Catering for the Cultural Identities of the Deceased in Roman Britain: Interpretative Potential and Problems
Natalie C. C. White
Introduction This paper aims to illustrate the interpretation potential of food remains within late pre-Roman Iron Age (hereafter LPRIA) and Roman burial contexts in Britain, by demonstrating that choices were made concerning particular foodstuffs interred with the deceased, and that they are likely to have been both socially and culturally driven. Whilst demonstrating the potential of this approach, it will also be necessary to address a number of problems associated with such data and its interpretation. Problems relate not only to the taphonomic factors affecting what goes into the archaeological record, but particularly to the objectives of archaeologists in the field, and the selection of particular data for written reports and subsequent publication. It is perhaps the latter which have made the data collection and analysis for this research the most difficult.
Why identity? The impact of the Romans upon the indigenous population of Britain and its transformation under Roman rule has been of considerable interest to scholars in the past, under the heading of ‘Romanisation’. Dynamic debate has recently seen the concept of ‘Romanisation’ dissipate in favour of new approaches and alternatives, particularly that of identity (see James and Millett 2001). Focusing on the concept of identity allows for a greater degree of individual agency; particularly that of the indigenous Britons, what Mattingly has called the ‘lived experience’ (1997: 13), that appreciates the unity and diversity of both public and private behaviour. Cultural change was not only multi-directional, but also unique according to geographical and chronological context.
Identities are a social construct; not fixed at birth but can, and do, change through deliberate or unintentional actions. Material culture can be manipulated in the creation, maintenance, and display of individual and social identities. After all, who people think they are, or who they project themselves to be is a fundamental part of human nature (Hill 2001: 12), and so identity is often a visible expression of choice. Choices are naturally dictated in part by wealth, but primarily by social and personal preferences. Britons were still able to make their own choices under Roman rule, choices that made them individuals and illustrated how they wished to be perceived both publicly and privately. As such, artefacts are no longer simply the passive products of people’s lives, but are increasingly being seen as having played an active role in sustaining existing social identities, and creating new ones.
Why food and identity? In examining the impact of the conquest, studies have tended to focus on the more traditional lines of material evidence, largely overlooking the potential of the bioarchaeological remains (Dobney 2001: 36). The value and significance of food within archaeology has been increasingly appreciated in recent years (see particularly Counihan and van Esterik 1997; Gumerman 1997). This is partly due to the improved methods of retrieval within the field, but perhaps more importantly because of the realisation that food (both past and present) is not consumed simply for its nutritional content. Whilst technology, nutrition, and symbolic importance can affect food choices, food is intrinsically social, and can be used to help define and maintain social relations and differentiations in the construction of identity (see Hawkes 2001; and Meadows 1995 for Romano-British examples). Consumption is conspicuous, and the choices made are not entirely economically logical.
You are what you eat; the once popular adage until the arrival of the perhaps less popular British television series by the same name, applies to both the physical and the social existence. Consumption can vary among individuals of differing age, sex, gender, ethnicity, religion, ‘status’, faction, and geographical location. Within a complex society, all of these factors can affect the differential access to goods, knowledge, wealth, and power. As such, it may be possible to analyse food within certain contexts as a form of communication similar to other types of material culture. After all, cuisines do not tend to transplant themselves in their complete and original form (Meadows 1995). Whilst some forms of opposition are overt, others can be more passive acts of resistance to changes imposed from the outside, and could be seen as statements of alternative values and identities. Choosing to maintain pre-conquest food traditions could thus be seen as a means of maintaining deliberate links with the past. Why food within burial contexts?
Establishing foods linked to specific individuals is extremely difficult within the majority of archaeological contexts. For this research, the focus has been on the food remains found within burials, as these provide a unique context in which the remains are found with a direct association with individuals. Additionally, burial ceremonies are both public and private affairs, a situation in which both public and private identities could have been expressed by the items accompanying the deceased.
