Category Archives: New Research

Millet in the Roman Diet

“If you want to waste your time, scatter millet and pick it up again” (moram si quaeres, sparge miliu[m] et collige)

A proverb scratched on a column in the peristyle of the House of M. Holconius Rufus (VIII.4.4) at Pompeii (Jashemski et al. 2002, 137).

Looking for the evidence of millet, a generic term for a large group of small seeded-grasses, that includes both Setaria italia (L.) P. Beauv. and Panicum miliaceum L., used during the Roman empire, circa 753 BC–610 AD, presents a number of challenges. Millets are only mentioned a handful of times in the ancient surviving texts, there are only a few well-documented preserved archaeological finds of millet and limited scientific evidence, including archaeobotanical (ancient preserved plant remains) and isotopic evidence (based upon plants using either C3 and C4 photosynthesis). All these lines of evidence are problematic in terms of their representativeness but together they offer a more complete glimpse into the growing understanding of millet and its use and importance in the Roman world.

In other plant taxa it may be problematic to attribute specific botanical species to ancient Greek and Latin names as it is difficult to trace the ancient version of the plant through to modern times. However, this is not the case with Panicum miliaceum or common millet as it has been recovered dating back to the end of the third millennium BC on European archaeological sites (Boivin et al. 2012; Dalby 2003, p. 99; Valamoti 2013) and Setaria italica (L.) P. Beauv or commonly known as Italian millet has been cultivated since the Bronze Age circa 2000 BC in Europe (Jashemski et al. 2002, p. 162). The wild progenitor or ancestor of foxtail millet, Setaria viridis (L.) P. Beauv, is well identified and shows clear morphological affinities with, and can interbreed with domesticated Setaria italica (L.) P. Beauv. (De Wet et al. 1979, p. 53; Zohary et al. 2012, p. 71).

Wall painting of common millet (left) and Italian millet (right) being eaten by two quails (NMinv. No. 8750) from Pompeii, Italy (photo by S. Jashemski p. 137 in Natural History of Pompeii). NB: It is possible to distinguish the two plant species and their similarity to modern species of common millet and Italian millet

Fig 1: Wall painting of common millet (left) and Italian millet (right) being eaten by two quails (NMinv. No. 8750) from Pompeii, Italy (photo by S. Jashemski p. 137 in Natural History of Pompeii). NB: It is possible to distinguish the two plant species and their similarity to modern species of common millet and Italian millet

Millet is most commonly known in ancient times as being used for fodder to feed livestock and other domesticated animals including birds (see Figure 1). However, millet was also eaten by humans. Millet can be boiled and made into a porridge or ground into a flour and made into a heavy flat bread. The Romans probably did both with it. From the writing of the ancient Greek physician and author Philotimus, we know that one way of preparing millet involved being ‘pounded when raw, ground finely and, after some water has been poured on, it is pounded once again, strained, boiled’ (Oribasius I.15.2).

Setaria italica (L.) Beauv. in autumn

Setaria italica (L.) Beauv. in autumn

Based upon the ancient texts it appears that millet was not the Romans’ favourite or first choice of flour for making bread with but it also wasn’t discarded by the Romans. For example, the presence of millet within the majority of properties with Insula VI.I and other elite houses within the city of Pompeii suggests that millet may have been consumed by the wealthy Roman owners and their servants and slaves. We now know that millets are rich in carbohydrates but poorer in digestible proteins than other cereals making it an excellent appetite satisfier to fend off hunger (Spurr 1986). Thus, common and Italian millet could have been used by the Romans in place of other cereal grains to make and/or bulk up breads and porridges, particularly in times of food shortages or crop failures in an to attempt to satisfy a starving stomach. Practically, millet filled a very useful place in the Roman diet. Millet was cheap to purchase and easy to grown alongside both summer and winter crops. Millet could have helped to hedge against famine in terms of its ability to grow in a wide range of less than ideal agricultural environments. This would have been a very important trait in an unpredictable agrarian world that was quickly exhausting its agricultural farmland (Fraser and Rimas 2010). As the ancient author Strabo (5.1.12) advised, ‘millet is the greatest preventive of famine, since it withstands every unfavourable weather, and can never fail, even though there be scarcity of every other grain’.

Traditional Roman foods were considered ones that the small farmer could grow cheaply on their small plots of land to sustain their families which included millets, pulses and vegetables. Millets were grown in Europe since the Bronze Age and possessed a hardy nature, capable of growing when and where other crops failed. These intrinsic attributes of common and Italian millet’s nature tie-in with traditional Roman values, connecting Romans with their perceived past as a conservative, hardy agrarian people living off the land. Based upon limited ritual evidence, common and Italian millet were likely traditional Roman foods that continued to be offered to the gods. Hence, common and Italian millet appear to fit into the model of the conflicted Roman psyche of traditional agrarian values and the reality of expanding new frontiers and increasing influx of foreign foods and ideas within the empire.

Thus, the evidence for millet reveals that millet was part of the Roman dietary assemblage, to varying degrees, throughout the Roman empire. Based upon the limited evidence to date it looks like millet consumption within Roman society was a more complex issue than the ancient sources alone would lead one to believe and millet consumption was closely tied to Roman social, economic and cultural values (Killgrove and Tykot 2013, p. 36). As more data is collected it is suspected that millets’ reputation and usefulness in the ancient world will become clearer and millets will move beyond being regarded simply as animal fodder.

