The archaeology of the Cottam area is of primary importance for the light it casts on the Anglian and Anglo-Scandinavian periods in the Yorkshire Wolds, and beyond that for the information it provides for the national context of settlement during these periods. It is argued that it provides a case study of a farming settlement which during the Anglian period was controlled as part of a royal multiple estate but as a result of estate reorganisation after the Scandinavian settlement then developed as an independent manor. First, however, it is necessary to trace its prehistoric antecedents.
The flint assemblage recovered through excavation and field collection demonstrates human activity in the immediate vicinity from the Mesolithic to the Bronze Age, but no definite habitation traces were recovered. The Wolds appear to have been extensively farmed from at least the middle Bronze Age and there are traces of numerous ploughed-out round barrows visible as crop marks in the fields adjacent to Burrow House Farm (Loughlin and Miller 1979, 86-9). Several have been excavated at various times in the past, including a group of three less than 1km south-east of Burrow House Farm, at Cowlam Wold. These were initially examined by Greenwell (1877, 213-4) and Mortimer (1909, 491-2) and then re-excavated by Brewster (Watts and Rahtz 1984). The sherd of collared urn recovered from the ditch in COT93.1 may represent a fragment of burial urn from another such ploughed out-burial.
It was in the late Iron Age, however, that the Wolds became fully settled. Mapping of crop marks from aerial photographs reveals a landscape dissected by ancient trackways and partitioned by extensive field systems (Stoertz 1997). Many of these form so-called 'ladder settlements' comprising series of rectilinear fields or paddocks defined by ditches and often fronting onto a trackway, with occasional settlement enclosures. That this landscape is pre-Roman is clearly demonstrated south of Wharram-le-Street, where the Roman road south of Malton cuts obliquely across the field systems and trackways. Where such ladder settlements have been excavated, as at Wharram Percy, a Late Iron Age date has been confirmed, although they have generally been shown to continue in use into Romano-British times (Beresford and Hurst 1990, 87-92). Excavation at Wetwang Slack suggests construction beginning around the end of the 2nd century BC and continuing throughout the Roman period, and possibly beyond (Dent 1983, 36-9). The spacing of the settlements is generally every half mile (1km), for example along the Thixendale valley, where there is a medieval village every mile (1.6km) and a Romano-British farm halfway between (Beresford and Hurst 1990, 92). The trackways look like cattle droveways and the paddocks may have served as animal enclosures although excavation of some of the medium-sized local villas and their associated corn-drying ovens shows that cereal crops were already important in the Roman period. The ladder settlement enclosure known as Cottam A, to the north-west of the present-day Cottam Grange Farm, appears to be an example of just such a series of farm enclosures and is linked to Cottam B by a sunken trackway which skirts the edge of the dry valley known as Philip's Slack. The finds assemblage recovered from Cottam A through field walking and metal detecting has a broad chronological range spanning the Late Iron Age to the Anglo-Scandinavian period but there is an emphasis on Romano-British finds and the density of Anglian finds is much lower than at Cottam B (Richards in prep). Excavation at Cottam A in 1996 demonstrated that it was indeed a Romano-British farmstead with very limited activity in the Anglian and Anglo-Scandinavian periods, limited to a very small number of structural features, and probably related to the continued use of a large quarry or watering feature (Richards in prep).
The immediate Post-Roman settlement pattern of the Wolds is much less well understood. Ancient pollen does not survive well on the chalk soils but most environmental archaeologists assume that the Wolds landscape remained largely cleared and that there was no extensive reafforestation. Nevertheless, there are no securely dated early Anglo-Saxon settlement sites from the Wolds (Watkins 1983). Early Anglo-Saxon cemeteries have been excavated on the western Wolds edge at Sancton (Myres and Southern 1973; Timby 1994), and to the east at Sewerby (Hirst 1985), but only isolated burials are known from the Wolds tops. There are no known early Anglo-Saxon sites in the vicinity of Cottam and it is probably significant that the metal detector collections from both Cottam A and Cottam B do not include material of the fifth-seventh centuries AD, although such finds are rare outside of cemeteries. Early Anglo-Saxon settlement evidence has been found immediately adjacent to the late Roman settlement at Elmswell, c.2km west of Driffield (Congreve 1938]). To the north the Anglo-Saxon cemetery and extensive settlement at West Heslerton (Powlesland et al. 1986) lies just off the chalk, at the southern edge of the Vale of Pickering.