Studies have shown that food was a significant feature of funerary rituals for ‘Romans’ (Helbaek 1956; Lindsay 1998), and this is reinforced by numerous classical written sources. Items of food and drink feature at several stages of the funeral and its associated rituals. For example, Cicero (De Legibus: 2.22.55–57) and Ovid (Metamorphoses: 6.566–570) discuss food within the actual funerary ritual, while Petronius (Satyricon: 65) and Tacitus (Annales: vi, 5) consider libations being poured upon the grave when the full period of mourning was over on the ninth day after the burial. Varro (De Lingua Latina: vi, 13), Ovid (Fasti: 2. 533–570), and Lucian (De Luctu: 19) describe annual commemorations of the dead, although only in jest from the latter.
Foodstuffs also appear to have been used in the rituals associated with the burial and commemoration within the provincial societies of the empire, such as the Netherlands (Lauwerier 1983; 2002), Mediterranean and central France (Bouby and Marinval 2004), North Africa (Burke 2001; Stirling 2004), and central Europe (Kreuz 2000). Though only touched upon briefly in terms of the noted presence of foodstuffs associated with burials during both the LPRIA and post-conquest period in Britain, there appear to have been similar funerary rituals to those seen within other contemporary provinces (see Bercui and Wolski 1973; Philpott 1991; Whimster 1981). Whilst we should not assume that all of these practices were the same, the archaeological evidence would appear to suggest recurring themes.
The only real attempt to quantify such data and examine its possible significance has been by Philpott within his comprehensive survey of Romano-British burial practices (1991). Whilst Philpott’s study has gone some way in looking at the species represented and their chronological distribution, the concept of identities was not addressed, and age and sex correlations have not been discussed at any great length. The range of published data has also advanced significantly since Philpott’s publication.
It is clear that some indigenous Britons, ‘Romans’, and peoples of the wider empire shared a practice of interring food with the deceased, and that so far within Britain, this avenue of research has been significantly under examined. Yet it is plausible that the food remains within burials provide direct evidence for how the deceased’s identity was expressed or reinforced, or perhaps the identities of those responsible for the burial. An examination of such remains may help in taking us beyond simple projections of a dichotomous ‘Roman’ and native identity, to perhaps exploring regional, gender, ‘status’, faction, and even religious identities.
The archaeological evidence: taphonomy This paper examines the primary offerings; those that were deposited during the rituals of disposal with the deceased within the grave. Here, items found within the graveshaft have been omitted owing to the complexities involved in assigning a direct relationship with the deceased individual. The identification and interpretation of these remains requires a careful consideration of the taphonomic processes that created them, which are the cultural and noncultural processes involved during the disposal, preservation, subsequent retrieval, and scholarly documentation of the archaeological record. These will all affect the data considerably. Examples will be discussed in relation to specific data below.
It must first be remembered that a number of food offerings are likely to have left no archaeological trace, but could have once been included. Meat served off the bone, liquid offerings (such as wine and oil), bread, cakes, and unburned fruits and vegetables are very unlikely to leave any archaeological trace in Britain. What survives will be largely dependant upon the different conditions of preservation associated with both cremations and inhumations, although the same offerings may have been included with both. Cremation burials have been found with both pyre goods (items that were cremated with the individual on the pyre) and unburnt offerings, whereas only the latter are found in association with inhumations. Animal and fish bone will generally survive in both cases with the soil conditions permitting. Botanical remains will only survive if they have first been subjected to heat, or if the remains have been waterlogged.
It is important to note that cremations rarely, if ever, contain the entire cremated remains of the individual, but a deliberately selected subset (or anthropogenic sample) (Kreuz 2000: 45; McKinley 2000: 66–67). It is often apparent that only the larger pieces of bone were gathered from the pyre site and subsequently placed within the burial container. At the eastern cemetery of Roman London, an average of 46%, but sometimes as little as 3%, of the human remains from the pyre were included within the burial (Barber and Bowsher 2000: 76). Thus, we are only retrieving a small fraction of all the cremated material, and one that often favours the larger pieces of bone, rather than the ash and smaller elements, such as fish bones and charred seeds. The quantity of pyre goods found within cremations should therefore be viewed as a minimum, as it is probable that if not all of the human remains were collected for burial, then not all of the pyre goods were collected either.
