Roman food has a reputation for its exotic – and sometime, frankly repulsive – dishes. The clichéd recipes, such as dormice in honey (glires melle ac papauere sparsos) or larks’ tongues in aspic (linguas alardum in aspide, also an album by King Crimson), are found satirical literature. The wealthy did eat dormice – Apicius’s well-known de Re Coquinaria (‘On Cookery’) includes a recipe for glires stuffed with a pork forcemeat – but the idea of eating larks’ tongues originates in Monty Python’s Life of Brian. According to Pliny (Historia Naturalis X.68), phoenicopteri linguam praecipui saporis esse Apicius docuit (‘Apicius taught that the tongue of the flamingo is of supreme flavour’), although Apicius did not include a recipe for the delicacy in his book.

Of course, such pretentious foodstuffs were not everyday fare for people living throughout the Roman world. At a time when refrigeration was many centuries in the future, exotic meats and vegetables would need to be preserved – by salting or pickling, for instance – before transport. Even flamingos’ tongues would probably lose their ‘supreme flavour’ during the journey from the Mediterranean to Britain.

Bread, made from wheat or barley flour, was a staple and had been for centuries before the Roman conquest. These same grains went into making the main drinks, ceruisa, a type of low-alcohol wheat beer, and curmi, a low-alcohol barley beer. Both types were sometimes sweetened with honey or flavoured with herbs such as henbane, a poisonous plant with psychoactive effects that can, however, cause convulsions and even death. Oats were used to make porridge or gruel.

Everyday vegetables included onions, wild garlic, turnips, parsnips, cabbage, seaweed, nettles, fennel, sorrel, parsley, spinach, beans and mushrooms. Most fruit was wild, such as blackberries, gooseberries and bilberries. We have previously seen how the local people, the Succii, took their name from the pigs that they bred for meat. They also reared cattle, sheep and goats, while chickens arrived from the Roman world long before Julius Caesar had been born. People seem not to have eaten much in the way of seafood or game before the Roman conquest in AD 43. Secondary products from domesticated animals included milk, cheese and sausages.

The ability to drink milk in adulthood and eat cheese is something that many humans lack; although born with it, most people lose it after breastfeeding ends in infancy and their bodies stop making lactase, the enzyme that allows them to process the sugar lactose, found only in milk. We refer to this as lactose intolerance, but it is really lactose tolerance that is the oddity. More than 85% of north Europeans (but not Icelandic people) are lactose tolerant, but the picture is very different across the rest of the world. In Africa, south and east Asia and among Native Americans (both in North and South America), fewer than 25% of people are lactose tolerant in adulthood. In northern Europe, prehistoric populations had been drinking milk since the arrival of agriculture about 6000 years ago. They developed a genetic mutation about 4000 years ago, which spread rapidly through the population. One suggestion is that during times of disease or famine, those without lactose tolerance would die as a reaction to the milk they were drinking; those with it would benefit from digesting the lactose in the milk.

In the century between Julius Caesar and Claudius, when Britons in the southeast of the island began to adopt Roman ways, they started importing new foodstuffs. They included vegetables such as peas, cucumbers, carrots and celery. Domesticated fruits included figs, apples, pears, cherries, plums, and dates. Newly planted vineyards, like one at Barkway, meant that local wines became available. People used new flavourings such as the herbs coriander, thyme, mint, basil, bay and dill, and the famous but pungent garum (a fish sauce like southeast Asian nam pla). Exotic imported items such as ginger, laser (galangal), pepper and olives added new flavours. People also began to develop a taste for seafood, and oyster shells are a common find on Roman sites, even in places a long way from the sea. Hunting game was a new pastime for the wealthy, so we begin to see venison appearing on the menu, at least for the better off.

Iron Age cheeses may have been almost as varied as modern types. Analysis of coprolites (human stools) shows that there was a type of blue cheese, while other evidence points to making cream cheese and hard cheeses. The pot seen here, dating from the last quarter of the first century AD, was first thought to be a strainer (and, indeed, was once labelled as such in a display at Letchworth Museum). However, it is a cheese press, used for draining the whey from the curds and compressing them to make a hard cheese.