Evidence for rural settlement in the Wolds in the Anglian period is also tenuous. York may have been a thriving and cosmopolitan trading centre from the early eighth century onwards (Kemp 1996; Hall 1994), but its hinterland was virtually unknown. During what is known as the 'Final Phase' of pagan burial, in the seventh and early eighth centuries, there was a trend to highly visible burials in the Wolds, often reusing prehistoric barrows, such as at Garton and Uncleby (Mortimer 1905). Few Anglian settlements have been excavated although post-built structures have been discovered at Thwing where occupation debris, including sceattas and stycas, indicates activity from the eighth century onwards (Manby forthcoming). Northumbrian stycas found within some of the buildings at West Heslerton suggest that this site was still occupied in the late eighth century. At Wharram Percy excavations have suggested there was rather extensive Anglian settlement throughout the guardianship area, with evidence for at least four sunken structures, a post-built hall, non-ferrous metalworking, an eighth-century smithy, and fragments of stone sculpture of the eighth century (Milne and Richards 1992; Stamper and Croft 2000). As witness to the late ninth- and tenth-century Danish settlement there are large numbers of Scandinavian type place-names but there has been scant archaeological evidence for Viking Age farmsteads in the Wolds. At Wharram Percy there is now good evidence for Anglo-Scandinavian activity in the South Manor area, including schist and phyllite hones and a ninth- or tenth-century sword hilt guard, close in appearance to a smaller guard from Coppergate (Stamper and Croft 2000). Most significantly, there is a strap end and belt slide decorated in the Borre interlace style, current in Scandinavia in the ninth and tenth centuries.
Cottam can now be added to this small sample of sites and allows us to begin to build a more complete picture of the range of sites current during these periods, and of their development.
It appears that although there is no suggestion of continuity of post-Roman settlement at Cottam A a number of the quarry pits remained open (possibly as watering holes) and the trackway which ran through that ladder settlement continued to be a major landscape feature and presumably therefore continued in use as a communications route into the eighth century AD. This can be the only reason that the Cottam B Anglian settlement was constructed almost astride it; the assemblages from extensive metal detecting and field walking are clearly focussed on the new enclosure and the trackway (10th century metal artefacts, coins, dress pins and strap ends). The reasons for the precise choice of this location on the trackway are unclear. Although the site is on the edge of the valley it commands good visibility in all directions - indeed, the twin towers of the Humber Bridge are visible on the modern horizon to the south. Nevertheless the site is slightly below the most exposed contours at the top of the valley and thus gains some small protection from the wind. The settlement does not appear to have been ideally situated with regard to water supplies although today there is a dew pond to the west of Burrow House Farm, some 100m from the settlement, and in the dry valley it can be assumed that Cottam Well would already have been providing a supply of fresh water.
Initially it appears that the settlement comprised a sub-rectangular fenced enclosure and a number of wooden buildings. The complete outline of the enclosure is revealed by crop mark mapping and geophysical survey. These show a trapezoidal enclosure with rounded corners whose longest side is on the west and which tapers towards the east. The enclosure is aligned north-south on the trackway and sits astride it and to one side, apparently controlling use of the track. The enclosure comprises several parallel ditches but excavation revealed that these may not be contemporary and suggests that the perimeter was recut on several occasions, but following the same alignment. The two sub-circular features, apparently at the north-west and north-east corners, may be contemporary features but could be earlier, possibly ploughed-out barrows. The enclosure is of a type not recognised elsewhere. Its curvilinear outline is reminiscent of the curvilinear enclosures categorised by Stoertz (1997, 69) as perhaps being post-Roman features. However, it is not as complex as those features; nor does it contain the negative features interpreted as Grubenhäuser seen on those sites. Fragments of at least two timber-halls, Buildings 1 and 2, were excavated. None of the excavated buildings yielded much dating evidence from structural features and their dating must rest upon their general spatial association with Anglian metalwork, and the absence of metal objects of an earlier or later date. The earliest datable object is a Northumbrian styca of Eadberht of 737 AD. Although we have noted a general background scatter of Romano-British pottery this is not focussed on the enclosure in the same way as the metalwork.
A similar site is known from Riby, Lincs (Steedman 1994) where excavation has revealed a number of sub-rectangular enclosures defined by ditches, here seen as foundation trenches for fences or hedges. At Riby the enclosures are also linked by droveways but despite domestic refuse being recovered from the enclosures, no post-hole buildings were found, although four possible Grubenhäuser were recorded. Occupation appears to have commenced in the sixth or seventh centuries, and to have ceased by the second half of the ninth century.