It is possible that small amounts of the edible food remains found within cremation burials were accidentally mixed if the pyre site was re-used, and so not related to the individual at all. Although if charred seeds were provided with oxygen and sufficient heat within a second or even third funerary pyre, the carbon would burn off, leaving only ash flecked with charred remains (Martin Jones pers. comm.). Furthermore, small numbers of charred and waterlogged seeds need not necessarily have been included within the pyre as deliberate food offerings, but could have been included as part of the fuel, deposited through animal waste, or otherwise have entered accidentally through the general cemetery ‘background noise’. Charred seeds are very robust, and could survive for long periods of time, they could even be the remains from earlier occupation deposits (Barber and Bowsher 2000: 70). Additionally, during the burning of the pyre, the vegetation in the surrounding area can be disturbed and even charred by the heat of the fire, being subsequently collected and placed in the grave pit unintentionally (see Kreuz 2000: 46). With all of these possibilities in mind, these remains are further dependant upon their subsequent retrieval, another taphonomic factor that needs to be addressed.
Excavation biases Within excavations, the bioarchaeological evidence has largely been overlooked in the past in favour of other items of material culture found within the burial context. Animal bone was rarely collected systematically, as this was not a primary excavation objective (Clason and Prummel 1977: 172). Whilst fish and smaller mammal bones may have originally been included within the burial, they are not only more fragile and less likely to survive, but are often likely to be less easily identified and retrieved unless the soil is sieved. It is often apparent, particularly within older cemetery excavations that only associated bone groups of what may be classed as the ‘more obvious’ animal bones were recorded, such as complete chickens or legs of lamb. At Ospringe, near the Roman town of Durolevum in Kent, excavated in 1925 (Whiting et al. 1931), a total of 172 cremations and 140 inhumations were excavated. Only one inhumation was recorded as having contained ‘partridge bones’ (owing to their small size in relation to the modern chicken), with some oyster shells. The primary foci of the excavation were the human remains and any associated pottery, and whilst it is very likely that further animal bone was present, it was neither recognised during the excavation, nor identified during later analysis.
Sampling, sieving, and the subsequent flotation of soil samples have only systematically been put into practice on archaeological sites in the last 15 years. The wide-spread sampling of nearly all burial contexts has only been identified since this pilot study has been carried out, within some published reports (Evans and Maynard 1997; Fitzpatrick 1997), and at the recent excavations at the Colchester Garrison and Hanford House sites (Phillip Crummy pers. comm.). Within the sample dataset examined here, the majority of sampling has only been occasional or sporadic (see the sampling in Barber and Bowsher 2000; Niblett 1999), the repercussions of which will be discussed below. All of these issues will need to be borne in mind in attempting a preliminary synthesis of the available data from a selection of Romano- British cemeteries.
A sample illustrating the interpretative potential of food offerings within burials A database of 248 burials was compiled for this pilot study, although only particular examples have been selected here in order to illustrate their interpretative potential. The focus here is solely upon the published cemetery reports, and only the details published within them. Whilst the work of the editors of the reports should not go without praise for what they have done to make such data available to so many, it will become apparent how restricted by their research agendas and analyses the data can be.
Food offerings according to the sex of the deceased In this section, three cemeteries will be used to illustrate that within certain areas, meat offerings were more frequently interred with individuals of a particular sex. Folly Lane, northeast of the Roman town of Verulamium, adjacent to contemporary St Albans, was excavated between 1991 and 1993 (see Niblett 1999). There was one apparently unique and prestigious burial, surrounded by a ditch bordering what has been interpreted as a ceremonial enclosure, and in a very prominent position over-looking the town. This main burial was dated to the period shortly after the Roman invasion. From the later first to mid second century A.D. a number of other cremations and a smaller number of inhumations gathered on the slope below. The main burial was the focus of some lengthy and rather complex mortuary rituals, involving some ostentatious burial goods; including several broken Italian wine amphorae, imported platters and beakers, and Samian vessels, implying feasting activities and imported tastes (Niblett 1999: 393–394). Unfortunately, only ‘selected contexts’ were sieved, with the only burial related sample coming from the main cremation (Niblett 1999: 4). Although 11 (39%) of the burials from this site contained animal bone, specific details concerning the species and quantities found in association with the main burial are notably absent from the published report.
The remaining sample of ten burials for which the species included were detailed was unfortunately too small to determine any patterns between the species present and the sex of the deceased. However, overall, the males do appear to have been more frequently interred with meat remains in comparison to the females (Fig. 1), although both were equally represented within the burial group. In fact no female was apparently buried with an ‘offering’ of an edible portion of a domestic animal within this burial group.
