The Late Roman and Early Byzantine Solidi of the Stiernstedt Ancient Coin Collection

Journal of Archaeology and

Ancient History

2020 Number 28

Editor: Karl-Johan Lindholm

Editorial Board: Assyriology: Olof Pedersén. Archaeology: Anders Kaliff, Neil Price. Classical Archaeology and Ancient History: Gunnel Ekroth, Lars Karlsson. Egyptology: Andreas Dorn.

Editorial history, illustrations: www.arkeologi.uu.se/Journal/jaah_28 ISSN: 2001-1199 Published: 2020-12-15 at http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-426054

The Late Roman and Early Byzantine Solidi of the Stiernstedt Ancient Coin Collection

Svante Fischer¹

1 docent.s.fischer@gmail.com

Department of Archaeology and Ancient History Uppsala University, Sweden

ABSTRACT

Svante Fischer 2020. Th e Late Roman and Early Byzantine Solidi of the Stiernstedt Ancient Coin Collection.

Journal of Archaeology and Ancient History. 2020, No. 28 pp 1–26. http://urn.

kb.se/resolve?urn=urn:nbn:se:uu:diva-426054

Th is is a study of 33 Late Roman and Early Byzantine solidi from the period 394-565 that are kept in the Stiernstedt Ancient Coin Collection. Th e solidi were acquired in the late nineteenth century by the co-founding president of the Swedish Numismatic Society, August Wilhelm Stiernstedt. After his death, the solidi along with 2,434 other coins were published as a coherent assembly, the Stiernstedt Ancient Coin Collection (Heilborn 1882). Th e entire collec- tion was acquired at a sale from the Bukowski auction house in Stockholm by the Swedish-Texan antebellum cattle baron and gilded age banker Swante Magnus Swenson the same year. Together with many other coins and various prehistoric objects acquired in Sweden, the Stiernstedt Ancient Coin Collec- tion was donated by Swenson in 1891 to the State of Texas under the name of the Swenson Collection. Th e Stiernstedt Ancient Coin Collection is currently kept at the Dolph Briscoe Center for American History at the University of Texas at Austin. Th e study concludes with a catalogue of the 33 solidi. In the commentary, I have tried to identify and recontextualize the solidi by compar- ing them to recorded hoards from Scandinavia and the European Continent as well as unprovenanced solidi in Swedish and European collections.

KEYWORDS

Late Roman Empire; Scandinavian Migration Period; Roman Solidus;

Wilhelm August Stiernstedt; Coin collection; 19th century antiquarianism;

Scandinavian Archaeology; Late Roman and Early Byzantine Numismatics

Svante Fischer

The Late Roman and Early

Byzantine Solidi of the Stiernstedt Ancient Coin Collection

Introduction

Th is study concludes a research project fi nanced by a 2018 grant from the Berit Wallenberg Foundation. ¹ Th e purpose of this research project was to document and publish a number of solidi that were acquired in the late nineteenth

century by Baron August Wilhelm Stiernstedt (1812–1880) (see Fig 1, Table 1). Th e solidi were included in his Stiernstedt Ancient Coin Collection, which is currently kept at the Dolph Briscoe Center for American History at the University of Texas at Austin.

1 Th e research project has also been supported by the Gunnar Ekström Foundation for Numismatic Research, Th e Sven Svensson Numismatic Foundation, the Swedish Society for Ancient Monuments, the Dolph Briscoe Center and the Department of Classics of the University of Texas at Austin. I wish to thank D. Alex Walthall and Ingrid Edlund-Berry of the Department of Classics, and Stephanie P. Malmros of the Dolph Briscoe Center. I am also much indebted to the helpful staff of the Dolph Briscoe Center for their generous assistance and stoic patience.

Fig 1. August Wilhelm Stiernstedt. Photograph courtesy of Riksarkivet.

