Markets in Early Medieval Europe

by Tim Pestell

  † Fig. 12.2 does not include coins, the distribution of which has been discussed elsewhere (Blackburn 1993b, 80–90; Ulmschneider 2000b, fig. 6).

  * The writer is indebted to the generosity of Gary Parkin on whose finds, records and advocacy with other detectorists this paper is based. Without his work we would know nothing of this important site. While the finds come from the parish of Barnetby, rather than Melton Ross, the site is referred to by the latter name, this being the closest settlement. The area of the site in the northern field is in different ownership and Mr Parkin was unable to obtain permission to detect.

  † I am indebted to Marion Archibald for her comments on some coins from the site shown on a photograph.

  ‡ While one burial rite tends to predominate in Lincolnshire Anglo-Saxon cemeteries there is usually some mixing with the cremation cemeteries containing inhumations and vice-versa.

  * National Monument 21250, listed as a ‘moated site with adjacent fish ponds’.

  * National Monument 32623, scheduled as ‘a large univallate hill fort’.

  * Wapentakes were district assemblies of free men. While the word is Scandinavian, the institution is far older and Sawyer (1998 134–6) has pointed out that most of the wapentakes in Lincolnshire have English, not Scandinavian, names. Yarborough Camp retained some importance into later times, acting as one of the mustering points during the 1536 Lincolnshire Rising.

  * The place name ‘Melton’ is derived from ‘middle ton’ (Cameron 1998, 87) showing that it, at least, was a central place, the –ton element unlikely to date to before eighth century. ‘Ross’ comes from the family who held the manor in the thirteenth century and whose name is linked to ‘Ross Castle’. ‘Barnetby’ is less interesting, Cameron suggesting that it comes from an otherwise unrecorded Anglo-Saxon personal name, Beornede, to which has been added the Danish suffix –by (ibid., 10).

  † The word furc appears in a number of contexts amongst which are furcae suspensivae in 1317 (Fisher 1968) and furcae judiciales in 1296 and 1305 (Latham 1965, 204). Elsewhere in the Rolls the word furc is used in conjuction with tumbrell and pillor, again suggesting an unpleasant function.

  * The best known Anglo-Saxon gallows site is that excavated at Sutton Hoo (Carver 1998, 138–42). These burials date from between the seventh and the eleventh centuries but in view of the exceptional nature of Sutton Hoo any parallels must be treated with caution. References to gallows, or at least to the feudal right to exact the death penalty, are not uncommon in medieval Lincolnshire and, on average, gallows appear to have been set about eight kilometres apart (Platts 1985, 53–5, fig. 19). The very survival of the Melton gallows suggests their importance.

  † Trial excavations carried out around the gallows by the University of Sheffield have shown the subsoil to be severely truncated with no remains surviving.

  ‡ This building is Norman with medieval additions but contains a curious key-hole shaped window bearing the figure of a cat-like animal and some crude interlace. It is difficult to date but the Taylors favoured a Saxo-Norman date on the basis of the way in which the window opening was formed (Taylor and Taylor 1980, 47–8, fig. 24).

  CHAPTER 13

  The Anglian and Anglo-Scandinavian Sites at Cottam, East Yorkshire

  Julian D. Richards

  Introduction

  This chapter is based upon fieldwork carried out at the so-called ‘productive’ site at Cottam (East Yorkshire) from i993 to i996, the first two seasons of which have been published (Richards i999a). On the basis of the results from Cottam, I have argued elsewhere (Richards i999b) that the term ‘productive’ site is unhelpful as it generally reflects a method of finds recovery, based upon the use of metal-detectors, rather than a distinct activity in the past. Somewhat controversially for a volume including the term ‘productive’ site in its title, in this paper I will attempt to outline and develop this argument. I will suggest that not all sites which have been labelled ‘productive’ are especially rich in finds; that they do not have to have been involved in artefact production; nor need they have been primarily markets. In fact, excavation of such sites is beginning to reveal that the term ‘productive’ masks a range of activities which themselves evolve through time. We should seek to refine our typology of Middle Anglo-Saxon sites by more detailed study of metalwork assemblages alongside other material. Most critically, we urgently require further excavation of sites which have attracted the ‘productive’ label.

