situne dei

Årsskrift för Sigtunaforskning och historisk arkeologi

2011

Redaktion: Rune Edberg och Anders Wikström

Utgiven av Sigtuna Museum SITUNE DEI

Eyvind Skáldaspillir’s silver –

refining and standards in pre-monetary economies in the light of finds from Sigtuna and Gotland

Anders Söderberg

t is said that in the first half of the 10th marks. This they sent to Eyvind; but Eyvind Icentury, when the court poet of King had the shoulder-pin broken into pieces and Hakon the Good, Eyvind Finnsson Skál- with the silver he bought a farmstead for daspillir, had composed a poem to the himself.” (Foote 1961:125f) honour of the Icelanders, it was decided to In other translations keypti sér bú með is reward him by making him a cloak-pin. variously translated as “bought cattle with it” or “obliged to…buy food for his house- ”Eyvindr orti drápu um alla Íslendinga, en þeir hold”. The phrase “…at hverr bóndi gaf ho- launuðu svá, at hverr bóndi gaf honum skattpen- num skattpening. Sá stóð þrjá penninga ing. Sá stóð þrjá penninga silfrs vegna ok hvítr silfrs vegna” has also been translated as í skor. En er silfrit kom fram á alþingi, þá réðu “each bonde [farmer] offered a scatt penny menn þat af at fá smiða til at skíra silfrit. Síðan of pure silver equal to three weighed penni- var görr af feldardálkr, en þar af var greitt es” which is supposed to be a reference to smíðarkaupit. Pá stóð dálkrinn fimm tigu mar- imported debased Anglo-Saxon pennies, of ka. Hann sendu þeir Eyvindi, en Eyvindr lét hög- which three were equal to one penny of fine gva í sundr dálkinn ok keypti sér bú með.” silver (Dasent 1861:405). (Heimskringla: Sagan af Haraldi konungi James Graham-Campbell (2007:216) gráfeld:18. Aðalbjarnarson 1941–1979:221f). and Gareth Williams (2007:182) draw attention to this episode from Heimskringla, “Eyvind composed a poem about the people as it sheds light on several different aspects of Iceland, for which they rewarded him by of the handling of silver. In different situa- each bonde giving him three silver pennies, tions the very same silver is treated as of full weight and white in the fracture. And counted coins when collected as tax; as when the silver was brought together at the jewellery and a status gift; and finally as General Thing, the people resolved to have weighed hack silver in a market situation. it purified, and made into a shoulder-pin, The story illustrates how different systems and after the workmanship of the silver was of exchange were adopted for different pur- paid, the shoulder-pin weighed some fifty poses, and how old and new traditions could

5 SITUNE DEI

Fig. 1. Cupellation of silver: 1) recycled silver is molten together with an excess of lead in a hearth lined with a bone ash material. 2) Oxygen is blown over the molten metal. 3) The hearth material absorbs the oxidised lead (litharge) and other impurities. Drawing by the author. – Kupellering av silver: 1) Silver för återbruk smälts samman med ett överskott av bly. 2) Syre blåses över den smäl- ta metallen. 3) De oxiderade föroreningarna absorberas av härdväggen. be interchanged in a way that probably was Refining silver a reality for 10th century Scandinavia. In Silver was refined by cupellation (fig. 1). addition to this, the story reflects a clear This technique first appeared c. 3000 BC in awareness among the Scandinavians of the connection with silver mining. The technol- value and different grades of silver. There ogy became widespread and was used on a are, as both Graham-Campbell and Williams show, aspects of value measured by weight, and of fineness of quality tested by pecking or carving as indicated by the quality criterion “white in the fracture” (hvítr í skor). There is also the small but im- portant verb skíra – to purify, to cleanse, to baptise (Vifgusson 1874:550). This word refers to the process of refining the silver to maximal purity and was performed before transforming it into the gift for Eyvind. Clearly, accumulated metal went through a full refining process before reuse. Even though the story was recorded some 300 years after the event, this alleged informa- Fig. 2. Hearth material (litharge cakes) from sil- tion about Viking customs is of great inter- ver refining hearths is heavy, due to a high lead est, as several finds of early silver re fining content, grainy and sometimes contains tiny waste have recently come to light in Scan- bone fragments, grey with a green zone close dinavia. The refining of noble metals has to the surface and slightly bowl-shaped in sec- tion. 12th century litharge material from Träd- long been associated to minting, but the gårdsmästaren, Sigtuna (find no 5952). Photo Eyvind story may also help us to place these by the author. – Härdmaterial från silverraffiner- finds in the context of the management of ing är påtagligt tungt p.g.a ett högt blyinnehåll, silver in pre-monetary Scandinavian grått och kornigt med ibland synliga smärre bullion economies. benfragment, lätt skålformat i profil och med en grönaktig zon närmast ytan.

6 ANDERS SÖDERBERG

Fig. 3. A fragment of Fig. 4. A scorifier a small scorifier from fragment, Sigtu- the workshop area na, late 10th near the Birka town century, Träd- rampart, probably gårdsmästaren, from the 10th century Sigtuna (find no (find no 1071). 10532) Drawing Drawing by the by the author. a u t h o r . – Fragment – Skärvelfrag- av en liten skärvel ment, sent 900- från verkstäderna vid tal. Sigtuna. Birkas stadsvall, troli- gen 900-tal.

Fig. 5. A small scorifier with gold droplets, early 11th century, Urmakaren, Sigtuna (find no 5195). Photo by the author. – Liten skärvel med gulddrop- par, tidigt 1000-tal. Kv Urmaka- ren, Sigtuna.

large scale, as for example in the antique 1999). In these cases the aim is to drive out silver mines of Larium, Attica, a few the copper and other alloying metals, in thousand years later (cf. Conophagos order to obtain silver of maximal and 1980). The basic principle of cupellation in known silver content. During the first the mining process is melting argentiferous millennium AD the method was further de- lead in a hearth lined with a porous materi- veloped into a technique for testing or al of bone ash and clay (fig. 2). Oxygen is assaying noble metal content in smaller blown over the melted alloy whereupon the samples. This could be considered man’s lead is transformed into molten lead oxide, first invention of a method of analytical litharge, that also absorbs oxidised copper chemistry. The technique came to be linked and other metal oxides. The molten oxide is with the use of precious metals in mining, then skimmed off or absorbed by the porous goldsmithing, as symbols of value and ob- hearth material. The silver, which doesn’t jects for payment and to test the silver con- react with oxygen, stays in a metallic state tent in copper and lead in market situations and is thus refined to a concentration of up before fixing the price (Martinon-Torres, to 99%. Nils Björkenstam (1989:149ff) of- pers. comm.). Fire assay is per formed as a fers a good description, in Swedish, of the process of two steps, wherein the sample is process as used in the 17th century silver melted together with an excess of lead in a mines of Piteå in northern Sweden. small ceramic vessel, a scori fier (figs. 3–5), The same technology was used to refine and the molten alloy is then cupelled in a recycled silver. In this process lead is added vessel, cupel (fig. 6), made from absorbent to the silver in question (cf. Rehren & Kraus bone ash. The sample is weighed before the

7 SITUNE DEI

Fig. 6. A small cupel, probably made from bone Fig. 7. A muffle for covering the test hearth, ash. Probably 10th century. Stray find from according to Agricola 1556 (after Hoover & Birka (find no SHM 8985:14). Drawing by the Hoover 1950). – Ett valv att täcka testhärden a u t h o r . – Litet kapell, troligen av benaska. Lös- med vid silvertestning enligt Agricola 1556. fynd från Birka. test, as is the silver remaining afterwards, is nothing until the early 11th century finds and the difference reveals the purity of the from King Olof Skötkonung’s mint on the tested silver. Renaissance test furnaces we- Urmakaren site in Sigtuna and the Hum- re easily put together by using a portable legården site in the same town. The next muffle (fig. 7), creating a space free from find dates from the mid-11th century, and charcoal for maximal oxi dation, and also comes from a silver workshop at Fröjel, for directing the toxic lead fumes away Gotland (Söderberg & Gustafsson 2006; from the assayer. A similar solu tion was Gustafsson & Söderberg 2007). Rich finds probably used in early medi eval Scandina- of cupellation waste have occurred in late via, but so far no traces have been identifi- 12th century contexts associated with mint- ed of these kinds of objects. Scorifiers and ing and exclusive jewellery production on cupels are generally found in medieval and the Trädgårdsmästaren site at Sigtuna Renaissance mints (cf. Wad sten 1979; Mc- (Kresten & Larsson 1996; Söderberg & Lees 1996) and they also seem to have been Gustafsson 2007). Apart from these there is the alchemists’ favourite vessels in their ar- an exclusive group of finds that deserves senal of containers used for different che- mention: ceramic vessels with no obvious mical and metallurgical processes (cf. Mar- admixture of bone ash, from building group tinon-Torres & Rehren 2005). 3 at Helgö. These vessels are rectangular, up to 100x150 mm in size, and flat with low Finds of cupellation waste brims. The size seems suitable for refining and some SEM/EDS analyses smaller amounts of silver, and analyses Let us now return to Eyvind Skáldaspillir show high contents of lead, tin and zinc and examine the treatment of the silver col- (Lamm 2008:200ff). A fragment of similar lected in his honour. There are some Scan- ceramic material was discovered during ex- dinavian archaeological finds of cupellation cavation of the Urmakaren site in Sigtuna waste, in the shape of hearth lining or (find no 2903, recorded as a crucible). litharge cakes. The earliest is from a work- The traces of cupellation that were found shop of the 7th century AD at Dagstorp, in the late Viking silver workshop in Fröjel Scania (Kresten et al 2000). After this, there can be considered as concrete and illumi-