Table 1. Solidi in the Stiernstedt Ancient Coin Collection issued 394-565. Inv. #PageFig, PlateRIC X/MIBDate (from)Date (to)RulerMintmarkMintDiam. (max)Weight (g)Die-axis 2400-216536Fig 17a-b, VIX 506451455Valentinian IIICONOBConstantinople204,436 2400-218728Fig 6a-b, Pl IIX 1287402412HonoriusRVRavenna20,094,5112 2400-218828Fig 4a-b, Pl IX 1206394395HonoriusMDMilan21,354,4312 2400-218928Fig 5a-b, Pl IIX 1206394395HonoriusMDMilan21,014,3911 2400-220832Fig 11a-b, Pl IVX 2021426430Galla PlacidiaRVRavenna22,865,165 2400-220933Fig 13a-b, Pl IVX 2005425455Valentinian IIIRMRome21,94,396 2400-221033Fig 12a-b, Pl IVX 2018430440Valentinian IIIRVRavenna20,494,456 2400-221234Fig 14a-b, Pl VX 3711425?Valentinian IIIIMITATION, RVGaul?21,484,415 2400-221135Fig 15a-b, Pl VX 3711425?Valentinian IIIIMITATION, RVGaul?21,264,3612 2400-221335Fig 16a-b, Pl VX 287441447Valentinian IIICONOBConstantinople21,954,416 2400-221541Fig 25a-b, Pl VIIIX 2816468472AnthemiusCOMOBRome21,584,45 2400-221642Fig 26a-b, Pl IXX 3212474477Julius NeposRVRavenna19,74,456 2400-223327Fig 3a-b, Pl IX 1205394402ArcadiusMDMilan21,194,4412 2400-224731Fig 10a-b, Pl IIIX 292441447Thedosius IICONOBConstantinople21,984,46 2400-224829Fig 7a-b, Pl IIX 27402408Thedosius IICONOB, BetaConstantinople20,644,455 2400-224930Fig 8a-b, Pl IIIX 202408420Thedosius IICONOB, SigmaConstantinople20,954,427 2400-225030Fig 9a-b, Pl IIIX 257431434Thedosius IICONOB, GammaConstantinople20,794,286 2400-225537Fig 18a-b, Pl VIX 510451456MarcianCONOB, GammaConstantinople21,184,426 2400-225638Fig 19a-b, Pl VIX 510451456MarcianCONOBConstantinople19,54,46 2400-225839Fig 20a-b, Pl VIIX 605462466Leo ICONOB, BetaConstantinople20,074,466 2400-225939Fig 21a-b, Pl VIIX 605462466Leo ICONOB, GammaConstantinople20,394,355 2400-226039Fig 22a-b, Pl VIIX 605462466Leo ICONOB, ThetaConstantinople20,474,195 2400-226139Fig 23a-b, Pl VIIIX 605462466Leo ICONOB, HetaConstantinople204,456 2400-226240Fig 24a-b, Pl VIIIX 616457472Leo ITHSOBThessalonica21,024,366 2400-226443Fig 28a-b, Pl IXX 929474491ZenoCONOB, DeltaConstantinople20,244,486 2400-226544Fig 29a-b, Pl XX 930474491ZenoCONOB; EpsilonConstantinople20,514,446 2400-226643Fig 27a-b, Pl IXX 3627474491ZenoCOMOB, RVRavenna20,314,446 2400-226744Fig 30a-b, Pl XX 930474491ZenoCONOB, SigmaConstantinople20,254,487 2400-226844Fig 31a-b, Pl XX 929474491ZenoCONOB, IotaConstantinople20,284,446 2400-226944Fig 32a-b, Pl XIX 929474491ZenoCONOB, IotaConstantinople19,64,446 2400-227045Fig 33a-b, Pl XI–491?AnastasiusIMITATIONGotland?22,274,46 2400-227145Fig 34a-b, Pl XI–491?AnastasiusIMITATIONGaul?21,464,426 2400-228847Fig 35a-b, Pl XIIMIB 74535565JustinianCONOBConstantinople20,984,366 2400-228947Fig 36a-b, Pl XII–527?JustinianCOMOBItaly?194,286

In 1849, Stiernstedt sent a solidus to the Swedish National Board of Antiquities, Riksantikvarieämbetet (RAÄ). It was struck for Constantine II in Aquileia (see Fig 2, p 53, Plate I). The coin was found by the farmstead of Ella near the lake Klämmingen in Gåsinge parish, Södermanland in the fall of 1848 by farmhands during potato harvest. It was acquired by the Swedish History Museum (SHM) and given the inventory number SHM 1500. It remains a highly unusual find as most such coins are found in hoards. The rationale behind the research project was to verify a lingering suspicion that at least some of the solidi kept at the University of Texas had been found in Sweden, too.

This proved to be the case. The solidi must be regarded as decontextualized artifacts that have been deliberately stripped of a major part of their historical narrative. It must be emphasized that collecting coins in nineteenth century Sweden was an ambivalent process where people were aware that finds should be reported but coins were considered far more important than archaeological finds. This paper aims to restore context to the solidi. But in order return an archaeological value to the coins, the Stiernstedt Ancient Coin Collection cannot be understood in isolation but must be integrated in a wider analysis.

The solidi constitute unexplored new source material from the Scandinavian Migration Period. They have to be recontextualized in much larger studies. To do so, one must first publish an updated catalogue of the solidi.

An initial obstacle for the research project was the fact that this very important coin collection has never been studied in any detail until recently.

