Archaeology Tuesdays – an enclosed farmstead at Dell Field, Codicote

In 1967, an aerial photograph taken by archaeologists based at the University of Cambridge revealed a previously unknown set of anomalies in Dell Field, south of Codicote. It shows an almost square set of ditches, extending away from the northeast corner to the east and from the southwest corner to the south. Another ditch runs from the southern end of the eastern side, apparently parallel with the northern ditch. For many years, the date and nature of the enclosure remained unknown, although its shape suggested that it might belong to the Late Iron Age or Roman periods (about 100 BC to AD 420). In 1989, North Hertfordshire District Council received an application to develop a golf course on the farmland between Hollard’s Farm on the B656 Codicote Road and the St Albans Road. Part of the permission required an archaeological evaluation of the site by trial trenching, which took place in March 1990.

The enclosure sits on the edge of a plateau with good views across the River Mimram valley to its west and south. The eight trial trenches examined the ditches on all four sides of the enclosure, as well as the ditches extending away from it; one targeted an area outside, to see if there was any evidence for activity in that area, while another extended into its centre. As well as digging trenches, the team of archaeologists from North Hertfordshire Museum carried out fieldwalking on the rest of the site, recording visible finds picked up from the surface, and recorded visible earthworks.

The underlying geology of the site was complex, with the weathered chalk bedrock disrupted by extensive areas of acidic silty clays. The clays appeared in bands running roughly west to east across part of the site, but as mottles elsewhere, making it challenging to identify archaeological features. These clays are evidence for the Pleistocene history of this area. During the Anglian Glaciation, 478,000 to 424,000 years ago, an ice sheet pushed westwards across this area from the direction of Ware. It tore up the underlying chalk, creating grooves and hollows in it. The clays known as the Ware Till filled these holes and remained after the ice sheet melted.

The ditches extending away from the enclosure look like part of an earlier system of land division. They were U-shaped and up to a metre of so deep. The sections across these ditches produced Late Iron Age grog-tempered pottery from their lowest fills. Grog is crushed pottery, used in making new pots as a way of helping water escape from the clay during firing. If trapped, microscopic droplets of water can expand as they turn into steam, making the pot explode and destroying others in the same firing. Grog-tempered wares are typical of the first century BC in Hertfordshire, continuing well into the first century AD. All of the types found were of first-century date. One of the sections also contained a sherd of flint-tempered pottery, more typical of the first half of the first millennium BC. As it came from one of the upper fills of the ditch, it is of no help in dating its construction. The ditches had been cleaned out and recut at some point during their use.

The enclosure ditches presented a very similar picture in some ways. They, too, had at least one recut. One of these recuts contained a very fine copper alloy brooch, as pictured here. More of this later. Unlike the ditches described previously, they were V-shaped with steep sides and at least 1.5 m deep. The lower fills had a mixture of grog-tempered wares and post-conquest types, including pottery from the Verolamium region kilns at Bricket Wood and part of an amphora. Some of this material dates from after AD 70 and it came from a layer that suggested nearby demolition. One of the trenches uncovered a large subrectangular feature, perhaps a pit or quarry, but there was no time to investigate it. The confusing character of the mixed geology made it impossible to identify postholes, so we do not know if any of the trenches revealed traces of buildings. A fragment of Roman roof tile from one of the sections, though, showed that there must have been a structure nearby. It is more likely that this lay inside the enclosure rather than outside it.

Away from the excavation trenches, almost 95% of the 437 finds picked up during fieldwalking were medieval or later in date. Most of the Roman finds, including roof tile, came from the field close to the enclosure. This probably means that the most intensive activity at this time lay in and around it. A couple of waste flakes from making flint tools came from the field on the eastern edge of the area surveyed. They perhaps date from the Early Bronze Age (about 2500-1400 BC) and, like the later material, tell us nothing about the enclosure.

The site does not exist in isolation. During the growth of Welwyn Garden City, several similar sites were uncovered and examined rapidly. Traces of buildings and hearths at these sites show that they were domestic enclosures, with substantial ditches (and, presumably, now-levelled banks). At Codicote, there are hints that the banks were deliberately pushed back into the ditches; there is similar evidence from two of the Welwyn sites, including one where fills pushed into the ditches had crushed complete pots in the bottom.

What were these sites and when were they in use? Although the V-shaped ditches look defensive, the pottery found in them shows that they were domestic. There are the remains of pre-conquest grain storage jars and post-conquest flagons, and there are bones from the main meat animals (sheep/goats, cattle, pigs and chickens), some of which have butchery marks. Oyster shells show that people were also eating seafood, while fragments of hare bone from the initial silting of the enclosure ditch may have come from a wild animal not used as a food source.

The pottery from all of them falls into a very specific period: the first three quarters of the first century AD. Welwyn archaeologist Tony Rook has suggested that they all fell out of use at the same time, perhaps following the rebellion of Boudica against Rome in AD 61. This is too early for some of the pottery from the Codicote site, which probably did not arrive in this part of Hertfordshire before AD 70. Indeed, the brooch already mentioned provides important clues to dating. An almost identical brooch from excavations at Baldock is of a type generally datable to the 60s AD. It is in near-mint condition, with no real sign of wear, showing that it cannot have been lost long after AD 70. At the same time, it comes from a recut through fills containing pottery no earlier than AD 70. As a result, we can suggest that demolition of this site – and probably the others, too – happened a decade or so after the Boudican revolt.

The scale of the ditches (and presumably the banks made from chalk quarried from them) shows that these enclosures belonged to people anxious to display their status. One, at Grubs Barn in Welwyn Garden City, lay very close to the high-status Welwyn Type burial on the Panshanger Estate excavated in 1965 and now displayed in the British Museum. Might this have been the tomb of one of the wealthy owners of this type of site? At least nine other enclosures are known within 5 km or so of that at Dell Field and they must have been a conspicuous part of the landscape. All that can be dated were abandoned at around the same time, which can hardly be a coincidence. As this was not a time of economic or agricultural collapse, we need to find some other cause. The sites were abandoned around the time that a village-like settlement was growing at Welwyn, southwest of where the Roman road crossed the River Mimram. Perhaps the most likely explanation is that land ownership became concentrated in the hands of one wealthy family, who encouraged independently-minded farmers to move from their showy enclosed farmsteads into a fashionably nucleated settlement. That way, the local authorities could make sure that a Boudica type rebellion would not happen again.

Written by Keith Fitzpatrick-Matthews