Roman Cheese Mould

The simplest way to make cheese is to heat milk and then add something to make it curdle. This could be sour cream, beer or vinegar. Next, the curds and whey need to be separated, by allowing the whey to drain away (and be stored for other uses in cooking). This can be done in a cloth bag, which makes a soft cheese. To make a hard cheese, it is necessary to press the curd to force out any remaining whey by putting it into jar, still wrapped in its cloth, and putting weights on top. This is how the pot we see here was used.
So, someone was making a hard cheese, but on what sort of site? We know from the Letchworth Museum Accessions Register that the vessel (described as a ‘Bowl Colander’) was ‘found with shards at 2 Chimneys Sandpit March 25th 1942. 4’6” deep’. The register contains two more relevant entries: on 18 August 1952, some pottery was accessioned ‘from Two Chimneys Gravel Pit’, having been found in that year and in 1939, while one from 19 November 1953 records pottery ‘found by F J Berry’.

Historic maps show a sand pit 175 m south of The Two Chimneys public house, which was still shown on a map surveyed in 1951. This must be the place, which is now covered by Wilbury Hills Cemetery. This is just inside Stotfold parish in Bedfordshire, although the pub became part of Letchworth in the 1960s. The Bedfordshire Historic Environment Record lists the site (number 508) and suggests that it may be a disturbed burial. This is unlikely, as Percival Westell’s note on the original discovery stated that ‘[t]he occupation area is approximately 250 ft’, suggesting that the area in which the pottery was found covered more than 75 m across, too large to be a single disturbed burial. Cheese presses are functional, everyday items that are unlikely to be put into a grave as a gift for the departed, as are the quern fragments that were also found there.

Westell’s original guess that it was an occupation site is likely to be correct. Although Bedfordshire County Council commissioned an evaluation of the site, which included geophysics and trial trenching, no further Late Iron Age or Romano-British finds were made. Presumably, everything on the cemetery site lay within the outline of the sand pit and, if the occupation area was larger, it extended east across Stotfold Road. Unfortunately, no aerial photographs show activity on this side, either.

The site fits into a pattern seen across Letchworth Garden City, where almost every hilltop has produced evidence for a Roman period farmstead or village. Examples include Caslon Way on the Grange Estate, the centre of Norton village, Spring Road, Sollershott West/High Avenue and the northern edge of the Jackmans Estate. As we saw last week, individual ‘sites’ are simply part of a wider landscape, and we are lucky to have a lot of evidence locally for land use in the first century AD. We also know a great deal about Romano-British foodways (a useful American term covering the production, preparation and consumption of food), thanks less to the survival of recipes for things eaten by the wealthy and more to archaeological evidence, including data from animal bones and plant remains, things often overlooked by those who can only see glittery metalwork.

Written by Keith Fitzpatrick-Matthews

Much as the media likes to focus on the idea of ancient sites, individual places that can be revealed through excavation, the reality is less clear-cut. Archaeologists have long been aware that discrete ‘sites’ or ‘monuments’ are part of broader landscapes, elements of which may still exist. Sometimes, it is more helpful to think in these terms, and I have been promoting the idea of the ‘Baldock Bowl’ for more than ten years. The ‘Baldock Bowl’ is distinctive landform, part of a drainage channel that formed during the Anglian glaciation (474,000-424,000 years ago) that was later blocked to the south. When you are inside the Bowl, you have the impression of being a hollow completely surrounded by hills. This is an illusion, as it contains the springs of the River Ivel, which flows north to the River Ouse. The surrounding hills are not of any great height but the gaps through them have channelled communication for millennia.

There is an enormous density of Neolithic ‘sites’ in the Bowl, and there is little purpose in trying to define the limit of each. Instead, we can see a continuum of activity, with areas of more concentrated repetitive actions of different types and other areas that seem not to have been sufficiently well utilised to leave archaeological traces. The activity includes tracks, flint mines, burials, settlement, pits and religious monuments. The number of ‘sites’ makes it impractical to think of them as discrete units, and we have to take a holistic approach to understanding what was happening here between about 4000 and 2000 BC.

