Ellen Munson
In October 2014, Phil Kirk was metal detecting on a field at Kelshall in the hills south-west of Royston, when he encountered a strong signal. Digging down about 15 inches, he found the top of something bronze. Thinking initially that it might be nothing more than a modern filter from a car, it turned out to be a complete Roman jug, missing its handle. On lifting it, he spotted the handle and bowl of a bronze patera (or trulleum). Next to this lay the battered bottom half of a large jug and last of all a third jug with a trefoil mouth in four or five pieces, apparently crushed by a large flint. All four vessels came out from a hole no more than 50 cm across. Quite by chance, the broken base of the third vessel matched a bronze object found a few months earlier and about ten metres away, which had been discarded under a nearby hedge as probably a twentieth century oil funnel, but which was the top half of the jug.
Phil recognised the importance of his find and contacted Julian Watters, who at the time was Finds Liaison Officer for Hertfordshire and Bedfordshire. He then contacted the museum to see if it would be worth investigating the site further. So, on a cold morning later that month, a group went up to the site and expanded the original hole to a one metre square. Early in the digging, the missing jug hand turned up, but no sign of the edge of the pit into which they had been put could be found, so we needed a larger excavation trench.
We returned to the site in November, when the farmer scraped the topsoil from an area three metres square. During the initial cleaning, the rim of a glass bottle became visible, then more shards of glass. Next, an iron lamp-holder and suspension bar turned up. Two layers of hobnails were the remains of shoes placed one on top of the other. Next, a bronze corner binding from a wooden box or tray turned up, followed by the other three.
On top of the decayed tray stood a shattered but otherwise complete shallow dish, about the size of a saucer, 14.5 cm in diameter. When first exposed, it seemed to be iridescent from decay, but as the excavator exposed more, it became obvious that it was multicoloured millefiori. The pattern was made from lozenges of fused dark purple, white, yellow, blue and red glass rods. Then a second dish with the same basic pattern turned up next to it. Both were covered with a thin fibrous carbonised deposit, perhaps the remains of a delicate cloth wrapping. Along with these were two shattered glass cups and a pair of blue glass handles, fragments of a lava object and a silver denarius of the emperor Trajan (reigned AD 98-117).
Next to the box stood a collection of glass bottles. The largest was hexagonal, 23 cm across, and contained the cremated bone of a probably middle-aged (40+) man and three worn second-century coins. A second bottle was octagonal, with two long sides, a type rarely found in Britain, and two square bottles. One has the letters IΛƧ (IAS) on its base, a type exactly paralleled at the Roman fort at Cramond, near Edinburgh, in a ditch dated AD 208-212. There was also a rectangular bottle. All these bottles were typical Romano-British products and represent the complete range available at the start of the third century. All were shattered because the pit – a grave – lay beneath a mound of large flint nodules that had collapsed onto them, but which had also protected them from the effects of ploughing.
The grave probably dates from a few years either side of AD 210, and all the finds apart from the millefiori dishes are typical of this period. The dishes are more unusual. Tests by the British Museum show that the glass was made in Alexandria, and the colours are right for a date around 200. Archaeologists have found sherds from similar vessels at Carlisle, Wroxeter, Inveresk, Frocester Court, Eccles (Kent) and London. The importance of the Kelshall examples is that they are the only complete vessels made from this mosaic glass so far discovered.
Each dish is about 14.5 cm in diameter, with straight sides rising at 30°-40° from a small flat base. The rim edges are plain and a little uneven, suggesting that the excess glass was sheared off while the vessels were still hot. They were not blown, but perhaps pressed into a mould. The pressing distorted the lozenge design, and one of the dishes appears to have been made by combining two or more sheets of the patterned glass. Millefiori consists of rods of different coloured glass arranged in a pattern, which are then heated to fuse them together. While still hot, the fused rods – about the size of a baked bean tin – are stretched into a long rod. This can then be cut into shorter lengths, which are then fused in the same way to create a pattern. With the final pattern created, the cylinder of glass can be cut into slices, ready to be reheated and pressed into moulds to make vessels. Some brooches also have millefiori decoration, often much finer than that used to make vessels.
What are these remarkable dishes doing in a grave on an exposed hilltop in Kelshall? Who had owned them? Where did that person live? One thing we can easily guess is that they were wealthy. Most Roman graves of this date might contain some pottery vessels and perhaps one or two glass vessels. Wealthier families could afford to put more things into the ground and some graves could be very ostentatious. This is the case here. Not only could they put in nine glass vessels (five bottles, two cups and two rare dishes), but also four bronze vessels (three jugs and a patera, a libation-pouring dish) and four coins.
