Reconstruction
At Vindolanda, a Roman auxiliary fort in northern Britain, food survives not as a single menu but as an unusually detailed archive of transactions, stores, discarded bones, charred seeds, cooking equipment and personal requests. The best-preserved early occupation phases date from about 85 to 120 CE, before Hadrian’s Wall was built nearby. Waterlogged conditions preserved thin wooden writing tablets whose ink records expose the work of feeding soldiers, officers, families, servants, animals and visitors on an imperial frontier.
Cereals formed the logistical foundation. One military account dated 97–105 CE lists measured quantities of soft wheat and halica, a cereal preparation translated as gruel. Another letter, dated 92–97, concerns hundreds of modii of grain loaded into carts associated with Britons, direct evidence that transport and dealings with surrounding populations helped maintain the fort. Archaeobotanical samples have identified hulled barley, spelt, bread wheat and possibly emmer, although samples from different excavation phases should not be treated as one simultaneous store. Querns ground grain into flour, while ovens and shared cooking areas reduced fire risks within timber buildings.
The tablets reveal much more than grain. An early shopping list requests bruised beans, twenty chickens, one hundred apples, one hundred or two hundred eggs, fish sauce and olives; a damaged reference probably mentions mulsum, honeyed wine. An account from 104–120 records a centurion acquiring 45 pounds of bacon plus 15½ pounds of bacon-lard. It also names Atrectus the brewer, probably a producer and seller of cervesa, the barley-based beer familiar in northwestern provinces. These are documented purchases and debts, not necessarily standard issues to every soldier.
Animal bones add evidence unavailable from ink. Cattle, sheep or goats, pigs and deer were consumed, but the precise proportions vary by context and period. Pierced cattle shoulder blades are reasonably interpreted as evidence that joints were suspended during smoking or another curing process, rather than direct proof of a particular recipe. Preservation by salting, drying or smoking would have helped stabilize supplies in a damp climate and during transport. A carefully stored cache of hazelnuts in an extramural household dated about 105–120 illustrates how local or long-established northern foods coexisted with imported Roman vessels and ingredients.
Vindolanda therefore documents a food system rather than a uniform ration: military accounting, private credit, local procurement, household storage, communal processing and long-distance commerce overlapped. Imported olives, wine products and fish sauce connected the frontier to wider imperial tastes, while British carts, regional grain, game and hazelnuts rooted eating in the local landscape. Its modern significance lies in this combination of written and biological evidence, showing that institutional provisioning has always depended on merchants, transport workers, processors and households as well as official policy.
Historical context
The Roman Empire extended from Britain to the Near East and North Africa. Vindolanda stood on the Stanegate frontier established during Rome’s consolidation of northern Britain in the 80s CE. Its early garrisons included auxiliary soldiers connected with units originally raised among the Tungrians of Belgic Gaul and the Batavians of the Rhine delta, alongside families, merchants, servants and other noncombatants. Roads linked the fort eastward toward Corbridge and wider provincial supply routes. Emperor Trajan ruled from 98 to 117 CE; Hadrian succeeded him in 117, and construction of the wall bearing his name began around 122. Vindolanda therefore predates Hadrian’s Wall, although it later participated in the wall-zone economy.
Evidence
Archaeological
StrongStratified fort and extramural deposits contain querns, ovens, cooking and serving pottery, storage features and exceptionally preserved organic objects. Individual finds belong to different occupation phases and do not represent one synchronous meal.
Archaeobotanical
ModerateEnvironmental samples identify cereals including barley, spelt and bread wheat, while a Period 4 extramural structure contained a stored hazelnut cache. Crop proportions vary by context, preservation and excavation phase.
Zooarchaeological
StrongLarge faunal assemblages directly document cattle, caprines, pigs and deer. Bones demonstrate animal use and butchery, but do not by themselves establish portion sizes, recipes or equal access across the community.
Written sources
StrongContemporary ink tablets explicitly record wheat, gruel, bacon, fat, beans, chickens, apples, eggs, olives, fish sauce, grain transport and a brewer. Their contexts and readings are published in scholarly editions.
Food identification
StrongMany foods are securely identified through explicit Latin vocabulary in the tablets and diagnostic plant or animal remains. Damaged words and broad taxonomic categories retain stated uncertainty.
Dating
StrongTablet and settlement dates are based on well-defined Vindolanda occupation periods and stratigraphic contexts. The combined card range nevertheless aggregates evidence from several phases between approximately 85 and 130 CE.
Preparation method
ModerateQuerns and ovens directly support grinding and baking or cooking. Smoking or curing of beef shoulders is a reasonable interpretation of pierced scapulae, not a directly observed recipe.
Geographic attribution
StrongThe tablets and archaeological remains were excavated at Vindolanda in Northumberland, approximately one Roman mile south of the later Hadrian’s Wall.
Historical interpretation
ModerateThe evidence strongly supports a mixed military, household and commercial provisioning system. Calling all documented foods formal military rations would overstate what the accounts and letters demonstrate.
Sources
- 1.Alan K. Bowman and J. David Thomas (1994). Tab.Vindol. 302. Ink writing tablet. The Vindolanda Writing-Tablets (Tabulae Vindolandenses II); RIB Online. romaninscriptionsofbritain.org/inscriptions/TabVHistorical primary source
- 2.Alan K. Bowman and J. David Thomas (1994). Tab.Vindol. 182. Ink writing tablet. The Vindolanda Writing-Tablets (Tabulae Vindolandenses II); RIB Online. romaninscriptionsofbritain.org/inscriptions/TabVHistorical primary source
- 3.Alan K. Bowman and J. David Thomas (2003). Tab.Vindol. 586. Ink writing tablet. The Vindolanda Writing-Tablets (Tabulae Vindolandenses III); RIB Online. romaninscriptionsofbritain.org/inscriptions/TabVHistorical primary source
- 4.Alan K. Bowman and J. David Thomas (2003). Tab.Vindol. 649. Ink writing tablet. The Vindolanda Writing-Tablets (Tabulae Vindolandenses III); RIB Online. romaninscriptionsofbritain.org/inscriptions/TabVHistorical primary source
- 5.Elizabeth M. Greene and Andrew Birley (2024). The Extramural Settlement at Vindolanda in the Early Second Century CE: Defining a Glocalized Environment on the Romano-British Frontier. American Journal of Archaeology 128(1). doi:10.1086/727259Scientific literature
- 6.Alan K. Bowman (2012). Vindolanda tablets. The Encyclopedia of Ancient History. doi:10.1002/9781444338386.wbeah10082Modern synthesis
Limitations
- moderatetitle
“Frontier Rations” may imply a standardized official ration. The strongest evidence combines military accounts, private shopping requests, debts, elite household consumption, environmental remains and civilian trade; not every listed food was issued uniformly to soldiers.
- minordate end
The exceptionally preserved first four occupation periods and most tablets emphasized here date approximately 85–120 CE. Evidence from 120–130 CE belongs to later Hadrianic contexts and should not be assumed contemporary with every tablet or food deposit.
