Markets in Early Medieval Europe

by Tim Pestell

  Appendix: the principal ‘productive’ sites in Britain

  Barham, Suffolk (50+ coins) (unpublished; 44 on EMC)

  Bawsey, Norfolk (124 coins) (unpublished; 40 on EMC)

  Bedford, near (19 coins) (unpublished; EMC)

  Bidford-on-Avon, Warwicks. (14 coins) (Wise and Seaby 1995)

  Brandon, Suffolk (20+ excavation coins) (unpublished)

  Caistor-by-Norwich, Norfolk (23 unpublished, 19 sold Christie’s 4.11.1986, lots 354–72)

  Cambridge, near (15 coins) (Blackburn and Sorenson 1984)

  Canterbury, near, Kent (19 coins) (Bonser 1997, 41)

  Carisbrooke (area), Isle of Wight (15+ coins probably from a single site) (Ulmschneider 2000a, 171–2)

  Carlisle, Cumbria (61 coins, various sites) (Pirie 2000, 49 and 76–7; EMC)

  Coddenham, Suffolk (33+ coins) (unpublished; some Sotheby 4.10.1994)

  Cottam, near Sledmere, N. Yorks. (40 coins) (Pirie 2000, 44 and 62)

  Flixborough, Lincs. (53 coins, excavation and detector finds) (Blackburn 1993b, 87)

  Hamwic (Southampton), Hants. (I34 coins from several excavation sites) (Metcalf 1988a)

  Hanford, Dorset (18+ coins) (unpublished)

  Hollingbourne, Kent (39 coins) (Bonser 1997, 41)

  Ipswich, Suffolk (c. 145 coins from several excavation sites) (unpublished, some on EMC)

  Lincoln, Lincs. (47 coins from various sites) (Blackburn, Colyer and Dolley I983; unpublished additions)

  London (158 coins from various sites) (Stott 1991; EMC)

  Malton, near, N. Yorks., site 1 (35+ coins) (Bonser 1997, 42)

  Malton, near, N. Yorks., site 2 (57+ coins) (Bonser 1997, 42–3)

  ‘North of England’, E. Yorks. or N. Lincs. (112 coins, many unrecorded) (Bonser I997, 43–4)

  Ravendale, West, Lincs. (12 coins) (Blackburn and Bonser 1984b; Blackburn I993b, 88 (‘near Grimsby’)

  Reculver, Kent (73+ coins) (Rigold 1975; Rigold and Metcalf 1984; Metcalf 1988c)

  Riby, near, Lincs. (28+ coins, few described) (Ulmschneider 2000a, 146; EMC)

  Royston, near, Herts./Cambs. border (116 coins) (Blackburn and Bonser 1986; Bonser 1997, 44; EMC)

  ‘South Lincolnshire’ (141 coins) (Bonser 1997, 41–2; EMC)

  South Newbald (‘Sancton’), S. Yorks. (126 coins) (Booth and Blowers 1983; Booth 1997)

  Thetford, Suffolk (various sites) (109+ coins) (Blackburn and Bonser 1984a, 69–71; EMC and unpublished)

  Tilbury, Essex (146 coins) (Bonser 1997, 44–5)

  Torksey, Lincs. (50+ coins) (Blackburn 2002)

  Whitby, Yorks. (169+ coins from 1920–8 excavations and other finds) (Rigold and Metcalf 1984, 265; Pirie 2000, 45 and 66–7)

  Whithorn, Wigtownshire, Scotland (66 coins from excavations) (Pirie 1997)

  York (150+ coins) (Pirie 1986a and 2000, 40 and 54–6; EMC)

  * The histograms illustrated in this article are intended to reflect the likely date of loss of the coins, rather than their date of striking. To achieve this a simple adjustment has been made to the data, assuming that one third of the coins struck in any 25-year period remained in circulation until the following 25-year period, with the proviso that no coins were carried past the years c. 760, c. 792, c. 850, c. 862, c. 865, c. 875, c. 880, c. 973, etc., when fairly comprehensive recoinages appear to have taken place. Some of the histograms are based on a smaller sample of coins than the totals indicated in the Appendix, for reasons of consistency, reliability of data or convenience.