Perhaps most significantly, the family did not put any ceramics into the grave. There was a distinct hierarchy to vessel use in the Roman world. Only the imperial family could use gold; other high-ranking families could have silver. Rich people would use bronze, which is what we see here. Aspiring families would use pewter. The silly idea promoted by so many popular works and television series that samian ware is ‘high status’ is nonsense: people who used pottery were rather low down in the pecking order. Yes, samian was more expensive than other earthenwares, but it was definitely not something that the rich would use, much less put in the graves of their loved ones.
The Kelshall family made a definite statement about its position in society during the funeral. They could afford the best quality bronze and glassware to put in the grave; perhaps not putting ceramics in was part of their ostentatious show of wealth. Some of the items were imported, and we can speculate that the cloth covering of the millefiori dishes was something expensive (fine linen or silk, perhaps). Large parts of the grave appear ‘empty’, but probably held items that have since decayed (clothing, wooden items and so on). After the contents of the burial were in the ground and the pit filled in, it was covered with a low cairn of carefully interlocking flint nodules.
Where might the family have lived? The closest Roman settlement of any size is the town at Baldock, only 8 km to the west and visible from the hilltop. There was also a settlement between Kelshall and Baldock, at Slip End, and villas at Steeple Morden, Litlington and elsewhere. The nearness of these places does not mean that the man buried here lived in any of them, though. The grave is in an area where aerial photographs, geophysics and random discoveries show that there was a lot going on in the past and specifically the Roman period.
Kris Lockyear of University College London used a cart-based magnetometer to survey part of the field where the burial was found. It showed that the burial lies inside a large, almost square ditched area about 90 m across, with ditches up to 4 m wide on three sides and open to the east, with hints of a fence on this side. There are other, smaller and irregular enclosures to both west and east of it, as well as buildings composed of postholes and rammed clay floors, pits and at least one well. There are traces of ditched roadways around the site. The scale of the enclosure does not look domestic, nor is it military or a cemetery type. Instead, the most likely explanation is that it is religious.
Unlike medieval churchyards, Roman temple sites are not often associated with human burials. At a temple in Wynn Close, Baldock, the precinct held the graves of several new-born babies, but this is unusual. Late Roman Christian graves often clustered around what were believed to be the graves of martyrs, which could then develop a religious building such as a chapel or church, as happened at St Albans.
The contents of the Kelshall grave include a bronze patera, used in religious rituals. Might the man buried in the grave have been an important patron of the temple or even its chief priest? In the Roman world, this was not necessarily a professional post, and local dignitaries would often serve as the priests in their communities. Perhaps the man buried with the costly goods was just such a priest, perhaps even a major patron of the temple who paid for its refurbishment or rebuilding.
The range of glassware is as remarkable as the dish pictured here. It includes every type that was available at the time the grave was dug. Some of the vessels were new, but the hexagonal bottle containing the owner’s ashes shows extensive wear on the base, where it has been taken on and off a shelf over some years. Does the range hint that the owner had made his wealth as a merchant trading in glassware? This is going beyond what we can reasonably infer from the contents of the grave. Even so, we ought to be able to speculate about such things, always acknowledging that they are merely guesswork.
Written by Keith Fitzpatrick-Matthews
they arrived at work, they would take it from the board where it was kept and put it into a box. Most were pierced to hang on hooks, but some, like this one, were not. The timekeepers would lock the box or take it away to the office as soon as a shift started, so that latecomers would have to report directly to them. In this way, the wages department would know exactly how much to deduct from wages to reflect how late a worker may have been. In some factories, even being a minute late would result in losing half a day’s pay. Tokens left on the board obviously belonged to absentees; managers would sack those who were absent regularly.About 700 m east of the village of Sandon lies a tree-covered mound, off Park Lane and north of Notley Green. According to the Royal Commission on Historical Monuments (England) in 1911, it is a ‘moated tumulus’. This is not a recognised term in archaeology and shows that the surveyor had difficulty recognising what it was. Indeed, there has long been speculation about its origins and purpose. It has been described as a prehistoric burial mound, an early medieval moot hill, a Norman motte, a medieval windmill mound or a combination of these.
It is almost circular, measuring about 26.2 m northwest to southeast, and 26.8 m northeast to southwest. The flat top of the mound measures 17.7 by 19.5 m and is raised a little over a metre above the surrounding ground level. A shallow ditch 4.3 m across surrounds the mound, with a gap of 4.9 m towards the northeast. Water collects in part of the ditch during wet weather. It was covered in trees by the time of the Tithe Award in May 1840; although trees are not shown here on Bryant’s map of 1822, he may have ignored small clumps in his survey.