The Cottam buildings are fairly typical of the Anglian period and may be compared with contemporary local post-built halls excavated at the South Manor site at Wharram Percy (Stamper and Croft 2000) - 6.9 x 4m; Flixborough (Loveluck 1998, 152) - 9 x 5.3m; along the Caythorpe gas pipeline in North Humberside (Abramson 1996) - 7.5 x 4m; and some 75 examples from West Heslerton (Powlesland 1998a). Further afield there are parallels from Brandon (Carr et al 1988) and Hartlepool (Daniels 1988). If it can be assumed that Building 1 was aligned east-west with external dimensions of 12 x 5m, then it is also similar in size to roughly contemporary structures at Chalton where the middle range of buildings are typically 8.5 x 5.3m (Welch 1992, 17).
At some stage, most likely in the early ninth century, the settlement was replanned with a more substantial ditched and banked enclosure, and the Period IIA post-built halls were replaced. It is difficult to define coherent building plans from the features assigned to Period IIB but it appears that there was at least one structure, Building 3, c.4 x 6m, within the excavated area. This structure was represented by a combination of post-holes and shallow beam slots. It is parallelled locally by similar rectangular timber buildings excavated at Aylesby, South Humberside, and there dated to an early medieval phase of the tenth to eleventh centuries (Steedman and Foreman 1995, 20-21). The majority of datable finds are of the mid ninth century and it may be assumed that most activity belongs to this period.
The Anglian settlement debris such as the ceramic lamp base and the thatch weights also help to confirm the presence of buildings during this period. In addition, domestic objects, such as the large quantity of iron knives, tend to support the suggestion that this was a settlement site. The precise nature of the settlement, however, is still difficult to determine. The area of land enclosed suggests that it was small with no more than half-a-dozen buildings in total. It is probably best seen as a single homestead.The wealth of finds is perhaps initially surprising when compared with sites such as West Heslerton and Cowdreys Down, particularly with respect to the gold sheeting recovered by the metal detector users. On the other hand, the density of copper alloy objects is not unusual for many sites of this period, and I have argued elsewhere that the description of Cottam as a 'productive site' is misleading and unhelpful (Richards 1999). Although over 34 strap ends have been recovered from Cottam B it has to be borne in mind that this is from an area of c.30,000m². The density of strap ends per m² is identical to that from Wharram Percy, and actually lower than that of Fishergate, Flixborough, and Whitby. Similarly, although there were 63 copper alloy dress pins recovered at Cottam B, the density of these finds is lower than that at Wharram Percy, Fishergate, Flixborough and Whitby. Even the density of ninth century coins at Cottam is lower than that of the other sites. When set aside the fact that the use of metal detectors both before and during the excavations at Cottam has led to a higher rate of recovery of small copper alloy objects than is probably typical one is forced to conclude that the rate of loss of these objects fits well within the continuum of sites. Therefore there is nothing particularly 'productive' about Cottam, or indeed other sites which have been given this label, which simply describes a method of finds recovery.
This begs the question of what the site was, and to try to answer this we must examine the scale and nature of trading and manufacturing activities. The animal bone assemblage is what might be expected of a production and consumption site. A range of domestic species is represented, the most common being the remains of sheep/goat, followed by cattle. The predominance of sheep is unsurprising in this generally dry upland area, a region ideally suited to pasturage. Higham (1987, 43) has suggested that the arable potential of the soils on much of the Yorkshire Wolds made them a prime area for Anglian settlement. The wide range of size of sheep was particularly noted, indicating a breeding population. The presence of wool combs indicates that the sheep were being kept for cloth production, as well as for their contribution to the diet. Pig, domestic fowl and geese appeared to be very poorly represented, although there is limited evidence for some variety in the diet, including salmon bones, eggshell, and deer bones. The nearest source of the salmon must have been either the East Coast, or the Humber Estuary; there are no rivers nearby. The presence of the deer is particularly significant as it probably indicates that the Wolds were more extensively wooded than they are today, and that there were at least some stands of trees within easy distance where hunting may have taken place. This is especially interesting given that it has been suggested that by Domesday there was little woodland at all in much of Eastern Yorkshire (Rackham 1994). The antler combs were also presumably made locally from shed antlers gathered in nearby woodland. The rather large dog bones may also represent hunting dogs; the cats could also have been domestic pets, but may of course have been skinned. The contribution of arable crops to the economy is more difficult to assess, given the poor preservation of botanical remains. Nevertheless the imported quern stones and the corn-drier (3023-4) do at least indicate the cultivation and processing of cereal crops. The corn-drier is parallelled by an example of similar date, comprising a depression consisting two circular scoops recovered at Addingham, West Yorkshire (Adams 1996, 168-70). In the Addingham example the channel linking the scoops was lined with two sandstone cheek stones. At Cottam it appears that the lining stones had been robbed when the drying oven was backfilled.