8 ANDERS SÖDERBERG

workshop had the competence to fulfil this ambition. Obviously the Gotlanders were well acquainted with the phenomenon of silver content and considered this an impor- tant quality when handling precious metals, just as did the people and craftsmen in the story about Eyvind. The cupellation material and casting waste from the Fröjel workshop have been analysed by Malgorzata Wojnar Johansson at the Archaeological Research Laboratory of Stockholm University (Wojnar Johans- son 2005a, b). The original surfaces of the samples were examined with Energy Dis- persive X-ray Spectroscopy (EDS) on the SEM. A few representative results are shown in table 1, where the values of phos- phorus (P) and calcium (Ca) indicate high Fig. 8. Mould fragments from the silver work- contents of bone material, together with shop in Fröjel, representing Stenberger’s type 3 arm-ring, find no 16889 (above) and an clay that is mainly represented by the pres- animal head arm-ring, find no 23675 (below). ence of Al, Si, Mg, Mn and Fe. The presence Photo and sketches by the author. of lead (Pb) mainly represents the lead used – Gjutformsfragment från verkstaden i Fröjel, in the refining process while Cu, Zn and Sn representerande Stenbergers armbygel typ 3 represent alloying metals that were separat- och en djurhuvudarmring. ed from the bullion during refining. Ni, As, nating illustrations of the Eyvind story, even Sb, Pt, Au and Hg probably represent trace though the find is about hundred years elements originating from the original sil- younger than the depicted event. The find ver ores. indicates that silver refining was the first These latter elements may contain infor- stage in the casting of Gotlandic silver arm- mation concerning the original source of the rings and penannular brooches. This activi- silver that was handled in the workshop. In ty is roughly dated to the mid-11th century particular, the percentages of arsenic (As), by 14C samples (lab. no. Ua-18941) and by platinum (Pt), gold (Au) and mercury (Hg) finds of mould fragments for 10th–11th may indicate an origin in the rich silver century arm-rings of type 3 (cf. Stenberger veins of Rammelsberg, Sachsen, that were 1958:110) together with the less common opened in the late 10th century. Rammels- late 11th–12th century arm-rings with ani- berg silver came to be of great importance mal heads (fig. 8). The fact that the silver not just for the German mint activity that was refined before producing these pieces, had expanded dramatically by the end of shows an ambition to produce objects of op- this century, but also for the north European timal or perhaps even pre-defined silver silver supply in general. German silver content, and that the goldsmiths in the reached the British mints in the form of re-

9 SITUNE DEI

Table 1. Values in atomic per cent from analyses (SEM-EDS) of cupellation hearth material (litharge cakes) from 11th century Fröjel (find no 20218) and from 12th century Trädgårdsmästaren in Sigtuna (find no 5952). Analyses were made on the inside surfaces, facing the molten bullion, and the outsides of the fragments. By Wojnar-Johansson 2005. – Värden i atomprocent från analy- ser av härdmaterial från silverkupellation från 1000-talets Fröjel och från 1100-talets Sigtuna.

Table 2. Values in atomic per cent from analyses (SEM-EDS) of a crucible, find no 25423, and two mould fragments, all from Fröjel. Find no 23675 represents the mould for casting animal head arm- rings shown in fig 8. Traces of As, Pt, Au and Hg indicate the use of silver similar to that refined in the workshop and raised levels of P and Ca in the crucible may indicate an admixture of bone ash to the crucible clay. Go Fröjel 2000, Bottarve 1:19. Analyses by Wojnar-Johansson 2005. – Värden i atomprocent från analyser av degel och gjutformar. Spårmetaller antyder användning av likartat silver som kupellerats i verkstaden. cycled German coins which served as the wards the use of German coin silver mixed raw material for Anglo-Saxon coins with silver of different origins, or German (Spufford 1988:74ff, 88ff). This new silver silver that had been re-alloyed in England. source also had an impact on the supply of Zinc just occurs as traces in German coins silver on Gotland, and compensated for the while Anglo-Saxon coin silver was usually decline in the import of Islamic silver that alloyed with brass, containing zinc (Hårdh had occurred by the mid-10th century. Ap- 1976:118). proximately 40,000 German coins have Also worth mentioning are similar com- been found on the island, imported between binations of trace metals, shown in analyses 990 and 1040. The combination of high of sandstone moulds for silver ingots from contents of gold and platinum in the refin- a late 10th or early 11th century house at ing waste from Fröjel, may possibly point Humlegården in Sigtuna (Söderberg towards an origin in Rammelsberg (cf. 2008:103). There are also slight similarities Kraume & Hatz 1961:19; Hårdh 1976:110). in samples of refining waste from a minting The high zinc contents in the analyses of the context related to the 12th century king crucibles and the moulds, could point to- Knut Eriksson, in Trädgårdsmästaren in the

10 ANDERS SÖDERBERG same town (table 1; cf. Söderberg & 1973; Hårdh 1996:58ff, 137ff; Gustin Gustafsson 2007). 2004:138; Kilger 2008: 286ff). It must be emphasised, however, that these analyses only show the element con- Small-scale chemical assaying and tent of metal oxides detected in small spot two inherited problems of interpretation checks on the surfaces of the samples, and The second aspect of cupellation, the ana- not the metallic content of the original al- lytical process of fire assay, is represented loys. As a result, no conclusions can be in Scandinavia from the late Iron Age by drawn concerning proportions of alloys. scorifiers from the workshop site at Helgö Another important point is the possibility (cf. Lamm 2008:191ff) and from the hill that imported silver of different ages and fort of Gåseborg (Kangur 2004). At these origins was mixed and processed together sites the scorifiers show signs of processing in the workshop. Thus, no unambiguous gold and electrum, a gold/silver-alloy. Hel- conclusions about the origin of the silver (or gö and Gåseborg were centres in the Migra- different “silvers”) can be drawn from these tion Period, which means that the skill in analyses, and the figures quoted in tables 1 handling a rather advanced analytical in- and 2 should be understood as guide lines, strument had already been well established rather than as indisputable truths. when large imports of Islamic silver were It would appear that the knowledge of introduced in the Viking Period. As gold cupellation was common among Viking sil- was always a rare and highly valued metal, versmiths and goldsmiths, just as seems to and thus usually handled in small volumes, have been the case in Provincial Roman we might conclude that the gold scorifiers cities early in the first millennium (cf. were used in metal analysis as well as being Rehren & Kraus 1999). Furthermore, it the goldsmith’s cupellation vessels for re- seems fair to assume this technology was finement of smaller amounts of gold to be important in contexts where silver served as used in his production. Silver was usually a means of exchange, as that concerns the handled in larger volumes and we might, ac- fineness of silver. As will be shown in what cordingly, consider that the silver scorifiers follows, the Scandinavians seem to have were mainly used as assaying vessels. This been concerned at an early stage with this is confirmed to some extent by their usual concept. The jewellery identified as being find contexts. produced in the Fröjel workshop are of Finds of bone ash cupels from the Viking types that may all be tied to customs of pay- Period are rare, but ceramic scorifiers are ment or the accumulation of economic relatively common. Perhaps this finds ex- wealth. The phenomena of Viking neck- planation in the early 12th century treatise and arm-rings often corresponding to spe- of Theophilus, where he describes a simpli- cific units of weight, and in many cases fied method executed solely in a ceramic were bundled together in silver hoards ac- vessel whereby the molten litharge was cording to units of weight, are well known skimmed off by hand during the process and give strength to the hypothesis that (Hawthorne & Smith 1979:96f). Neverthe- these were produced as means of payment less, bone ash cupels were probably used in or accumulators of value (cf. Lundström the period, and Theophilus does describe

11 SITUNE DEI them in another chapter concerning gold re- the late 1960s (1969:116f; Lamm 2008: fining. Raised levels of calcium indicate 195ff). This confusion has been accentuat- bone ash layers on scorifiers in Kaupang, ed by the mistaken interpretation of scori- Norway, as early as the 9th century (Peder- fiers at Fyrkat in Denmark in the 1970s, sen 2010:198) and together with a stray find which by reference to earlier identifications from Birka (fig. 6), this suggests an earlier of similar finds, were labelled soldering use of fully developed cupels than usually trays (Roesdahl 1977:51ff). This mis-iden- expected; the problem being that archaeol- tification has had a great impact on the in- ogists have not been able to identify them so terpretation of South Scandinavian finds far. This is further confirmed by finds of ever since. bone ash material identified as cupels from According to analyses made by Birgitta a smithy and a goldsmith’s workshop in Vi- Hulthén, the scorifiers from Fyrkat had borg Søndersø, Jylland, dated to 1018–1030 been exposed to temperatures of (Jouttijärvi & Andersen 2005:359). That 800–900° C from above and 500–600° C workshop was tied to activities connected to from below. 800–900° C is sufficient for King Cnut the Great, under whose rule Vi- melting silver and lead together. For in- borg became one of several mints. Another stance, 925/1000 sterling silver, containing contemporary find of bone ash material, 7.5 % copper, melts within this range. High- from the mint of King Olof in Sigtuna, is too er temperatures are demanded for full assay fragmented to determine with certainty above the melting points of lithargic lead whether we are dealing with sherds of cu- oxide (950°) and fine silver (960°). This pels or the lining of larger refining hearths may provide some valuable information (find nos 1547 and 1743; recorded as cru- about the Fyrkat scorifiers, as it may be an cibles). indirect indication that the process was fin- Scorifiers have until recently been usu- ished in cupels made from bone ash that ally described in Scandinavian archaeology were never found or identified by the ar- as “heating trays” or “soldering trays”, chaeologists. If this is the case, the Fyrkat and so too in the excavation records of the finds may suggest a fully developed fire-as- Sigtuna mint where they are recorded as saying process by the end of the 10th centu- soldering trays and sometimes as crucibles. ry, and this may indirectly support an early Despite this, we can without reasonable dating for the undated stray cupel from Bir- doubt determine that most of them were ka. used as trays in fire assay. This is confirmed The lower heat markings registered for by analyses of early 10th century finds from the undersides of the Fyrkat scorifiers might York (Bayley 1992:748ff), Viborg Sønder- be explained if they stood on a slightly cool- sø (Krongaard Kristensen 1990:344) as er surface while that heat was primarily di- well as Birka and Sigtuna (Bayley 1979; rected towards the molten alloy uppermost Söderberg 2004 & 2008) where lead, silver in the trays, as is the case in assaying. This and copper have been detected. This gives makes sense, since temperature levels in cause for a reassessment of “heating trays”, high-temperature charcoal furnaces tend to in keeping with an interpretation as assay- be rather more local than one might expect, ing trays made by Kristina Lamm already in