A similar pattern was noted among the burials at Trentholme Drive in York, on the edge of the Roman town of Eboracum that was excavated in 1953 (Wenham 1968). Only 17 (4%) of the burials are recorded as having contained animal bone, which is likely to be largely due to the date of the excavation and the contemporary research agenda. The large and unique in situ cremation does not have the animal bone associated listed in any detail, only the oyster shells are noted. For the remainder of the burials, the small size of the sample prevented any meaningful associations or patterns being drawn. There were no apparent correlations between the sex of the deceased and particular species for example. However, the majority of the animal bones were again associated with male burials (Fig. 2), while both were approximately equally represented within the cemetery. Very few of the burials could be dated, and those that were, were of the second and third centuries A.D approximately.
4 36% 2 18% 1 9% 4 37% Male ?Male Female ?Female Juvenile Indeterminate Figure. 1: Sex distribution among meat related offerings within the burials at Folly Lane.
8 50% 3 19% 1 6% 4 25% Male ?Male Female ?Female Juvenile Indeterminate Figure. 2: Sex distribution among food related burials at Trentholme Drive.
Species preferences, chronological changes, and anomalies within the animal remains Here, cemeteries with sufficient numbers of identified animal bone are discussed in terms of their regional species preferences, chronological patterns, and any other anomalies noted. What is already becoming apparent within this realm of research is that certain species were significantly more popular and more frequently included than others.
At the King Harry Lane cemetery, excavated between 1965 and 1968, the opposite side of Verulamium to Folly Lane (see Stead and Rigby 1989), animal bones were found within 87 of the 388 LPRIA and early Roman cremation burials that were not ‘mislaid’ after the excavation (22%) (Stirland 1989: 240). Although no soil samples were taken, that such a high percentage of animal remains were detected among the human is likely to be due to the osteologist removing any of animal origin for later analysis by a zooarchaeologist. From A.D. 1–60, pig bones are by far the most frequently occurring species on the site (Fig. 3), which is indicative of a clear preference for pork as a funerary ‘offering’ or component of a graveside meal. It is equally possible that this could be a local or chronological preference, perhaps even relating to both.
0 10 20 30 40 50 60 70 Bird Cattle Chicken Dog Ox Pig ?Pig Sheep/Goat Sheep-size Figure. 3: Frequency of species in the burials at King Harry Lane.
The burials from Poundbury camp, near the Roman town of Durnovaria Durotrigum, in modern Dorset (see Farwell and Molleson 1993), provide the ideal opportunity to examine change over time within a localised context. Excavations between 1966 and 1987 revealed a primarily late Roman cemetery, with a small number of burials dated to the LPRIA. The area was clearly in use as a burial ground during that early period, until the third and fourth centuries A.D. when it was transformed into a formally arranged burial ground, after which the zone excavated was abandoned. Unfortunately no soil samples were taken from the site, and the grave fills were not sieved, which is unsurprising considering the date of the excavation. However, animal bone was recorded within some of the burials.
In the Iron Age, six burials (11%) were interred with an ‘offering’ of food. In the late Roman period, only 13 burials (1%) appeared to have contained animal bone. However, none of the Roman burials were from what the editors classed as the ‘main late Roman cemetery’ (Farwell and Molleson 1993: 41). Again, whilst there were only a small number of burials from the Roman period, the majority were located near the northern periphery of the main cemetery, and haphazardly aligned along its boundary ditch. This seems to be indicative of differing burial customs, possibly representing a different faction within the population. Poundbury is not the only cemetery within which burials containing food remains were apparently marginalised and on the periphery of the main burial area. At both Watling Street, excavated in London during 1996 and 1997 (see MacKinder 2000), and St Johns Abbey grounds, Colchester, excavated in 1972 (Crummy et al. 1993), it would appear that the food related burials were similarly segregated.
However, as the burials at Poundbury are at least two hundred and fifty years apart, chronological changes according to species preference within a relatively local context can be examined. In the LPRIA, pork and sheep or goat offerings were chosen, with no particular preference for either (Fig. 4). But by the third and fourth centuries A.D., no pig remains were found at all (Fig. 5). Whilst sheep or goat offerings have remained consistently popular, a taste for pork appears to have been replaced by a preference for chicken.