Some local metalworking may have taken place, and although there is little evidence for iron smelting, pieces of fuel ash slag probably indicate the presence of the smith who would have been required to manufacture and maintain farm implements. Possible fragments of metallic ore were recovered during excavation (sf168). A number of small fragments of copper alloy and lead slag are presumably residue from casting processes. The diversity of copper alloy objects recovered might also reflect non-ferrous metalworking, either as raw materials gathered for melting down and re-casting, or as products. The small piece of gold sheeting (sf118) should also perhaps be seen in this context as it is difficult to interpret as a finished artefact. It was found within the area of the south enclosure, adjacent to the trackway. The silver finger ring (sf036) was found within the area of the south enclosure, in the vicinity of Building 2. Such items indicate that Cottam was part of a wider network, although the balance of evidence suggests that it was not free to import goods from the wic sites. Although the Mayen lava quernstones attest to some trading contacts all the honestones recovered from this phase of activity are from sandstone available locally in the Wharram area. Similarly there is no evidence for imported pottery, either for foreign imports as seen in York and (Mainman 1993) and Flixborough (Loveluck 1998, 154) or Ipswich ware from the South, despite there being over 250 sherds at Flixborough (Loveluck 1998, 157). The only imports from south of the Humber are the four silver wire strap ends, outliers of a type which is otherwise known mainly in East Anglia. Thomas (1996, 83) sees their presence in the Yorkshire Wolds as the result of sea-borne trade along the east coast.
Generally therefore, the artefactual evidence suggests a low level of trade, just as the dietary evidence suggests quite a low level of subsistence. In contrast to the contemporary settlement at Flixborough, Cottam appears distinctly impoverished. This is the same situation to that observed at Wharram Percy where there was little trickle-down from the thriving wic site in York (Richards 2000). O'Connor (1994, 141) has proposed that the occupants of the Fishergate site were somehow inhibited from exploiting a wider range of resources. At Cottam we see the reverse of this, with the inhabitants of a typical rural site of the period unable to benefit from the imported wine and other luxuries that were penetrating the urban markets. This underlines the fact that the wics were highly developed ports of trade, maintained by a ruling elite in order to control the movement and possession of prestige goods. At this stage Cottam and Wharram Percy may both have been excluded from direct contact with this developing market.
One possible explanation for this is that Cottam was itself under royal control during the Anglian period. The Anglo-Saxon Chronicle records that Aldfrith, king of Northumbria, died in AD 705 'on Driffelda' (Garmonsway 1953, 41). This is generally interpreted as Driffield, only 10km south of Cottam, and the association of the Driffield area with a Northumbrian king and later references in Domesday Book to a large composite estate of Driffield, held by the eleventh-century earls of Northumbria, has led to the conclusion that Driffield was a royal vill of the Northumbrian ruling clan (Brooks 1966, 58). Loveluck (1996) has suggested that there may have been a substantial royal landholding in this part of the Yorkshire Wolds, with antecedents in the wealthy Late Roman villa estates and important early Anglo-Saxon cemeteries in the area. Some have sought the caput or administrative centre of the vill in Driffield itself (Eddy 1983, 40) but Loveluck argues that we should understand 'Driffelda' as an estate rather than as a single settlement in the Anglo-Saxon period (Loveluck 1996, 43).
The presence of a silver sceatta coinage in East Yorkshire from the late seventh and early eighth centuries provides a further indication of the development of limited monetary exchange in the region at an early date. A small number of early eighth century sceattas are known from York but far larger numbers have been found in two concentrations: the first on the Humber Estuary, and the second in the Driffield area (Loveluck 1996, 44). It has been suggested that the presence of eighth-century coins at North Ferriby indicates a wic trading and craft working settlement, at a point where estuarine traffic met the cross-Humber traffic which eventually gave the two Ferriby's their names (Higham 1993, 169). Other late seventh and early eighth century sceattas, recently found at different locations on the shores of the Humber, may indicate periodic trading at beach sites rather than at a specific coastal centre (Loveluck 1994, 311). The second group consists of coins found west of Driffield, including eight in a male burial near the Tatton Sykes monument in Garton Slack (Grantham and Grantham 1965, 357), north-east of Driffield at Thwing (Rigold and Metcalf 1984, 264), and three examples at Cottam (105, 120, 132).
