is suspected that porotic hyperostosis is related to anaemia, possibly iron defi- ciency anaemia, though it has also been suggested that it may be associated with high parasite loads. Nutrient losses due to diseases that cause diarrhoea have also been cited as a possible cause. A direct relationship has been observed between infection and porotic hyperostosis, possibly because iron may be diverted to help fight infection. Scholars like Stuart Macadam consider that porotic hyperostosis is a reflection of the interaction between customs, diet, parasites and infectious diseases. Trace element and amino acid analyses have revealed lower iron levels in the bones of archaeological specimens diagnosed as having porotic hyperostosis. These changes, which are also described as cribra orbitalia when they present as a collection of small apertures or pitting on the orbital roof, are considered to be more likely to occur during childhood. It has been argued that active lesions are only found in individuals between the ages of six months and twelve years of age.109

It is possible to distinguish between active and healed lesions. Active lesions have been defined as those which exhibit a sieve-like appearance. They display porosity, which is interspersed with increasingly thin bridges of bone. Healing cribrotic lesions can be recognized by the smooth lamellar texture of the orbital plate with bone filling the peripheral pores. The usual locations of porotic hyperostosis are the orbital surface of the frontal bone or parietal bones. The lesions are usually bilaterally symmetrical. In extreme cases, all the bones of the cranial vault can be involved.110

One hundred and thirty adult and adolescent skulls that were complete enough for assessment were examined for evidence of porotic hyperostosis.111 No significant lesions were observed on the cranial vault. Some degree of pitting, or cribra orbitalia was observed on the orbital roofs of 90 per cent of the skulls. Most of the cases (about 59 per cent) displayed minimal but unequivocal porosity, which could be seen with the naked eye. A number of cases (about 28 per cent) were more equivocal with barely discernible pits, which could only really be seen with the aid of a hand lens. Only about 3 per cent of the sample exhibited a medium degree of expression with coalescing pores. There was no evidence at all of cribra orbitalia on about 7 per cent of the sample and approximately three per cent of the orbits in this sample were too incomplete to assess. All the cases exhibited lesions that could be described as ‘healing’, which is consistent with an interpretation of incomplete remodelling of the bone after an episode of anaemia a considerable time prior to death.112

It is dif ficult to assess the pathological importance of such slight bony changes, though it is notable that they occur with a high frequency. In the past, cribra orbitalia was considered to be a good stress marker. It was interpreted as indicative of poor adaptation by a community to its environment. There has been a shift in attitude in the last decades and the current wisdom is that it is an indicator of successful adaptation to an environment as the presence of lesions reflects the workings of a healthy immune system. Iron-deficiency anaemia has been reinterpreted as a result of an adaptive response by an individual to the total pathogen load of the environment.113

It should be noted that the relationship between porotic hyperostosis and cribra orbitalia has been questioned and it has been proposed that they may not just be different manifestations of the same process. Histological examination of bone structure in the orbital roofs of skeletons displaying cribra orbitalia does not always support a diagnosis of anaemia. Alternative diagnoses include inflammation and osteoporosis.114 Despite the dissenting views, most scholars still tend towards anaemia as the major interpretation for these lesions115 and point out that, regardless of the diagnosis, healing lesions suggest evidence of the body dealing with pathological environmental pressures.

Interestingly, Bisel did not record cribra orbitalia for the Herculaneum victims but did observe evidence of porotic hyperostosis on cranial vaults of the sample she studied. She interpreted porotic changes that she observed on these skulls as slight, healed anaemia. She considered that it was fairly common in the Herculaneum sample, with an overall frequency of 34.1 per cent (n = 98) as compared to 8 per cent in the modern American sample. She suggested that the most likely cause of the cases of porotic hyperostosis to have been heterozygotic thalassemia. It is worth mentioning the influence of Bisel’s mentor, J. Lawrence Angel, on her interpretations. Angel undertook a considerable amount of work in the Mediterranean region. After World War II he developed a particular interest in palaeopathology and palaeodemography. He published his theories about the development of thalassemia in relation to malaria and was inclined to interpret any evidence that may have reflected anaemia as thalassemia. Bisel conceded that there was considerable controversy surrounding the diagnosis of the different types of anaemia from the skeletal evidence. She, nonetheless, chose to follow Angel’s contention that the best way to determine the type of anaemia encountered should be based on a consideration of the ecological context and the entire population, rather than the appearance of any single individual. This view would not be supported by the majority of palaeopathologists as specific types of anaemia have a unique appearance on the bone. The changes associated with thalassemia are very distinctive; for example, the calvarium, when viewed in section, displays no clear distinction between the inner and outer tables and when examined radiologically, the bone trabeculae resemble the bristles of a brush. Most scholars acknowledge that Angel’s interpretation of porotic hyperostosis in the archaeological skeletal record is probably incorrect and that these bony changes are more consistent with anaemia resulting from nutritional deficiency.116

Unlike Bisel, Capasso recorded 12 cases of cribra orbitalia, along with 12 cases of porotic hyperostosis involving the vaults of the skulls of the Herculaneum sample he studied. Sixteen of these victims were male and seven were identified as female. The frequency of porotic hyperostosis reported by Capasso is about 7.5 per cent, which is considerably lower than that diagnosed by Bisel. In the light of the suggestion that they are different conditions, it is notable that there was only one individual that presented with both disorders.117

Osteophytic change

Osteophytic changes are bony changes, such as lipping that can be observed on the articular surfaces of bones. They are associated with bone proliferation, most commonly around the joint margin. Contributory factors include various diseases, diet, trauma and advancing age. The two most common types of arthropathies or degenerative diseases linked with osteophytic change are osteophytosis of the spinal column and osteoarthritis of the synovial joints. Osteophytosis involves growth of bone from the margins of the vertebral bodies as a result of the fibrous capsule of the intervertebral discs. Osteoarthritis is non-inflammatory and characterized by bony lipping and spur formation in response to the destruction of the articular cartilage of a joint. These changes are most commonly observed in load bearing joints, most notably the knees, hips and spine.118