 Charlene Murphy, UCL


Boivin N, Fuller D, Crowther A (2012) Old World globalization and the Columbian exchange: comparison and contrast. World Archaeol 44(3):452–469

Dalby A (2003) Food in the ancient world, from A to Z. Routledge, London

De Wet JMJ, Oestry-Stidd LL, Cubero JI (1979) Origins and evolution of foxtail millets (Setaria italica). J Agric Tradit Bot Appl 26(1):53–64

Fraser EDG, Rimas A (2010) Empires of food: feast, famine, and the rise and fall of civilizations. Free Press, New York

Jashemski, WF (2002) The Vesuvian sites before A.D 79: the archaeological, literary, and epigraphical evidence. In: Jashemski WF, Meyer FG (ed.) The natural history of Pompeii. Cambridge University Press, Cambridge, pp 6–28, chap. 2

Keenleyside, A, Schwarcz, H, Stirling L, Lazreg, NB (2009) Stable isotopic evidence for diet in a Roman and Late Roman population from Leptiminus, Tunisia. J Archaeol Sci 36:51–63

Killgrove K, Tykot RH (2013) Food for Rome: a stable isotope investigation of diet in the imperial period (1st–3rd centuries AD). J Anthropol Archaeol 32:28–38

Murphy, C (2015) Finding Millet in the Roman world. Archaeological and Anthropological Sciences. DOI 10.1007/s12520-015-0237-4.

Murphy, C, Thompson, G and Fuller, DQ (2013) Rubbish, Refuse and the Romans, the Archaeobotanical Assemblage of Insula VI.I, Pompeii. Vegetation History and Archaeobotany, 22 (5), 409-419. doi:10.1007/s00334-012-0385-8

Murphy, C (2011) Pompeii, A Changing City, the Archaeobotanical Analysis of Insula VI.I. Unpublished PhD thesis, Institute of Archaeology, UCL.

Spurr MS (1986) Arable cultivation in Roman Italy c. 200 BC–c. AD 100. Journal of Roman studies, monograph no. 3. Society for the Promotion of Roman Studies, London

Valamoti SM (2013) Millet, the late comer: on the tracks of Panicum miliaceum in prehistoric Greece. Archaeol Anthropol Sci. doi:10.1007/s12520-013-0152-5

Zohary D, Hopf M, Weiss E (2012) Domestication of plants in the Old World: the origin and spread of cultivated plants in west Asia, Europe and the Nile Valley, 4th edn. Oxford University Press, Oxford

Ancient Sources

Oribasius (1997) Dieting for an emperor: a translation of books 1 and 4 of Oribasius’ Medical compilations. Brill, Leiden

Strabo (1924) In: Jones HL (ed) The Geography of Strabo. Harvard University Press, Cambridge

The place of plants in pre-Islamic Saharan trade: cultivation and consumption of plant foods by the Garamantes

The Garamantes, a tribal group who occupied Fazzan, southern Libya from c900BC to  AD500, held a strategically important position within a complex trade network linking sub-Saharan and east Africa with the Mediterranean and Nile valley.  Through contact with diverse cultural groups they evolved agricultural, material, burial and building traditions which reflect both their desert origins and their trading links. Plant remains and artefacts associated with food and consumption recovered during the Fazzan Project , particularly excavations at the ancient capital of Garama (Old Garma), led by David Mattingly of Leicester University, indicate that this ‘fusion’ culture can also be seen in the agricultue and diet of the population.

Aerial view of excavations at Garma showing the final Garamantian phases . (Copyright Toby Savage)

Settled agricultural communities emerged in the first half of the 1st millennium BC, with a material culture evolved locally from the preceding late pastoral Neolithic period. The crops cultivated were more Mediterranean in character: emmer wheat, barley, and a small amount of bread type wheat, as well as the fruits grapes and figs. Dates, which are more suited to the desert oasis environment, were also cultivated, but may have been introduced to the region somewhat earlier (dates have been found in late pastoral Neolithic graves). Crops were probably grown in small garden plots, likely to have been regularly heaped with manure, perhaps with shade provided by the date palms, and irrigated by water, possible drawn from wells.

Towards the end of the first millennium cultural and archaeobotanical evidence indicates a new wave of influences, contemporary with the emergence of complex trading societies both north and south of the Sahara (Greek and Phoenician settlements, the beginnings of the Kingdom of Ghana, and city state of Djenné-Djenno, and the Kushite/Meroitic Kingdoms). The introduction of a new irrigation technique, foggara, originating from Persia and introduced via the Egyptian Oases, must have revolutionised arable production. Foggara technology enabled an increase in the speed water was taken to the fields and the area of land irrigated, as well as enabling year round cultivation and the introduction of two significant summer crops from sub-Saharan Africa: sorghum and pearl millet. Sorghum arrived in Nubia (southern Egypt/northern Sudan) at a similar time but appears to have been a different type to that found in Garamantian Fazzan, and which eventually evolved into the durra type grown both in Nubia and Fazzan today. Slaves may have been used to construct and maintain the new irrigation system and it is possible that they brought crops and associated artefacts with them. New pottery types included flat doka, ceramic platters which are thought to be associated with making flat, pancake type bread, and in Sudan are assumed to be associated with sorghum. The doka disappear from the site of Garama in the Islamic period when bread ovens started to appear, suggesting a change in bread technology associated perhaps with new social influences. Quern stones were flat, saddle quern types, and are found on both settlement sites and within burials, suggesting a particular significance. Other cultural evidence for sub-Saharan contacts include a lip plug found in the grave of a young woman. From the Mediterranean region Punic beads and Punic/Hellenistic amphora have been found on settlement sites, indicating the import of consumable goods. Graves were marked by upright stone stele and offering tables, which may have contained food.


Agriculture in Fazzan: adding manure to garden plots (top) and irrigating winter cereals (bottom). Copyright Ruth Pelling.

SEM images of Pennisetum glaucum (pearl millet) (left) and Sorghum bicolour (sorghum) (right) from the site of Tinda B. Copyright Ruth Pelling.

SEM images of Pennisetum glaucum (pearl millet) (bottom) and Sorghum bicolour (sorghum) (top) from the site of Tinda B. Copyright Ruth Pelling.