It is not possible to reconstruct the boundaries of the Anglian estate of Driffield. The Domesday estate is not likely to provide an accurate guide, as constituent parts of the Anglo-Saxon vill may have become independent manors by the Anglo-Scandinavian period. Loveluck proposes that the estate is likely to have contained a number of major settlements and that in addition to the caput there would have been other communities with different skills and responsibilities. He suggests that Elmswell may have provided iron-working whilst Skerne was a fishing port (Loveluck 1996, 45). On this basis it seems perfectly reasonable that the vill boundaries may have extended to include the rich agricultural and hunting land available in the vicinity of Cottam as well, which is therefore best seen as part of a multiple estate (cf Jones 1965). Corn was dried and food processing was undertaken using German quernstones imported under royal control. Animals were bred and grazed on the surrounding pasture. The inhabitants were relatively wealthy but during the Anglian period their contact with the wider world was nevertheless filtered and controlled through the caput at Driffield.
One of the most unusual finds from the final stages of the Anglian period and certainly one of the most important dating features is COT93.1: pit 1073, from which the adult female skull was recovered. The skull was already of some antiquity when it was incorporated in the pit fill. The radiocarbon date is 1295±60 B.P.,giving a calibrated date range of 684 - 775 AD with 68% confidence limits, or 647 - 877 AD with 95% confidence limits. The skewed curve makes a date in the second half of the seventh century or first three-quarters of the eighth century AD most likely. The skull may therefore have been contemporary with the first phase of Anglian settlement but was considerably weathered when buried, and had become detached from its lower jaw. It was also deposited about mid way down in the pit, resting upon the primary collapse and fill. The interpretation of the skull is difficult as there are few parallels but it is probably best seen as that of an execution victim as discussed by Reynolds (1998). Weathered and jawless skulls are known from formal execution cemeteries such as Roche Court Down, Wiltshire and more locally Walkington Wold, East Yorkshire, where there are a number of secondary execution burials in a prehistoric barrow with eleven skulls buried separately from the skeletons including seven cases following a period of rotting, the lower jaws having been lost (Bartlett and Mackey 1973). Reynolds ascribes both sites to the Late Saxon period and cites support from land charters for heafod stoccan (head-stakes) and heafod treow (tree-stakes). There are also examples of unusual burials from mid and late saxon settlements at Cheddar, Jubilee Hall and Bull Wharf (London), Hamwic, and Yarnton, Oxfordshire (A.Reynolds pers comm), and Kintbury, Berkshire (S.Ford pers comm). Whilst there is no evidence for a formal burial ground at Cottam only a small area has been excavated. One intriguing interpretation of the pit in which the skull was found might be to view it as the substantial hole in which the gallows post itself stood, and its proximity to a trackway is parrallelled with other execution sites (A. Reynolds pers comm.). Whilst it is tempting to regard the watervole and frog bones as evidence for owl pellets deposited by a bird perched atop the gallows, the fact that complete frog skeletons were discovered argues against this. The other finds from the pit comprise Anglian domestic waste including two fragments of decorated comb (sfs26, 56), a copper-alloy garment hook (sf34), a fragment of sandstone hone (sf101), and two iron teeth from wool or flax combs (sfs23, 42), as well as salmon vertebrae worked into beads. There were also a number of fragments of non-local stone (sfs31, 32) although the function of these is unknown. During this stage it appears that the pit was lying open with the skull resting on the lower fills. It seems to have acted as a trap for a number of small animals, including frogs and water voles. The presence of frog skeletons suggests that there may have been standing water in the pit at some stage, although the habitats of water voles are not thought to be specific to wet environments although they are unlikely to have been prevalent in an area subject to active human habitation. The upper fill, sealing the pit, contained a silver penny of Aethelberht of Wessex of 858-c.862/4 (sf13). It is proposed, therefore, that by the end of the ninth century this area of the settlement had been abandoned.
Occupation did not cease altogether, however, but appears to have shifted to an area some 100m to the north, and to the east of the trackway, where a farmstead was now constructed over the earlier field boundaries. The field walking evidence clearly demonstrates that Torksey type ware (of the late ninth century and early tenth century onwards) and York ware (of the tenth century) is concentrated in this area. The later metalwork is similarly concentrated to the north (Figure 26), with most but not all of the objects datable to the late ninth and tenth centuries in this area, including the spearhead (sf138), the so-called Norse bells (25, 180), the Borre style buckle (sf101), and the Jellinge style brooch (sf018). The lack of late ninth century coins should not come as a surprise as there appears to have been a general decline in usage of low denomination coinage following the Viking settlement. This need not correspond with economic decline; people may have found alternative means of conducting transactions, represented at Cottam by the lead weights (Blackburn 1993). The build up of coin finds in the late eighth century and their decline towards the end of the ninth century is typical of both Lindsey and Southern England and is thought to reflect fluctuations in the volumes of silver available (Blackburn 1993, 81).
