Resurrecting Pompeii (31 page)

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.
⁴³
Unfortunately, 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.

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