12 ANDERS SÖDERBERG sometimes giving surprising variance be- (Hultén 2006:41), and to provide more sol- tween different locations in the same fire. id ground for further research, all earlier The glazed upper surfaces of the trays finds would need to be re-examined. from Fyrkat contained large quantities of Another interpretative problem is pro- lead, and the metallic droplets contained sil- vided by the custom in British archaeology ver with a certain content of copper, both of using the term “cupel” for all early me- characteristic signs of cupellation. Lead, on dieval scorifiers since it has been presumed the other hand, is not used in solder and that bone ash cupels were not used in the moreover, analysis of 65 silver ornaments early Middle Ages. It has been suggested from Birka suggests that the Viking gold- that the early medieval vessels used for fire smiths usually preferred a chemical solder assay may just as well be described as cu- (different copper compounds) when dealing pels, no matter the material (cf. Bayley with delicate filigree and granulation jew- 1992:749; Lamm 2008:191). The surpris- ellery, rather than a metallic solder (differ- ingly early finds of bone ash cupels from ent alloys). A chemical solder was probably Viborg Søndersø, and the one from Birka more suitable for the purpose and does not that may date to the 10th century, tell a dif- contain any other metals but copper ferent story and show that we should rather (Duczko 1985). employ modern terminology that clearly Results from commissioned soldering separates cupels from scorifiers even when experiments by Birgittta Hulthén have been dealing with early finds. submitted when interpreting scorifiers as soldering trays. These experiments showed Scorifiers as indicators of high-status temperature levels of c. 300° C at the top of sites and minting contexts the trays (Hulthén 1995:23). When solder- There are examples of ceramic scorifiers ing small joints with a blowpipe, tempera- from 9th and 10th century Kaupang in Nor- ture levels are extremely local and hardly way (Pedersen 2010:194), Hedeby (Dresch- affect surfaces beyond a centimetre from er 1981:182), Staraja Ladoga, Gnezdovo the point of heat, and Hulthén’s results seem and Novgorod (Eniosova & Mitoyan to harmonise well with this. This also means 2005:328), in 9th century finds from the that soldering by this method could never Black Earth at Birka (Bayley 1979; Jakobs- create the dramatic vitrification we usually son Holback, personal communication) and see on the upper sides of the Viking Period at workshops close to the Birka town ram- scorifiers, even if the silversmith used a part (Söderberg 2004). There also are the fluxing agent when soldering. These scori- previously mentioned 10th century finds fiers usually show signs of being affected from Fyrkat in Denmark and those from the from temperatures of up to 1000° C. Ac- 11th-century workshop in Viborg Søndersø. cordingly, this means that there are great We find them at Brough of Birsay, Orkney numbers of finds that have been interpreted (Curle 1982:42) and in Dublin, Lincoln and as “soldering trays” on the basis of very York, and as far as can be seen, almost weak indications. There are, though, a everywhere that Scandinavians were active smaller group of low-fired ceramic trays of in the Viking Period. Obviously, the testing similar size that could derive from soldering of silver was an important metallurgical

13 SITUNE DEI process, but was it primarily tied to the markably dense find. The fort of Trelleborg goldsmiths’ production of jewellery or used in Zealand, Denmark, also produced finds in connection with exchange and the circu- of this type as well as the fort of Borgeby in lation of goods? Or both? The common size Scania; those at the latter being interpreted of the scorifiers – usually within the range as soldering trays (Brorsson 1998). of 40 to 70 mm in diameter – suggests a use for testing samples rather than for refining Refining, assaying, minting bulk silver. As the silver must be molten to- and weight manufacture gether with a considerable surplus of lead, in 11th Century Sigtuna these small vessels would hardly be suitable The excavation in 1990 of the early 11th- for refining silver sufficient for an object. century mint of King Olof on the Ur- However they do seem suitable for the makaren site in Sigtuna (Malmer et al. analysis of small samples of silver. 1991; Malmer 2010), revealed significant As with the early 10th-century finds numbers of scorifier fragments, most of from 16–22 Coppergate in York, which may them found in waste deposits in the mint be connected with mints of Æthelstan, (fig. 9). There are slight indications that dur- “King of All England” (Bayley 1992: ing the minting periods of Olof and his son 799ff), scorifiers are generally found in Anund Jakob (c. 995–1030/35) the use of minting contexts thoughout history; all the scorifiers was very firmly tied to the mint. way up to the 17th century Royal Mint at According to the few finds related to this Helgeandsholmen in Stockholm (cf. Wad- minting phase in the nearby Trädgårds - sten 1979). In a mint, the refiner purified in- mästaren site, scorifiers may have been in- coming bullion before alloying it with cop- frequently used in activities in “downtown” per or copper alloys into a legal percentage. Sigtuna during this period. It is not until the The scorifiers and cupels were important decades after the subsidence of the Sigtuna tools for controlling silver content before mint in the 1030’s that the practice of as- sending the metal to the moneyers. Accord- saying seems to flourish in workshops on ingly, it seems tempting to tie the scorifiers the common town plots in the Trädgårds - and the cupels in the Viborg Søndersø mästaren site (Söderberg & Gustafsson smithy to the mint of Cnut the Great, espe- 2007). This becomes comprehensible if we cially since bone ash cupels seem to be an consider that the right to assay precious exclusive technological speciality at this metals was probably reserved for the staff time and indicative of an activity of high so- of the mint and regarded as a regal right, and cial dignity. Finds of scorifiers may also be that the court lost control over the adminis- worth mentioning in connection with dis- tration of precious metals after the decline cussions concerning whether the Danish of the mint. 10th century circular forts – the trelleborgar Cross-comparisons with the results of – hosted the production of late Nordic coins other Sigtuna excavations must be made, (cf. von Heijne 2004:80f). Normally scori- though, before establishing this as more fiers are found occasionally and in small than a suposition, and we must await the numbers, but 160 fragments were collected publication of important reports before this in the circular fort of Fyrkat and this is a re- can be done. Nevertheless, since jewellers

14 ANDERS SÖDERBERG

Fig. 9. The medieval town plot on the Urmaka- ren site and the early 11th century mint. The distribution of scorifiers in layer 4, mainly con- centrated to the mint’s waste deposits, where also six discarded weights and two die impres- sions in lead of King Olof’s “long cross” coins were found. Distribution map super imposed on original plan by Ros 2009. – Stadsgården i Kv Urmakaren. Spridningen av skärvlar är tätast i avfallsdepåerna inne i mynthuset där även sex vikter och två stampavtryck på bly av Olofs ”long cross”-mynt påträffades.

do not show up in Trädgårdsmästaren until quality precious metal jewellery tied to the the mid 11th century, in the shape of pro- elite and the Danish Jelling court (Callmer duction of dress accessories and exclusive 1995:60f, 67). When jewellery production bridle-fittings, it seems fair to suggest that appears at all in late 10th and early 11th- most of the metal craft and manufacture of century Sigtuna it shows up in the shape of prestigious jewellery was closely tied to the delicate high-status craftsmanship, as in the court during the reigns of Kings Olof and find of a die for the production of brooches Anund Jakob. This impression correlates and a delicate gold pendant on the well with southern Scandinavia, where the Tryckaren site, not far from the mint and old tradition of mass-manufacture of cop- close to the supposed location of the king’s per-alloy brooches for widespread distribu- manor (Tesch 2007:96). The die suggests tion had declined by the second half of the local manufacture on the site and both ob- 10th century, after the collapse of the sys- jects are of the south Scandinavian Hid- tem of central trading centres for long-dis- densee style. Finds such as these are gener- tant trade. This breakdown was followed by ally considered as exclusive goldsmithing a stronger focus on the manufacture of high- products for courts closely connected to

15 SITUNE DEI

Danish traditions and to Danish politics, and indeed the manufacture of Hiddensee jewellery is in evidence at the forts of Trelleborg and Borgeby (Jansson 1991:279ff; Svanberg 1998:116ff). These links between Sigtuna and Denmark may be further strengthened by a soapstone mould with a carved image of Christ found when digging in Sigtuna’s Stora Gatan (main street) not far from the Tryckaren site. The carving is in many ways strikingly similar to the image of Christ on King Harold Blue- tooth’s runestone at Jelling (Floderus Fig. 10. The distribution of finds of type B2 1928:103f). At the Urmakaren site, the pres- weights, probably introduced by the beginning ence of goldsmiths is indicated both before of the 11th century. The distribution of type B1 as well as during the minting phase (cf. (middle) is similar. After Steuer 1984. – Fynd- distribution vikter av Steuers typ B2, antagen fig. 5). introducerad vid inträdet till 1000-talet. An exclusive town plot on the Hum- legården site in Sigtuna, dated to c. short period of production probably can be 990–1010 (Ljung & Wikström 2008:88f), counted in their thousands. These finds are has produced a few scorifiers as well as extremely important keys to further under- stone moulds for silver ingots. Furthermore, standing weight production in the Baltic cupellation waste of litharge-soaked bone area in the late Viking Period. Clearly, their ash was found in layers belonging to the manufacture was undertaken as major and subsequent building phase in the adjacent administered projects. plot, which together with the finds from Ur- In this context, maps correlating the find makaren make up one of the two finds from spots of truncated iron weights in the Baltic Sigtuna that can be stratigraphically tied to region (Steuer 1984:286f), may shed light the 11th century. The activities on the site on questions of their distribution. Steuer’s were of an exclusive nature, probably con- type B1 (middle type), considered to have nected to King Olof’s court, which is indi- first appeared at the end of the 10th century, cated by the fact that a blacksmith’s work- and type B2, at the beginning of the 11th shop here manufactured weights simultane- century, both show their densest concentra- ously with the extensive manufacture of tions in the Mälar area and southern Finland weights at the mint on the Urmakaren site. (fig. 10). If we study these maps in the light The two workshops probably formed parts of the large finds of manufacturing waste of one and the same production project. It is from Sigtuna, it seems fair to suggest that hard to tell if the weights were produced pri- the town was one of the more important pro- marily for use in the town or for external ducers of iron weights in the Baltic region distribution, but it seems fair to consider at this time. Despite the massive minting hinterland distribution while the amount of project in the beginning of the century, the weights produced during what was a very silver weight-based economy lingered on