Although the crop marks had indicated fewer settlement traces associated with these finds the magnetometer survey has revealed a clear settlement plan with a number of regularly planned sub-rectangular enclosures. Excavation of COT95 confirmed the existence of tenth-century settlement although truncation has made it difficult to do more than identify areas of clusters of post-holes, particularly to the north of the excavated area, at the head of a ditched trackway. Given that structures may also have rested upon sill-beams the lack of coherent building plans is disappointing but should not be surprising. Excavation also recovered evidence for a gateway structure, however, which is of more interest. The majority of the boundaries of the tenth century farmstead were ephemeral features, represented by shallow ditches which we must assume held wooden fences. Whilst they may have suffered from some plough damage it is extremely unlikely that there could have been differential truncation to the extent that they were ever substantial features. The entrance way was a very different affair, comprising a massive external ditch, an internal rubble bank, itself possibly topped by a timber palisade, and a timber superstructure represented by massive post-holes either side of the gate. The gate faced south, towards the recently abandoned Anglian farmstead, and the trackway leading to Cottam A and beyond. It is difficult to interpret it as a defensive structure as any potential attackers could very easily have out-flanked it and stepped over the flimsy ditch and fence to either side of the farmstead. Rather it has to be seen as a status symbol, with a major, possibly arched and decorated, timber superstructure which would have been clearly visible to those coming up the valley or using the trackway.
The existence of a gatehouse may have had a special significance for contemporary visitors to the site. In a discussion of lordly residences in England before the Norman Conquest, Ann Williams (1992, 225) cites a passage in an eleventh-century text written by Archbishop Wulfstan of York (1002-23) in which he describes how a ceorl may thrive to thegnhood:
And if a ceorl prospered so that he had fully five hides of his own land (agenes landes), [church and kitchen], bell [house] and burh-geat, seat and special office in the king's hall, then was he thenceforward entitled to the rank of a thegn.
Williams suggests that the use of the word burh-geat implies that the gate-house was the most prominent feature of the defences, and cites a number of sources to suggest that the manorial enclosures may have been hedged, fenced or ditched (1992, 227).
There are few parallels for the tenth-century farmstead at Cottam, although there were enclosures and traces of a possible entrance structure at Maxey (Addyman 1964, 38) and the Late Saxon settlement at Little Paxton comprised a ditched and gated enclosure (Addyman 1969, 67). Occupation at Little Paxton was dated to the late ninth - early eleventh centuries and comprised two formal enclosures defined by ditches, palisades and fences; within the enclosures pits and post-holes were recognised but there were no coherent building plans, suggesting these were mainly of ground sill-beam construction. There were contemporary ground sill buildings at St Neots (Addyman 1972). At Little Paxton, the entrance was defined by two substantial gate posts, some 10 feet apart, set back from the ditch, and with a third post providing a catch post in the centre (Addyman 1969, 67). Boundary ditches and enclosures are also recorded at Late Saxon sites at Cheddar (Rahtz 1979), Thetford (Davison 1967), North Elmham (Wade- Martins 1980) and Goltho (Beresford 1987).
It is tempting to regard the replanned tenth-century settlement at Cottam as Anglo-Scandinavian. A number of the artefacts, including the Jellinge-style brooch, the Borre style buckle, and the so-called Norse bells, are in distinctive Viking styles. The Norse bells are particularly interesting. There function is unknown although they could have been used as decorative horse harness, or could have been worn as charms, although no religious significance is known. Such miniature bells are not known from Scandinavia but they have been found exclusively in the Scandinavian colonies overseas. Batey has discussed the form with respect to the recovery of a similar bell with ring-and-dot ornament, from the Scottish site at Freswick Links (Batey 1988). She also records that examples of small copper alloy bells are known from the Isle of Man, with one from Meols, and that there are also three from Iceland (Batey ibid; Sveinbjarnardottir pers comm). Several are now known from the Danelaw, including one from Lincoln (Batey ibid, 214), a second from a context of c. AD 975 from 16-22 Coppergate, York (Richard Hall pers comm), a third from Goltho in a context dated c. AD 950-75 (Beresford 1987, 176). A further three are known as metal detector finds from Lincolnshire (Kevin Leahy and Caroline Paterson pers comm) and there have been recent metal detector finds from Norfolk, from Fincham and Cranwich (Helen Geake pers comm). Their dating, by analogy with other finds, must fall within the tenth century. Like the hogback tombstones (Lang 1984) it may be appropriate to regard them as a 'colonial monument', created specifically to express Scandinavian ethnicity in an overseas context.