  Cottam: the archaeological evidence

  Excavation reveals that the site at Cottam is in fact not just one ‘productive’ site but three, developing and shifting through time (Fig. 13.1). Each of the sites is related to an ancient droveway which follows the side of a typical Wolds dry valley. Although the droveway apparently ceased to function during the medieval period it is still visible on recent aerial photographs as a linear crop mark which continues northwards up the side of the valley and then joins a major east-west route, the ‘High Street’, running along the Wolds top. If the alignment of the droveway is projected to the south then it would intersect a second major east-west route, the Roman road running from Bridlington to York. It appears from the positioning of Iron Age square barrows adjacent to the Cottam droveway that it goes back at least as far as the Iron Age, and is of a type found throughout the Yorkshire Wolds (Stoertz 1997). It certainly continued in use during the Roman occupation and survived to influence the Anglo-Saxon settlement pattern.

  FIGURE 13.1. The location of the sites at Cottam. Evidence from fieldwork demonstrates that settlements shifted through time.

  The southernmost settlement, adjacent to Cottam Grange Farm, is known as Cottam A; the two settlements to the north, adjacent to Burrow House Farm, have been described collectively as Cottam B. Cottam A begins as a standard Romano-British ‘ladder’ settlement with a series of rectilinear enclosures developing on either side of the droveway, from which metal-detectorists recovered a standard range of brooches and coins as might be associated with a reasonable high-status Romano-British Wolds farm. However, they also recovered a relatively small number of Northumbrian stycas, strap-ends and dress-pins, numbering some twenty Anglo-Saxon and Anglo-Scandi-navian objects in total (D. Haldenby, pers. comm.). They also found a chalk thatch or net weight, possibly depicting an incised ship (Richards 1995).

  Excavation found few clear Anglo-Saxon settlement traces other than a few post-holes reflecting ephemeral structures. It seems as if the main focus of the site in the ninth and tenth centuries was a large quarry hole, which may have been used as a watering hollow by herdsmen following the line of the droveway from the sites at Cottam B, to the north (Richards, in prep.).

  Cottam B was first discovered by a group of metal-detector enthusiasts in i987. Over sixty pieces of eighth-and ninth-century date were found over the following two autumn seasons, during approximately 200 man-hours of searching by five metal-detector enthusiasts (Haldenby 1990, 51), at a recovery rate of one object per three hours of detecting. The importance of the site was appreciated at an early stage and the non-ferrous metal finds were systematically plotted (Haldenby 1990, 1992 and 1994). No attempt was ever made to recover non-metallic artefacts, other than unusual finds, although the presence of pottery and bone has been acknowledged by the metal-detector enthusiasts and substantiated by the results of fieldwalking (Didsbury 1990; Richards 1999a, 15–16). Metal-detecting continued during 1990, although the rate of recovery had by now decreased to approximately one artefact per six hours of detecting (Haldenby 1992, 25).

  The detected finds from Cottam B include some sixty-eight dress pins, thirty-four strap-ends, seven rings, four brooches (including one with Jellinge-style decoration), eight lead weights, over thirty-five iron knife blades, and two so-called Norse bells. There are also some nineteen Roman coins, three eighth-century sceattas, and twenty-two ninth-century stycas.

  All the finds recovered have been found in ploughsoil, close to the surface. The site has been regularly ploughed for cereal cult
ivation but, on at least one occasion, it has also been ‘subsoil ploughed’ for the planting of potatoes, resulting in disturbance of material to a depth of c. 0.4m (i5 inches), which led to the recovery of additional metalwork (D. Haldenby, pers. comm.). Several of the metal items were quite corroded, having suffered from agricultural disturbance, whereas much appears to have only been ploughed up in recent years and was still in a good state of preservation. During the subsequent excavation most of the objects recovered were found resting in the archaeological layers immediately at the base of the ploughsoil. In certain places the ‘tram-lines’ left by the sub-soiling equipment could be seen cut into this layer, providing a graphic demonstration of the site formation processes which create a ‘productive’ site (Richards 1999a, illus. 19).