The 1st and 2nd centuries AD represent the high point of Garamantian society, coinciding with the maximum extent and wealth of the Roman Empire as well as the flourishing of sub-Saharan kingdoms. It is likely that the Garamantes grew wealthy due to their access to water and strategic geographical position at the centre of a complex trading network which took gold, salt, slaves and wild animals (and later cotton and rice) to the north, and consumable goods like oil and wine as well as ceramics, glassware and worked jewellery, to Fazzan, and maybe beyond. In this period large quantities of imported vessels appear in both grave assemblages and on settlement sites: amphorae, glassware, sigilata, oil lamps and more utilitarian ceramic coarse-wares.  Rotary querns appear for the first time, initially imported basaltic lava examples, soon replaced by locally produced sandstone versions. Deep mortars also emerge, better suited to pounding millets and other small seeds, as well as non-food items. New food crops which appear include imported fruits such as pomegranate, almond and cucumber or melon (the seeds are difficult to distinguish if preservation is poor). A new wheat type, durum wheat, appears which is the variety associated with pasta and couscous. Fish bones of Mediterranean origin have even been recovered, perhaps imported with garum fish paste. Camel also appears at this time, while large deposits of seeds of wild, thorny shrubs only usually consumed by camels, suggest some tethering or stabling of the animals within the settlement. Ceramic fine-wares were mostly recovered from grave assemblages, often accompanied by quern stones (rotary and saddle querns), glass, amphorae, and oil lamps, perhaps related to the funerary rite. Stele and offering tables continue to mark the graves. In contrast the ceramics from within the settlement sites tend to be of more utilitarian imported coarse-ware as well as locally produced pots. New types of coarse-ware include casseroles and saucepans with lids. It is tempting to suggest a link with the introduction of durum wheat and perhaps a development of more sauce based couscous type dishes.

Towards the end of the Garamantian period the population and the wealth of imported goods declined. The range of exotic imported fruits dwindled and sorghum disappears from the archaeobotanical record. Interestingly cotton, a crop first domesticated in the Indian sub-continent, but also possibly domesticated in sub-Saharan Africa, appears in the archaeobotanical record during the later phases of the Roman Empire, perhaps to fill demand for cotton cloth at a time when the decline in global trade networks reduced the availability of imported Indian cloth. The disappearance of sorghum could be related to the reduction in sorghum growing population, particularly if it was associated with sub-Saharan slaves. When sorghum was reintroduced in the post-medieval period it was the durra type, which may have evolved in Nubia, and which is today known by its variety name ‘durra’, suggesting both name and crop were late introductions. Unlike sub-Saharan Africa, sorghum is not today cultivated as a bread grain or beer, but is more commonly use as animal feed, or a snack food (as pop-corn) or occasionally mixed with barley flour to produce bazeen, a flour, salt and water paste, eaten with meat and gravy. This is typical of a late adoption of a crop or food type into a pre-existing culinary tradition, rather that its introduction as part of a new cuisine or with a migrating population.

Ruth Pelling, Historic England

This discussion draws on a range of specialist and archaeological evidence generated by the project team, all of which is published in a series of volumes:


Mattingly, D. J., Daniels, C. M., Dore, J. N., Edwards, D. and Hawthorne, J. 2003. The Archaeology of Fazzān. Volume 1, Synthesis. The Society for Libyan Studies/Department of Antiquities, London.

Mattingly, D. J., Daniels, C. M., Dore, J. N., Edwards, D. and Hawthorne, J. 2007. The Archaeology of Fazzān. Volume 2, Gazetteer, Pottery and Other Finds. The Society for Libyan Studies/Department of Antiquities, London.

Mattingly, D. J., Daniels, C. M., Dore, J. N., Edwards, D. and Hawthorne, J. 2010. The Archaeology of Fazzān. Volume 3, Excavations of C.M. Daniels. The Society for Libyan Studies/Department of Antiquities, London.

Mattingly, D. J., Daniels, C. M., Dore, J. N., Edwards, D. and Hawthorne, J. 2014. The Archaeology of Fazzān. Volume 4, Survey and Excavations at Old Jarma (Ancient Garama) carried out by C. M. Daniels (1962–69) and the Fazzān Project (1997–2001). The Society for Libyan Studies/Department of Antiquities, London.

Hunting in the Roman world: anthropology, animal bones and ancient literature

By guest blogger Dr Martyn Allen (University of Reading)

Hunting in the Roman world is a rarely studied and poorly-understood phenomenon, and it is normally considered to have been quite unimportant to the Romans. This was a society which, of course, was based upon agriculture. Its wealth and complexity was founded on its ability to produce and transport large quantities of food to support both urban and rural populations, as well as a geographically-widespread military with high demands. As with all cultures, animals play a central and fundamental role, providing sources of food, means of transport, and offering companionship, practices which are often key to understanding those cultures. Due to the substantial economic importance of domestic livestock across the Roman Empire we could be forgiven for overlooking the relationships between people and wild animals. However, if these relationships are to tell us anything about Roman society, we must think more carefully about what it actually means to hunt: who is doing it, where they are doing it, and why. Hunting in agricultural societies is very different to that in hunter-gather societies. Firstly, for the latter group, there is no such thing as a ‘wild animal’–the classification exists only in our consciousness as the opposite of a domesticated animal–and, secondly, the ways in which hunter-gatherer groups and farmers behave towards and think about, animals are very different. A hunter-gather’s relationship with animals is one based upon kinship and trust, whilst the farmer-livestock relationship is more exploitative and unequal (Ingold 2000; Willerslev 2004). Hunting in agricultural societies must then also be seen in different terms.

Farming was first developed as a means of producing food, larger quantities of it in fact, leading to the development of more complex societies and settlements, such as towns and cities, a shift which largely eliminated the need to hunt animals for food. And yet, people did not stop hunting. The reasons why are much debated, because although game animals tend to be eaten they are not required for survival as they would be for hunter-gatherer communities (except perhaps in extreme circumstances, but not as a general rule). In addition, there is considerable debate surrounding what we actually classify as ‘hunting’ within farming communities. Anthropologists would argue that a hunt must involve an animal which is free to run away, or at least is perceived to be ‘free’ (Cartmill 1993, 29). The ‘running away’ bit is important, because the animal must not be under any obvious form of control or restriction. The hunt must also involve direct violence between the hunter and their quarry, albeit usually a short and very final act. There are, however, notable exceptions, such as trapping (fur-bearing mammals, for example). It is debatable of course, but this is not strictly a form of hunting under these criteria.