16 ANDERS SÖDERBERG

there are no obvious mint-related finds from the site, such as dies or die impressions, to support such a suggestion. There are con- temporary English examples of the simulta- neous running of several mints within the same town, such as in late 11th-century Winchester where different moneyers inde- pendently administered the mints by com- mission from the king, with personal re- sponsibility for obtaining, storing and refin- Fig. 11. Central Sigtuna and the different sites ing the bullion (Biddle 1976: 398ff). At that (modern housing blocks) mentioned in the paper. Finds of craftworking from the late 10th time Winchester had old minting traditions and early 11th centuries. The supposed site of that are in no way comparable to the condi- the king’s manor is in the northern part of the tions in King Olof’s Sigtuna but the phe- St. Gertrud block. – Centrala Sigtuna och de nomenon should be noted, not least since kvarter som nämns i texten. Förmodad plats för English moneyers possibly were personally den tidiga kungsgården är den norra delen av kv. S:t Gertrud. present and active in Sigtuna. The mint may very well have been decentralised and and was an important concern for central spread over several workshops, similar to control. The Sigtuna mint represented a that obvious for the manufacture of weights. transitional stage between old and new sys- Probably the production of weights and tems of exchange. Perhaps the mint was not coins went hand in hand, making a decen- even intended to introduce a fully devel- tralised “landscape of royal workshops” in oped monetary economy in the strict sense, the areas surrounding the St Gertrud block since the court was obviously aware of the (fig. 11), the presumed location of the king’s need to perpetuate the old custom of pay- manor. Accordingly, it becomes rather one- ment by weighed silver (cf. Kilger in press). dimentional to speak of the workshop in the The type of workshop that existed at Urmakaren site as the mint. Judging from Humlegården is difficult to ascertain in de- the finds, rhetorically speaking, the mint tail, but both there and at the Urmakaren building could just as well have been a mon- mint the manufacture of weights was con- eyer’s own smithy for the manufacturing of fined to an extremely short period – perhaps dies and weights, storing and assaying of even weeks, judging from the very distinct bullion and goldsmithing. It would make and well-defined waste deposits and the the subject a lot more intruiging to consider Humlegården site showed one single, mas- the Urmakaren mint as one of several work- sive primary deposit of manufacturing shops involved in a larger context, includ- waste. Accordingly, the workshop can hard- ing minting and the production of goods ly have been built primarily as a “weight connected with exchange and prestige for smithy” since its activities probably lasted the court. for at least a decade. Perhaps it might even In this context it is worth mentioning the have been a second mint working simulta- manufacture of high-lead content glass neously with that in Urmakaren, although (perhaps too with English reference accord-

17 SITUNE DEI ing to crucible types and chemical compo- sition of the glass) that took place in the Humlegården smithy by the beginning of the 11th century – a prestigious production obviously performed by itinerant craftsmen returning to the site several times and build- ing a new furnace for each visit (Söderberg 2008). The glass melting crucibles from Humlegården show similar designs to con- temporary finds from Lincoln and Glouces- ter. As in Sigtuna, there is a slight tendency Fig. 12. Fragments of melting bowls, i.e. cera- in Lincoln for a contextual connection be- mic shells in which copper alloy coatings were tween the production of means of payment, brazed onto truncated spherical weights. Late silver processing and glassmaking. The ma- 10th–early 11th century. Sigtuna, kv Humlegår- jor part of the glassmaking finds from there den 2006 (find no 2438). Photo by the author. – Fragment av smältkulor, keramiska skal i vilka were excavated at Flaxengate, which is also ytskikt av kopparlegering lötts fast på sfäriska the site of most of the silver refining waste järnvikter. Kv Humlegården, Sigtuna. and an 11th century coin die (cf. Bayley 2008a:23, 27ff & 2008b).Traces of glass its evidence for weight manufacture, glass manufacture, including crucibles sherds manufacture and casting of silver ingots. (find nos 1825 & 1842, recorded as pot- The possible presence of a South Scandina- tery), are also present among the finds from vian royal goldsmith on the Tryckaren site the exacavations of the Urmakaren site in is of great interest in this context, too, as is Sigtuna. the partly excavated Trekanten site close to The finds of refining material of bone the king’s manor (fig. 11). Excavations in ash from the Urmakaren mint are extreme- Trekanten in the 1920’s revealed 11th-cen- ly sparse, well below 30 grams, and thus we tury bronze workshop finds (Floderus 1928; should probably exclude that bullion was Arbman & Floderus 2005) and waste from processed here. This may also be confirmed weight manufacture that could be primary by the sparse finds of crucibles from the depositions. A few scorifiers were found as mint. Metal was not melted here to any larg- well as a large litharge/bone ash fragment er extent. It may be that the few fragments (find no 1069f; 165 grams), that unfortu- represent rather the use of bone ash cupels nately is an undated find collected from the for assaying. Accordingly, we must search excavation dump. Despite this, the site re- elsewhere for sites where the bullion may mains highly interesting and there may still have been prepared and perhaps the money- be unknown workshops belonging to royal ers took care of this in workshops of their contexts buried in the neighbourhood. own, as they may have been personally re- The production of weights in the Sigtu- sponsible for the treatment of the bullion as na workshops is indicated by large volumes was the case for their contemporary English of melting bowls (Schmelzkugeln) – ceram- colleagues. Perhaps the Humlegården town ic remains of brazing packages in which plot could be a candidate in this matter, with thin copper alloy coatings were brazed onto

18 ANDERS SÖDERBERG

Fig. 13. Sigtuna, Urmakaren site. Distribution of melting bowls in layer 4; mainly representing the minting phase. Densest occurrence located to the waste deposits in the mint and in the passage outside. Distribution map super- imposed on original plan by Ros 2009. – Stadsgården i Kv Urmakaren med mynthuset (A 79). Spridning av smältkulor. Spridningen är tätast i avfallsdepåer i mynthuset och i passa- gen utanför.

the iron weights (fig. 12). The fragments are lie still buried in unexcavated soil. Accord- recorded in the Urmakaren site records as ingly, the early 11th century production of Schmeltzkugeln and crucibles and several weights in Sigtuna was remarkably large. papers refer to “hundreds of crucibles found On the Urmakaren site, the main finds de- in the mint”. It must be stressed that these rive from the waste deposits in the mint are not crucibles at all, but the remains of (fig. 13), where several discarded weights packaging material from a brazing process. were also found (fig. 14). As mentioned above, the occurrence of cru- cibles for melting metal was surprisingly The control of weight and silver low on the site, just 200 grams in the layers content and some aspects of connected with the mint and its following trade security phases, while there were 9000 grams of One scorifier doesn’t make a mint. Not even melting bowls, roughly equivalent to the 160 scorifier fragments would make one. manufacture of c. 300–400 weights. So far, On the other hand it seems far-fetched to waste of this sort from the manufacture of imagine a mint without the occurrence of c.1000 weights have been found at different scorifiers and cupellation waste, since these sites in the town, and even more probably processes are associated with contexts

19 SITUNE DEI wherein control of the the degree of fine- ness of precious metals is essential. Never- theless we must remember the small group of shallow low-fired vessels mentioned ear- lier that could have been used for soldering, and that a goldsmith may have been cu- pelling smaller amounts of precious metal to obtain a particular purity, such as in the manufacture of metallic solder. Still, it is hard to ignore the occurrence of scorifiers Fig. 14. A badly corroded truncated iron weight and cupels in contexts allied to gauging val- of Steuer’s type B1(middle), from the mint’s ue, to the exchange of goods and to minting. waste deposits (find no 4009). A single beaded The link between the analysis of silver and circle surrounds the weight markings on the various methods of payment has, for obvi- flat end. Photo by the author. – Korroderad ous reasons, been central throughout histo- sfärisk järnvikt av Steuers typ B1, mellanvari- ant, från avfallslagren i mynthuset i Kv Urma- ry and this circumstance has served as a karen, Sigtuna. springboard for the following hypothetical discussion. ings on coins (cf. Kilger 2006) and a few Scorifiers appear as part of the analytical finds of touchstones – “lydian stones” – at equipment in mints, but they also appear in ports of trade (cf. Van Es & Verwers contexts associated with the metalworkers’ 1980:167f; Moore & Oddy 1985). The craft and with trade in pre-monetary con- scorifier was probably also a part of this texts, such as the 8th–10th century trading “pillar stone”, supporting the practical use port of Birka. In these cases interpretation of silver bullion as a means of exchange in of function can be problematic, as the finds the Viking Period. usually appear in contexts where jewellery Ingrid Gustin (2004) emphasises the im- making and the operation of payment may portance of creating trust and a sense of se- both have taken place. Nevertheless it is im- curity in trade during the Viking Period. In portant to remember that the use of unmint- this era, market trade was probably consid- ed silver as a mean of exchange indicates ered as problematic and associated with the that silver was managed not just according risk of being swindled; trade between to weight, but also according to its purity in strangers at an open market was not pro- a very conscious way. tected by the norms that regulated exchange Viking weights and balance scales are between counterparts within one’s own well known to archaeology, as is their po- group. Accordingly, there was a need for tential to measure to levels of very fine ac- rules, common norms and procedures and curacy (cf. Herschend 1989:190). The other even designs that signalled reliability and pillar stone for a metallic means of pay- security. The use of normative weights may ment, the means to control the purity of pre- have been one of many important compo- cious metals, has so far been noted mainly nents for creating security and trust in this as ocular inspection by means of mechani- matter. cal markings, such as peck marks and carv-

20 ANDERS SÖDERBERG

In this context, it is important to stress weights – may have been one of several the possibility of a monopolisation of ways of securing the market peace that was weight manufacture in Sigtuna (Gustin a basic condition for maintaining a trading 1997; Söderberg 2008:122f), as well as sim- site, and one of several different ways of ilar indications at a 10th century workshop emphasising the the local ruler’s protection. site near the Birka town rampart, excavated The early medieval arrangement of produc- in 1987–1996 (Holmquist Olausson 1993; ing weights according to standards drawn Holmquist Olausson & Fennö Muyingo up by the court, goes back at least to the 8th 1995; Söderberg & Holmquist Olausson century royal decrees of the Frankish em- 1996), where 1374 grams of melting bowls pire, stating that exact and identical weights were found, mainly recorded as crucibles. should be used in each town and monastery The find spot was interpreted as a bronze (Gustin 2004:180). This may be the tradi- foundry in the excavation report, but should tion from which the late Viking Scandina- rather be considered a waste pit belonging vian petty kingdoms sought their models, to a smithy. The site with its several house and the power over measurement and terraces was, before the workshop phase, weighing may very well have been consid- settled with long houses and is interpreted ered a regal matter in the 10th and 11th cen- as a high status settlement (Holmquist turies. Thus, weight control may have been Olausson 1993:90ff) and it is seems likely at least as important as the novelty of mint- that the site’s connection with the upper lev- ing, once the king had founded Sigtuna by els of society was still intact during the the end of the 10th century. workshop phase in the 10th century (Hed Viking balances had the potential for Jakobsson 2003:197ff). This may be con- strong accuracy and this was also the case firmed by the production of high status jew- with fire assay. Accordingly, the Viking ellery as indicated by a find of a mould for Scandinavians possessed the necessary a disc brooch (Holmquist Olausson tools for performing accurate testing of pre- 1993:105), glass bead making and weight cious metals. An important ritual of trust manufacture, a combination that suggests a may have been to visit an assayer when close kinskip to the royal workshops in ear- dealing with larger transactions or assign- ly Sigtuna. The manufacture of weights was ment and, as the payee, to show a willing- a late event at the site, probably dating to the ness to undergo a spot test of the silver to be middle or later part of the 10th century, to assigned. As the assaying procedure is judge from the fact that the waste pit ap- somewhat complicated and takes a little peared immediately under the topsoil and while to perform, it is unlikely to have been that the melting bowls came from the upper used in smaller transactions, but the larger level of the pit-fill. the amount of silver about to change own- The manufacture of weights in Birka ership, the larger the risk, and accordingly, may have been an administered project sim- the stronger the need for procedures for cre- ilar to what we see in King Olof’s Sigtuna ating confidence; procedures that were half a century later. Monopolising and stan- well-known and trusted by both parties to dardising the production of means of pay- the assignment, even if strangers from dif- ment – in this case in the shape of normative ferent parts of the known world. Perhaps the