  By plotting the distribution of the metal finds it became apparent that there were two main concentrations at Cottam B (Fig. i3.2). These clusters are believed to be real, as the surrounding fields were also intensively detected and did not yield these densities of artefacts. Activity was clearly focused on these two areas. The southern concentration coincided with an enclosure visible as a crop mark, but there was also a general spread of metal artefacts to the north of this. The Cottam B southern crop-mark enclosure fits within a type identified by Stoertz as a curvilinear enclosure complex:

  Each site comprises a nucleated cluster of distinctive elongated and irregular curvilinear enclosures, tightly grouped and occasionally overlapping. They contain faint and superimposed traces of small internal features; rectangular pits are also found in association with these enclosures. The perimeter ditches appear to have been recut many times. Although linear features may be present in the vicinity of curvilinear enclosure complexes, the two are not directly related (Stoertz 1997, 55).

  Although Cottam is not as complex as many of these examples it can be seen to comprise a single module which is often included within a larger complex. Stoertz identified eleven curvilinear enclosure complexes within the RCHME Wolds survey area, including Butterwick, Rudston, Huggate, Low Caythorpe, Willerby, Garton, West Lutton and East Lutton. Of these most clearly defined examples, two are located on slopes in similar situations to Cottam, while the rest occupy low-lying positions, on the floor of the Great Wold valley. At each of these sites rectangular pits are visible as crop marks and Stoertz suggests that these represent Anglo-Saxon Grubenhäuser. On this basis she has suggested that the whole class of sites may be post-Roman. To date, rectangular pits have never been identified on aerial photographs of Cottam, but one was excavated in the southern enclosure (cut 3026: Richards 1999a, 31–3). Although the fill was clean and no dating evidence was recovered, it has been suggested that it was a Grubenhaus (D. Powlesland, pers. comm.). Certainly Grubenhäuser have been recovered in Middle Anglo-Saxon contexts at West Hesler-ton and at Wharram Percy (Milne and Richards 1992) and appear to be used quite late in the Anglo-Saxon period, at least in East Yorkshire. It is quite plausible therefore that the curvilinear enclosure complexes of Butterwick type identified by Stoertz are in fact Middle Anglo-Saxon sites (see Leahy, this volume), and may represent a stage in the evolution of rural Anglo-Saxon settlement towards enclosed sites with droveways and individual farmstead enclosures, such as that excavated at Catholme (Losco-Bradley and Wheeler 1984). It would be interesting to know if ‘productive’ sites from elsewhere in Eastern England were associated with similar enclosures, although the Wolds may be exceptional for the quality of its crop-mark evidence.

  FIGURE 13.2. The distribution of metal-detector finds (marked by crosses) superimposed on crop-mark features of the Cottam B enclosures. The modern field boundaries of Burrow House Farm are to the east.

  A similar site is known from Riby, Lincolnshire (Steedman 1994) where excavation has revealed a number of curvilinear 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.

  Excavation at Cottam B has revealed that, as well as various pits and a corn-drier, the southern enclosure includes a number of postbuilt structures, in use over at least two phases. The settlement debris includes thatch weights and ceramic lamps, which provide further support for the existence of residential accommodation, but would not have been recovered by metal-detector survey alone. In addition, there were several whetstones which had clearly been used in the sharpening of metal tools, possibly including scythes and sickles as well as knives. These may have related to metalworking; there was limited evidence for copper-alloy metalworking, in the form of some tiny fragments of molten lead, but greater quantities of fuel ash slag, as might be associated with a smithy. There was also a weathered female skull in a pit, sealed by a layer including a coin of Æthelberht of Wessex, 858–62. The pit also acted as a trap for frogs and voles and probably dates to the abandonment of the enclosure, which is therefore placed in the late ninth century. It seems unlikely that the skull would have been left lying in a pit while adjacent buildings were occupied. I have suggested that it may represent the victim of an execution killing of a type investigated by Andrew Reynolds (1998a). Kevin Leahy (this volume) has noted the possible judicial function attached to some ‘productive’ sites, specifically Melton.