Importantly, it is the act of hunting which is important, rather than the production of meat. It is a social performance which sends an array of messages about the identity, or identities, of the people involved. Much ancient Greek literature concerning hunting discusses it as a form of military training, not only in a practical sense but also metaphorically, where the quarry are associated with a human enemy, and the hunting landscape (e.g. woodland, etc.) is perceived as a foreign land (Cartmill 1993, 32). This shows us that it is not only the relationship between people and wild animals which is important, but also their connection to the place where hunting is undertaken. Hunting takes place in the wild, within ‘nature’, beyond the domestic realm. Compared to ancient Greek literature, perhaps surprisingly, Roman writers were less intent on placing the same emphasis upon hunting. This is a cultural difference, but there are hints of Roman attitudes towards the natural world in the writing of its poets and in iconographic depictions of some of its elite citizens. More specifically, some of the early Emperors (particularly the Hadrianic ones) appear to have been keen on associating themselves with hunting, or at least the idea of hunting. Domitian, in particular, is to be found in numerous reliefs and carvings where he can be seen slaying wild beasts (Tuck 2005, 239). The contemporary writer Suetonius (Domitian 19) suggests that this was not merely a symbolic gesture, but was enacted by Domitian in the game park on his estate outside Rome.

However, not everyone in the Roman world was an advocate of hunting. Not because of animal welfare as might be the case today, but because hunting in this context was seen as transparent and overly-extravagant. Occasionally game was not even killed during the end of the chase, but first driven into nets and then presumably slaughtered (perhaps not even by the hunter, but by slaves). Such a practice may have removed one of the most important aspects of the hunt: the violent killing of a free animal. Pliny the Younger, poking fun at some of his notable contemporaries, humorously derided that he captured animals, not with spear and lance, but with pen and notebook in hand (Epistulae 1.6, quoted in Anderson 1985, 100). It was the chase, the show of horsemanship and mastery over the wild beast, and not necessarily the kill that was always important in the mind-sets of some. Despite the ambiguous nature of the hunt in the Roman world, as shown in its literature and its iconography, it would appear to have been an important device which demonstrated elite identity and social power within the landscape. It represented the authority of the emperor as protector of the Roman state and its people. But, how much hunting actually took place in the Roman Empire: how much meat from wild animals actually made it to the dining table and in whose house? And, was it really hunting, or merely hollow demonstrations by the wealthy few?

In Britain, archaeological excavations have produced an astonishing number of animal bone assemblages from an array of towns, military sites, and rural settlements all occupied during the period of Roman occupation (historically taken as AD43-410). However, the remains of wild animals in these assemblages are generally rare, normally occurring in very low quantities compared to the bones of domestic livestock. In a comparatively large assemblage, animal bones from wild species would normally register somewhere between 0.5%-2% of the total identified. An assemblage with a wild component of around 5% would be seen as exceptional, and only very occasionally do they occur in a greater frequency than this. But we should expect this. Domestic livestock were of major economic significance in Roman Britain, not only for meat, but also for dairy, wool, leather, horn, fat, and bone, not to mention the importance of manure and the use of cattle and horses for ploughing and traction–vital components of an agricultural society. It makes complete sense that the remains of domestic livestock dominate the faunal assemblages from Roman farms and towns. However, as I have sketched out above, the importance of wild animals in agricultural societies does not lie in their role within the economy, which would have been negligible at best, but in their symbolic importance. Deer bones found at late Iron Age and Roman sites in Britain are recovered in greater frequencies on villas and military sites, compared to other types of settlement (Allen 2014, 177), indicating that venison was more commonly eaten at higher-status sites. As is suggested by the historical accounts, the archaeological evidence does appear to show that deer more frequently hunted by those with landed wealth and by military groups, people with a level of social and economic power.

There is also evidence that some elite groups in Roman Britain furnished their estates with deer parks. Fishbourne Palace in West Sussex produced bones of fallow deer, a species which, although common today, is not native to Britain and was only properly introduced after the Norman Conquest. Radiocarbon dating and isotopic analysis (looking at chemical signatures) of fallow deer teeth from Roman Fishbourne showed that one individual had been imported during the mid-1st century AD, whilst a second had lived its entire life in Britain (Sykes et al. 2006). The results of this important analysis could be inferred as evidence for the maintenance of a breeding herd of fallow deer at the site at least into the 2nd century AD.

Unfused red deer Cervus elaphus femurs from Roman Fishbourne (photo by M Allen)

Unfused red deer Cervus elaphus femurs from Roman Fishbourne (photo by M Allen).

The analysis of red deer bones from Fishbourne, which were comparatively numerous, indicates that these animals may have been deliberately managed (Allen 2014, 178). Numerous red deer femur (thigh) bones came from skeletally-immature animals, whilst most other red deer bones were found to have been fully developed. The femur bone in the red deer is known to be one of the last to fully mature, which suggests that this species were generally killed and eaten at Fishbourne when within a restricted age range. Measurement of red deer bones has also indicated that most were from particularly large animals compared to those from other sites, suggesting that male stags were a focus for slaughter (Allen 2014, 179). Such a cull pattern would imply that red deer of particular age and gender were selected for killing rather than randomly caught in the wild, i.e. a deliberate management strategy was being employed. In addition, the recovery of some exceptionally young roe deer remains may also support this interpretation. Radiographs of a number of roe deer mandibles from Fishbourne have shown that some were culled almost as new-borns (neonates). This may indicate that neonatal roe deer were considered a delicacy by the inhabitants of the palace, and consumed soon after birth, but we must also think of this evidence in terms of management strategies. Roe deer were certainly hunted at Fishbourne, as attested by the remains of older animals, but the selection and removal of young would have enabled population control for risk management as well as the maintenance of a healthy herd.