In view of the dif ficulty of establishing the cause and kind of osteophytic change from disarticulated skeletal material,119 this study concentrated on the identification of specific arthropathies that could be diagnosed from a minimal number of bones from one individual. The main interest in the identification of arthropathy was to determine what effect it may have had on the ability of individuals to escape and whether there was evidence of any age-related arthropathy that might provide some insight into the Pompeian lifespan. Virtually every different type of articular surface was represented by at least a few examples of osteophytic change of varying degree. The degree of osteophytic change was routinely recorded, using a four-point scale.120

Osteophytic changes were observed in a number of bones. The cases observed in the femur and left humeral sample provide a reasonable indication of the frequency of osteophytic change in the overall long bone sample. A sample of 320 left and right femora in the Forum Bath collection were inspected. Eleven cases of polishing or eburnation, resulting from friction between the articular surfaces as a consequence of destruction of the intervening cartilage, were observed on the distal condyles.121 Osteophytic change in the form of lipping was observed on the heads of some femora.122It is possible that osteoarthritis may account for the major changes observed on the head of the femur that was also associated with significant shortening of the neck,123 described above in the section on trauma. It is notable that less severe degrees of osteophytic change were not prevalent in the sample.

Marked eburnation was observed on a number of the humeri in the sample, for example on the distal end of seven left bones124 and on the proximal end of one left bone125 in the hundred left humeri in the Forum Bath collection. A lesser degree of osteophytic change was observed on a small number of other bones in the sample.

The only arthropathy that presented in the sample that could be unequivocally identified from disarticulated material was diffuse idiopathic skeletal hyperostosis, also known as DISH or Forestier’s disease. This is because it has a distinctive morphological appearance. DISH is characterized by ossification of the anterior longitudinal ligament, particularly along the right anterolateral aspect of the thoracic vertebrae. This ossification is continuous and has been described as resembling dripping candle wax. It generally affects the thoracic vertebrae. It is usually only diagnosed as present when at least four contiguous vertebrae are fused. The cause of this disorder is unknown but there does appear to be some association with obesity, especially in early life, and diabetes. It occurs more frequently in males and there is a strong correlation between advancing age and the presence of DISH. It is unlikely to be observed in individuals in modern populations who are under fifty years of age.126

Two likely cases of DISH were observed in the Pompeian collection. One case in the Sarno Bath collection is very clear127 and involved four fused thoracic vertebrae. The other128 (Figure 8.9) is more equivocal and comprised a group of two and a group of three fused thoracic vertebrae. These vertebrae were all contiguous; it is possible that there had been at least partial fusion between the groups and that they had been separated as a result of post mortem damage. It is clear that all the bones were involved in the same disease process, with additional bone growth of the same thickness along the

Figure 8.9 An apparent case of diffuse idiopathic skeletal hyperostosis (DISH) observed on thoracic vertebrae associated with TdS NS: 1

right anterolateral aspect of all the affected vertebrae. The appearance of the pathology was far more consistent with a diagnosis of DISH than any other disorder of the vertebrae, such as the autoimmune joint disease, ankylosing spondylitis. The latter case was one of at least four individuals that were excavated in 1986 and were stored in the same box. The other bones that could be associated with this person were the skull and mandible, a left femur and the pelvis.129 Osteophytic change130 was apparent on the articular surfaces of all the bones that could be associated with these vertebrae. The skeleton presented as slightly mid-range but female. The skeletal age indicators, based on the pelvis, skull and teeth, were consistent with an older adult, at least in the fifth or sixth decade.

With the exception of the two cases of DISH, it is dif ficult to assess the osteophytic change observed in this sample. Some of the cases probably do reflect age-related osteoarthritis, though consideration must also be given to the possibility that some result from trauma (see above) or occupational stress. It has been suggested that there is a tendency for individuals to develop premature degenerative joint disease in complex communities which feature specialization in craft and trade as a result of excessive strain from intensive and repetitive activities.131 Whatever the cause, pain and lack of mobility associated with osteophytic changes might well have slowed down or discouraged individuals from making good their escape from the eruption. It is notable, however, that the incidence of eburnation132 in long bones is not particularly high, with an overall frequency of 3.4 per cent in the Forum Bath femora collection (3.1 per cent in the left and 3.8 per cent in the right femoral sample) and 8 per cent in left humeri.

The presence of DISH is merely indicative that there were individuals who survived into older adulthood. It has been asserted that DISH is not a disease but a reflection of the ageing process. It is not considered to be a true arthropathy because it does not involve the cartilage or the synovium. It appears to result more from excessive bone production at joint margins. Despite its appearance, it is not particularly debilitating. This is because it tends to involve the thoracic vertebrae. This means that anterior flexion, which mostly involves the lumbar segment, is largely unaffected. In addition, the disc spaces and the facet joints present as normal. It is hard to correlate the decision to remain in Pompeii during the eruption with the presence of an age-related disorder, like DISH. While afflicted individuals would hardly have been advantaged, escape would have still have been possible. The presence of osteophytic change in the Pompeian sample is low and cannot be used to support the notion that only people with infirmities were unable to escape the eruption.133

Bisel recorded a slight to moderate degree of vertebral arthritis in 47.5 per cent of the males in the Herculaneum sample and 36.4 per cent of the females. She explained the sex difference in frequency to be the result of males having engaged in heavier work than females, though no evidence was presented to substantiate this claim. She compared the combined frequency of 42.5 per cent for vertebral arthritis in the sample with a rate of 67 per cent in that recorded for a modern American sample. She explained the disparity as the result of the American sample being composed of older individuals. She concluded that the American data reflect degenerative disease due to age rather than stress, which she invoked as the most likely cause for vertebral changes in the Herculaneum victims. The diagnosis of a case of DISH in her sample is consistent with this interpretation.134

Capasso also reported high rates of osteophytic change to the vertebral columns of the victims in his sample. He diagnosed five cases of DISH, all in male skeletons, though it should be noted that Becker questioned the accuracy of the diagnosis of this disorder in all these cases. Osteophytic changes were recorded at articular surfaces of bones in 69 of the victims and he noted that these changes were more prevalent in males than females at all locations with the exception of the region around the knee, where they were equivalent.135

Hyperostosis frontalis interna

The discovery of bilaterally symmetrical deposits of bone overgrowth on the inner table of the frontal bone of a skull in the Pompeian sample led to the investigation of the available crania for evidence of hyperostotic change. Complete crania were examined through the foramen magnum with the aid of a torch and direct observations were made of the surfaces of incomplete specimens in the sample of 360 skulls that were well enough preserved to enable observations to be made. Forty-three skulls were identified as having some degree of hyperostotic change.136 These bony growths have been diagnosed as hyperostosis frontalis interna (HFI), a disorder generally associated with post-menopausal women that has arguably been described as a syndrome with a suite of signs and symptoms, including obesity, hirsutism, non-insulin dependent diabetes and headaches. HFI can be diagnosed from the skull alone, which means that the disarticulated nature of the sample does not have a significant effect on the confidence of diagnosis.