0 1 2 3 4 5 Cattle Chicken Dog Pig Sheep/Goat Sheep Figure. 4: The frequency of species at Poundbury prior to, and following the conquest.
0 1 2 3 4 5 6 7 8 Cattle Chicken Dog Pig Sheep/Goat Sheep Figure. 5: The frequency of species during the third and fourth centuries A.D. at Poundbury.
0 1 2 3 4 5 6 7 Chicken Eggshell Sheep-size Small rodent Figure. 6: The species frequencies during the third and fourth centuries A.D. at Lankhills.
The sample from Lankhills, excavated between 1967 and 1972 was again small (nine burials, only 2% of the total) (Clarke 1979). However, all of the burials were from the third and fourth centuries A.D. and display a similar preference for chicken at this time (Fig. 6). Initially, the animal bone data from the eastern cemetery of Roman London was going to be used here to illustrate another possible example of a post-conquest preference for chicken. However, a reassessment of the report has brought to light a number of cremations containing animal remains (both burnt and unburnt) that have been excluded from the burial catalogue (see Barber and Bowsher 2000: 72), and so were thus excluded from the initial study. What is clear, however, is that 16 (3%) of the inhumations and at least 76 (62%) of the cremations contained animal bone with a direct association to the skeletal remains. Neither the animal bone found adjacent to human remains, nor those found within grave fills have been included within this study, as the precise associations between these bones and specific graves were not published within the report. From the information provided in the burial catalogue, it is tempting to say that chicken offerings appear to have been favoured over other species from the late first century at the earliest through to the fourth century A.D. (Fig. 7). However, the cremation burials that were listed within the report were included because the bones they contained were associated bone groups, such as pig forelimbs and chicken skeletons, or because they contained multiple species (see Barber and Bowsher 2000: 73–74). Single occurrences of ribs and other bones do not appear. As such, the data may be biased towards the wealthier sector of society offering more extravagant ‘meals’. With the inclusion of associated bone groups, it is also not surprising that chicken may appear to be more frequently represented when considering the meat to bone ratio.
It is interesting that all of the bones were recovered from cremation deposits by sieving, whereas no soil was sieved from the inhumations. We may also note that no fish remains were recovered from an inhumation context. As both forms of disposal overlap chronologically, it is therefore highly probable that this anomaly is illustrative of the method of archaeological retrieval, particularly considering the fact that the fish bones identified within the cremation burials, and detailed within the report, are all unburnt. From the cremation burials listed, it would also appear that the earliest burial to contain saltwater fish dates to the late second century A.D. at the very earliest, at least c. 100 years after the area was initially used as a place of burial.
0 5 10 15 20 25 30 Bird Cattle Cattle-size Chicken Eel Flatfish Frog/Toad Goose-size Herring Horse Pig Plaice/Flounder Sheep/Goat Sheep-size Small rodent Smelt Unidentified animal bone Unidentified fish bone Whiting Figure. 7: The frequency of species of animal and fish bone within the inhumations and an uncertain percentage of the cremations at the Eastern cemetery of Roman London.
Interestingly, cremation CB835, an unsexed adult, may have contained a local variety of garum, or more likely its bony by-product hallec. Fish sauce of this kind was a popular flavouring of meat dishes, and used within a vast proportion of the recipes by the Roman gourmet Apicius (see Flower and Rosenbaum 1958). The burial contained four different varieties of local saltwater fish in addition to the goose-sized bird bones and sheep-sized mammal fragments, although it is important to note that the exceptional preservation in this case is likely to be due to the waterlogged conditions in which the burial was found. A more detailed examination of the bones may help to illustrate if the fish had been processed in any way prior to burial. Although garum is traditionally thought of as a Mediterranean product, and reflective of a Mediterranean liking for fish, there is evidence that local versions were being manufactured in the port of Roman London, using herring and species of flatfish (see Milne 1985: 87), just like the species found within CB835. Regardless of whether an imported or local commodity, the presence of fish sauce or fish would certainly indicate a liking for ‘Roman’ tastes, even if used within a British cuisine.