Radiograph of neonatal roe deer Capreolus capreolus from Roman Fishbourne (photo by M Allen).

Radiograph of neonatal roe deer Capreolus capreolus from Roman Fishbourne (photo by M Allen).

It could be argued that deer were being slaughtered solely to keep local numbers of wild deer down, perhaps if they were being destructive to local farmland. However, if this was the case we should see similar evidence from other sites. Predominantly, this is not the case; deer management was not practiced by the vast majority of the rural population. On the other hand, if deer were being kept within an enclosed space, such as a game park, they would need to be consistently managed and maintained, firstly to minimise damage to the local environment, and secondly to provide suitable animals for the hunt, i.e. young stags. If we accept that the zooarchaeological evidence represents the presence of a managed deer park, it is helpful to view it against the ancient literature, because it is here that we find the attitudes of Roman elite groups towards wild animals and how this was reflected in what it meant to have been a member of the elite classes during that period. Of course, much more work is required on this desperately under-studied subject, particularly with regards to what the evidence means in terms of the relationships between different people, with animals, and also with the landscape, but when placed in context it is truly amazing what an animal bone can tell us!

If you want to know more about Martyn and his work, please take a look at his profile on our Guest Bloggers page.


Allen, M.G. (2014) ‘Chasing Sylvia’s Stag: Placing Deer in the Countryside of Roman Britain’ in Baker, K., Carden, R. and Madgwick, R. (eds.) Deer and People, Windgather, Oxford, pp.174–186.

Anderson, K. (1985) Hunting in the Ancient World, University of California, Berkeley.

Cartmill, M. (1993) A View to a Death in the Morning: Hunting and Nature through History, Harvard University Press, Cambridge.

Ingold, T. (2000) ‘From trust to domination: an alternative history of human-animal relations’ in Ingold, T. (ed.) The Perception of the Environment: Essays in Livelihood, Dwelling and Skill, London, Routledge, pp. 61–76.

Sykes, N. J. White, J. Hayes, T. and Palmer, M. (2006) ‘Tracking animals using strontium isotopes in teeth: the role of fallow deer (Dama dama) in Roman Britain’, Antiquity 80, 948–959.

Tuck, S. (2005) ‘The origins of Roman imperial hunting imagery: Domitian and the redefinition of virtus under the principate’, Greece & Rome 52, 221–45.

Willerslev, R. (2004) ‘Not Animal, Not Not-Animal: Hunting, Imitation and Empathetic Knowledge among the Siberian Yukaghirs’, Journal of the Royal Anthropological Institute 10.3, 629–652.

Late Iron Age food at Silchester

The foods that we choose to eat and avoid are strongly tied up with our identity. Yet the foods being eaten in one part of the world can change due to the movement of people, who bring their food customs with them, and/or the adoption of new foods by resident populations. A town at the edge of the Roman Empire at the end of the first century BC provides a glimpse of how and why foods changed in the past.

Spelt wheat grains

Iron Age diet in Britain was pretty dull, consisting mainly of spelt wheat and barley made into porridge or bread, some meat, and probably some collected wild plants, such as black mustard. Towards the end of the first millennium BC, transport amphorae containing wine and olive oil began to make their way to settlements in south-eastern Britain. The types of crockery used by people living in oppida (large “proto-urban” nucleated settlements) also changed, with plates and bowls becoming more common. Other distinctive food remains are also found at these sites which were not commonly used during the Iron Ae, such as oysters. Whilst around 50 new plant foods were imported to Britain after AD 43, whether any of these were adopted before the Roman invasion was not known.

Late Iron Age occupation at Insula IX ©LisaLodwick

Late Iron Age occupation at Insula IX ©LisaLodwick

Ongoing excavations in Insula IX at Silchester have revealed numerous Late Iron Age wells containing waterlogged sediments, providing the best potential in southern Britain for investigating pre-Roman dietary change. Sediments from the bottom of “Well 10421”, in the centre of the excavations were sampled, and processed with a flotation tank to extract the waterlogged plant remains. The ceramics from these well fills were dated to c. 20/10 BC – AD 10/20 – the period of the first occupation within this area of the oppidum. A few months later, when the plant remains were studied under a microscope, amongst the typical seeds of weedy plants like docks and chickweed were several very exciting items.

Two olive stone fragments were present – the first time that olives had been found in Late Iron Age Britain. Olive stones have been found only from major towns and high status sites in Roman Britain, but would have been eaten regularly in the Mediterranean. Several seeds of celery were also found. Celery is native to Britain, but only grows in coastal areas, and very rarely at inland salt springs. The seeds from Silchester (not a salt spring) are much more likely to have been imported from the continent to be used as flavourings. A few coriander seeds were also found in the samples, which is definitely not native to Britain. Written evidence, such as Apicius, tells us that these flavourings were used in Roman cooking, and their addition to the Iron Age store cupboard may have provided a welcome spicing up to the staples of bread and porridge. These flavourings and olives may have only been eaten by the leaders of Silchester in feasting events.  Future study of the plant remains should tell us how widespread these new foods were at the Late Iron Age settlement.

A - Celery seed, B - coriander seed, C - dill seed, D - olive stone ©LisaLodwick

A – Celery seed, B – coriander seed, C – dill seed, D – olive stone ©LisaLodwick

We now know that the foods eaten at Late Iron Age Silchester were changing, as well as the crockery that the food was eaten off of. But does this get us any closer to understanding the identity of the residents of the oppidum? The interesting thing is that we are not yet sure who was living at Silchester, and the range of foods being eaten can fit easily into different scenarios.