Classi fication of the degree of bony involvement was largely based on the system proposed by Henschen.137 This system is roughly equivalent to the one more recently proposed by Hershkovitz et al.,138 which is also based on the degree of tumour development and the area of bone involvement.

The features observed on 40 of the 43 crania in this sample concur in all respects with the descriptions of hyperostosis frontalis interna in the literature.139 The majority of cases that were identified as having bony changes consistent with HFI displayed only very slight deposits of additional bone on the inner table of the frontal bone. Seven cases were described as slight, five as moderate (Figure 8.10), and one skull exhibited pronounced bony changes (Figure 8.11). There was only one skull in this series, which demonstrated extensive bony tumorous swellings (Figure 8.12).140 The three most

Figure 8.10 Inner table of the frontal bone of a skull (TdS 80) displaying moderate hyperostosis frontalis interna (HFI)

Figure 8.11 Inner table of the frontal bone of a skull (TdS 28) displaying pronounced hyperostosis frontalis interna (HFI)

Figure 8.12 View through the foramen magnum of a skull (TF SND) displaying extensive hyperostosis frontalis interna (HFI)

equivocal cases were not included with the diagnosed cases of HFI. The bony growths associated with these cases tended to be rather asymmetrical, though it has been suggested that HFI can present as asymmetrical in the early stages of development.141 Also, some of the cases may have appeared asymmetrical because the skulls were incomplete. When it has been possible to reconstruct skulls with this disorder from fragments found in the Sarno Bath collection, the expression does appear both greater and more symmetrical.

Hyperostosis frontalis interna is anatomically characterized by the occurrence of bilaterally symmetrical, benign bony tumours and marked thickening on the internal surface of the frontal bone. The midline is generally spared, which gives it a ‘butterfly-like’ appearance. When viewed in crosssection, the new bone is mostly revealed to be cancellous and integrated with the diploë, or spongy tissue between the inner and outer tables of the cranial bones. Diagnosis of this disorder is based on the existence of such growths, irrespective of their degree of expression. HFI has been classified by some scholars as a symptom of various syndromes but current wisdom is that it is a distinct pathological entity. It has been described as the most obvious skeletal indicator of an endocrine disorder of uncertain aetiology, or origin, and there has been considerable discussion about its clinical manifestations. It has been suggested that HFI is associated with hyperlactinaemia. This results in an increase of adrenal androgen production, carbohydrate tolerance and hypothalamic hormone secretions, all of which would contribute to the occurrence of obesity, hirsutism and diabetes. Other suggested contributors to the development of HFI include estrogens, dietary phytoestrogens, parathyroid hormone, calcium modulating agents and neuropeptides.142

Based on unselected samples in hospitals, the frequency of hyperostosis frontalis interna in modern European populations is generally thought to vary between 5 per cent and 12 per cent, though some scholars suggest the incidence might be as high as 15 per cent.143The frequency of HFI in the Pompeii sample, at between 11.1 per cent and 11.9 per cent, is consistent with this range.

Surveys of the incidence of HFI in modern Western populations indicate that HFI is most commonly associated with females.144 It has been, perhaps rather extravagantly, suggested that the correlation between the occurrence of HFI and women is so high that its presence alone should be sufficient as a basis for sex attribution from archaeological skeletal material.145

It is most frequently observed on post-menopausal women, with a reported incidence of 40–62 per cent amongst this group in modern populations, and has been associated with adiposity and male-type hair growth patterns, though the three features do not always occur together. Various other signs and symptoms can accompany HFI, the most common being headaches. HFI-associated headaches apparently are hormonally induced and not related to excess bony development. The degree of bony change to the inner table of the frontal bone does not necessarily reflect the degree of development of other signs and symptoms, though it has been suggested that the severity, like the frequency, appears to be age dependent.146 Nonetheless, it probably would not be reasonable to draw specific conclusions for the variation in expression observed in the Pompeian sample.

There is some suggestion that there may be a genetic component in the manifestation of this disorder. It has been detected in four generations of one family, and in individuals from archaeological contexts who possibly were related but, as yet, there have been no definitive inheritance studies.147

The Pompeian skulls diagnosed with HFI were examined to determine whether the presence of this syndrome had any bearing on the survival prospects of affected individuals in the AD 79 eruption or whether these cases just reflect the normal incidence of HFI in the ancient Pompeian community.

Attribution of sex and estimation of age-at-death

Since HFI is both sex and age-related, it was necessary to establish both for the sample, but there can only be limited confidence in sex and age-at-death attributions based solely on skulls, especially ones that are not well preserved.

The majority of the skulls148 exhibited more female than male characteristics, which is consistent with the greater female prevalence of this disorder that has been observed in modern populations. One skull was completely mid-range, five skulls appeared to be more male than female and one skull presented as male. Another skull was so incomplete that no diagnostic features were retained and no sex attribution could be made.149 It is possible that some of the six skulls, which display predominantly masculine traits, may be female. Virilism, in the form of the development of masculine facial features, has been associated with HFI and such changes could be detectable on the skull. In addition, it is not uncommon for the skulls of normal older females to develop male traits.150 It should also be reiterated that the confidence levels for sex determination solely from skulls in the general Pompeian sample were not very high (see Chapter 6).