  CHAPTER 4

  Variations in the Composition of the Currency at Different Places in England

  Michael Metcalf

  Bede tells us that when Earconberht became king of Kent in 640, his daughter Earcongota was a virtuous young lady then serving God in the monastery built by the Frankish princess Fara (i.e. Faremoutiers-en-Brie):

  At that time, because there were not yet many monasteries founded in England, numbers of people from Britain used to enter the monasteries of the Franks or Gauls to practice the monastic life; they also sent their daughters to be taught in them and to be wedded to the heavenly bridegroom; they mostly went to the monasteries at Brie, Chelles, and Andelys-sur-Seine. (Historia Ecclesiastica, 111, 8)

  Earcongota died at Faremoutiers, and Bede actually speaks eulogistically of her as ‘that gold coin’ (aureum illud nomisma) which had come from Kent.

  The list of Merovingian and other foreign gold coins found in England now runs to about 140 single finds from c. 530 to c. 670 (Rigold 1975, 653; Metcalf 1989, 267; Coin Register 1987-, ad loc.). It includes no fewer than five examples from the mint of Loco Sancto (Lieusaint-en-Brie), compared with only two from the far more prolific mint of Paris nearby. Three of the five are from Folkestone, one is from near Lincoln, and one (Coin Register 1989, No. 58) is from Heston, Middlesex. This is the only case which comes to mind, from seventh-to eighth-century England, where the distribution pattern of single finds of some particular variety of coin seems to reflect a non-commercial context. The emphasis on Folkestone is intriguing for the early history of the Kentish ports, but unfortunately it is not clear from the record whether they might not be from a single discovery, or even whether two of the finds may not be the same coin reported twice. Even so, four single finds or even two single finds and a hoard from this relatively obscure small Merovingian mint may be judged significant. There is only one French single find of Loco Sancto in Lafaurie’s list, against eight from Paris (Lafaurie 1961, 275). I cannot resist quoting what Gough’s edition of Camden’s Britannia says about Folkestone, a delightful example of laconic brevity:

  In this towne there is a mayer; and this Lord Clynton’s grantfather had there of a poore man a boote almost full of antiquities of pure gold and silver. A cony drawing his yerth betwyxt Folkestan and Hythe did cast up antique money’ (Gough 1789, i, 246).

  If I had got my hands on that rabbit, I would have put it out to stud.

  In this paper I wish to address the specific question of whether it is possible to distinguish different categories of market sites by the nature of their coin finds, and the composition of the currency found at different sites. Were there any ascertainable differences between the currency circulating on the one hand in the English wics, and on the other hand in inland centres such as ‘productive’ sites or markets of the second rank, such as might imply a different character to monetary affairs in the smaller centres? The coins in use inland might, in principle, differ from those in the wics in a variety of ways. They might be slightly inferior coins, for example they might be more worn, or the heavier coins might have been culled. Neither of these differences, if they existed, would be easy to prove, although Frands Herschend has demonstrated the phenomenon in Scandinavia (Herschend 1989, 373). The most obvious and straightforward difference would be if the range of types represented inland (remember that there are more than a hundred different types of sceattas) was slightly different from the range encountered in the wic, either because certain coin types were not carried inland to the same extent, or because certain types were actually minted at inland centres.

  This exercise, by its very nature, involves comparing one assemblage of finds with another. The comparison has to be statistical, whether one is comparing histograms of weights, or proportions of different types in a sample. So how do we define the samples we are to compare? Because there are so many different coin types, most of them scarce, we may sometimes need to group them, in order to create a sample of a statistically decent size. Those coins excavated from a wic offer a neat, ready-made sample from a clear and restricted topographical context, and those from an inland ‘productive’ site which has been conscientiously reported, without the selective withholding of common varieties by the finders, should provide a satisfactory, straight comparison. But the smaller inland centres are more difficult to handle statistically: the numbers, almost by definition, are not going to be large enough to be statistically adequate. We will leave that problem on one side for the moment. First, it is ne
cessary to find some differences (if there are any), and only then can one dream up hypotheses to try to explain them, and look for ways of testing the hypotheses.