One possibility is that some of the residents of Silchester were immigrants from Northern France, fleeing from political unrest (Fulford and Timby 2000, p 546). Not only does Silchester have exceptional amounts of imported material culture from the continent, but links have been drawn between the COMMIOIS whose name appears stamped on coins, and the historical figure of Commius, a king of the Beglic Atrebates in northern Gaul (Creighton 2006). These political refugees could have carried some of their foods with them across the channel. Celery and coriander have been found from several sites in northern Gaul in this period, such as Damary in the Aisne valley (Bakels 1999). Excavations at Nemetecum, the capital of the Atrebates tribe in Northern France, have produced several similar foods to those found at Late Iron Age Silchester including coriander seeds, hazelnut shell, hulled barley and spelt wheat (Derreumaux and Lepetz 2008).

The second option is that the residents of Late Iron Age Silchester mostly moved to the site from the local region of central-southern Britain. Faced with the need to forge a new group identity, they made use of cross-channel contacts and obtained new ingredients, alongside new types of pottery, wine and olive oil. The flavourings which they managed to get hold of were those also being adopted in Northern France (celery and coriander) – these might have been the most commonly available, the easiest to transport (as dried seeds), or those which were easiest to combine with the Iron Age cuisine of stews and porridge.

The third option is of course somewhere in between – some occupants of Silchester may have travelled back and forth to north-western Europe or even Italy itself, where they may have acquired a taste for, and access to, new flavourings, whilst other residents of Silchester were locals, who adopted these new foods to emulate the lifestyles of their leaders.

Trying to interpret what the presence of ‘new” or “luxury” foods mean is tricky. Once analysis of the Late Iron Age settlement, objects, animal bones, fish bones and plant remains from the Silchester ‘Town Life Project’ is complete, it will be possible to explore more fully what the diet and lifestyle of the residents were, and we may be able to better evaluate whether these people were locals adopting foreign luxuries, or immigrants trying to remind themselves of home. Who ever the Late Iron Age Callevans were, the use of new food flavourings was clearly an important  aspect of their lifestyles.



Excavations within the Silchester Insula IX have now come to an end after 18 years, with post-excavation work now concentrating on interpreting the excavated features, finds and environmental evidence. Details of how to visit the site can be found here. The Town Life Project excavations are run by the University of Reading, and the research discussed in this paper was funded by the AHRC.



Apicius, C., Grocock, W., & Grainger, S. (2006). Apicius : a critical edition with an introduction and an English translation of the Latin recipe text Apicius (pp. 85–115). Totnes: Prospect.

Bakels, C. (1999). Archaeobotanical investigations in the Aisne valley, northern France, from the Neolithic up to the early Middle Ages. Vegetation History and Archaeobotany, 8, 71–77.

Creighton, J. (2006). Britannia: the Creation of a Roman Province. Cambridge: Cambridge.

Derreumaux, M., Lepetz, S., With, Jacques, A., & Prilaux, G. (2008). Food supply at two successive military settlements in Arras (France): an archaeobotanical and archaeozoological approach. In Stallibrass, S. and Thomas, R. Feeding the Roman Army. The Archaeology of Production and Supply in NW Europe (pp. 52–68). Oxford: Oxbow Books.

Fulford, M., & Timby, J. (2000). Late Iron Age and Roman Silchester: Excavations on the site of the Forum Basilica, 1977, 1980-86. London: Society for the Promotion of Roman Studies.

Lodwick, L. (2014). Condiments before Claudius: New Plant Foods at the Late Iron Age Oppidum at Silchester, UK. Vegetation History and Archaeobotany 23: 543-549doi:10.1007/s00334-013-0407-1 (An open access version of this paper can be found here.)

Roman food at the AEA Conference

Environmental archaeologists spend their time studying the remains of animals, fish, plants, insects and snails from archaeological sites. The Roman world has left behind a large amount of archaeology, from monuments to rural farmsteads, and environmental archaeologists have been working hard to study the bioarchaeological remains from these sites for decades. The Association for Environmental Archaeology is the international umbrella organisation for environmental archaeology, and they hold annual conferences where specialists get to discuss all of the latest advances in the field.

This is what a meeting of environmental archaeologists looks like!

This years meeting was held in Plymouth at the start of November, and as always, some of the research presented explored food and agriculture in the Roman world. Here’s a quick run down of the some cutting-edge Roman themed environmental archaeology research!

Husbandry and agriculture in a changing environment: bioarchaeological evidence from Althiburos (Tunisia), from the late Bronze Age to the Roman period (11th c. BC – 5th c. AD)

Silvia Valenzuela-Lamas (University of Sheffield) and Dani Lopez

This paper presented the results from excavations at Althiburos in Tunisia, the Roman city of Numidia. Compared to north-western Europe, we have little environmental-archaeological data from Northern Africa, or the the Iron Age period before the Roman Empire. Silvia talked about long term changes from the Bronze Age to the Roman period at the site, and interestingly showed that cattle decreased in importance through time.

Of bantams and broilers: two-thousand years of chicken breeding in London

Richard Thomas (University of Leicester), James Morris, Matilda Holmes

If you ever wondered about chickens in the Roman world, then your in luck, as there is currently a big interdisciplinary project investigating human-chicken interactions. Chickens were introduced to Britain on a small-scale in the Late Iron Age, but became more common in the Roman period. By undertaking lots of measurements on chicken bones from London, this research showed how chickens have changed in size over time, and what this meant for how they were used in different periods.

Synthesis and the City

James Morris (University of Central Lancashire) and Matilda Holmes

Another paper on London, this discussed the pros and cons of analysing the very large amounts of zooarchaeology (animal bone) data from London.  So many excavations have taken place in London, that differences in the types of site excavated, the type of sampling that was undertaken and how the animal bones are counted make it difficult to find and interpret patterns in how animals were used in the past. In short, this paper showed you can get over these problems, and a cool insight for Roman food is that there was very little cod eaten in Roman London.