The animal bones discovered within burials seem to present different characteristics when compared with bone assemblages from contemporary urban and rural sites. Although regionally specific, most sites in Britain throughout the LPRIA tend to display a stronger predominance of sheep and sheep or goat bones than any other species (Albarella in press; King 1999: 178; 2005: 331). However, the beginning of the Roman period saw a change in the frequency of the main domestic species. There appears to have been an increase in the importance of cattle and pigs at the expense of sheep (Albarella in press.), which Anthony King believes correlates with the apparent nature of ‘Romanised’ sites, and is similar to patterns in Gaul and Germany (1999: 178–179; King 2005: 331). The choice of meat within the context of burial might therefore illustrate one’s allegiance to relatives and the past, or perhaps to the future and new regimes. Subtly reinforcing your loyalty through culinary choices at the funeral could have been a very public, yet passive statement of cultural allegiance. This could explain the consistent popularity of sheep or goat at Poundbury, with some members of the community deliberately choosing not only to bury or be buried with meat offerings, but also to identify with more traditional fare. A more detailed analysis at regional and local levels, and comparisons with contemporary domestic assemblages might illustrate this more clearly. It is also possible that the meat offerings may have had some kind of religious significance. At Harlow in Essex, a sharp decline in pig and sheep or goat bones can be seen after the conquest when a late Iron Age shrine was succeeded by a temple (King 2005: 335). The rural shrine at Uley in Gloucestershire, whilst exhibiting a dominance of sheep or goat bones (mainly thought to be goats), displays a marked increase in the frequency of chicken bones, from 40 in the late first century A.D. to 604 after the building of the Romano-Celtic temple between the second and third centuries A.D. (King 2005: 333; Woodward and Leach 1993: 260). Whilst it should be noted that the Romans did not bring the chicken to Britain, chicken keeping certainly intensified during the Roman period (Umberto Albarella pers. comm.). According to finds from the temple, the deity worshipped was Mercury, as he appears in depictions on both cult statuary and altars, perhaps more interestingly in association with cockerels and sheep or goats at his side (see Woodward and Leach 1993, especially pages 89– 102). However, whilst it is tempting to see associations between such assemblages as plausible explanations for trends and patterns noticed, at this level of analysis, over generalisations are a little rash, and a contextual analysis at a local level is needed. Animals were clearly being used in a unique manner at many sites, which is likely to have impinged upon the choices made in everyday life and at funerals.
Botanical remains The eastern cemetery of Roman London comprises 11 different sites that were excavated between 1983 and 1990, and published as one cemetery report (Barber and Bowsher 2000). The cemetery contained the most recently excavated sites included within the analysis for this paper, and as such, it demonstrates the most sophisticated excavation techniques. However, the frequency of soil samples taken varied according to each separate excavation (see Table 1). The greatest numbers were taken from site ‘D’, the final excavation conducted within the area. Whilst a large number of soil samples were taken from burial related contexts at site ‘D’, a much smaller number were later studied. The published report does not mention why some samples were chosen for subsequent analysis over others. Communication with the Museum of London Archaeological Service (MOLAS), and visits to the site archives concluded that the samples were processed selectively due to the constraints of time and money (Andrew Westman pers. comm.).
Table 1: Samples taken, and those later analysed from the individual sites within the Eastern cemetery of Roman London. SITE D F K C I H A Number of samples taken at excavation: 100+ 17 10 8 18 3 1 Feature type of samples deemed worthy of study Urned cremations 13 1 – – – – – Unurned cremations 13 2 5 2 – – – Pyre deposits 30 1 – – – – – Cremation debris 4 9 – – – – – Grave fills etc. – – 1 1 2 – – Pit fills etc. – – – – 3 3 – Total: 60 13 6 3 5 3 0
0 5 10 15 20 25 30 Indeterminate cereal Lentil Wheat Barley Vetch/Tare/Vetchling/Pea Oat Pea Emmer/Spelt Bread/Club Wheat Wild Strawberry Elder/Danewort Figure. 8: Frequency of plant species within charred samples at the Eastern cemetery of Roman London.
From the 36 samples taken from the cremation burial deposits that were later analysed, 33 (92%) contained some kind of edible plant remains. Waterlogged edible plant remains were also found within six of these cremations. With regard to the cremated material, there were no apparent distinctions according to the sex of the deceased individual. This is most likely due to the very low number of burials that could be sexed successfully, and is unfortunate. From the ten that were sexed, neither males nor females were favoured in particular, nor were the plant species represented biased towards either sex. Fig. 8 illustrates the frequency of charred species occurring within the 33 cremation samples.