21 SITUNE DEI will itself to undergo the ritual might have when checking objects of fine silver and served as proof of the paying party’s hon- this quality is probably what the Heims - esty. Christoph Kilger places the custom of kringla refers to in Evvind’s story as “white ocular testing by pecking into a similar con- in the fracture” (hvítr í skor). It would hard- text of creating confidence, and emphasises ly have been used when handling coins the test was a symbolic social act – “the since they were never struck from unal- practice of carving was understood as sign- loyed silver. ing a contract, irrespective of the silver con- There is another aspect to the necessity tent” (Kilger 2006:464). The important part of showing “white in the fracture”. Fine sil- of the assay may have been the ritual itself ver shows the whitest tint of any metal in a and the agreement between the two parties, cut, with a slightly warm and noble hue. In- therefore it was not necessary to test every creased debasement turns the alloy colder single piece of silver within a transaction. and bluer in hue, which is already clearly Ocular testing techniques were the visible at the alloying ratio of sterling silver; every day methods used by craftsmen, mer- 925/1000. To an eye well acquainted with chants and the common man. The touch- silver these differences are easily detected stone, used for testing gold content by ex- by a simple ocular test. In early medieval amining the colour left by the metal after Scandinavian sources, fine silver was called scratching the stone, was probably a tool re- skírt silfr or brennt silfr – refined or burnt served for goldsmiths (cf. Moore & Oddy silver (Vifgysson 1874:528; Naumann 1985). The peckmarks and carvings we find 1987:377). These terms obviously refer on coins reveal silver testing associated back to the purification process in fire that with transactions, where judgement was the metal had undergone – cupellation. De- made according to consistency under the based silver was known as bleikr silfr, pale edge of the knife and the colour of the un- silver (Naumann 1987:377), or blá-silfr and corroded metal. The hardness and colour of grá-silfr, blue silver and grey silver respec- silver alloys change with the increase in tively (Vifgysson 1874:528), where the ex- content of base metals. Pecking also leads pression blue silver obviously refers to the the eye under the surface of the coin, thus blue hue in the fracture of alloyed silver. showing the real nature of the silver by re- Thus, the terminology concerning pure sil- vealing possible surface enriching, whereby ver relates directly to the refining process. the silver content of the surface has been de- The silver is actively purified – i.e. treated liberately enhanced in order to hide the use in a refining process that common people of debased silver (Hydman 2009, personal were acquainted with. The different terms communication). Moreover, fine silver has for debased silver, on the other hand, relate the following property: if pecked, fired red to colours and hues and thus obviously re- hot and cooled down, the peckmark will late to experiences from ocular testing still be metallic white because of the silver’s methods. Thus, while we see obvious ter- reluctance to react with oxygen, while de- minological hints of chemical refining be- based silver will turn black from the same ing well known in early medieval Scandi- treatment. This phenomenon could have navia, we don’t see similar linguistic signs been taken advantage of as a testing method of chemical testing. The common testing

22 ANDERS SÖDERBERG

(cf. German “Blicksilber”) – whereby the metal suddenly opens its eye and “looks up” are traditional professional terms for the finishing phase of the cupellation process, and this convincingly emphasises the depth of its visual impact. Perhaps this is also one of the reasons why the process later came to Fig. 15. A late Nordic coin with crosses pre- be an important part of the Renaissance al- sumably inspired by Continental and Byzantine chemist practice, as it truly could be con- coins. Harold Bluetooth (?), late 10th century. sidered an illustration of ideas of transfor- After Lagerqvist & Nathorst-Böös 1981. mation from low to high, from impure to – Nordiskt mynt, halvbrakteat, sent 900-tal. noble. If fire assay was an exclusive method method was pecking and other ocular meth- connected with upper social levels, it is ods. Fire assay, with scorifiers and cupels, once again worth stressing that the produc- may have been an exclusive technology tied tion waste from the Birka town rampart to specific contexts and professions. Even if workshops also contained a few scorifier archaeology comes across the vessels that fragments. Obviously, assayers had been were used in this process, it doesn’t show present in this workshop context where any impact on medieval Scandinavian col- weights were mass-produced under the pre- loquial language in the way that the refining sumed control of the local king. This is just process does. another example of the production of means Chemical testing may, accordingly, not of payment going hand in hand with knowl- have been a part of common everyday silver edge in assaying. The different disciplines exchange but enjoyed an exclusivity that also have in common that they require well- perhaps was tied to larger transactions with educated artisans – craftsmen with spe- silver between parties of a higher social lev- cialised mathematical knowledge including el. It is likely that the ability to consult an the understanding of the phenomenon of assayer who borrowed his technology from fractions, as well as weight-smiths who the continental mints, made a significant would have needed insight into the phe- symbolic impression in 9th, 10th and 11th nomenon of density. These metalworkers century Scandinavia. The process must also were probably specialists possessing have appeared rather magical to the specta- knowledge akin to the skills needed by the tor in a way that may have strengthened its staff in the Continental mints of the period. symbolic value. It is certainly a dramatic ex- perience to watch purified silver appear in Late Viking Gotlandic exchange of the molten oxide, when the rather dirty- silver under the influence looking litharge surface opens up and a of monetary ideas shiny mirror-like silver surface appears at It’s easy to consider pre-monetary systems the end of the process. It is almost as if sil- of exchange as “primitive” in the worst ver is created out of primordial soup. The sense of the term – uncontrolled and resting Swedish terms “blickar” and “blicksilver” upon a low cognitive level. The refinement

23 SITUNE DEI of Viking weighing technology already shows what a mistake this would be. More- over, the customs of refining and assaying precious metals show that the pre-monetary exchange systems in Scandinavia were al- ready resting upon two of the three corner stones of the monetary ideal; the concepts of weight and purity. Only the third founda- tion stone was missing – the standardised Fig. 16. The reverse of Fig. 17. The reverse an Æthelred penny of of an Æthelred penny unit with an implied additive value and car- the “long cross” type, of the “helmet” type, rying the king’s image to assure its quali- 997–1003. From the 1003–1009. From the ties, which would later make weighing and Hultungs coin hoard, Hultungs coin hoard, assaying superfluous at the circulation lev- Bunge parish, Gotland Bunge parish, Got- el. The introduction of the penny was not at (f46.2). Photo by Dan land (f20.2) Photo by Carlsson. – Frånsida Dan Carlsson. – all dependent primarily on knowledge, but av Ethelredmynt av Frånsida av Ethelred- on organisation, decision-making, and on a long cross-typ, mynt av helmet-typ, political power to bear and defend the proj- 997–1003. 1003–1009. ect itself (cf. Williams 2007). Thus there were several transitional episodes between symbols shown in figures 15–17 seem to gift and display economies, bullion or hack- correspond well to the symbols on Got- silver economies and the monetary systems landic arm-rings (figs. 18 & 19) where they present in the Viking world. One such are considered as a ”symbol type C” by episode was evident in the increasing vol- Thunmark-Nylén (Thunmark 1974:28). ume of fragmented silver in South Scandi- The type-C symbol seems to be closely as- navian hoards during the 10th century (cf. sociated with Stenberger’s type 3 arm-ring Hårdh 1996:84ff). The fragmentation is ex- (cf. Stenberger 1958:110), a type that was plained as a sign of an increasing use of sil- cast in the silver workshop in Fröjel (cf. fig. ver as a weighed means of payment, which 8). The symbols may have been punched af- is also illuminated by the story of Eyvind ter casting, and it seems reasonable to sug- Skáldaspillir, who hacked his cloak pin into gest that this happened as a result of some smaller pieces in order to buy merchandise. sort of administrative control. The symbol A hint of an ambition to actually provide depicted in figure 19, a square with curved Gotlandic silver with a third foundation sides and groups of rings or dots at each cor- stone, after weight and purity, may be rep- ner, could have originated from the symbol resented in the symbols on the reverse of that is included in the reverse cross of the late Viking arm-rings. Sometimes there are English helmet pennies introduced by different constellations of punched crosses Æthelred II in 1003 (fig. 17). The type was and rings, often reminiscent of symbols fa- also struck under King Olof in Sigtuna from miliar from late Nordic coins (fig. 15) as c.1005, while the long cross types were well as from the Anglo Saxon long cross struck by Æthelred II from 997 and by Olof and helmet pennies (cf. Thunmark in Sigtuna shortly thereafter (Malmer et al. 1974:28ff; Malmer et al. 1991:15ff). The 1991:16, Malmer 2010:52ff).

24 ANDERS SÖDERBERG

Fig. 18. A punched symbol on the reverse of an Fig. 19. A punched symbol on the reverse of an arm-ring of type 3, reminiscent of crosses on arm-ring of type 3, similar to the symbol on the Nordic coins and contemporary European pen- reverse of Æthelred’s “helmet” pennies. From nies. Kyrkljuves, Vänge parish, Gotland (find no Lilla Rone, Lye parish, Gotland (find no SHM SHM 7259) Photo by Yliali Asp, SHM. – Punsad 8315). Photo by the author. – Punsad marker- markering på insidan av armbygel av typ Ab 3. ing på insidan av armbygel av typ Ab 3, snarlik symbolen på frånsidan av Ethelreds helmet- mynt.