Agriculture and rural society in Roman Britain: new perspectives from developer-funded archaeology

Martyn Allen (University of Reading)

Elsewhere in our blog we’ve mentioned the large amount of data produced by developer-funded archaeology, and how it can tell us about Roman food. This paper present an ongoing research project which is collecting together archaeological evidence from all rural settlements excavated in Roman Britain. Some of the emerging food-related findings include the presence of complex rural settlements, with multiple enclosures for different animals and crops.

Transporting food plants: a network and spatial analysis approach for the integration of archaeobotanical and landscape data of Roman Britain

Alexandra Livarda (University of Nottingham) and Hector Alex Orengeo

We’ve known for quite a while that archaeobotany (the study of plant macro remains from archaeological sites) can tell us what people ate in the Roman world, but this exciting new research also aims at using archaeobotanical data to find out how food was transported in the world. The paper combined data on Roman transport routes in Britain with the find spots of exotic plant foods, did some complicated network and spatial analysis, and then showed how trade routes changed over time.

The AEA hold biannual conferences on environmental archaeology – you can snoop on their activities and the later research in the field by checking them out on twitter @envarch and follow the latest news on their website.

Grain milling in the Roman world

One of the most iconic sights in Pompeii are the bakeries with their donkey mills. These hourglass-shaped mills were one of several types in use in the Roman world to mill grain ready to make bread. The donkey mills have a solid lower stone, called a meta, which was bell-shaped. Over the top was a hollow upper stone, which was shaped like an hour glass; this was called the catillus. The catillus functioned like a hopper into which the grain would be poured for grinding. The stone used is hard and abrasive to help with the grinding, often basalt. As their name suggests, these mills were probably rotated by donkeys. In addition to these extant examples, there are also numerous depictions on tombstones and sarcaophagi. These come not only from Italy, but also from North Africa, the Aegean and Gaul, which attests to how widespread they were in the Roman world.

One of several bakeries in Pompeii where grain was milled on site using donkey mills.

One of several bakeries in Pompeii where grain was milled on site using donkey mills.

One of the objects we have selected for our exhibition comes from a different kind of mill: a rotary hand quern. These are also made from hard, abrasive stone to help grind the grain. Rotary querns are flatter than donkey mills, but also have an upper and lower stone. These kinds of mill would have ground smaller amounts of grain and would have been sufficient for a domestic household.

At the other end of the scale are watermills, which could grind grain on an industrial scale. Studies of Roman watermills have suffered in the past from the traditional view of technological stagnation in the ancient world. It was thought that slave labour and lack of interest in investment in new technologies caused a lack of water milling technology. As a consequence, for a long time, all water mills were thought to date to the early medieval period at the earliest. This idea was challenged by Orjan Wikander in the 1980s, who started to find archaeological evidence for watermills belonging to the Roman period.

There are two main types of watermill that are distinguished by the orientation of the wheel: vertical or horizontal. The vertical wheel requires a right-angle gear and turns a horizontal wheel shaft on the other end of which is a vertical cogwheel. This cogwheel drives a horizontal cogwheel and the millstones above. The horizontal wheel does not require a gearing system and is driven by water conducted either by a steep chute onto oblique paddles on the wheel or from a nozzle at the base of an arubah penstock (which functions like a tube delivering water to the wheel), or exceptionally, as a turbine.

Schematic drawing of a vertical-wheeled mill with a right-angled gear.

Schematic drawing of a vertical-wheeled mill with a right-angled gear.

Schematic drawing of a horizontal-wheeled mill with an arubah penstock.

Schematic drawing of a horizontal-wheeled mill with an arubah penstock.

In most parts of the Roman world, mills are of the vertical-wheeled type. In some cases, as at Barbegal in France, there were 16 wheels in a single installation. This multi-wheeled mills give a strong impression of the scale of grain milling that was occurring in the Roman world. They suggest that grain was being milled at a near-industrial scale and so that there was a large market for grain and bread. This market may have been related to the annona, a system of free grain (and later bread) given out to the populace – this is where the phrase ‘bread and circuses’ comes from i.e. that these were given to the people to keep them happy and peaceful.

In the Near East, the majority of the 25 known watermills seem to be of the horizontal type with an arubah penstock. The arubah penstock is usually a 6-10 m high stone tower that contains a column of water. This column of water is let out through a narrow orifice at the bottom of the tower, so the wheel is driven by a jet of water under pressure. It seems that this type was used in the Near East because it is a particularly efficient design to use when there is a limited flow of water.

Dating watermills is fraught with difficulty, so most of the Near Eastern mills I looked at in my PhD thesis were only tentatively dated to the Roman period. A case in point is the mill complex at Lejjun, which was thought for several years to date to the Roman period, but excavation subsequently revealed that it dates back only as far as the 19th-century AD Ottoman use of the area (McQuitty, A. “Watermills in Jordan: technology, typology, dating and development,” in Amr, K., Zayadine, F. and Zaghloul, M. (eds) Studies in the History and Archaeology of Jordan 5 (Amman: Dept of Antiquities, 1995), 745-751.). In some rare cases, the mills can be dated from literary references, for example the mills at Amida were mentioned by Ammianus Marcellinus (18.8.11) and so can be firmly dated to AD359.

For more info and a list of all possible Roman watermills in the Near East, see my book (Zena Kamash 2010 Archaeologies of Water in the Roman Near East (Gorgias Press) and

For a full account of grain mills and milling, in the ancient world, see LA Moritz 1958 Grain Mills and Flour in Classical Antiquity – old, but still good on donkey mills and rotary hand querns.

Finding evidence for fish-breeding and fish-salting sites in the Roman Near East

When people think about fish in the Roman world, they often think about garum: a sauce made from salted and fermented fish that is similar to Thai fish-sauce (‘nam pla’). As well as this, we do also know of recipes for fresh fish from Apicius, for example baked bream and mackerel stuffed with mint, honey and hazelnuts. We also know that Romans ate salsamenta: salted fish. How much do we know, then, about where fish were bred and salted? In this blog post, I am going to focus on this question for a few sites in the Roman Near East (broadly the modern Middle East).