For a brief history of the collection from a Texan perspective, see King (1965), Kroll (1978); for a general numismatic description, see Castellano (2018a, 2018b). Only one of the solidi, coin 2400-2270, has been briefly mentioned in passim in the field of Late Roman and Early Byzantine numismatics. This was due to the fact that it belongs to a group of die-linked solidi from Etelhem Parish, Gotland, an integral part of the Scandinavian solidus horizon (Fagerlie 1967, 70, fn 84).

In the course of my examination of the solidi, I have compared them to the corpus of recorded Late Roman solidi in Scandinavia along with most of the published recorded hoards from Italy and the European Continent.

In particular, I have attempted to determine the find places of the solidi by comparing the distribution pattern of various solidus types. This endeavor was possible due to the considerable regional variation of recorded solidus finds in both Scandinavia and the European Continent. Thankfully, the recorded data from the main solidus contexts shows that various areas of Scandinavia and the European Continent were very different from each other. Certain types of solidi related to political events and military campaigns have circulated in separate networks and along specific importation routes during certain periods, resulting in very different solidus finds in various regions.

Stiernstedt and his Coin Collections – A Short Biography

Baron August Wilhelm Stiernstedt may be described as a man of consequence.

He was a state attorney and politician who served in the house of lords in the Swedish estates-general 1844–1866. In addition, he was a royal chamberlain 1849–1880, state herald of Sweden 1855–1880, and Swedish herald of arms in ordinary 1855–1861. He was a co-founder of the Swedish Numismatic Society (SNF) and served as its first president 1873–1880. Stiernstedt published a number of works on numismatics and heraldry, receiving a reward from the Royal Swedish Academy of Letters, History and Antiquities (KVHAA) for his scholarship on Swedish bronze coinage (Stiernstedt 1863–64). He married well and subsequently settled on Hesselbyholm, the estate of his wife’s family near Strängnäs. Despite this at least superficially impressive career, Stiernstedt was described by the secretary of the SNF, the prominent Swedish academy member and poet Carl Snoilsky (1841–1903), as a ruthless person who employed unethical methods to enlarge his personal collection. This probably had to do with how Stiernstedt was able to acquire Snoilsky’s own collection when the latter filed for divorce and moved abroad (Lagerqvist 2009, 465). In fact, Stiernstedt was able to assemble at least two major coin collections.

The first collection was posthumously published in 1881 and consisted of 3,356 Scandinavian coins (Stiernstedt 1881). The second was the collection of ancient coins. After Stiernstedt’s unexpected death of typhoid fever in Pojo, Finland in 1880, his estate was liquidated. The surviving relatives had the two collections auctioned off in two major sales in 1882 and 1884. The Scandinavian collection was sold in 1884 to the affluent collector Herman Frithof Antell of Finland after prolonged negotiations. It was eventually acquired by the National Museum of Finland (NMF) in 1893 after Antell’s death.

The second collection, the Stiernstedt Ancient Coin Collection of 2,467 Ancient, Greek, Roman and Byzantine coins, was published in 1882 as a rather terse French-language catalogue with an impressive title: Catalogue descriptif de la collection des monnaies antiques, grecques, romaines et byzantines de feu Baron A. W. Stjernstedt; avec une appendice ; monnaies des rois Ostrogoths et Vandales en Afrique. The catalogue was written by Otto Henric Heilborn (1856–1909) on behalf of the Bukowski auction house in Stockholm. It is uncertain but quite unlikely that Otto Heilborn added anything of scientific value to the catalogue. Heilborn eventually became a prominent Stockholm businessman and banker, but also a collector and amateur scholar. He had a successful career in banking, being elected to the executive board of Stockholms Enskilda Bank after its reconstruction in 1902. As was customary for the well-connected élite of the time, Heilborn joined the ranks of the corps consulaire, serving as consul general of Mexico to Sweden 1895–1909. Heilborn was also a member of the Swedish Society for Anthropology and Geography. In addition, he had a colonial business enterprise in German Cameroon in the 1890’s. Although it is

abundantly clear that Heilborn was a very erudite person, it seems reasonable to assume that he wrote the catalogue at the age of 26 in his function as a relatively junior financial manager of the Stiernstedt estate or the Bukowski auction house rather than as a professional numismatist.

Stiernstedt’s Ancient Coin Collection was probably far from complete when he died in 1880. Given the impressive size, it must be regarded an unfinished work in progress. It was an assemblage that consisted of many different collections and acquisitions. One possible source may have been the collection of Gustav Daniel de Lorichs (1785–1855), the Swedish chargé d’affaires in Spain 1814–1853. Some of Lorichs’ collection was acquired by the Royal Coin Cabinet in Stockholm (KMK) in 1863 and Stiernstedt may have been involved in this transaction (Castellano 2018b, 157). There are seven coins with inked inventory numbers belonging to another collection.