This custom has a history, judging from spect it may be relevant to ask what was the Hiberno-Norse arm-rings of the late 9th and primary significance of equal-armed cross- early 10th centuries. Kilger associates these es to the common man in late Viking Scan- symbols with the crosses on contemporary dinavia – Christianity or, possibly in addi- coins such as Anglo-Saxon pennies or even tion, security and trustworthiness in ex- Merovingian tremisses that are supposed to change? have set the weight standards for these arm- Lena Thunmark-Nylén discusses the rings (Kilger 2008:286). The phenomenon phenomenon and rejects the idea that the is also reminiscent of the use of Islamic coin symbols relate to geographic locations of symbols on Scandinavian weights in the the production sites, as the distribution pat- mid-Viking Period, including the Arabian tern for arm-rings with identical obverse word bakh (meaning choice, first-class, decorations shows that these products were first-rate) (Sperber 1996:96ff). The trust- usually not distributed as widely as were the worthiness of coins was in this way trans- arm-rings with identical reverse symbols. ferred to the weights (Kilger 2008:307ff). Common symbols punched on the reverse There is also a late Viking weight with a must have been used in several different cross symbol that is obviously influenced workshops on the island. However, she does by the crosses of late Nordic coins not deny a possibility that the symbols may (Lagerqvist & Nathorst-Böös 1981) and a be tied to silver purity (Thunmark similar symbol is engraved on a copper case 1974:31f). There are not yet any series of for a set of scales from Uppland (now in the analyses available to confirm whether these Kalmar County Museum; KLM 2694). symbols concern silver quality, site of ori- There seem to have been certain traditions gin or master of the workshop site, but the of providing credibility by the transferred custom of punching symbols in this way use of coin symbols in this way; in this re- does remind us of modern hallmarking. It is

25 SITUNE DEI also worthwhile mentioning that the sym- central control (Jonsson 1994:194ff). In the bols sorted under Thunmark-Nylén’s type mints of Winchester and in Worchester the C, obviously influenced by the reverse sym- moneyers were independent entrepreneurs bols of coins, began to show up simultane- working under laws strictly stipulating ously with the introduction of the Æthelred weight and silver content, and a continuous pennies in fig. 16 and 17 around 1000 AD purity control of the bullion was a central (cf. Thunmark 1974:30). Thus, the custom part of the production line (cf. Biddle of marking the silver in this way may imply 1976:398ff; Symons 2006:546ff). Later, in a certain wish to adapt the administration of the 13th century, the mint in Venice, Italy, Gotlandic exchange silver to the contempo- reached high levels in the formalized han- rary attempts to introduce coin economies dling of noble metals. The bullion was kept on the Scandinavian mainland. A tradition- under continuous quality control by the al “primitive” system of exchange seems to mint’s staff, and all refining in the city had meet an advanced administrative approach to be executed by the state refinery or by the at a turning point between old and new in mint, even for the precious metal used by Scandinavian customs of exchange the Venetian goldsmiths (cf. Stahl Perhaps this should not be considered as 2000:289; 326ff). unexpected – the fact that silver bullion was Eleventh century society on Gotland refined before casting the arm-rings, as was could hardly be compared to that of Eng- done in the Fröjel workshop, may by itself land, and even less to a Venetian organisa- be a strong hint of an administrative ap- tion two centuries later, but it seems likely proach towards silver. If cupellation wasn’t that the punched symbols on the arm-rings just connected to an ambition to be able to – to be considered more as ornamented in- distribute silver of the highest possible pu- gots than as jewellery – hint at something rity, but was also aimed at a slightly higher happening to the organisation of exchange. level by the desire to actually reach stipu- A somehow official approval of purity and lated standards, this would perhaps suggest weight, confirmed by a symbol, may sug- a further approach towards the routines of gest the presence of an authority with cer- contemporary mints. English mints had a tain ambitions for centralisation. long-standing tradition of keeping stan- dards, as well as of periodic recoinages. We Hence strength and security find these ingredients in King Alfred’s The silversmiths of Gotland did not execute coinage in the late 9th century (Blackburn their work at random. The custom of refin- 1995:106f). As mentioned, there are also ing the bullion in combination with punch- obvious material traces of quality testing ing symbols on the reverse of their prod- from early 10th century York in the shape of ucts, possibly suggests a wish to formalise cupellation- and assaying waste – litharge their production under the influence of cus- cakes and scorifiers – tied to King Æthel- toms associated with minting in England stan’s minting. In 973, King Edgar’s reform and Germany as well as on the Scandina- enforced a standard coinage over the king- vian mainland by the beginning of the 11th dom. The production was decentralised to century. several regional mints, but operated under

26 ANDERS SÖDERBERG

The knowledge of weighing, refining Hårdh, where the neck- and arm-rings gen- and assaying implies the ability to maintain erally show zinc contents below 0.5% and proportional alloy standards. Can the idea the filigree jewellery mainly shows no zinc of standards have been an integral part of content at all (Hårdh 1976:120ff). Pure cop- the production and use of silver objects in per was not used to alloy silver in Scandi- the Viking Period? The average silver con- navia and England in the Viking Period; in- tent in southern Swedish neck- and arm- stead copper/zinc-alloys were used for this rings of the 10th and 11th centuries, as purpose (Hårdh 1976:117). analysed by Hårdh (1976: 114ff, 120ff), all The custom of debasing coins to secure show high percentages of silver. The work profit for the king by a successive reduction of Hårdh may reflect different “standards” of the silver content in each new edition relating to different categories of objects, without altering its weight and nominal val- where analysed neck- and arm rings – the ue, did not make its appearance in Scandi- classic “primitive money” groups – show navia until the late 11th century. In the mid- average Ag-percentages of 90,6 and 92,6 %. 11th century Norwegian coinage of King This hits close to the percentages in pre-mid Harold Hardrade, roughly contemporary 10th century Islamic coins as well as to with the workshop in Fröjel, the practice western European coins of the early 11th seems to have been in full swing. This century. Analyses of four 10th century arm- seems to be supported, along with modern rings from Birka show similar if slightly analyses (Elfver 2007), by the Icelandic higher percentages, c. 93–96 % (Arrhenius saga Morkinskinna, which relates that the et al. 1973:154ff). These figures seem to re- king’s guardian, Halldór Snorrason, pre- flect different silver standards for exchange ferred payment in fine silver rather than in purposes, while the average percentages in King Harald’s coins which he knew con- status objects such as filigree jewellery gen- tained a very high content of copper (cf. erally lie between averages of 96,8 and 97,7 Williams 2007:178). % and even up to 99±1,0 % as the highest The coins of early 11th century Sigtuna content. This may probably have been con- hold a fair silver content and do not show sidered as fine silver – “skírt” or “brennt” signs of debasement. They show an average silver – and perhaps this was considered a content of 89,7 % for the first decades of common standard for fine jewellery when minting (10 analyses), close to the results of the reward for Eyvind Skáldaspillir was analyses of 16 German Otto-Adelheid coins manufactured. struck in Goslar (Malmer 2010:44) and It is obvious that Islamic silver was not slightly below the contents of the early 10th re-alloyed in Scandinavia before casting it century Islamic and late 10th century An- into objects. Islamic coins, and objects glo-Saxon coins. Furhermore, the average made from re-melted Islamic coins, are gen- percentage in the analysed Sigtuna pennies erally alloyed with pure copper and contain shows a slight tendency to correspond clos- very sparse zinc percentages. This can be er to the silver content in the neck-rings studied in the Scandinavian 10th century rather than to the arm-rings analysed by jewellery from Birka (Arrhenius et al. Hårdh, which perhaps suggests an adjust- 1973:156f) and in the analyses made by ment to traditional neck-ring standards – or

27 SITUNE DEI simply that they were struck directly from this way have been transferred from the dis- imported coin bullion of German origin. tribution level and back to the producer, Minting was in itself an innovation for who confirmed it by punching symbols be- the Scandinavians, and it seems likely that fore releasing the arm-rings into circulation. this was performed not just under the influ- ence of western European coinage but also Conclusions with regard to traditional Scandinavian ap- There seem to be a few discernible Viking proaches to silver. Thus deliberate debase- silver standards, one or two “currency stan- ment was perhaps out of the question in the dards” for arm- and neck-rings that fairly early Sigtuna coinage since it may have closely follow foreign coin standards, and a challenged traditional codes of honesty and standard for fine jewellery with the quality honour. This, together with the simultane- of fine silver. We do not know for certain if ous royal maintenance of the weight econo- silver was routinely refined before being my, may make the Sigtuna coinage rather a cast into objects for exchange, or if the finds transitional stage than part of a fully devel- from the Fröjel workshop uniquely reflect oped monetary economy (cf. Kilger in current trends belonging to the mid-11th press). The differences between the Danish century. This was an era when the large in- and Swedish systems of exchange in the flow of silver to Scandinavia was ebbing 11th century, may in some respects be illus- and when obtainable bullion may have been trated by Steuer’s distribution maps for the of poor quality, which could explain the B1 (middle) and B2 weight types. The need for refining before use in order to ob- weights are hardly represented at all in Dan- tain a traditional currency standard. ish soil except in the Hedeby area and From the available sources, we might sparsely in Scania and Blekinge (cf. fig. 10). carefully draw the following conclusions: In terms of monetisation, Danish minting of silver used for fine jewellery was refined in Sven Forkbeard and Cnut the Great was far order to maintain its high purity, while arm- more developed than the Svealand/Sigtuna and neck-rings were possibly cast directly minting ever became since the use of from imported coins without pre-treatment weights remained deeply integrated with – except during periods when the quality of Svealand’s exchange customs. the imported silver was less reliable and re- The formalized handling of silver that fining would have been necessary. We can- we experience from 11th century Gotland not yet tell if the late 10th and early 11th may in some respects also be considered a century moneyers of Sigtuna refined their transitional stage on the way to monetisa- bullion on a regular basis or if they simply tion. The refining of bullion, the adjusting struck the coins directly from delivered bul- of weight and the application of punched lion, but we do know that they controlled its “hallmarks” inspired by the monetary quality by fire assay. sphere that most Gotlander’s were probably Analyses were made possible by financial contribu- well acquainted with, may have lent a sym- tions from the Non-Ferrous Metals Research Com- bolic assurance of security and reliability at mittee of Jernkontoret’s Historical Metallurgy Group. the exchange level. The responsibility for English language revision by Fiona Coffey & Uaininn ensuring weight and silver quality may in O’Meadhra.