Painting of fish in the market building at Pompeii, where fish bones were also found in the drain.

Painting of fish in the market building at Pompeii, where fish bones were also found in the drain.

Identifying fishponds and fish-breeding sites archaeologically is not very easy as most vats and ponds could have other uses as well. We do, though, have a useful written source, Columella, who tells us that jars (‘cells) set into the walls were essential in a fish-breeding pond (De Re Rustica 8.1.3, 17.1-6). Identifying this special feature is much easier and has actually been done for some sites in the Roman Near East.

Such jars, set horizontally into the walls, were found at Khirbet Sabiya (300 jars: Ayalon 1979:175-177, 179), Caesarea (6th century: 60 jars (Gibson 1991:41) and Tel Tanninim: 16 Gaza amphorae (Stieglitz 1998:63-5)) and Sataf (Mango 2002:325). A similar arrangement was found in the south-eastern reservoir at Andarin in Syria. Instead of jars or amphorae rectangular recesses were found at the base of the reservoir walls, possibly numbering 200 (Mango 2002). The example from Andarin was very large measuring 61 m long x 61 m wide x 3 m deep because it also acted as a reservoir for irrigation supplies. The freshwater n this pond makes it possible that catfish were bred there. Catfish were the only freshwater fish bred in fishponds cited in ancient sources, such as Apicius and Pliny (André 1981:109-113); in addition, catfish bones have been identified in the assemblage retrieved from the cistern Andarin bathhouse, though not from the reservoir itself (Mango 2009:75).

Columella also recommends that the water for the fish should circulate (De Re Rustica 8.1.3, 17.1-6), which is what we find at most of these sites. The examples from Dor seem to have been fed by seawater (Raban 1995:343). A pipeline branch (Channel E) from the Caesarea High Level aqueduct channel A fed the reservoir at Tel Tanninim (Stieglitz 1998:57-8). Spring water from a spring flow tunnel at Ein Bikura, Sataf fed a pool with two rows of ceramic jars in its sides with mouths towards the pool (Gibson 1991:41). In all of these cases the nature of the water supply means that a constant supply of circulating water was ensured (see Columella De Re Rustica 8.1.3, 17.1-6).

There are also examples from the Near East where we don’t have jars, but think that fish might have been bred in these installations. These are some paired rock-cut tanks from Dor (Raban 1995:343) and Beirut (Thorpe 1998-9: 36-38), which had no recesses or jars. It has been suggested that the Beirut tanks, which were lined with opus signinum (waterproof concrete), might have been holding tanks for holding the catch after the return of the fishing vessels. This is because the vats were in close proximity to a cove to which they were linked by a flight of steps. A similar function is also possible for the Dor installations as they were also situated very close to the coastline.

Alternatively, it is possible that these were vats for the production of salsamenta where the flesh of the fish was cut up and salted. While there is evidence that fish were salted in the eastern Mediterranean, archaeological evidence has been lacking so far, prompting the idea that maybe the process was done in large ceramic jars, such as pithoi or dolia (Wilson 2006; Curtis 1991:112-8, 129-147). Salting vats in the western Mediterranean (in particular Spain, Portugal and Morocco) were remarkably similar and almost universal in construction, though they varied in size and depth (Trakadas 2005:69-72). While the rectangular or square tanks in the western Mediterranean were usually built of brick or rubble, rock-cut examples are known from Portugal at Punta de l’Arenal and Praia de Angeiras. These Portuguese examples were not joined together as was usual elsewhere, which is similar to the possible salting vats from the East. The western examples were faced with opus signinum, as at Beirut. So, the pairs of vats at Beirut and Dor illustrate some, but not all, of the common traits found at the sites in the western empire. It is possible, but not definite, that these installations may be, then, the first fish-salting sites identified from the eastern Mediterranean.

For more detail about these sites, see my book: Archaeologies of Water in the Roman Near East by Zena Kamash (2010; Gorgias Press)



André, J. L’Alimentation et La Cuisine à Rome (Paris: Belles Lettres, 1981).

Ayalon, E. “The jar installation of Khirbet Sabiya,” Israel Exploration Journal 29 (1979), 175-181.

Gibson, S. “The Sataf project of landscape archaeology and the Judean hills: a preliminary report on four seasons of survey and excavation (1987-1989),” Levant 23 (1991), 23-54.

Mango, M. “Fishing in the desert,” Palaeoslavica 10/1 (2002), 323-330. Mango, M. “Baths, reservoirs and water use at Androna in late antiquity and the early Islamic period,” in Bartl, K. and al-Razzaq Moaz, A. (eds) Residences, Castles, Settlements. Transformation Processes from Late Antiquity to Early Islam in Bilad al-Sham (Rahden: Deutsches Archäologisches Institut, 2009), 73-88.

Raban, A. “Dor-Yam: maritime and coastal installations at Dor in their geomorphological and stratigraphic context,” in Stern, E. and Berg, J. Excavations at Dor: Final Report (Jerusalem: IES, 1995), 285-354.

Stieglitz, R. R. “A late Byzantine reservoir and piscina at Tel Tanninim,” Israel Exploration Journal 48/1-2 (1998), 54-65.

Thorpe, R. “Bey 007: the souks area: preliminary report of the AUB/ACRE project,” Bulletin d’Archéologie et d’Architecture Libanaises 3 (1998-1999), 31-55.

Trakadas, A. “The archaeological evidence for fish processing in the western Mediterranean,” in Bekker-Nielsen, T. (ed.) Ancient Fishing and Fish Processing in the Black Sea Region (Aarhus: Aarhus University Press, 2005), 47-82.

Wilson, A. “Fishy business: Roman exploitation of marine resources,” Journal of Roman Archaeology 19 (2006), 525-537.