Archaeological knowledge of the area east of Letchworth Garden City, on the northwestern edge of the Baldock Bowl, began in 1957. Margaret McFarlane, then Assistant Curator of Letchworth Museum, discovered human remains in the footings for a kerb on the new Blackhorse Road. They proved to be part of a small cemetery dating from about AD 600, much later than the sites forming the Neolithic landscape. When work began in earnest on factory building in the following year, Miss McFarlane’s successor John Moss-Eccardt investigated large areas. By 1973, the final season of his work, he had investigated almost 100 ha of land. In 1988, an area to the north came up for development and is now Kristiansand Way and Talbot Way. Nine years later, from 1997 to 2000, an extension to the east end of Works Road, on the other side of the railway from these sites, revealed yet more archaeological remains. Finally, between 2010 and 2013, the Norton Community Archaeology Group investigated several sites in Hundred Acre Field, beyond the end of Blackhorse Road, uncovering further elements of the prehistoric landscape.

The earliest activity in the landscape lay close to the Ivel Springs, on the northern edge of Baldock and on the boundary of the historic parish of Norton. They formed the focus for the south-eastern end of a track at Nortonbury, although its course and destination to the north-west are unknown. Moss-Eccardt described the Nortonbury track as a cursus monument, consisting of parallel ditches and internal banks; at 7 m wide, it is much narrower than any other example of the type and is probably something else entirely. It does not seem to be a so-called ‘bank barrow’, as there is no evidence that the central area between the ditches ever held a mound. It seems instead to be a route along which people travelled to and from the springs but on a much narrower scale than the classic cursus type.

A short distance to the west, during the investigation of Norton henge in 2013, late Neolithic houses were found beneath its bank. They are horseshoe-shaped, with a windbreak on the east side of the door, the direction of the prevailing wind. Their plan is identical with those discovered by Mike Parker Pearson at Durrington Walls about five years earlier, although those at Norton do not have surviving floors. The evidence for them consists of trenches dug into the ground that once held upright overlapping timber planks. The walls they formed would probably have been covered in daub. The excavators found at least five such buildings, including one complete example. The builders of the henge bank carefully dismantled the structures before piling up the chalk, which fell into the trenches that had held the timbers. At Durrington Walls, these sorts of buildings are dated about 2600 BC, and the Stonehenge visitors’ centre boasts reconstructions of several, complete with internal fittings such as shelves.

At Norton, the houses are probably earlier than at Durrington Walls, as finds from the pre-henge activity include Peterborough ware type pottery and leaf-shaped arrowheads, Early to Middle Neolithic types. This may push the start of the activity back before 3000 BC. When the outer ditch of the henge was built, it cut through an earlier ditch, but only one short stretch could be examined, so we do not know if it was part of an enclosure containing the houses.

To the south, there are more houses at Works Road, although they consist of posthole constructions. It was challenging to produce plans of complete buildings, although at least one rectangular building is visible. The western part of the area investigated contains a mass of postholes, which show that building and rebuilding were happening here over a long time. They were effectively undatable, although associated with the same type of Middle Neolithic pottery and flintwork as at the pre-henge settlement.

At Blackhorse Road, a dense cluster of postholes in the southeastern part of a large D-shaped enclosure probably also represent many buildings. This area is complicated by a statement passed orally to the writer by a prominent academic (who shall remain nameless) that students digging there ‘invented’ postholes deliberately to confuse John Moss-Eccardt. The least said about this, the better!

What this evidence shows is that a large area across the eastern side of Letchworth Garden City saw an immense amount of activity in the centuries around 3000 BC. This early activity all seems to be domestic, apart from the ‘cursus’ at Nortonbury, where its general direction, leading away from the Ivel Springs in the opposite direction to the settlement suggests a different purpose. Given the later importance of the springs in what seem to be ritual activities, we may suspect something similar at this early date. It is also worth remarking that the ‘cursus’ is leading towards the edge of the Baldock Bowl. Did it lead people in from outside?

The henge was built in the centuries after 3000 BC, making it possibly as old as the first phase at Stonehenge (about 2960 BC). Its builders dug a ditch 60 m in diameter, 5 m across and about a metre deep, with steep sides and a flat base. The chalk taken from it was piled up inside, leaving a gap of about 3 m and creating a bank over 2 m high. Both bank and ditch had a gap facing exactly due east, towards the equinox sunrise on 21 March and 21 September. The location of the henge also meant that this line also crossed the Ivel springs. In the centre of the entranceway was a line of three irregularly shaped pits that were deliberately backfilled with clay. Inside the bank, people were setting small fires on which they burnt polished stone axes among other things, and smashed pottery.