28 ANDERS SÖDERBERG

Sammanfattning den period då Kung Olof myntade i Sigtuna, på- Eyvind Skáldaspillirs silver – raffinering träffas fynden framförallt i mynthuset. Det är och standard i förmonetära ekonomier i rimligt att anta att även förmonetär närvaro av ljuset av fynd från Sigtuna och Gotland skärvlar, dvs före Sigtunamyntningen, varit kopplad till testmetoder för betalningsmedel. Kupellationsprocessen har en lång historia. Den Det finns en nära koppling mellan kupellation, bör ha uppfunnits i Mellanöstern runt 3000 f Kr silvertestning och betalningsmedel genom och används vid extrahering av silver ur silver- historien. En annan koppling mellan fynd av haltiga blymalmer. Senare har processen även skärvlar och betalningsmedel visas av verk- använts vid raffinering av legerat silver för åter- städer vid Birkas stadsvall där även storskalig bruk, samt utvecklats till en analytisk process produktion av vikter ägde rum, en typ av pro- som används för test av silverkoncentrationen i duktion som senare också ägde rum i kung Olofs silverprover alltifrån medeltida myntning till mynthus i kv Urmakaren och i en smedja i kv modern ädelmetallhantering och även av alke- Humlegården i Sigtuna. Vid handel och silver - mister under renässansen. Blybemängt härdma- överlåtelser föreslås i artikeln att testmetoder terial av benaska från silverraffinering dyker genom kupellation kan ha förekommit som en upp i Sverige redan i vendeltid i en vid en stor- tillitsskapande rutin knuten till större transaktio- mannagård i Dagstorp i Skåne. Senare finner vi ner på högre nivåer i samhället, ett komplement fynd i det tidiga 1000-talets myntning under till gängse okulära testmetoder. kung Olof Skötkonung i Sigtuna och i en verk- Myntekonomin tog god tid på sig att slå rot i stad från 1000-talets mitt i Fröjel socken på Got- Svealand och på Gotland. Ändå finns tidigt teck- land, som producerade silverarmbyglar av spe- en på monetära influenser av vilka omsorgen av cifikt gotländska typer. silverhalten var en. Av fynden från Fröjel att dö- Förmodligen har även kemiska testmetoder ma lade 1000-talets gotlänningar möda på att för ädelmetaller använts tidigt. Keramiska raffinera silver inför tillverkning av armbyglar skärvlar är representerade redan i folkvand- som fungerade som värdeackumulatorer. Till- ringstid på verkstadsplatsen i Helgös husgrupp sammans med noggrannheten i vikingatida vik- 3. Senare har de en tendens att dyka upp i vi- tmätning kan metoden även ha gett skandina- kingatida högstatussammanhang under 800-, verna verktyg att upprätthålla fasta metallstan- 900- och 1000-talen, som i Birka, Kaupang, He- darder, något som i någon mån bekräftas av till- deby, Staraja Ladoga, Gnezdovo, Novgorod, gängliga silveranalyser som grovt förefaller Sigtuna, Dublin, York, i Fyrkat och Viborg Søn- kunna särskilja tre skilda legeringsgrupper – ar- dersø. På ett fåtal platser, i Birka och 1000-talets mringar, halsringar och filigranarbeten. Andra Viborg Søndersø, har även hittats tidiga exem- tecken på hög omsorg om silverkvalitet åter- pel på s.k. kapell, små kärl av benaska som an- finns som språkliga avtryck. Medeltidssvenskan vänts tillsammans med de keramiska skärvlarna skiljer skírt silfr eller brennt silfr – raffinerat el- vid ädelmetallanalys. ler bränt silver, dvs finsilver, från bleikr silfr, Det tidiga 900-talets York och 1000-talets blekt silver, blá-silfr och grá-silfr, blått och grått Sigtuna och Viborg Søndersø var myntorter och eller legerat silver. Av intresse är att termerna för det verkar rimligt att anta att användningen av finsilvret refererar direkt till kupellationsproces- tekniken på dessa platser intimt hängt samman sen, medan termerna för legerat silver hellre re- med myntningen. I början av 1000-talet, under fererar till okulära erfarenheter och således till

29 SITUNE DEI

vardagliga testmetoder som pecking, dvs att Bayley, J. 2008a. Lincoln: Evidence for metal - hacka märken i silvret med kniv för att komma working on Flaxengate and other sites in the åt att studera silvrets färg i rena okorroderade city. English Heritage. Research Department ytor, något som för ett vant öga kan berätta Report Series 67–2008. Portsmouth. mycket om silvrets renhet. (http://services.english-heritage.org.uk) Ett annat tecken på inflytande från mynteko- Bayley, J. 2008b. Lincoln: Evidence for glass- nomi är förekomsten av gotländska armbyglar working on Flaxengate and other sites in the med punsade symboler på insidorna, som på- city. English Heritage. Research Department minner om symboler på frånsidorna hos samtida Report Series 68–2008. Portsmouth. engelska och kontinentala mynt samt hos de se- (http://research.english-heritage.org.uk) na nordiska halvbrakteaterna från 900-talet. Det Biddle, M. (ed.). 1976. Winchester in the early förefaller rimligt att anta att dessa symboler Middle Ages: an edition and discussion of the markerar en kontrollrutin i verkstäderna som an- Winton Domesday. Oxford. slutit till någon form av administrativt system Blackburn, M.A.S. 1995. The London Mint in kopplat till distributionen – en kontroll av vikt the reign of Alfred. In Blackburn, M.A.S. & eller silverkoncentration eller kanske både och. Dumville, D.N. (eds.): Kings, Currency and En eventuell auktoriserad kontroll ses här som Alliances: History and Coinage of Southern en trygghetsskapande rutin för att säkra mark- England in the Ninth Century. Woodbridge. nadsfrid och tillit till värdebärarna och denna Björkenstam, N. 1989. Framställning av silver kontroll kan ha bekräftats genom symboler lå- under 1600-talet – en översikt anpassad till nade från mynt som gotlänningen givetvis var malmer och metoder som kan ha använts i Sil- väl förtrogen med. bojokk. In Andræ, T. et al. (eds.): Silvret från Nasafjäll. Arkeologi vid Silbojokk. Stock- References holm. Aðalbjarnarson, B. (ed.). 1941–1979. Snorri Brorsson, T. 1998. In the workshop of the Viking Sturluson: Heimskringla 1. Íslenzk fornrit Age goldsmith – Gold and silverwork at Bor- XXVI–XXVIII. Reykjavík. geby in Scania, southern Sweden. Fornvän- Arbman, H. & Floderus, E. 2005. Vattenled- nen 93. ningsgrävningen i Sigtuna 1925: arkeologisk Callmer, J. 1995. Hantverksproduktion, sam- rapport redigerad och med en inledning av hällsförändringar och bebyggelse. Produksjon Rune Edberg. Sigtuna. og samfunn. Beretning fra 2. nordiske jernal- Arrhenius, B., Linder Welin, U. & Tapper, L. dersymposium på Granavolden 7.–10. mai 1973. Arabiskt silver och nordiska vikinga - 1992. Varia 30. Oslo. smycken. Tor 1972–1973. Uppsala. Conophagos, C. 1980. Lelaurium antique et la Bayley, J. 1979. Examination and analysis of so- technique Grecque de la production de me crucibles from Birka. Ancient monuments l’argent. Aten. Laboratory Report 3458. London. Curle, C. L. 1982. Pictish and Norse finds from Bayley, J. 1992. Anglo-Scandinavian Non-Fer- the Brough of Birsay 1934–74. In Murray, rous Metalworking from 16–22 Coppergate. H. K. & Shepherd, A. N. (eds.). Society of In Addyman, P. V. (ed): The Archaeology of Antiquaries of Scotland Monograph Series 1. York 17/7. London. Aberdeen.

30 ANDERS SÖDERBERG

Dasent, G. W. 1861. The Story of Burnt Njal. Gustin, I. 1997. Islam, Merchant or King – Who Edinburgh. was behind the making of the Viking Age Drescher, H. 1983. Metallhandwerk des 8.-11. weights? In Andersson, H. et al.(eds.): Visions Jh. in Haithabu auf grund der Werkstattabfäl- of the past. Trends and Traditions in Swedish le. In Jahnkuhn, H. (ed.): Das Handwerk in Medieval Archaeology. Stockholm/Lund. vor- und frühgeschichtlicher Zeit. Teil II, Gustin, I. 2004. Mellan gåva och marknad. Archäologische und philologische Beiträge. Handel, tillit och materiell kultur under Göttingen vikingatid. Malmö. Duczko, W. 1985. Birka V. The filigree and gra- Hawthorne, J. G. & Smith, C. S. 1979. Theophi- nulation work of the Viking Period. An analy- lus; On Divers Arts. The Foremost Medieval sis of the material from Björkö. Stockholm. Treatise on Painting, Glassmaking and Metal - Elfver, F. 2007. General debasement during the work. New York. reign of Svend Estridsen? The coinage of Hed Jakobsson, A. 2003. Smältdeglars härskare Lund c. 1035–1050 and some metal analyses. och Jerusalems tillskyndare. Berättelser om In Fransson, U. et al (eds.): Cultural Interac- vikingatid och tidig medeltid. Stockholm. tion between East and West: archaeology, von Heijne, C. 2004. Särpräglat: vikingatida artefacts and human contacts in northern Eu- och tidigmedeltida myntfynd från Danmark, rope. Stockholm. Skåne, Blekinge och Halland (ca 800–1130). Eniosova, N. & Mitoyan, R. 2005. Metal-mel- Stockholm. ting crucibles from Medieval Russian Towns. Herschend, F. 1989. Vikings following In Kars, H. & Burke, E. (eds.): Proceedings of Gresham’s Law. In Larsson, T. B. & Lund- the 33rd International Symposium on mark, H. (eds.): Approaches to Swedish Pre- Archaeometry, 22–26 April 2002, Amsterdam. history. A spectrum of problems and perspec- Amsterdam. tives in contemporary research. Oxford. van Es, W. A. & Verwers, W. J. H. 1980. Exca- Holmquist Olausson, L. 1993. Aspects on Birka. vations at Dorestad 1. The Harbour: Investigations and surveys 1976–1989. Hoogstraat I. Amersfoort. Stockholm. Floderus, E. 1928. Några brons- och silver - Holmquist Olausson, L. & Fennö Muyingo, H. smedfynd från det äldsta Sigtuna. Fornvännen 1995. Rapport från en arkeologisk undersök- 23. Stockholm. ning vid Birkas stadsvall 1995. Björkö, Adel- Foote, P. 1961. Heimskringla : Sagas of the sö sn. Raä 118 Uppland. AFL, Stockholm Norse Kings / by Snorri Sturluson; translated University. Stockholm. by Samuel Laing 1844; rev. with intro. and no- Hoover, H.C. & Hoover, L. H. 1950. Georgius tes by Peter Foote. London. Agricola: De Re Metallica, translated by Graham-Campbell, J. 2007. Reflections on ‘Sil- Herbert Clark Hoover and Lou Henry Hoover. ver Economy in the Viking Age’. In Graham- New York. Campbell, J. Williams, G. (eds.): Silver Eco- Hulthén, B. 1995. Ceramics and Clays at nomy in the Viking Age. Walnut Creek. Ancient Högom – A Ceramological Study. An- Gustafsson, Ny B. & Söderberg, A. 2007. En tikvariskt arkiv 77. Stockholm. senvikingatida silververkstad i Fröjel. Got- Hulthén, B. 2006. Helgö, moulds and routes. ländskt arkiv 2007. Visby. Monographs on ceramics 3. Lund.