  Æthelberht At that stage settlement at Cottam shifts to the north where, although it was invisible as crop marks, magnetometer survey revealed a series of sub-rectangular farm enclosures, with a massive entrance way with bank and ditch, and possible gatehouse. By the tenth century the settlement form appears to have evolved to match the Catholme model more closely (Fig. 13.3). The gatehouse reminds one of the need, recorded by Archbishop Wulfstan of York in an eleventh-century text, for a thegn to possess a burh-geat (Williams 1992, 225–7).

  FIGURE 13.3. The Anglo-Scandinavian farmstead at Cottam B, seen as a magnetometer plot. The entrance is at the south with enclosures on either side of the trackway.

  This settlement shift and the replacement of the Anglian enclosure by an Anglo-Scandinavian enclosure is reflected in the distribution of pottery recovered by fieldwalking. All the Torksey ware, which is introduced in the late ninth and early tenth centuries, is found in the northern area.

  The metal-detector evidence shows the same pattern. Objects datable to the eighth and ninth century are predominantly, but not exclusively, found in the southern group, while those of the later ninth and tenth century are predominantly but not exclusively in the northern group (Fig. 13.4). Most of the stycas were also found in the southern group; this would also be consistent with a settlement shift to the north as coin usage on rural sites in Northumbria is believed to decrease in the tenth century (Blackburn 1993b). The lead weights were originally published as spindle whorls (Haldenby 1990, 59; 1992, 36) but should now be seen as weights, of a type recovered from Dublin (P. Wallace, pers. comm.). Their distribution is focused entirely upon the northern enclosures, also reflecting the reversion to a barter economy. Some artefact types, such as the dress pins, appear to be chronologically undifferentiated, each type occurring in both the south and north enclosure areas. When other categories of find are grouped by type, interesting differences in distribution begin to emerge, reinforcing the value of using locational information for plough zone finds in order to develop artefact typologies.

  FIGURE 13.4. The distribution of lead weights, Norse bells, and other objects attributed to the tenth century, superimposed on the crop-mark enclosure and the suggested northward settlement shift.

  The strap-ends fall into two broad categories. These groups can be described as zoomorphic (those employing animal motifs, including Trewhiddle-style strap-ends), and geometric (those not employing animal motifs). These broad types have been further sub-divided into twenty-one groups by Haldenby (1997; 1998a; 1998b) on the basis of
a study of some 400 strap-ends recovered by detectorists. It is apparent from the distribution that, with the exception of a single outlier, the strap-ends with geometric ornament are confined to the area of the south enclosure, while those decorated with zoomorphic ornament are found in both concentrations (Fig. 13.5). The zoomorphic strap-ends with Trewhiddle-style ornament are also found across both areas. This would suggest that geometric types went out of fashion at the end of the ninth century, and were not used in the tenth century. This is at odds with Haldenby’s view that the geometric Arc and Step type (his Group 19) is late in the sequence, but it is proposed that the evidence from horizontal stratigraphy from the excavations at Cottam B should lead to a revision of that typology, based as it is on the assumption that the generally lower level of technical accomplishment seen on the Arc and Step type is indicative of a debased and later production date. On the other hand, the presence of both examples of the Animal Mask type (Haldenby Group 13) in the northern group might lend some support to his other proposal that these are an Anglo-Scandinavian type produced in York (Haldenby 1998a, 40). Finally, four of the Cottam strap-ends – each in the southern group – have been recognised as members of a distinctive type which have traces of inlaid silver wire ornament, generally set in rectangular panels. These four are seen as outliers of a distribution which is focussed on East Anglia (Thomas 1996). Their coincidence with the south enclosure at Cottam B supports the ninth-century dating of the type.