An Early Iron Age pit in Area 13 (Site ELV091) contained a large and distinctive assemblage of pottery deriving from at least three vessels, fresh flint-knapping waste which may derive from the manufacture of flint quernstones (for grinding grain or other seeds into flour), and animal bone that might perhaps best be described as ‘feasting waste’. As this feature was revealed in a narrow strip for a new emergency access track, it is unclear exactly how it fits into the pattern of contemporary settlement and other land use in the area. However, other Early Iron Age pits and settlement-related features have been found during fieldwork carried out in the adjacent Center Parcs complex, and the pit in Area 13 is probably an outlying feature associated with this.

By the Middle Iron Age (c. 400 BC) a farmstead had been established on the Elveden Bypass section of the road route (Areas 10 and 11; Sites ELV085 and 086), represented by grain storage and rubbish pits, small livestock enclosures for sheep and/ or cattle, and some surviving fragments of boundary ditches which may have demarcated parts of an early field system. One group of pits formed a distinct ‘arc’, suggesting that they surrounded a lost roundhouse of which no other trace survived. There seems to have been a period of economic expansion and population growth here during the Middle to Late Iron Age (c. 300 BC – AD 0), as indicated by the continuity of the settlement described above, the establishment of an entirely new, separate farmstead several hundred metres to the south, and, in all probability, the existence of several other settlements/ farms outside the limits of the road strip, which are hinted at by the distribution of features and finds. Each of the two identified farmsteads comprised a core domestic area with one or more roundhouses, surrounded by an infield containing clusters of food/ grain storage pits and stock enclosures, giving way to less intensively managed outfield agricultural land beyond this. In one of the later Iron Age settlements (Site 1; south end of Area 10, ELV085) the outfield appears to have been open/ unenclosed and was probably primarily used for grazing livestock (most likely sheep given the limited availability of water and nutrient-poor soils). In contrast, the settlement to the north (Site 2; middle and north end of Area 10 and Area 11, ELV085 and ELV086) had agricultural land subdivided by systems of field boundary ditches, at least some of which went through several stages of development. At both sites the more outlying areas seem to have been the preferred location for a range of non-mundane or what might be called ‘ritual’ activities, including burying structured/ placed deposits in pits (for example, a large disused storage pit containing several articulated animal burials and a human skull) and the dumping of large quantities of burnt flint arising from an as yet unknown process.

In Site 2 there was a limited but tightly-defined distribution of Romanising/ Belgic-type pottery (characteristic of the last few decades BC and early 1st century AD), centred on an early-1st-century AD roundhouse. Significantly, the distribution of diagnostically 1st-century Roman pottery is concentrated in precisely the same area, indicating continuity of occupation in the decades either side of the Roman Conquest. The absence of any ‘latest’ Iron Age pottery in Site 1 and the lack of subsequent Roman occupation there suggests that this farmstead was abandoned some time before the Roman Conquest and Roman annexation of Iceni territory.

The impact of the Roman Conquest and Boudican Rebellion in the Breckland is a key question. The evidence from the excavation shows elements of continuity: as discussed above, one of the settlement foci (the core of Late Iron Age Site 2) has evidence for occupation throughout the 1st century AD, including the rebuilding of a roundhouse in the late 1st to early 2nd century AD (the retention of roundhouse architecture also pointing towards some cultural continuity). Struck flint had certainly still been used for tools during the Late Iron Age and there is limited – but convincing – evidence that some flint-working continued at the site into the Romano-British period, albeit at a low level. Although in the absence of clear evidence for linings, the identification of particular holes in the ground as ‘storage pits’ is somewhat moot, it is certainly the case that similar large and deep pits with steep to vertical-sided profiles and flat bases occur widely both in the later Iron Age phases of the site and well into the Romano-British period, suggesting that food/ surplus grain continued to be stored in ‘traditional’ fashion long after the Roman Conquest. There is no evidence for raised granaries or barns for storing crops until the later Roman period (c. mid-3rd century AD onwards).

However, there are also changes between the Iron Age and Roman phases: although there are some elements of continuity in the way that the landscape was organised, there does seem to have been a comprehensive effort in the late 1st to early 2nd century AD to divide up a large tract of the landscape with a new, regularly aligned ‘grid’ of boundary ditches forming a series of adjoining rectilinear fields and enclosures with interconnecting trackways, all following a common orientation. The early Roman period also saw the appearance of distinctive new forms of building, specifically two ‘sunken-floored’ structures, each of which comprised a large shallow hollow with a screen or partial roof on one side, which seem to have been used for a range of craft/ industrial/ agricultural processing activities. These might perhaps be taken as evidence of increasing economic specialisation and a shift away from the essentially subsistence-level agriculture of the Iron Age towards more market-orientated surplus production. Overall, the evidence suggests that at least some of the ‘native’ Iron Age inhabitants were still living at the site in the early Roman period but what is not clear is whether they continued to farm the land for their own benefit or whether they now labored for a Roman — or at least outside — landlord.

A crucial consideration in weighing up the different strands of evidence for continuity and change, and their implications for the fate of the native British population, is the relatively coarse-grained nature of the available ceramic-based dating evidence. The elements assigned to the early Roman period (late 1st to mid-2nd century AD) could be the result of gradual developments that took place over several generations as the ‘native’ inhabitants of the landscape adapted themselves and their farmstead to the new challenges and opportunities posed by Roman rule (for example, the road network and access to new markets in the form of the ‘small town’ at Icklingham, in the Lark valley, and the presence of non-productive consumers, such as the army stationed at nearby Pakenham), rather than being rapid changes that were imposed wholesale on the population by new Roman landlords in the immediate aftermath of the Conquest/ Boudican Revolt.