The Mount lies in a field called Woodley Yards by 1910, although in 1840 it was called Knotley Field. A field east of it was known as
Knotley Mill Field, to the south of which was Mill Field. Knotley is a variant of Notley, the name of the green south of the site, itself recorded as Knott Green in 1676. Mill Field is over a mile north of Mill End in the parish, so the names can hardly be connected.
The Hertfordshire folklorist William Blyth Gerish (1864-1921) recorded a gruesome story about The Mount. A house belonging to a wealthy man stood on the top of it and a local boy overheard some men plotting to burgle it. They spotted the boy, captured him, and threated to flay him (remove his skin) if he told anyone what he had overheard. The brave lad did tell the owners what he had heard, so they could defend their property. The attempted burglary was foiled, but the thieves caught the boy and carried out their threat. The boy survived the ordeal, saying that the most painful part of being flayed was having the skin removed from his fingertips and his toes. Letchworth Museum’s archives (now in North Hertfordshire Museum) have a typescript bearing the name of R J Kingsley of Nelson in New Zealand dated May 1904, which may be the source of Gerish’s published tale.
The East Hertfordshire Archaeological Society organised a trip to The Mount on 24 July 1929. The visit reignited interest in it. The headmaster of Sandon School, James W Sherlock (1898-?), suggested that it was a ‘Late Bronze, or Early Iron Age’ burial mound in 1932. He reported that the owner of the site, Joan Bowman of Sandon Bury, had given permission to excavate in the near future. As it is a Scheduled Ancient Monument, Percival Westell of Letchworth Museum got permission from the Ministry of Works to undertake investigations.
He organised a cross-shaped trench across the centre of the mound between April and September 1933. Local volunteers, including many residents of the village who had attended an adult education class run by Mr Sherlock, did the bulk of the work. Westell was rarely on site, and Mr Sherlock kept him informed of progress by letter. Around the time of the excavation, a schoolboy discovered a Romano-British melon bead from the field surface near The Mount, raising the possibility that it was of Roman date.
The excavators began by sinking a cross-shaped trench across the centre of The Mount, labelling each arm as a separate trench (1 to the north, 2 to the south, 3 to the east and 4 to the west). Each branch was three feet (0.9 m) wide and none went deeper than six feet six inches (2 m). Two extensions made later were labelled 5 and 6. At the centre of the mound, the deepest part, the top deposit was a layer of loamy topsoil, one foot (0.3 m) this. Beneath this was a layer of chalky clay two feet six inches (0.75 m) thick above a band of loamy clay-with-flints six inches (0.15 m) thick. The lowest two feet six inches (0.75 m) consisted of boulder clay containing glacial erratics. This was the underlying natural subsoil.
The first finds were made immediately beneath the topsoil, with those at the lower levels found in an area of 16 by 3 feet (4.9 by 0.9 m). The excavators found 757 potsherds, four small copper alloy objects, 200 iron nails, a knife blade, a key and other iron objects, animal bones (cattle and pig, including teeth) and oyster shells. On 27 April 1933, Mr Sherlock wrote to Westell mentioning the discovery of a coin. According to the Royston Crow of 21 April 1933, all the finds made by that time were made at depths of between six inches and five feet (0.15 to 1.5 m) but no deeper.
Six feet (1.8 m) down, the excavators discovered two wooden beams, arranged at right angles. They extended away from the sides of the trenches in which they were found (1 and 3) and numbered these extensions Trenches 5 and 6. The beams were made from oak, 16 feet (4.9 m) long and one foot (0.3 m) square and at each end, there were recesses to take angled uprights. The excavators correctly identified the beams as a cross-tree, the base elements of a wooden post mill, the earliest type of windmill. They lay at the bottom of a pit cut through the layers of the mound, apart from the topsoil, and into the underlying subsoil; it was the fill of this pit that contained all of the finds. The cross trees were no longer in their original positions, one lying on top of the other in such a way that they were disconnected at the mortices that originally held them together.
Although Percival Westell reported that all the finds went on loan to Letchworth Museum (accessioned as 1935.6922), there are only 13 potsherds in the collection. Perhaps the loaned material was returned (to Mr Sherlock or Mrs Bowman?) or it suffered from a ‘rationalisation’ of the archaeological collections in the early 1970s. However, the thinning out of archaeological material involved mainly disposing of unfeatured body sherds, yet only one of the ten handles, none of the twenty-five base sherds and none of the thirty-six glazed or decorated sherds remains in the collection.
However, Gerald Dunning of the British Museum examined the entire collection of pottery in the 1930s, after visiting the site. He identified about 75% of the material as being of thirteenth-century date, with some extending into the fourteenth and fifteenth centuries and a few residual Romano-British sherds. He reported that the most common form was cooking pots, followed by bowls and jugs. In a letter to Westell dated 30 May 1933, he stated that ‘I see no reason for dating the mound earlier than 13th or even 14th century’.