Charred cereal grains were the most frequent offering; grain was present within 30 of the burials. Pulses, such as lentils, peas and beans were within 16 of the samples, with local varieties of pea and bean being just as common as lentils (each being found within ten cremations). Lentils could have been cultivated on the warm, light soils in the south of England, but are most likely to have been imported from the Mediterranean or Near East at this time (Barber and Bowsher 2000: 70). These are, therefore, likely to have been a more expensive offering. Peas and beans have been recovered from Iron Age sites in England, and grow well in this country (ibid.). At the eastern cemetery, they were believed to have been deposited raw and dried onto the pyre, as they would have been softened by cooking and be unlikely to survive in a recognisable state (ibid.). Local fruits and nuts were found within only two of the cremated samples, but within all waterlogged burials. Within these, there was evidence of imports such as fig, poppy seeds, and grape from the late second to third century A.D. at the very earliest.
The apparent widespread preservation of some kind of edible plant remain within the samples analysed strongly suggests their use as frequent pyre goods. It would be unusual for this quantity of remains and the species present within the majority of samples to simply be pyre fuel, or randomly occurring waste, particularly so far removed from a domestic context. Evidence suggests that pulses such as lentils, peas, and Celtic beans have been found at other Roman sites in London, but often identified tentatively and in very small numbers (Barber and Bowsher 2000: 70). As such, it may be safe to assume that these were deliberate offerings, and perhaps unique in comparison to the everyday fare.
These findings beg the question of how many other burials are likely to have contained some kind of food offering, even if only a handful of grain or beans. According to Grant (2004: 383), the role of meat is often over-emphasised in the past when looking at faunal assemblages, and would only have contributed a minor proportion of the overall diet in relation to fruits, vegetables, and pulses. Perhaps this pattern could also be evident within the context of burial.
Discussion and Further Reflections Whilst the analysis so far is only a pilot study, patterns and trends are apparent. This study illustrates the as yet untapped potential of the bioarchaeological evidence within the burial context, despite its taphonomic constraints. In relation to the animal bone remains, there appears to be a bias towards including meat offerings within the burials of males, although other cemeteries within the country and during differing chronological periods may paint quite a different picture when added to the database. Chronologically, species preferences appear to change from the LPRIA through to the Roman period within certain regions. It may be possible that the inclusion of pork represents an indigenous practice within burials as something special in comparison to the apparently more regularly consumed sheep or goat.
It is difficult to examine chronological changes within such a small sample of botanical remains from the eastern cemetery of Roman London, although it would appear that ‘exotic’ or imported goods were introduced and available for a few members of the contemporary society, and deemed worthy of burial with a friend or relative. A greater variety of meats certainly seem to have become available after the conquest, and there appears to have been a greater exploitation of, or at least an increased desire for marine fauna. Both the ‘exotic’ fruits and an increasing desire for the flavour of fish could be regarded as illustrating an influx of Mediterranean tastes. Although being so near to the port of Roman London, it is equally possible that these goods are illustrative of an organised trade and fishing network. Nevertheless, they imply a growing market, with some people choosing to spend their wealth on more unusual fare.
What is certainly apparent is that a great deal of further work could, and indeed should be done in this area, particularly within a more contextually specific approach where possible. Since this research has been conducted, other cemeteries have come to light within the Dorchester area. It is hoped that future work will illustrate a more regionally specific picture of species related to chronological changes. Further research is starting to indicate that Poundbury, Watling Street, and St Johns Abbey grounds in Colchester are not the only later Roman cemeteries to have burials containing food offerings segregated, or on the periphery of other burial areas. It would appear that people were prepared to practice their own particular funerary rituals on the edge of a society with distinctly different habits, maintaining and reinforcing their separate group identity both publicly and privately. It is therefore not entirely implausible that people were also reinforcing their connections with the past, present, or future through food choices and preferences at such an event.