Though the skulls were generally too incomplete to enable a full set of observations to be made, it was still possible to see general trends in terms of relative age-at-death. The majority151 could be identified as adult, with most of the individuals tending towards older ages. Two skulls lacked sufficient diagnostic features to establish age-at-death but did not appear to be juvenile. Six skulls could be identified with certainty as adult but they were too incomplete to enable further assessment. Seven skulls presented as consistent with an age estimation of, at least, the third decade, ten the fourth decade or older, the minimum ages of a further 14 were consistent with the fifth decade and four individuals with, at minimum, the sixth decade.152

It must be remembered that the ages attributed to nearly all of these cases reflect minimum ages-at-death as it was not possible to build up a complete score due to the poor preservation of the skeletal remains. It is therefore likely that most of the individuals in this sample were chronologically older than their estimated ages. Further, age-at-death attributions are notoriously unreliable for adult skeletons, especially when based solely on features of the skull. At best, the age markers that were employed could be used to seriate the sample. The results certainly demonstrate skewing towards the older age range, which would be expected for a sample of individuals with hyperostosis frontalis interna.

As for sex, the correlation between maturity and HFI is so high that it has been suggested that adult age can be reliably estimated for archaeological skulls that display some degree of HFI.153 With some qualifications, the results obtained from the Pompeian sample for sex and age are consistent with a diagnosis of HFI.

Alternative diagnosis

Differential diagnoses have been considered for the archaeological cases of bony lesions that have been interpreted as hyperostosis frontalis interna.154 The features that have been described for at least 40 of the cases of hyperostosis in the Pompeian sample are characteristic of HFI with clear boundaries, which limit overgrowth to the inner table of the frontal bone, an unaffected midline and overall bilateral symmetry. Other disorders that are associated with additional cranial bony growth, such as Paget’s disease, senile hyperostoses, Leontiasis ossea and acromegaly, are easily distinguished from HFI as they do not tend to be confined to the frontal bone and they involve both tables. Leontiasis ossea is further excluded as a possible interpretation because it results in the destruction of the frontal sinuses. In all the cases in the Pompeian sample where the bone was complete enough to assess, the frontal sinuses appeared normal.

The lesions observed in the Pompeian sample, including the more equivocal cases, are too large to be diagnosed as pregnancy osteophytes. These appear as a thin chalky layer of surface parallel periosteal bone, most commonly on the outer table, though those that are observed on the endocranial surface are usually found in the frontal region. The growths are usually less than 0.5 mm in thickness. These changes, along with an increase in the weight and density of the cranial vault during pregnancy, have been attributed to changes in pituitary hormone secretion.

Caffey ’s disease, or infantile cortical hyperostosis, could not be considered a reasonable alternative diagnosis as it only occurs in young children. It affects the skeleton of infants in the first year of life and is characterized by a large deposit of layered periosteal woven bone on one or more bones. The bones that are most frequently involved are the mandible and the clavicle. The skull generally suffers no bony change.

Fibrous dysplasia is not implicated as an alternative diagnosis because it presents quite differently to the observed hyperostotic changes in the Pompeian cranial series. The cause of fibrous dysplasia is not understood. It can occur in single or multiple bones and is often confined to one limb or one side of the body. It is more commonly found in females and is manifested as faulty differentiation of the parts of the osteogenic mesenchyme into fibro-osseous tissue. The lesions associated with this are characterized by fine trabeculae of woven bone.

Another unlikely alternative diagnosis is other types of neoplasm or tumour. These tend to be more destructive than HFI and usually involve both tables.
Though it appears likely that the diagnosis of HFI is the most reasonable option, it is important to note that as the sample was disarticulated, it was impossible to assess the Pompeian sample for post-cranial pathological involvement. No clear alternative diagnosis is apparent for the three cases that were marked by some degree of asymmetry and tubercules.

Additional pathology

All bony changes to the skulls that could be attributed to a pathological cause were recorded, both to confirm that the diagnosis was correct and to establish whether there was any linkage between HFI and additional pathology that was present on these skulls. The range of observed pathology included porotic hyperostosis, ante mortem tooth loss, dental abscess, interdental alveolar resorption, button osteoma, osteophytic change and trauma.

No well-developed cases of porotic hyperostosis were observed in this sample and those cases only presented as cribra orbitalia (see above). Of the 37 skulls that were sufficiently preserved to enable observations to be made, 32 exhibited minimal to slight pitting, with only one case demonstrating a medium degree of expression. Four skulls did not display any sign of cribra orbitalia. The incidence and degree of expression observed on the skulls diagnosed with cribra orbitalia was consistent with that observed in the overall skeletal collection and does not appear to have any particular relationship to the presence of HFI in the sample.

Since the causes of various dental problems are related, it is appropriate to consider dental pathology as a whole. Some degree of pathology was observed on the majority of the surviving maxillae, though only eight cases were sufficiently preserved to enable dental observations to be made. Post mortem tooth loss prevented the assessment of caries or enamel hypoplasia in this sample. Slight alveolar resorption was observed on the maxillae of six individuals, one displayed medium resorption and no sign of resorption was visible on the remaining maxilla. Six of the eight maxillae displayed evidence of loss of between one and five teeth some time prior to death. Abscesses were recorded in the maxillae of five of the eight cases where observations could be made. Evidence of one abscess was observed in four cases and three abscesses were evident in one case.

Dental pathology in these cases could have resulted from various causes. It is worth considering, however, that such dental problems can be exacerbated by age-related changes to teeth, such as attrition, and in most populations occur with increasing frequency as an individual ages.155 It is impossible to draw conclusions from such a small amount of material, beyond the inference that the presence of such pathology may suggest older individuals, which, in turn, would be consistent with the identification of HFI.

Two cases of button osteoma were observed in this sample. These usually occur on the outer table of the cranial vault and present as a smooth lump of compact bone with a maximum diameter of two centimetres. The occurrence of these benign tumours is independent of HFI. They were probably asymptomatic and had no impact on the individuals who had them.156

Osteophytic changes to the articular surface of the temporal part of the temporo-mandibular joint were discerned on four of the seven skulls that were sufficiently preserved to enable assessment. It would be impossible to attribute a specific cause for these changes, though they could be associated with increasing age, some degree of malocclusion or tooth loss. The possibility that these changes might be age-related is also consistent with HFI.