  It is at least obvious that stark differences will not be found. At most there will be differences of degree, or tendencies, lying at the edge of what is statistically significant. The wics are certainly the best starting point: we have several well-excavated sites, from which at best a hundred or more sceattas have been recovered. Each wic is different in the range of types represented. Hamwic has a far higher proportion of the locally minted Series H than occurs anywhere else; York has a high proportion of the locally-minted Series Y. Ipswich, London, the east Kentish wics of the Wantsum Channel all tend to be more varied, with no predominant type. And indeed, even Hamwic has yielded a wide variety of English types from different mint-places, together accounting for about the same proportion of the total as Series H. Monetary circulation in England was compartmented into half-a-dozen big regions such as south Wessex, the Midlands, or East Anglia, but it was not tightly compartmented: coins from mints throughout England circulated throughout England, certainly in the primary phase (to c. 710), and to a lesser extent in the secondary phase (Metcalf 1998a, 276). In practice it is apparent that each of the wics was the gateway from the North Sea or the English Channel, into its hinterland: the mix of sceatta types in the inland parts of that region was broadly similar to what we find at the wic. That was because the minting of coinage was concentrated in the wics or in the coastlands. The currency needs of the inland parts were met by coinage being drawn in from the coastal mint-places, doubtless in a context of inter-regional trade. It follows from all this that we cannot legitimately or usefully compare one inland site with another in a different region.

  Very generally, and in proportional terms, we can say that roughly 10 percent of the sceattas minted in England were minted in York, 30 percent in East Anglia, 50 percent in Kent and London, and 10 percent in Hamwic. Money originating in the south-east became widely diffused through the Midlands, whereas the other regions were rather more self-contained. At the ‘productive’ site near Royston, for example, Series R, from East Anglia, is noticeably lacking.

  On top of this pattern, a good 20 percent of the English currency consisted of foreign sceattas, minted in the Rhine mouths area (and one percent or less consisted of the equivalent Merovingian silver coins). The Frisian coins entered England at all the major points of entry along the North Sea coasts and along the south coast. From the west, and from the Irish Sea coasts, there were virtually no inflows; and there were no English mints in the west. That, very briefly, is how English monetary circulation operated in the eighth century: recognizably much the same patterns as in the eleventh century. A priori it seems a forlorn hope to imagine that there might be significant differences between the inland centres and the wics. But let us not prejudge that question.

  Inland we find a certain number of astonishingly ‘productive’ sites. Two or three adjoining fields will sometimes yield a hundred or more stray finds of sceattas. The eighth-century equivalent of a car-boot sale was perhaps a periodic fair held regularly at a traditional fairground. People would know where to come, in the same way that they knew the particular fields or pasture where the hundred court or the wapentake was held. We also find, throughout the countryside, stray finds in ones and twos, very widely distributed. In all, we have a data-base of over 2,360 single finds (plus a substantial number of coins from hoards, which cannot be used as evidence in the same way). The art of this research lies in deciding, or discovering, what constitutes a fair comparison. In southern Wessex, for example, we can say that at Hamwic Series H accounts for almost exactly 50 percent of the secondary sceattas, and ask how that figure compares with what we find in the hinterland. The scattered single finds are too few to give a statistically secure sample of the currency available in any particular locality. They need to be collected into larger groupings somehow before they can be compared, as to the proportions of different types, with the wics, or indeed with the ‘productive’ sites. How do we draw boundaries so that the stray finds so grouped neither obscure significant concentrations of sceatta types by lumping them together with material for which the context was different, nor create artificial significance by picking and choosing?

  We can use a form of regression analysis in which we count up the total number of single finds within a circle of, say, twenty-five kilometre radius, and calculate the frequency of a particular type, for example Series H, as a percentage of that total. By repeating that procedure at numerous localities, we can create a map which is dotted with percentage figures. We can then interpret that information by drawing contour lines on the basis of the percentages, to reflect the districts where the type being studied was relatively more or less frequent. This procedure does not measure absolute quantities of coins: Series H could, in principle, be relatively frequent in a district where there was little coinage. A map of Wessex constructed in this way (Fig. 4.1) shows rather clearly that money from Hamwic was diffused into the hinterland westwards but not eastwards or northwards. The same technique has been applied to the finds from East Anglia (Metcalf 2001), in an attempt to discover the mint-places of Series R, and also to other sceatta types, the mint-places of which are debateable.