Between about 2700 and 2500 BC, the henge saw major changes. The original bank was circular, as was the ditch outside it. This configuration is known as a ‘formative’ henge. By about 2500 BC, all newly built henges were oval, and Norton henge is the only formative type to change its shape to reflect the new fashion. The bank was reshaped, partly by cutting it back inside, opposite the entrance, and a new, shallow ditch dug inside it, while the outer ditch was left to silt up. A layer of chalk rubble was laid down as paving inside the inner ditch and through the entrance; we do not know how far this track led, but it perhaps took users down to the springs. At the centre of the henge, an oval pit contained the combined cremated remains of a new-born baby, a child and at least one adult.

Meanwhile, at Works Road, a smaller henge-like oval monument, conventionally referred to as ‘hengiform’, had a central grave. The ditch was little more than a gully, no more than 0.8 m wide and 0.45 m deep, defining an area about 10.5 m by 8,5 m, the long axis running west to east, with a gap at each end. Anna Rohnberger of the University of Reading kindly examined the skeleton that lay in a contracted position in the grave at its centre. She found that, based on the teeth, it was a child aged between 4 and 5½ years at the time of death. The backfill contained Neolithic flintwork.

Norton Henge was modified again, between about 2300 and 2000 BC. A massive post now stood close to the centre, while a low inner bank, made from material cleaned from the inner ditch, may have held a ring of smaller posts. A new pit cut through the paving in the entrance held the cremated remains of a child, while a square pit near the centre held a small collared urn. This style of pottery originated in the second half of the third millennium, and was often a container for cremated bone. At Norton, the urn was empty. Did this symbolise the ‘death’ of the henge?

At Works Road, a group of pits dates from this period. The largest cut through the southern part of the hengiform monument; it was an elongated lozenge shape, 9.5 m long, with a maximum width of 3.9 m and 2.12 m deep, perhaps extended several times. The lower fills were all chalk rubble that seems to have gone into the pit shortly after it was first dug. Their layering showed that they entered the pit from the northwest, the site of the hengiform. The chalk rubble contained struck flints and a piece of antler. Above this, there were natural silts, containing Neolithic pottery, animal bones, flints and a grinding stone with a semicircular pebble worn at one end. The grinding stone has a groove along its long axis, which the worn end of the pebble fits precisely. Nearby, another but much smaller pit contained Late Neolithic pottery, struck flint, and a second grinding stone with a cubic rubber (cushion stone), both carefully laid on the base of the pit with an antler pick placed on top of them.

What are we to make of these? Cushion stones are uncommon, but found across Europe, where they seem to be tools used in metal working. More specifically, they were used in making gold sheet, which was used to make jewellery, dress items and decorated wood and bone objects. The well-known Amesbury Archer, an exceptionally rich Early Bronze Age burial found near Stonehenge, had one in his grave. At Works Road, the stone and the grinders hint at early working of metal before the close of the third millennium BC. Several more grinding stones came from inside Norton henge. The careful placement of a cushion stone and grinder in a pit at Works Road, beneath an antler pick (perhaps even the very tool used to excavate the hole), suggests that they were an offering to the Earth.

The density of activity and the hints at wealth in the early centuries of metal working make this area in east Letchworth Garden City very important indeed for understanding the Late Neolithic/Chalcolithic transition in the region. We have more than a millennium of activity, and this account has barely scratched the surface of what was found. I haven’t mentioned the Blackhorse Road flint mines, with a puppy sacrifice at the bottom of each, containing pottery up to 1000 years old when they were filled in. I haven’t mentioned the Neolithic house on Clothall Common, surrounded by evidence for flint tool production. I haven’t mentioned the L-shaped ditch at Works Road, which replaced an earlier line of pits. Except that now I have!

