The valuable contribution of controlled metal-detector survey, with the plotting of all finds, has been raised to a new level by this study. The relatively small-scale 1993-5 excavations, in conjunction with field-walking, plotting of metal-detected finds, and geophysics, have demonstrated settlement shift from the Anglian to Anglo-Scandinavian period, revealing a new form of Anglo-Scandinavian farmstead, and changes in economy. They have also helped us to start to refine the typologies of a number of early medieval artefact types.
A further two decades of detecting have now doubled the number of finds, and while they support our earlier broad conclusions, they also lead to some significant advances in our understanding of the Anglian to Anglo-Scandinavian settlement transition.
Firstly, with over 390 early medieval non-ferrous metallic finds now logged the amount of material recovered at Cottam B and on other such Middle Saxon 'productive' sites seems remarkable, compared to other periods. This is a widespread phenomenon that has been much debated (Ulmschneider 2000; Pestell and Ulmschneider 2003; Naylor 2004; Hutcheson 2006). When one considers, however, that the site may have been occupied for two centuries then the average rate of loss may have been one or two items a year, which seems more acceptable. Loosely secured dress pins, and strap-ends that were generally stitched to leather belts, may have been regular losses and their low value may have meant that such rates of casual loss were acceptable (Haldenby and Richards 2010, 1154). Similarly the large number of lost stycas attest to a fully monetised economy in which most people carried coins of relatively low value, and which were used in day-to-day transactions. The fact that coins were regularly changing hands does support the view that productive sites were places where economic activity took place, either as markets, for at least some of the time, or as local tax collection centres, or that they were simply places where transactions were regularly taking place between residents or with itinerant traders.
Secondly, it has now become clear that the metal-detecting evidence indicates two broad phases of Anglian activity, with an initial 8th-century focus on the settlement enclosure, and a 9th-century expansion of activity alongside the trackway northwards, in the area that later saw the development of the Anglo-Scandinavian farmstead. Although we have considered that this expansion may have been associated with metal-working, we have limited evidence for that, and there is also a full range of 9th-century finds from the new area, including stycas, pins and strap-ends. We therefore conclude that the expansion may have been caused by the development of a dedicated periodic market area, north of the settlement enclosure. We propose that this 9th-century change in activity coincides with the rebuilding of new timber structures within the enclosure, witnessed in the 1993 excavation (Richards 1999a, 34). This suggests, more broadly, that many 'productive sites' may have begun as settlement enclosures, and taken on more functions as they evolved, whereas others, such as South Newbald, have no settlement evidence, and must have been dedicated market sites (Leahy 2000, 77).
Thirdly, one of the most important discoveries to come out of the further plotting and refinement of dating of late 9th and 10th-century artefacts is that there are two stages of Viking activity at Cottam B. We still believe that in the late 9th century the Anglian enclosure was replaced by an Anglo-Scandinavian farmstead a little further north. However, we now believe that there was an initial phase of transformation of the site, which not only saw the abandonment of the Anglian enclosure, but also saw the melting down and weighing out of bullion metals over a large area. This phase was probably very short-lived and it is tempting to associate it with that part of the Great Army operating in Northumbria under Halfdan after the seizure of York. Although it is on a much smaller scale, as might be expected on a small-scale rural site, the activity is comparable to that seen on the winter camps at Torksey in 872-3, and the riverine site north of York, which has already been associated with Halfdan's move towards the Tyne in 874-5 (Williams 2015). It therefore seems appropriate to term this a 'Viking' phase, in the full sense of the word. Maybe it was the Anglian market trading or the tax collection activities that attracted the Vikings to Cottam, but whatever the cause it appears to have led to a complete change in the fortunes of the settlement. Unlike the Torksey or the north of York temporary Viking camps, at Cottam the Vikings returned, or maybe even stayed – it is impossible to say which – and built their own farmstead, on top of the Anglian market area. Given the hybrid culture that develops in this new farmstead, which we still believe was only occupied for a couple of generations at most, we term this a phase of 'Anglo-Scandinavian' settlement. Although the precise date of its foundation is impossible to determine it should certainly be seen in the context, if not the year, of Halfdan's land partition.
Finally, the further plotting of finds, combined with our understanding of the horizontal stratigraphy of the settlement, has allowed some new conclusions to be drawn regarding artefact chronology. In particular it indicates a 9th-century date for the main strap-end series (Thomas Class A) whereas the spatial differentiation of one form of strap-end (Thomas Class B4) suggests an Anglo-Scandinavian date for this type. Similarly 'collared' pins, tentatively assigned an 8th-9th century date, can now be seen to be largely 9th century, while the distribution of the collared pin with plate heads indicates that they emerged, or at least flourished, around the time of the demise of the Anglian settlement. On the other hand, chip-carved and gilt metalwork has largely been dated to the 8th century on stylistic grounds alone. The fact that seven examples are found in the southern area provides rare stratigraphic support for the view that they do not continue into the 9th century. Strong support is also provided for the Anglo-Scandinavian dating of 'Norse' bells by the three examples from the northern area. The more tentative Anglo-Scandinavian date of other artefact types, including the strap guide, and copper sheet rings with ring-and-dot decoration, is similarly supported. They are also found in the northern area at Cottam B, and on other Viking Age sites. In addition, the chronology at Cottam suggests that while ring-and-dot ornament was popular on dress pins of the Anglian period, during the late 9th and 10th centuries its usage was extended to a much broader range of artefact types. This may be in part because it was straightforward to execute on mass-produced dress accessories, but it can also be seen as a hybrid Anglo-Scandinavian cultural marker. Our full dataset is made available alongside this article so that others may be able to refine these typologies further.
Internet Archaeology is an open access journal. Except where otherwise noted, content from this work may be used under the terms of the Creative Commons Attribution 3.0 (CC BY) Unported licence, which permits unrestricted use, distribution, and reproduction in any medium, provided that attribution to the author(s), the title of the work, the Internet Archaeology journal and the relevant URL/DOI are given.