Controversy arose from the identification of three potsherds from low down in the trenches. Fortunately, they survive in the museum’s collection. Gerald Dunning concluded that they were from a single vessel, which he dated to the seventh to ninth centuries. Westell seems not to have been satisfied with Dunning’s dating and he approached O G S Crawford, M O’Reilly, L A Curtis Edwards and Sir Cyril Fox for further opinions. Their estimates of the date of the pottery ranged from Pre-Roman Iron Age to High Medieval. Puzzled, but realising that the date of these sherds might throw some light on the date of the mound, Westell left a note with the sherds: ‘Please keep these 3 sherds separate & kindly date if possible. ?Iron Age or Romano-British? Found 6 ft down in Sandon Mount resting on undisturbed, original ground level. ?Urn containing cremated interment. ?Do these a/c for origination of the Mount? ?Tumulus.’.
Dunning’s published description of the sherds was wrong in several ways. Although he was correct to say that the outer surface of the sherds is black, the ware itself is a reddish-brown and the surface has been blackened by smoke action, suggesting that they are from a cooking vessel. His dating is also wrong. These sherds are identical to the Museum Service’s reference sherds of St Neots-type Ware, a fairly rough fabric tempered with fossiliferous shell and a slightly soapy feel to the surfaces. This ware was produced in the south and east Midlands, centring on the towns of Cambridge, St Neots, Bedford, Northampton and Oxford. Several kiln sites exist, including St Neots and Olney Hide. The date is firmly Saxo-Norman (AD 850-1200).
Westell also changed the stratigraphic position of the sherds between the original draft of the text and the publication. In the published version, Dunning states unequivocally that they came from ‘the old ground level below the mound’, presumably because this is what Westell had told him. However, the typescript contains the crossed-out phrase ‘found close to, if not upon, the original ground level’, suggesting that while they were indeed found at a considerable depth, they were nevertheless contained in the mound material.
The excavators found a coin towards the northern end of Trench 1, at a depth of three feet (0.9 m). Mr Sherlock sent it by post to Letchworth Museum on 27 April 1933. He described how it had been ‘encased in clay’ and ‘covered in a greenish deposit’. However, we do not know what the coin was as it is not mentioned in Westell’s report. There is a medieval coin from Sandon, donated by a Miss Field, which is an issue of Edward IV, dated 1464×70 and minted in London. Although it is possible that this is the one from The Mount, Miss Field’s two other donations of coins (a Roman coin of Antoninus Pius and a late medieval jeton) both came from Roe Green.
The excavation at the Mount shows that it was undoubtedly the site of a sunken post mill, the pottery suggesting that it was raised in the thirteenth century. Although Westell was determined to demonstrate that the mound pre-dated the mill, there is little reason to believe this to be the case. The pit dug through the mound down to the level of the cross-trees could have been dug after the mound to insert them or, perhaps more likely, to remove timbers at the time of its demolition. If the latter, the ceramic finds suggest that it happened in the fifteenth century at the earliest.
Westell angered Ruth Pym of The Settlement, who read his account of the excavation in The Times. In a letter of 12 November 1933, she said ‘I have seen the Times article & find it almost unconceivable that you should have let the Settlement down so completely – the “elder scholars” were in the Settlement Class, which gave you the opportunity of making the whole excavation – that is where you should have paid your debt to us & given us our due place – no wonder you did not show me the draft report.’ He was evidently taken aback by this and the published version of the report acknowledges ‘Miss Ruth Pym and members of Sandon Adult Educational Settlement Class’.
Westell did his best to create an air of mystery about The Mount by creating a ’manufactroversy’ around the three sherds found low down in the excavation. Having received an answer from Gerald Dunning that he did not like – that they were medieval – he sought the opinions of other experts and we no doubt he was delighted that they did not agree. It is fortunate that they remain in the museum collection, it is clear that they are contemporary with the other pottery found, albeit at the early end of the range.
Westell also tried to establish the mill on the mound as one of the earliest in England, if not the earliest. The earliest documented mill in England is from 1185 at the ‘lost’ village of Weedley in Yorkshire, but they were known in the Arab world before AD 800. The earliest types in this country are the post mill, as at Sandon, in which the entire body of the mill pivoted around a central post to allow the miller to catch the best wind direction. There is a documentary record of a mill in Sandon as early as 1222, on land belonging to the manor of Gannock. Whether this is the windmill at The Mount (which is close to Gannock Green) or the lost mill that gave its name to Mill End is uncertain. The mill at The Mount was probably known as Knotley Mill, to judge from the nearby field names.
Written by Keith Fitzpatrick-Matthews