This need not necessarily mean that the inhabitants of the tenth-century farmstead were Viking settlers who had dispossessed their Anglian predecessors. Indeed, the impression is one of localised settlement shift rather than discontinuity. There is absolutely no evidence for a violent takeover; as has been discussed above, the single female skull in the pit is much better seen as the victim of indigenous execution practices. On the other hand, the ostentation of the new farmstead with its ranch-style entrance is perhaps suggestive of the 'nouveau-riche' farmers of the Wolds, or of old farmers wanting to appear to be fashionable.
The late ninth and tenth centuries was a time of tenurial change, due to the Scandinavian settlement. A number of former large, often ecclesiastical, estates were being fragmented and passing into private ownership (Richards 1991, 30-31). In Yorkshire the Anglo-Saxon Chronicle informs us that in 876 "Healfdene shared out the lands of the Northumbrians, and they [the Scandinavians] proceeded to plough and to support themselves". Healfdene inaugurated a process of estate seizure and re-allocation which is reflected in the placename evidence. Tribute collection gave way to a more intensive and localised manorial economy as represented in the Domesday Book. The Yorkshire Wolds are one of the main areas where we believe that Scandinavian landlords took control of former large estates. In the former East Riding of Yorkshire 48% of place-names are Scandinavian influenced. This cannot denote completely new settlements filling in gaps as we now know that most of the landscape was already extensively farmed. Existing settlements must therefore have been renamed after a redivision of the land by the new Scandinavian aristocracy. If Cottam had previously formed part of a multiple estate controlled from a caput in the vicinity of Driffield then the fragmentation of that estate and the establishment of local manors would fit into a pattern suggested to have prevailed throughout much of the Danelaw (Jones 1965).
Although the settlement at Cottam remained small the character of activities also changed during the Anglo-Scandinavian phase. From being largely aceramic there is now a rich diversity of pottery types, reflecting widespread trading contacts. From south of the Humber there is Torksey type ware and Maxey type ware; and from the premier city of the Viking kingdom there is York ware. There is a pewter disc brooch, analogous to examples recovered from Coppergate, and jet from the north-east coast (sfs199-200). Similarly, the excavated Norwegian ragstone (sf40) comes from COT95; both examples recovered during field collection come from the northern part of the area. As noted above, the introduction of Norwegian stones into England has been linked to the Scandinavian invasion and settlement (Ellis 1969, 49). There are lead weights, similar to those from Viking Dublin, exclusively from the tenth century occupation area (Figure 5). Their usage probably reflects exchange transactions where the face value of coinage cannot be trusted and there has been a reversion to the use of precious metals as bullion. Unfortunately the poor state of preservation of the animal bones at COT95 makes it difficult to compare the animal exploitation patterns, other than to say that sheep remained dominant. It appears therefore that during the early tenth century the farmers of the East Riding may have begun to break out of the subsistence cycle, in which the elite consumed the agricultural surplus, with little market penetration other than via the households of the aristocracy (Higham 1993, 210). In the same way that merchants in York were now free to range more widely in search of more interesting additions to their diet (O'Connor 1994) the inhabitants of Cottam may have been freed from royal control and allegiance to the vill of Driffield and were now able to carry out their own transactions with the blossoming towns. In both eastern and southern England there was now a higher level of mercantile activity and a richer material culture.
Nevertheless, occupation appears to have been relatively short-lived, perhaps spanning some 50 years or a single generation only, from the late ninth to early tenth centuries. The absence of late tenth century coins can be taken as evidence that the site was no longer occupied by this date; elsewhere stray coin finds increase markedly during this period (Blackburn 1993, 83). It is proposed that at that point settlement shifted again, probably to the site of the deserted medieval villages at Cowlam or Cottam, although this theory is as yet untested by excavation.