Evidence of trauma was only observed on one skull. This pathology had features consistent with changes that can be observed on healing trephinations (see above). Trephination, a surgical procedure involving the removal of a portion of the cranial vault, was performed for a number of reasons in antiquity, including the excision of splinters from fractured skulls, alleviation of the symptoms of epilepsy and the treatment of headaches.157 There is no apparent reason for the performance of this operation, though the presence of hyperostotic changes may be invoked to postulate that the trephination might have been undertaken to relieve headaches that have been associated with HFI. If this were the case, the operation, though successful in that there was healing well with no sign of secondary infection, would not have been of assistance, as such headaches would have a hormonal origin.

To summarize, all the pathological changes observed on these skulls could be interpreted as either consistent with HFI or as independent entities. No pathology that conflicted with the diagnosis of HFI was identified in the forty crania that were unequivocally diagnosed.

Features consistent with HFI in Pompeian wall painting?

A wall painting from the Casa di Adone Ferito (VI, vii, 8) depicts a naked hermaphrodite surrounded by three attendants, one of which appears to be a woman sporting a rather luxuriant beard.158 The interpretation of the bearded individual in this painting is problematic and it could variously be explained as either a male in female garb, or as a woman with a disorder. This painting was described at length in the nineteenth century by Raoul-Rochette.159 He described the figure in question as being sexually equivocal. He considered that the beard was consistent with a male attribution but the dress and the long curly hairstyle that fell down over the shoulders were indicative of a female. Further, the mirror that was held in the right hand of this figure was interpreted as a feminine accessory. The style of dress worn by this figure, with one shoulder bared and a piece of material covering the top of the head, was claimed to be one that was favoured by women and certain effeminate men. Despite the fact that this figure is barely taller than the female attendant on the far right of the image, Raoul-Rochette cited the height as an additional argument for a male attribution. He concluded that this figure represented a male dressed in female clothing as a reflection of certain Eastern religions where a deity could have both male and female attributes. The figure of the hermaphrodite was used to support this view as Hermaphroditus, the child of a union between Aphrodite and Hermes, has often been identified with Oriental deities with combined sex.160

A certain degree of caution should be exercised with the use of art for the diagnosis of medical conditions because such interpretations are based on the assumption that art works are literal images of what the artist observed. This is obviously not always the case as attested by the mythical elements that can be observed in this painting, like the hermaphrodite and the Eros in the foreground. This hermaphrodite is generally interpreted as Hermaphroditus, who is commonly encountered in Pompeian wall paintings.161 Nonetheless, it is possible that elements of such works were informed by the experience of the artist. A good example can be seen in the depiction of pygmies in wall paintings and mosaics across the site. The pygmies in the numerous Nilotic scenes that have been found in Pompeii and other Campanian sites are all achondroplastic dwarfs.162 It is likely that the artist who set the standard for the depiction of pygmies had never seen a real pygmy but had seen an achondroplastic dwarf. An argument can therefore be mounted that the bearded attendant in the painting from the Casa di Adone Ferito could well have been a representation of a person that the artist had actually seen. In view of the number of cases of HFI that have been observed in the Pompeian skeletal sample, it is tempting to consider the possibility that this painting may represent an individual displaying a sign of this syndrome.

Survival prospects of individuals with HFI

The multiple occurrence of HFI in this sample raises the question of the survival prospects of affected individuals trying to escape from the AD 79 eruption. It is likely that these victims were in the older age range. There is also the possibility that at least some of the cases were associated with obesity. Although HFI has been linked with various disorders, such as noninsulin dependent diabetes, headaches and psychiatric problems, no causal relationship has been established.163 As a result, it would not be reasonable to assume the existence of these diseases from the skeletal record.

Obesity was certainly known in Campania in this period, as evidenced in the writing of the younger Pliny. In his letters to Tacitus describing the eruption of Mt Vesuvius and the death of his uncle,164 he mentions that both Pliny the Elder and his mother were somewhat corpulent. His description of his relatives does not suggest that their obesity was at all remarkable for that period. It is possible, but by no means certain, that adiposity was a contributing factor to the respiratory failure that apparently resulted in Pliny the Elder’s death.165 Pliny the Younger’s account of his flight from Misenum includes a passage where his mother entreats him to leave her behind as she was old and stout and would have possibly threatened his survival by slowing down his escape.166 In the end, however, she did accompany her son and they both survived.

Although obesity, old age and other possible disorders would certainly not have advantaged individuals in their attempt to flee from Pompeii, there is no compelling evidence that, in the absence of more serious pathology, older, fat people would have been much less likely to escape than the rest of the general population.

If HFI had been a factor, which in fluenced survival prospects, one would have expected to find much higher percentage rates in the sample. Instead, the observed frequency of HFI suggests that the Pompeian skeletal sample reflects a normally distributed population rather than a sample biased by a high incidence of pathology.

Hyperostosis frontalis interna in the archaeological record

Until comparatively recently, few cases of HFI were reported in the archaeological literature and the interpretation of these cases is telling.167 Armelagos and Chrisman reported a female Nubian skeleton dating to about 300 AD. Gladykowska-Rzeczycka also recorded a case of HFI in a skeleton that was interpreted as that of an elderly female and was found in an archaeological context in Poland. Anderson identified a case of HFI in an AngloSaxon skeleton from Kent. The authors of these papers conclude that HFI only occurs rarely in archaeological remains, possibly as a result of a shorter lifespan in antiquity. Anderson suggested that the case he described, whom he aged between 30 and 40 years, may be a reflection of the earlier onset of menopause in antiquity, despite the fact that there is no biological justification for this argument.168

Hyperostosis frontalis interna is not of great clinical signi ficance but it is a good health indicator as an individual has to survive long enough for it to be expressed. It can be difficult to detect in complete skulls and it is possible that scholars did not attempt to look for it because it may have been unexpected due to the often held assumption by physical anthropologists that lifespans were shorter in antiquity. The potential of HFI as a longevity descriptor has been recognized in recent years and it has been more actively investigated in the archaeological record. As a result, the last decade has seen a number of new cases of HFI reported from archaeological contexts, though most of these only involve one or two individuals.169 Hershkovitz et al. surveyed a number of these cases and also concluded that HFI was rare in historic and prehistoric populations. They argued that in the Western world, the ‘industrial revolution’, in the mid-eighteenth century, marked the turning point in the increase in frequency as a reflection of increased longevity in females.170 The frequency of HFI in the Pompeian skeletal record and the identification of the disorder in 12 out of 37 adults from Pueblo Bonito in New Mexico challenges these claims.171

Small sample sizes have led to some interesting claims about HFI, including the suggestion that it was more common in males in antiquity.172 The majority of cases in the multiple findings of HFI in the skeletal record in Pompeii and Pueblo Bonito tend to be female, which indicates that the sex-related nature of this disorder has not changed over time. It is possible that the apparently higher prevalence of males in isolated archaeological cases may be an artefact of small sample size and sex identification based on incomplete skeletal remains.