Much of the activity in the area shows a concern with things underground. Dozens of enigmatic pits with no obvious practical function, the focus on the Ivel Springs, the careful and unusual burial of infants and children, perhaps hint at a belief in chthonic (underground) deities. At the same time, activities inside the henge were cut off from the landscape, perhaps hinting at a fascination with the sky and, at the very least, the sun. We should never think that Neolithic religion was focused on just one thing, such as ‘mother goddesses’. The evidence from east Letchworth Garden City shows that people were interested in earth, water, sky and, perhaps, even fire, the four elements of pre-modern Europe.

Written by Keith Fitzpatrick-Matthews

If you browse auctioneers’ or metal detectorists’ websites, even Etsy or eBay, you may come across things referred to as ‘Celtic woad grinders’. The example seen here comes from Baldock. Leaving aside the fact that the term ‘Celtic’ is today used mainly as a linguistic rather than art historical label and that archaeologists today call them simply ‘cosmetic grinders’, what evidence is there for how their owners used them?

The Baldock cosmetic grinder was found with a pair of tweezers in a ditch excavated on Clothall Common that was infilled some time between AD 180 and AD 200. The ditch formed part of the western boundary of a large trapezoidal enclosure with a roundhouse in its southern part. Although interpreted at the time as a house, the range of later finds from the enclosure – a pit with 33 iron spearheads, votive axeheads and spears, part of a large bronze statue and an iron rattle – suggests that this was religious rather than domestic.

The grinder consists of a crescent-shaped copper alloy trough with a horned animal head (cow or bull) at each end and a central ring underneath. One of the animal heads has a blunt-ended muzzle, the other is more pointed and open-mouthed. Each side of the trough is decorated with three shield-shaped panels containing enamel inlay, red in the centre and light turquoise either side. The top of the grinder has a deep V-sectioned trough with a slightly squared off base, which seems to derive from wear. Its capacity when full would have been about 1.8 ml.

When Reginald Smith first described three examples from Wroxeter in 1918, he was misled by the position of the loop into thinking that they were pendant charms associated with horses as the loops were evidently made to hang from a cord. He compared them with a serrated nose-band or cavesson, used for controlling unruly horses. He distinguished a type with a central suspension loop and one with an end loop; by the 1930s, he had also recognised a ‘solid’ type. The identification as pendants or amulets remained standard until the 1980s, although Ian Stead and Val Rigby had recognised in the 1970s that the troughs and solid pieces belonged together. About twenty complete sets have been found.

During the 1980s, Ralph Jackson recognised that the two elements formed either a rocking pair or a sliding element along the trough part. He concluded that they must have been grinding implements, a miniature mortar and pestle. Their association with other items and personal care, such as tweezers, and their small size made it likely that they were used to grind mineral-based pigments. It was from this work that the wrong idea of the ‘woad applicator’ or ‘woad grinder’ arose: the dye extracted from woad by steeping and boiling the leaves does not need to be dried and can be used as a liquid. The dried powder does not need grinding like minerals.

It is not even certain that Julius Caesar’s description of the Britons (de Bello Gallico V.14) mentions woad. His text reads omnes uero se britanni uitro inficiunt, quod caeruleum efficit colorem (‘Indeed, all the Britons dye themselves with uitrum, which produces a blue colour’). The Latin word uitrum usually means ‘glass’. The first person to translate it as ‘woad’ was Arthur Golding in 1565 and experiments with the plant have shown it not to be effective as a skin decoration. The usual Latin term for woad was isatis (from Greek ἰσάτις). Caesar’s description of ‘blue’ Britons may even derive from the traditional associations of blue in the Classical world with barbarians, ghosts and death and not reflect reality.

A second quote, from Propertius’s Elegies Book II, that has been taken to refer to woad, does not mention uitrum or isatis but just describes the Britons as infectos, ‘stained’, without mentioning a specific dye or colour. Moreover, Pliny (Historia Naturalis XXII.2) describes how the Britons use a plant known as glastum in Gaul, which by which they are aethiopum colorem imitantes (‘imitating the colour of Africans’), so dark rather than blue; the Celtic word glastos means ‘blue’ and survives in Welsh as glas, which refers to a range of colours from sky blue through greenish blue to grass green. Ovid (Amores II.16) talks of uiridesque Britannos (‘and the green Britons’), while Julius Solinus (de Mirabilibus Mundi XXIII) mentions tattooing: inde a pueris uariae animalium effigies incorporantur (‘from childhood, they embed in their bodies various images of animals’). Claudian (de Consulato Stilichonis II.248) gives the personification of Britannia picta genas (‘painted cheeks’, often but unnecessarily translated as ‘tattooed cheeks’).