  FIGURE 4.1. Regression analysis showing the frequency of Hamwic coinage within Wessex, mapped in contours.

  With this much encouragement from the study of English coin types, we turn back to the inland market centres where sceattas have been found, and to a regression analysis of the foreign sceattas. The wic was an interface; or, in the words of Bede, London was ‘an emporium for many nations who came to it by land and sea’ (Historia Ecclesiastica, ii, 3). But what of merchants from the Rhine mouths area, who brought so much money into England, especially in the primary phase, up to c. 710? Might not some of them have thought to cut out the middle-man, by travelling straight to the interior of England, rather than by buying at a wic? If that had happened, the proportion of continental sceattas might, contrary to rational expectation, be higher inland than at the wic. Practical experience suggests that it will be better to treat the primary and secondary phases separately. We will look in the first instance at the proportion of primary porcupines to all primary finds (Fig. 4.2). Regression analysis can be applied to a statistically very adequate sample of 153 porcupines from 96 localities, out of 744 primary sceattas from 292 localities. Porcupines were pervasive: there was no part of England which they failed to reach, or from which they were excluded. In more detail, however, the picture is complex and difficult to interpret. That should not surprise us, because we are trying to isolate tendencies, in a human situation which was multifarious: some merchants may have decided to go inland, while others opted for coastwise trading, or for buying and selling at a wic. The only scientifically acceptable procedure is first to discover discrepancies, and then to try to explain them, through a hypothesis which seems to stand up to scrutiny.

  FIGURE 4.2. Regression analysis showing the proportion of primary porcupine sceattas compared to all primary sceatta finds, mapped in contours.

  The average proportion of primary porcupines, for England as a whole, is 20.6 percent. East Kent and much of East Anglia (where the evidence is numerically strongest) are close to the average. The south coast reaches 30 percent, as does Humberside. Hamwic is the focal point for diffusion westwards into Dorset but not eastwards – as we have already seen with Series H. London, the lower Thames valley, and Essex are distinctly below average, at 11–12 percent. The percentage figures marked in Fig. 4.2 are derived from all the primary sceattas lying within a circle of eighteen kilometres radius, that is, an area of 1,000 square kilometres. The size of the circle makes no difference to the argument. It is a matter of expediency: a larger sample permits a smaller circle which still includes enough single finds to be statistically indicative. Too large a circle is topographically vague. In the present case they may number anything from fifteen or twenty coins,
up to eighty. If the total is smaller, the result will be less reliable, because even one porcupine more or fewer will alter the proportion appreciably. This applies to the fringes of the circulation-area of sceattas, but also to certain empty areas between the main concentrations that have been mentioned.

  So far as one can judge from small numbers, there is a real contrast: the proportions of porcupines are genuinely low in the in-between areas. One can see how near to the edge of statistical significance the evidence is, even in the maximally favourable case of comparing 20 percent of the data, i.e. porcupines, against the whole. For many other types, the statistics are still very inadequate, even with an overall sample of 2,360 finds. But juxtaposed with the ‘empty’ areas there are also two or three ‘hot-spots’. The Oxford region exceeds 40 percent, on the basis of a fairly adequate sample, while Bidford-on-Avon, from where we have a sample that is utterly reliable and complete, is close to 50 percent. If money had been carried to those two districts from London or the lower Thames, should not one have expected figures closer to 11–12 percent? Bearing Dorset in mind too, the glimmer of a hypothesis comes to mind, involving sheep on chalk uplands. One recalls that the abbess of St Peter’s, Gloucester, at a date between 735 and 767, acquired land for use as a sheep-walk, and that the bishop of Worcester in 743/5 was freed from payment of toll on two ships at London (Sawyer 1968, No. 88). Finberg went so far as to suggest that Mercian bishops and abbesses drew an important part of their revenues from the traffic in wool (Finberg 1957, 12). Whether the same was true at an earlier date, for example between 695–715, we can only speculate, as documentary evidence does not exist.