Few cases of HFI have been reported from Italian sites. These include a female skull from Isola Sacra, dating between the first and third centuries AD and two sexually indeterminate individuals from the Early Middle Age necropolis of Vincenne Campochiaro in Molise. These individuals were cautiously interpreted as female.173

Given the large number of cases identi fied in the Pompeian sample, it is surprising that only one possible case of HFI has been reported from the Herculaneum sample that has been studied. Capasso mentions that two out of 162 individuals he studied presented with hyperostotic changes to the skull, but only one of these conforms to a possible diagnosis of HFI. He noted that there was evidence of hyperostotic change in the skull of an older female that was consistent with a hormonal disturbance that is typical of menopause. He estimated that this individual was about 50 to 55 years of age when she died. Unfortunately, he did not describe the pathology in great detail and did not refer to standard sources for classification, so assessment of his diagnosis is problematic.174 It is extremely difficult to account for such a significant difference in the finds between these two sites. If this is correct, it suggests that there may be real distinctions between the Pompeian and Herculaneum samples and that they are not interchangeable, even though they are contemporary, geographically very close and all the victims died as a result of the same event. It is possible that Capasso, whose study included radiographic analysis of each skull, may not have observed HFI, especially if it were in the early stages. It has been noted that HFI can be difficult to detect on x-rays.175 It would certainly be worthwhile to reinvestigate the Herculaneum crania with direct observations on the bone to ascertain whether the frequency of HFI really is so low.

It is probable that the cases of HFI in the present study re flect the incidence of this syndrome in the ancient Pompeian sample. The observed frequency of HFI in the 360 Pompeian skulls that were examined is between 11.1 per cent and 11.9 per cent, which is in the upper end of the range for the incidence cited by the literature of this disorder in modern populations. It should, however, be reiterated that the reported frequency of HFI in modern populations is probably an underestimate as it is generally sub-clinical and is often only incidentally reported. Since the presence of HFI in the Pompeian skeletal record is consistent with its apparent incidence in a modern population, it appears more likely to be a reflection of the frequency of its occurrence in the Pompeian sample than a major contributing factor to the demise of individuals when Mt Vesuvius erupted. This suggests that the total Pompeian skeletal sample is representative of a normally distributed population with no bias towards pathology.

Biochemical analysis for the determination of diet

Trace element analysis was conducted on the essentially complete skeletons retrieved from the House of Julius Polybius (IX, xiii, 1.3) in an attempt to establish the diet of the victims. It would not generally be considered appropriate to attempt biochemical analysis on the compromised and disarticulated Pompeian sample as no evidence of the archaeological context remains and it would be impossible to account for possible diagenesis.

Bisel, Capasso and Petrone et al. undertook independent trace element analyses of the Herculaneum material. Bisel pioneered trace element study on skeletal material from the region. Both human and non-human bone was analyzed by atomic absorption spectroscopy for calcium, magnesium, phosphorus, strontium and zinc. Soil samples taken from the same region as the bones were analyzed for calcium, magnesium, strontium and zinc. She considered that strontium would provide more information about diet than the other trace elements. This is because the strontium/calcium ratio in bone is said to indicate the proportion of animal or vegetable protein consumed during life. Calcium and strontium levels in plants are equivalent to those in the soil. As a result of this, herbivores have a higher level of strontium in their bones than carnivores. This scheme is complicated by the fact that sea creatures naturally contain high levels of strontium, which makes it difficult to separate a largely seafood diet from a vegetarian one. She argued that the data she obtained for this seaside town implied that terrestrial animals did not provide the main protein source for the Herculaneum sample and concluded that the Herculaneum population probably relied on vegetables, seafood or a combination of the two. Subsequent studies on skeletons from Pompeii and Herculaneum produced largely comparable results.176

Unfortunately, more recent research casts doubt on the results of all these studies. While standard techniques were applied, trace element analysis has been superseded by stable isotope analysis, which produces much more reliable results. As yet, there has not been any attempt to establish whether stable isotope analysis would be appropriate for the skeletal remains of the victims of the AD 79 eruption.177 Current studies of refuse from the kitchens and latrines of Pompeii and Herculaneum may ultimately provide the best source of information about the ancient Campanian diet.178

Lead content in the bones

It has been claimed that lead poisoning was the primary cause for the demise of Roman civilization.179 Though the arguments for this contention are somewhat spurious, it continues to influence popular thought to the extent that scholars embarking on skeletal analysis of Roman remains are compelled to, at least, test the hypothesis that Roman populations suffered from significant lead exposure.180

Lead is a toxin which can affect all human organs if absorbed in signi ficant quantities. When ingested, it is stored in bone and other tissues. The effects of continued exposure to lead are cumulative. It has been cited as a factor in retarding intellectual development in small children. Large doses can cause anaemia, intestinal spasm, nerve weakness or paralysis, kidney damage, changes of mood, convulsions, coma and death. In theory, the presence of lead in bone should increase with exposure and one would expect to see higher lead content with age. It can be leached out, however, as bone is remodelled. Leaching can also occur with age as a result of decalcification due to osteoporosis, or as a result of starvation or serious illness. A number of scholars, including Gilfillan, have claimed that the ingestion of lead produces sterility, miscarriage, stillbirth and a high level of infant mortality. Laboratory studies on animals indicate that there is a relationship between the presence of lead and decreased gonadal function and hence fertility, but there is no conclusive evidence that this is true for humans.181

Gil fillan argued that the use of lead, especially in vessels for cooking and the manufacture of wine, was responsible for the decline in the birthrate of the Roman aristocracy from the second century BC. He considered that this was responsible for the so-called fall of the Roman Empire.182 By this he meant the decline of the Roman aristocracy, to whom he attributed all the major cultural achievements of the Roman world. That this happened at all is questionable and is apparently based on the Ius Trium Liberorum Lex Papia Poppaea of Augustus. This law, dated to AD 9, was designed to provide incentives to Roman citizens to produce no less than three offspring. Small families were discouraged by the imposition of penalties. This law has been interpreted as proof of a declining level of fertility among the Roman upper classes. A decrease in birth rate, however, can be due to a number of causes. For example, planned families of small size are increasingly common in modern, urban affluent societies to ensure that standards of living are maintained.