In the absence of direct evidence, perhaps the only way to resolve this conundrum is to extrapolate back from how the estate developed in the mid- and later Roman period. Certainly by the late Roman ‘phase’ (c. mid-3rd century onwards) and probably starting some time earlier, carbonised plant remains, as well as the presence of six large and well-built crop-drying ovens, a large barn and a raised granary, shows that the farm had developed into a ‘processing hub’ geared towards the drying, storage and, presumably, onward transport of wheat and barley grain. Crops were not grown on any scale at the site, rather they were brought in a semi-cleaned state for further processing from a wide area of the surrounding landscape (the animal bone assemblage awaits full analysis – a key question is whether there is evidence of similar large-scale processing of livestock on the Roman farm). This implies that, by this time, the site was just one part of a far larger estate or network of estates. Indeed, the constraints of the distinctive Breckland landscape in which the farm is located mean that it probably could not have been this productive except as one specialised component of a larger network: the soils are too nutrient-poor to sustain high-yield crops and water is too scarce to water large herds/ flocks of livestock. Yet the location perhaps has some advantages over more fertile sites in the river valleys and fen edge to the west as a processing point for agricultural produce: the heathland has abundant heather which was burned as fuel in the crop driers, and an isolated location away from centres of population may have been desirable to avoid fire risk.

Was this shift from later Iron Age subsistence farmsteads to specialised processing point within a large, surplus-producing agricultural estate a ‘homegrown’ development, resulting from the hard work and increasing prosperity of the original, now ‘Romanised’ (for want of a better term) inhabitants? Theoretically, they may have prospered and been able to gradually buy up other farms or land in the surrounding area. However, the general lack of evidence for wealth or status on the Roman site (fairly low number of coins, few rich dress fittings or other signs of status, a coarseware-dominated pottery assemblage with a very low proportion of fine tableware, very limited oyster shell) does not fit particularly well with a scenario of increasing prosperity. The agricultural produce processed at the site presumably generated considerable income but this wealth does not appear to have come back to the estate’s inhabitants.

It seems more likely that, at some point in the early or mid-Roman period, the farm was taken over or bought up by an outside individual or institution. The beneficiaries of its output were either absentee landlords living elsewhere or, potentially, an institution such as the army or the state itself. There are two strands of evidence which potentially conflict with such a reading. First, there is circumstantial evidence – in the form of a very large (30m across) flint quarry pit – for a large masonry/ part-masonry building having been constructed on the farm in the mid-Roman period (around the turn of the 3rd century). The assemblage of ceramic building material from the site is extremely small and could easily have arrived at the site as hardcore rather than having been used in buildings on the site, but there is one large fragment of wall plaster from a rubbish deposit that is too fragile to have moved far from its original setting and certainly derives from a well-built building in the vicinity. However, a building with flint wall footings (or part-flint walls) and plastered walls need not necessarily indicate particularly high status and could easily have been the dwelling of estate workers or, perhaps, an estate overseer or manager. The few identified buildings at Elveden that are likely to have been domestic are all of relatively modest size and of timber construction, evidenced by postholes and post-trenches/ beam slots. Given the fact that this farm was only part of a larger estate or estate network, why would one necessarily expect the owner to live here given its isolated position?

The second strand of evidence is a coin hoard (ELV065) discovered a few years before the excavation just outside the limit of the road corridor but in what appears to be the core of the farm. This comprises 621 bronze radiate coins deposited at or after the reign of Allectus (AD 296). The possible marking of the hoard, which had been buried in a pottery vessel in the top of a ditch, with a small flint cairn, suggests that it was intended to be recovered. Its presence might reasonably be taken as evidence that some of the income generated by the estate’s agricultural output did indeed flow back there. However, this does not necessarily imply that the people working the estate were the beneficiaries – the hoard could equally represent wages for the estate workers or an installment of profits to be transported on to an absentee estate owner. Neither the inferred presence of a well-constructed building nor the presence of some portable wealth need imply that the people working the Roman estate were the ones who benefited from its productivity.As indicated above, there are in general few indications of any great prosperity accruing despite the farm’s obvious large-scale surplus production. The questions over ownership of the Roman estate and who benefited from its output are perhaps over-laboured here, but it is important to stress how this site does not easily fit simplistic categorisation as a Roman ‘villa estate’, with all the misleading connotations that could bring.

The latest coins from the site are Theodosian issues from the last two decades of the 4th or possibly the first years of the 5th century; there are one or two other small finds and vessel forms of diagnostically late 4th or potentially 5th century date. A rapid ‘winding down’ of activity on the farm in the late 4th to early 5th century seems probable given its highly specialised economic function: such specialisation can only exist in the context of a complex economy with good transport links, peaceful political conditions, and populations of non-productive consumers e.g. urban centres or state employees. Nevertheless, it is possible that people still continued to live at the site and farm the land at a more subsistence level into the 5th century. Such continuity is almost always impossible to identify on late Romano-British sites given the cessation of large-scale coinage imports in the first decade of the 5th century and the likely end of mass production of pottery at the major manufacturing centres at the same time or soon after, two changes in material culture which render ‘latest Roman’ and ‘sub-Roman’ occupation extremely difficult to characterise or date. Even on sites with clear evidence for intensive activity extending into the very late Roman period (e.g. Colne Fen, Earith, Cambridgeshire) and extensive programmes of radiocarbon dating, definably 5th-century occupation has proved elusive.