These texts, written a long way from Britain, are evidence that the Roman aristocracy perceived the Britons as unusually decorated. Solinus’s animalium effigies must be tattoos, as he describes how they grow as the person with them grows; none of the other descriptions are clear about whether the staining is permanent or temporary. However, the cosmetic grinders are evidence that many Britons did stain their skins with coloured materials.

Other cosmetic grinders come from a range of sites dating from the Late Iron Age through to the end of the Roman period, all but four of the more than 1000 known examples having been found in Britain. Most date from the first and second centuries AD, although they sometimes turn up in early medieval graves among collections of ancient Roman objects. They are most common in East Anglia, the Ciuitas Icenorum in the Roman period, although they turn up throughout Britain apart from North Wales, Cornwall and Scotland. They also occur most often on urban sites, including so-called ‘small towns’ like Baldock, particularly in burials and religious contexts.

In a major study published in 2010, Ralph Jackson divided the mortar elements into thirteen types: ten with a central loop, two with an end loop and two examples of three joined into a triangle with no loop. He published evidence that they were case in bronze from lead models (there is an example from Skipton Street, London), with finishing ranging from almost non-existed to smooth polishing, punched decoration and elaboration of the terminals. The wide variation of styles is partly a result of different treatment at the finishing stages, and it is possible that decoration was added following instructions from the buyer.

He also showed that many examples were heavily worn, showing intensive and prolonged use. Some examples have repairs, meaning that they were probably not cheap to replace. The groove or trough of the mortar element often shows wear in the base, as with this example from Baldock. Crescent-shaped pestles with end-loops are always damaged on the convex underside, while those with centre loops show the greatest wear at the centre of the convex side. Rod-shaped pestles have wear at the tips and some are even shortened.

Some examples seem to have entered the archaeological record as votive gifts at temples. At Wicklewood in Norfolk, this was certainly the case, and it is likely that the Baldock example entered the ditch from a clearing-out of temple gifts. Nevertheless, their wider distribution shows that they were relatively common items associated with grooming and body care.

On 18 September 2019, the BBC ran a news story about how a curator at English Heritage had recently recognised the three examples Reginald Smith had described in 1918 as pestles and mortars. Their reporter had evidently not seen any of Ralph Jackson’s publications, beginning in 1985! More worryingly, the report said that they ‘allowed women to achieve their look using charcoal, soot and chalk without importing expensive products like kohl’.

We do not know that these cosmetic grinders were used exclusively by women: one of the sets found in a grave was with a definitely male skeleton. The idea that only woman in Roman Britain would wear makeup is a good example of projecting not only current social norms back into the ancient past but also of assuming that what was standard in Roman Italy – where much of the surviving literature was written – was also standard in the far north of the empire. We know from this literature that some men did wear makeup, but men who were vain about their personal appearance (such as the Emperor Otho) were often mocked. Some writers even complained about women who wore makeup, suggesting that their morals were loose.

We have no way of knowing if men in Late Iron Age and Roman Britain did wear cosmetics and, if they did, what sorts they used. Indeed, Gillian Carr suggested as long ago as 2005 that these peculiarly British cosmetic grinders are evidence that attitudes to makeup were very different here from the rest of the Roman world (perhaps except for Egypt). She even speculated that the pigments were being used to make dyes for tattooing rather than (or along with) temporary makeup. As we have already seen, the literary evidence shows that the Roman upper classes viewed the Britons as having eccentric, even terrifying attitudes to colouring their skin.

This is yet another object that highlights how our own sexist attitudes can unfortunately pervade our interpretations of the past. No, these cosmetic grinders are not ‘hugely important to understanding women’ in Roman Britain, despite the statement by English Heritage’s Cameron Moffett. Instead, they underline the remoteness of Roman Britain and how little we understand of how different it was from today’s United Kingdom.

Written by Keith Fitzpatrick-Matthews

1 2 3 69