31 SITUNE DEI

Hårdh, B. 1976. Wikingerzeitliche Depotfunde Kraume, E. & Hatz, V. 1961. Die Otto-Adel- aus Südschweden. Probleme und analysen. heid-Pfennige und ihre Nachprägungen. In Lund. Häuernick, W. & Hatz, G. (eds.): Hamburger Hårdh, B. 1996. Silver in the Viking Age. A Beiträge zur Numismatik Band V, Heft 15–17, Regional-Economic Study. Stockholm. 1961–63. Hamburg. Jansson, I. 1991. År 970/971 och vikingatidens Kresten, P. & Larsson, L. 1996. Kemisk analys kronologi. In Iversen, M. (ed.): Mammen. av fynd 1910b från Kv Trädgårdsmästaren. Grav, kunst og samfund i vikingetid. Højbjerg. Analysrapport 16-1996. Riksantikvarieämbe- Jonsson, K. 1994. The coinage of Cnut. In tet, UV GAL. Uppsala Rumble, A. R. (ed.): The Reign of Cnut: King Kresten, P., Hjärtner-Holdar, E. & Stilborg, O. of England, Denmark and Norway. Leicester. 2000. Vendeltida metallurgi i Dagstorp, Jouttijärvi, A. & Andersen, L. M. 2005. Affald Skåne. Analysrapport 5-2000. Riksantikvarie - fra metalbearbejdning. In Iversen, M. et al. ämbetet, UV GAL. Uppsala. (eds.): Viborg Søndersø 1018 – 1030. Arkæo- Krongaard Kristensen, H. 1990. Spor efter logi og naturvidenskab i et værkstedsområde guldsmide fra vikingetiden i Viborg. Kuml fra vikingetid. Viborg. 1988–89. Kangur, M. 2004. Metallhantverket på Gåse- Lagerqvist, L. & Nathorst-Böös, E. 1981. En borgs fornborg, Järfälla sn i Uppland – en vikt från Vikingatiden. Fornvännen 76. studie av metallurgisk keramik. CD-uppsats i Lamm, K. 1969. Svensk konstindustri för 1500 laborativ arkeologi 2004. AFL, Stockholm år sedan. In Holmqvist, W. & Granath, K. E. University. Stockholm. (eds.): Helgö, den gåtfulla ön. Stockholm. Kilger, C. 2006. Silver handling traditions Lamm, K. 2008. Cupels and crucibles from buil- during the Viking age – some observations ding group 3, with a contrubution by Justine and thoughts on the phenomenon of pecking Bayley. In Clarke, H. & Lamm, K. (eds.): and bending. In Cook, B. & Williams, G. Excavations at Helgö XVII, Workshop part III. (eds.): Coinage and History in the North Sea Stockholm. World c. 500–1250. Leiden. Ljung, C. & Wikström, A. 2008. Stadsgårdar, Kilger, C. 2008. Wholeness and Holiness: vägar, hustyper och funktion. In Wikström, A. Counting, Weighing and Valuing Silver in the (ed.): På väg mot Paradiset – arkeologisk un- Early Viking Period. In Skre, D. (ed.): Means dersökning i kvarteret Humlegården 3 i Sigtu- of Exchange. Dealing with silver in the Viking na 2006. Sigtuna. Age. Kaupang Excavation Project Publication Lundström, L. 1973. Silverringar som pengar Series, vol. 2. Aarhus. under vikingatiden. In Rasmusson, N. L., Ol- Kilger, C. (in press). Hacksilver, weights and rog, F. & Lagerqvist, L. O. (eds.): Numisma- coinage: The Anglo-Scandinavian bullion coi- tiska meddelanden XXXI. Stockholm. nages and their use in Late Viking Age Socie- Malmer, B. 2010. Den svenska mynthistorien. ty. In Proceedings from symposium in Aarhus, Víkingatiden ca 995–1030. Stockholm. december 5–6, 2008: Silver Economies, Mo- Malmer, B., Ros, J. & Tesch, S. 1991. Kung netisation & Society in Scandinavia, 800– Olofs mynthus i kvarteret Urmakaren, Sigtu- 1100. na. Sigtuna.

32 ANDERS SÖDERBERG

Martinon-Torres, M. & Rehren, T. 2005. Alche- Spufford, P. 1988. Money and its use in medie- my, chemistry and metallurgy in Renaissance val Europe. Cambridge. Europe: a wider context for fire-assay re- Stahl, A. M. 2000. Zecca. The mint of Venice in mains. Historical Metallurgy – Journal of the the Middle Ages. Baltimore. Historical Metallurgy Society. Vol 39 No 1 Stenberger, M. 1958. Die Schatzfunde Gotlands (2005). der Wikingerzeit. 1, Text. Stockholm. McLees, C. 1996. Itinerant craftsmen, perma- Steuer, H. 1984. Feinwaagen und Gewichte als nent smithies and the archbishop’s mint: the Quellen zur Handelsgeschichte des Ostsee- character and context of metalworking in raumes. In Jankuhn, H. et al. (eds.): Archäolo- medieval Trondheim. Historical Metallurgy – gische und naturwissenschaftliche Unter- Journal of the Historical Metallurgy Society. suchungen an Siedlungen im Deutschen Vol 30 No 2 (1996). Küstengebiet. Vol 2, Handelsplätze des frühen Moore, D. T. & Oddy, W.A. 1985. Touchstones: und hohen Mittelalters. Bonn. some aspects of their nomenclature, petro- Svanberg, F. 1998. Exclusive Jewellery, Borge- graphy and provenance. Journal of Archaeo- by and Western Scania c. AD 950–1050. logical Science 1985, 12. Forn vännen 93. Naumann H. P. 1987. Warenpreise und Werte- Symons, D. J. 2006. The moneyers of the Wor- verhaltnisse im alten Norden. In Düwel, K. chester mint, 1066–1158: some thoughts and (ed.): Der Handel der Karolinger- und comments. In Cook, B. & Williams, G (eds.): Wikingerzeit. Untersuchungen zu Handel und Coinage and History in the North Sea World Verkehr der vor- und frühgeschichtlichen Zeit c. 500–1250. Leiden. in Mittel- und Nordeuropa. Vol 4. Göttingen. Söderberg, A. & Holmquist Olausson, L. 1996. Pedersen, U. 2010. I smeltedigelen. Finsmedene Rapport från en arkeologisk undersökning vid i vikingetidsbyen Kaupang. Avhandling for Birkas stadsvall – kompletterande undersök- Ph.D-graden. Universitetet i Oslo. Oslo. ning 1996. Björkö, Adelsö sn. Raä 118 Upp- Rehren, T. & Kraus, K., 1999. Cupel and Cru- land. AFL, Stockholm. cible: The Refining of debased silver in the Söderberg, A. 2004. Metallurgic ceramics as a Colonia Ulpia Traiana, Xanten. Journal of key to Viking Age workshop organisation. Roman Archaeology 1. (ed. J.H. Humphrey). Jonas 14 – Journal of nordic archeological Portsmouth. science. Stockholm. Roesdahl, E. 1977. Fyrkat – En Jysk Vikinge- Söderberg, A. & Gustafsson, Ny B. 2006. A borg. II. Oldsagerne og Gravpladsen. Køben- Viking Period silver workshop in Fröjel, Got- havn. land. Fornvännen 2006/1. Stockholm. Ros, J. 2009. Stad och gård. Sigtuna under sen Söderberg, A. & Gustafsson, Ny B. 2007. Från vikingatid och tidig medeltid. Uppsala. prestigevarugjutning till myntning, tidigme- SHM 8985:14 http://mis.historiska.se/mis/sok/ deltida metallurgi i kvarteret Trädgårdsmästa- fid.asp?fid=417812 ren, Sigtuna. Situne Dei 2007. SHM 7259 http://mis.historiska.se/mis/sok/ Söderberg, A. 2008. Metall- och glashantverk. bild.asp?uid=28669 In Wikström, A. (ed.): På väg mot Paradiset – Sperber, E. 1996. Balances, Weights and Weig- arkeologisk undersökning i kvarteret Humle- hing in Ancient and Early Medieval Sweden. gården 3 i Sigtuna 2006. Sigtuna. Stockholm.

33 SITUNE DEI

Tesch, S. 2007. Sigtuna – det maktpolitiska och Williams, G. 2007. Kingship, Christianity and sakrala stadsrummet under sen vikingatid och Coinage. Monetary and Political Perspectives tidig medeltid (ca 980 – 1200). In Perlinge, A. on Silver Economy in the Viking Age. In (ed.): Människors rum och människors möten. Graham-Campbell, J. & Williams, G. (eds.): Kulturhistoriska skisser. Stockholm. Silver Economy in the Viking Age. Walnut Thunmark, L. 1974. Stämplar på gotländskt vi- Creek. kingasilver. Gotländskt arkiv. Visby Wojnar-Johansson, M. 2005a. Analys av två Vifgusson, G. 1874. An Icelandic-English fragment av kupellationshärdväggar, RC6:3 Dictionary based on the ms. collections of the och 57/238. Arkeologiska Forskningslabora- late Richard Cleasby. Oxford. toriet, Uppdragsrapport 19. Stockholms Uni- Wadsten, T. 1979. Om proberkonsten på Hel- versitet. geandsholmen under 1600-talet. Fornvännen Wojnar-Johansson, M. 2005b. Analys av en de- 1979/3. Stockholm. gel och 2 fragment av gjutformar från Fröjel, Gotland, Äga Bottarve 1:19. Arkeologiska Forskningslaboratoriet, Uppdragsrapport 29. Stockholms Universitet.

34