There is now evidence that a number of medieval villages on the Wolds were planned (Harvey 1982). The origins of these planned villages is still subject to debate, but there are three contesting views. Sheppard has suggested that they were creations of the twelfth century (Sheppard 1966, 74-5), but it has been suggested that such major reorganisation at such a late date would almost certainly have generated some form of documentation. It has also been suggested that William of Normandy's 'harrying of the North' in 1068 led to considerable devastation in Yorkshire (Kapelle 1979), and that this would have been an appropriate time for extensive re-planning. However in recent years there has been a move to link regular planned villages with planned Open Field systems, arguing that both were more likely to have taken place during the Scandinavian period in the ninth or tenth centuries (Harvey 1981, 1982, 1983). It has been suggested (Hurst 1984; Richards 2000) that Wharram Percy was laid out as a planned village in the tenth century. A similar process of nucleation has also been observed further afield, as at Yarnton on the Upper Thames, where a well organised mid-ninth century settlement of timber halls situated in enclosures was relocated to the site of the present village in the Late Saxon period (Allen et al 1997). It is proposed, therefore, that the Cottam B tenth century settlement was short-lived because rapidly developing manorial circumstances quickly created a need for a new planned and enlarged village which was probably created on the other side of the dry valley under the Cowlam DMV.
The former village of Cowlam was ploughed out in 1972, although T.C.M. Brewster was able to carry out a series of rescue excavations for the East Riding Archaeological Research Committee during 1971-72. After Brewster's death in 1984 English Heritage commissioned Colin Hayfield to compile the excavation report (Brewster and Hayfield 1988). Brewster examined the remains of four structures within a courtyard farm complex, demonstrating that the farm had remained in occupation until the late seventeenth century, a date which correlated with the documentary evidence for desertion of the village.
Brewster's excavations focussed on one farmstead where the earthworks suggested the presence of a courtyard farm. He examined four areas, totalling some 1,348.5m2, excavating down to the natural chalk and investigating any pits or post-holes that he noticed (Brewster and Hayfield 1988, 36). In Area 1, Hayfield identified a timber structure, Building A, some 5.18m wide by 9.75m in length. This was not recognised during the excavation and there was no attempt to define floor levels or occupation levels. The building was defined by two lines of post-holes, comprising the northern and eastern walls. It is undated, but was sealed by a stone structure, Building C, which was abandoned in the seventeenth century. A sherd of Cistercian ware was found in the fill of a post-hole but it is uncertain if this feature is anything to do with the first timber building; the general pottery assemblage from the early levels in Area 1 included a sherd of a Late Saxon shell-tempered cooking pot (Brewster and Hayfield 1988, Fig.8, No.1). In Area 2, there were two east-west ditches or slots under the southern part of the medieval stone structure, Building D. Apart from a number of flints the artefacts from these slots were exclusively of early medieval date, including a sherd of a Torksey-type bowl, two sherds of Late-Saxon gritty wares, and a sandy fabric cooking pot (Brewster and Hayfield 1988, 60, Fig.21). In Area 5, there were a number of undated post-holes but also a circular pit from which a Torksey-type bowl was recovered (Brewster and Hayfield 1988, 80, Fig,35).
In summary, although Hayfield's report concentrates upon the late desertion of Cowlam village, there are some clues as to its origins. Hayfield concludes that the pottery recovered from the pits and quarries "suggests that the site was occupied from as early as the eleventh century; indeed a number of residual late-Saxon sherds amongst these groups hint at even earlier activity" (Brewster and Hayfield 1988, 86). The Late Saxon sandy fabrics are similar to those recovered from tenth- and eleventh-century deposits at Lurk Lane, Beverley (Armstrong et al 1991), and the Torksey type fabrics are common from the ninth to the eleventh centuries (Barley 1981). In contrast to Cottam B, the lack of any Anglo-Scandinavian copper alloy metalwork might be thought to rule out ninth or tenth century occupation, but given the relatively small area excavated and the recovery techniques employed it is perhaps not so surprising. Targeted metal-detecting was employed over the ploughed-out remains of Cowlam village over six days in late 1998 and early 1999. This revealed a background scatter of Iron Age and early Roman material, but there was a complete absence of any artefacts datable to AD 400-900. However, four probable tenth-century metal artefacts were recovered, comprising a third bell, an anthropomorphic buckle, and two strap ends decorated with ring-and-dot ornament, as well as eleventh- and twelfth-century and later medieval finds (David Haldenby and David Hirst pers comm). These finds support the idea of a settlement established at Cowlam in the tenth century, following the abandonment of Cottam B. Although Hayfield does not point to any early medieval structures at Cowlam, Building A is a possible candidate. If so this is of some significance for the east wall of Building C is seen as being a direct replacement of Building A (Wrathmell, in Brewster and Hayfield 1988, 106), suggesting remarkable continuity. Similar continuity is seen between the earliest chalk-footed peasant longhouses at Wharram Percy, and their timber predecessors (Richards in Rahtz in prep).