With funerary practices involving food evident in so many other contemporary provinces within the empire, there may also be a great deal of potential for provincial comparisons concerning foodstuffs and the manner in which they were used. Thus, it would appear that these remains do provide many avenues for potential research, if people take the time and trouble to deal with them both on and off the site, and subsequently within published articles. Greater attention should be paid to botanical remains, particularly in the field and postexcavation. Systematic attention to plant remains has been proved successful in analyses in France (see Bouby and Marinval 2004). The Gallo-Roman graves in the Rhone valley and Mediterranean France received many more fruit offerings, particularly imported fruits, and baked goods than the graves in central France (Bouby and Marinval: 82). This might be illustrative of a stronger continuity with pre-existing Iron Age practices, which the authors have attributed to a greater Mediterranean influence in the south of France (ibid.). If cremation burials and waterlogged inhumations were sampled more consistently, some surprising results could become apparent.
What is certainly evident is that this angle of research is largely dependant upon the excavator’s aims and methods on site during the excavation. This is often a problem brought about by the lack of time and resources within developer-funded archaeology. However, the eastern cemetery of Roman London is a prime example of the potential of sieving burial contexts. If the data has been excavated and later analysed, it is very often published, but each excavation report is entirely unique in its format, as it is structured according to what the author(s) feel(s) to be necessary. For burials of great ‘excitement’, such as at Trentholme Drive (Wenham 1968), Folly Lane (Niblett 1999), and Watling Street (MacKinder 2000), detail concerning the animal bone was omitted from the report. As already discussed in relation to the eastern cemetery of Roman London (Barber and Bowsher 2000), detail concerning the animal bone found within burials is not always listed as an inventory, even though the frequencies of species were discussed in general (see also McWhirr, et al. 1982 for similar issues). With so many other finds taking centre stage within the more recent reports, this is perhaps not surprising, and may reflect that food remains are still not being taken seriously in terms of its interpretative potential within Romano-British archaeology.
In an ideal world, much greater detail with regard to the animal remains associated with burials would be preferred. When animal bone is included within a published report, it is often not reported upon in great detail. The presence or absence is simply noted. If the age, sex, and portions of the animals were included, then this may open up many new lines of inquiry. Particular portions may have been reserved for particular members within a community for example, or species of a particular sex and age. With specific regard to cremations, if bones were recorded as being burnt and unburnt within the same cremation, they ought to be separated within the report. It is highly probable that burnt and unburnt bones within cremations are representative of two entirely different activities. The unburnt accompaniments may be the remnants of a funerary feast at the graveside, or the remains of continued graveside celebrations from which a portion was set aside for the deceased person, whilst burnt remains may have been gifts for the afterlife, or a portion of a funerary meal at the graveside. The location of the inclusion within the burial may also be significant. The example of the decapitated individual from Baldock whose head has been removed and replaced with a chicken is a little unusual (Keith Matthews pers. comm.), but indicative of the placing of some food remains not simply being random.
The idea behind this paper was to use material culture (food remains within burials) as a means of identifying human agency; not simply to identify difference, which was a criticism of the term ‘identity’ within so many TRAC sessions this year. At this early stage, it is perhaps important to identify differences and similarities within the data, only then can any further discussions concerning the wider implications of these anomalies ensue. By targeting the identities of the peoples of Roman Britain, this study has gone beyond simply dividing people by what they were not doing, by examining to a greater extent what people were doing. What is clear is that with a greater attention to these remains at all levels of analysis, from the site to publication, the interpretative value of food remains within Romano-British and LRRIA burials could really be exploited to its full potential. Some people were actively choosing to bury dead friends and relatives with foodstuffs, while others were not. When food items were buried with the deceased, choices were clearly being made. By trying to access the thinking and practice behind these choices, it may indeed be possible to examine aspects of cultural identity.
Department of Archaeology, University of Cambridge
Acknowledgements Particular thanks go to David Mattingly and Marijke Van der Veen, who advised, supported and encouraged the initial research at the University of Leicester. Additional gratitude for directing subsequent research, and for his valuable input on reworking this into both a spoken and written format goes to my supervisor Martin Millett. Ali Klevnäs and Ben Croxford should also be thanked for both having patiently read through earlier drafts, and to the latter go additional thanks for persuading me to give this paper in the first place. Finally, without the AHRC, none of this research would have been possible; I am both indebted and truly grateful.