Regardless of whether skeletal lead content has any fertility-related implications, lead levels in bones can still provide useful information about a society with a partially lead-based technology.183 It was not possible to determine lead levels in the Pompeian skeletal sample as there was no way to account for diagenesis, such as the potential leaching of lead into or out of the bones during the period in which they were buried.

Bisel examined samples of cortical bone, taken in most cases from the tibia, of 138 skeletons. Soil samples were analyzed for diagenetic purposes and concluded that, for the majority of cases, there was no reason to believe that the lead levels she observed in the bones were the result of anything other than exposure or ingestion during life. She did not find any correlation between age and lead levels in her sample. To test the assertion that significant levels of lead in the bodies of a large portion of a population would have resulted in increased sterility, a lower birth rate and a higher incidence of stillborn infants, Bisel attempted to find some correlation between bone lead levels and parity in women (see below). The Herculanean mean lead level for males was higher than that obtained for females. The higher lead level in the bones of the Herculanean males may have been a reflection of gender-based occupational differences but there was insufficient evidence to provide proof. There was no apparent separation into distinct social groups.184

Bisel surveyed the ancient literature and the physical evidence in an attempt to establish possible sources of lead contamination and the extent of lead usage in the Roman world. She cited Ovid, Plautus, Celsus and Pliny the Elder for the various medicinal and cosmetic applications of lead in the Roman world, as well as the use of lead acetate to sweeten wine. Apparently, lead was used for the treatment of eyes, ulcers, fissures of the anus and haemorrhoids and, when mixed with vinegar, was employed to create a fairer complexion.185

Bisel also mentioned Vitruvius ’ reservations about the use of lead for water pipes to indicate that the Romans were aware of the dangers of exposure to lead. Bisel noted that the employment of lead pipes at Herculaneum and other towns in the region of Mt Vesuvius would not have presented a significant health risk to their inhabitants. The hardness of the water in this area meant that the interior of the water pipes was coated with a deposit of calcium carbonate, which provided a barrier against the lead. The people most in danger of lead ingestion would have been those directly involved in mining, smelting and the manufacture of lead objects.186

The possible impact of lead on parity and evidence for palaeofertility in Herculaneum and Pompeii

Bisel estimated parity by applying Angel ’s technique, which was based on the now discredited assumption of a direct correlation between the degree of distortion to the dorsal rim of the pubic symphysis and the number of pregnancies that had come to term.187

Bisel calculated the mean number of births from the 37 adult women she examined to be 1.69. She claimed that this level of parity would not enable the population to be maintained and argued that this problem would have been exacerbated by the additional loss of individuals as a result of stillbirth or juvenile mortality. The mean number of babies born to women over the age of forty was determined to be 1.81. Bisel considered that women in this age group would be no longer able to bear children. The correlation between the observed lead levels in bones and parity was found to be very weak. This led her to consider other explanations for the apparent low birth rate, such as the practice of abortion, contraception, homosexuality and abstinence.188

Capasso also invoked changes to the pubic symphysis, as well as the preauricular region to establish the fertility of the sample of 40 female pelves he studied. He concluded that the mean number of births per woman was between 2 and 2.7 and that the total number of pregnancies that had come to term for women who were aged between 45 and 50 years and, apparently at the end of their fertile period, was between 4 and 5.189

While Henneberg and Henneberg also interpreted so-called pits of parturition on the dorsal surface of the pubic symphysis as indicative of the number of children a woman had borne, they used a different method to establish the fertility and birth rate for the ancient Pompeian sample. They calculated the total fertility rate on the basis of a series of assumptions, including a stationary population with a death rate that was reciprocal to newborn life expectancy. They arrived at a result of between 6 and 7 children being born to each woman surviving to menopause, with more than half of these failing to reach sexual maturity.190

The notion of establishing fertility rates for the ancient Pompeian and Herculaneum populations is very appealing but, unfortunately, these figures are not credible as none of the described skeletal techniques can provide this information. It is unfortunate that a number of scholars persist in using socalled pits of parturition to determine fecundity despite the fact that most academics would not consider them to be a reliable indicator (see Chapters 1 and 6). Further, it is unlikely that the Hennebergs’ assumption that the AD 79 population of Pompeii was stationary is correct (see Chapter 4). It is, therefore, not possible to develop any arguments about the demography of these sites based on any of this work.

Lifestyle indicators

Establishment of lifestyle is a very optimistic branch of skeletal identi fication and scholars are often required to stretch the evidence to provide interpretations. It is largely based on the premise that habitual use of bones and muscles can cause specific development that may give some indication of the activities that the individual engaged in during their life. Interpretation is stymied by the fact that a number of different activities can produce similar skeletal alteration or development. It is therefore extremely important to have a good knowledge of the cultural context of the skeletal sample under investigation but, even with that knowledge, it is difficult to isolate a single cause for the majority of changes. Most archaeological material would be better served by documentation of observed skeletal modification with a suggested range of possible activities that could have produced such alterations rather than just one explanation. The problems associated with attempting to establish status from archaeological skeletal material are discussed in Chapter 1.191 These studies are difficult enough to justify with complete skeletal preservation so no attempt was made to undertake such a study on the compromised Pompeian sample.

Both Bisel and Capasso attempted reconstructions of the lives of various individuals in the Herculaneum skeletal samples that were available to them. A few of the cases that Bisel presented have already been discussed in Chapter 1. These and reconstructions of the health, occupation and status of several others were published posthumously by her daughter.192

Two of the female victims studied by Bisel were considered to have been prostitutes on the basis of observed changes to the pubic region of their pelves. Similar skeletal alterations were alleged to have been visible on the pelvis of a modern American prostitute. Both individuals presented with very degenerated pubic symphyses and small bony outgrowths along the ventral surface of the rami.193 The suggested career path of these individuals is rather spurious, especially since it is based on a random observation on one individual from an entirely different cultural context. No proof is presented to justify why these changes would be indicative of the activities of a prostitute.

A more convincing case is mounted for industrial use of the teeth of an immature individual. An adolescent with a suggested age of about 16 years was observed to have worn right central and lateral maxillary incisors. Based on ethnographic analogy, the wear pattern was found to be consistent with that observed on the teeth of fishermen who hold the bobbin of cord with their teeth when they make repairs to nets. Well-developed musculature was claimed to provide corroborative evidence.194 Capasso identified 18 cases of what he considered to be evidence of industrial usage of anterior teeth in the sample he studied. Like Bisel, he interpreted the wear as consistent with that produced from working with paraphernalia associated with fishing. He argued that fishing was a male-dominated industry and since 15 of these individuals presented as male, his conclusion was further supported.195

Tooth wear can result from habitual activities, like industrial or recreational use of teeth. Activities that could produce distinctive wear patterns include the use of teeth to assist in mending or breaking threads, the holding of nets, opening hard objects or, in more modern populations, clenching around a pipe. The patterns for particular usage are consistent and, especially if one knows the culture of the population under investigation, can be recognized and interpreted with a fair degree of confidence.196 Nonetheless, such interpretations must be made with due caution as there can be more than one cause for specific wear patterns.

Capasso identi fied one of the skeletons that he examined as that of a boxer because it displayed pathology that he considered to be consistent with a long-term practitioner of this sport.197 This individual was described as a very robust male of about 35 to 40 years of age. The skeleton displayed a number of healed injuries, including fractures of the rib, the nose and the fifth left metacarpal, as well as evidence of a number of traumatic insults to the right side of the body. These injuries, the fractured metacarpal in particular, led Capasso to the conclusion that this person engaged in violent activities during their life. The fractured nose, with a deviation towards the left side, was argued to be typical of the type of injury sustained by boxers. Capasso illustrated this point with an ancient bronze image of a boxer with a broken nose and a radiograph of a modern boxer with a nasal fracture.198 He also suggested that bony changes observed on the hands, especially on the metacarpals and phalanges, was consistent with continuous use under force of the flexor muscles of the fingers. Of all the evidence presented, perhaps the changes to the hands could probably be argued to support this hypothesis, but it should be noted that this individual had numerous other healed injuries, for example to the left tibia and right femur, that could not be specifically related to a pugilistic career. The broken nose is hardly compelling evidence as they are not solely produced by suddenly meeting with a fist. The evidence of trauma on this individual does not necessarily reflect professional activity and could be explained as the result of a number of other activities or events.

While reconstruction of the lives of victims is an extremely attractive proposition, these few examples highlight the limitations associated with interpretation of the skeletal record for this purpose.


Despite the dif ficulty in obtaining more than a vague indication of general health from the skeletal record, especially that of the disarticulated Pompeian sample, it is still possible to gain some insight into the health status of the Campanian victims.

The evidence indicates that there was no apparent skewing towards people with infirmity in the Pompeian sample, which is consistent with the findings in the Herculaneum skeletal sample. There appears to be regional continuity in stature from antiquity to the modern era and the height difference between males and females indicates that these individuals were not exposed to major stresses in the form of poor nutrition and illness in the growing years. Cases of healed trauma and healing porotic hyperostosis suggest that the individuals in the sample had robust immune systems, though the dental data indicate that there may have been some underlying health problems in the form of systemic infection or more serious soft tissue pathology. The presence of DISH and hyperostosis frontalis interna demonstrates that individuals were able to survive long enough to succumb to these age-related disorders.

It would have been preferable if the Pompeian skeletal sample had not been disarticulated with the resultant loss of information. The identification of various disorders, like brucellosis and tuberculosis in the Herculaneum sample highlights this, as entire skeletons are required for diagnosis. Regardless, the compromised Pompeian sample not only yielded results that corroborated a number of the findings from Herculaneum but it also provided evidence of the presence of disorders not observed in the Herculaneum sample, most notably the substantial number of cases of HFI. Hyperostosis frontalis interna proved to be a very important discovery because the frequency of its occurrence in the sample is comparable to that in a modern Western population. This suggests that Pompeian lifespans were probably comparable to those in modern Western populations. This is important to note as there is a tendency for scholars to kill off our ancestors at relatively young ages, probably because of their expectations. Age-at-death results, which are based on techniques that systematically underage, as age is diffi- cult to establish for adult skeletons, are supported by the use of life tables that provide mean ages-at-death. These are always depressed by the rate of infant mortality. Of equal importance, the frequency of occurrence of HFI supports the notion that the Pompeian skeletal sample reflects a random sample of a normally distributed population.

The lack of observed cases of HFI in the Herculaneum sample invites further investigation. There are a number of possible explanations for this. One is that there may be significant differences between the samples of victims at these sites, either because they reflect different populations or because they experienced the lethal stage of the eruption in different ways. The latter explanation is unlikely as it would be difficult to argue that the Herculaneum experience of the event would have spared large numbers of older females. Another possible explanation for the difference in the observed number of cases at the two sites is that this disorder may have not been identified by researchers studying the Herculaneum sample. HFI is not readily apparent on complete skulls, and while Capasso did x-ray every individual that he studied, it is not easily detected on this medium.

The disparities between the results of the work of different scholars who have worked on the Herculaneum sample, most notably Bisel and Capasso, indicates that there is a need for further examination of this material, using standard techniques.

Lifestyle indicators are notoriously dif ficult to identify and interpret. The Pompeian sample was too compromised to enable a study to be undertaken and the results of such work on the Herculaneum sample can be questioned.

If you find an error please notify us in the comments. Thank you!