(Four of these have sequential numbers in two segments: 3093-3094 (Coin 2400-718 Trajan, and Coin 2400-749 Hadrian), and 3102-3103 (Coin 2400- 848 Sabina, and Coin 2400-874 Antoninus Pius). The remaining inventory numbers are: 3112 (Coin 2400-1114 Septimius Severus), 3116 (Coin 2400- 1223 Severus Alexander) and 3131 (Coin 2400-1896 Constantine I). As late as 1873, Stiernstedt sold three of the Roman coins to the Strängnäs coin collection (Hedlund 2012). Most likely, Stiernstedt hoped to use some of the collection to get hold of other more valuable coins in the future. Many of the solidi in the collection may have been kept by Stiernstedt as future bargaining chips. There was probably no reason for him to be sentimental and hold on to six worn and mutilated solidi in the name of Zeno from some non-reported Swedish hoard. This realization forces us to accept that what is preserved in the Stiernstedt collection today is not necessarily representative of what once passed through Stiernstedt’s hands. In addition, that Stiernstedt may have been able to procure solidi from the established collections of the KMK and the SHM, the Lund University Historical Museum (LUHM), and the Uppsala University Coin Cabinet (UUM) even if there are no records of such transactions. It is no secret that Jonas Hallenberg (1748–1834), the KVHAA secretary responsible for managing the KMK up to 1817 was by his own admission blatantly incompetent and lazy (Hallenberg 1804, Hildebrand 1971, 22). During his tenure at the KMK, Hallenberg refused to acquire all reported coin finds and left behind a disorganized mess that his successor Bror Emil Hildebrand (1804–1886) had to clean up for many years.

Stiernstedt was a knowledgeable collector in certain fields, but not in regard to Late Antiquity. There are errors in the auction catalogue of Heilborn (1882) than can possibly be attributed to Stiernstedt, notably the faulty attribution to Valentinian II (375-392) of a RIC X 506 struck in Constantinople under Marcian (450-457) in the name of Valentinian III (425-455). The three authorities employed for the classification are the French scholar Théodore Edmé Mionnet (1770-1842), author of Description des médailles antiques, grecques et romaines (1806–30, in 17 vols.), De la rareté et du prix des médailles romaines (1815; 3d ed. 1847), and the Danish scholars Christian Ramus

(1765-1832), author of Catalogus numorum veterum graecorum et latinorum Musei Regis Daniae (1816), and Christian Jürgensen Thomsen (1788–1865), see Brock (1866-1869). These were probably used by Stiernstedt and it is likely that the Heilborn 1882 catalogue simply copied the notes kept by the former. Despite not being an expert, Stiernstedt must have been quite aware as to which solidi were rare and more valuable than others in his own time.

Besides Ramus (1816), Mionnet (1806–1830, 1847) and Thomsen’s collection published by Brock (1866-1869), he probably used the works of Hallenberg (1804), Sabatier (1862), and Brock (1874) to ascertain what was at hand in Swedish and Danish collections. It also seems likely that Stiernstedt was in contact with a contemporary circle of Danish collectors and numismatic scholars which included Thomsen, Ludvig Müller (1809–1891), Christian Fredrik Herbst (1818–1911) and Peter Brock (1842–1906). They were both museum curators and private collectors (with all ethical complications that such double standards imply) and quite knowledgeable about Late Roman coinage. Several of them may well have assisted Stiernstedt in various ways although this can only be confirmed by further study beyond the scope of this paper.

In 1882, all coins listed in the Heilborn catalogue were acquired by Swedish-Texan antebellum entrepreneur and later gilded age philanthropist Swante Magnus Swenson (1816–1896) at the price of 4,500 Swedish Crowns.² Swenson and his business partner and maternal uncle Swante Palm (1815–

1897) were two affluent collectors with a clever investment strategy. They had left their home in Barkeryd Parish, Småland for America in search of wealth and opportunity. They were determined businessmen, but also self-taught intellectuals eager to diffuse learning to others (Edlund-Berry 2017). Upon arrival in the then independent Republic of Texas and the United States in the late 1830’s and early 1840’s respectively, they soon Americanized their names from Svante Jeansson and Svante Svensson to Swante Palm and Swante M.

Swenson.

When Palm and Swenson had acquired enough wealth from steamboat trade on the Mississippi and landholdings in Texas, they diversified their investment portfolios by buying books, coins and other antiquities at auctions in Sweden and shipping them to Texas where the value immediately increased upon arrival. Both wisely stayed out of national American politics during the Civil War 1861–1865, but they were nevertheless involved in both Austin and Texan state politics before and after the Civil War as elected politicians and philanthropists. Especially Palm was very active, he donated books and money to public libraries and supported Lutheran congregations financially. Despite changing their names, Palm and Swenson deliberately kept their Swedish citizenship; Palm was even being granted diplomatic status by the United

2 This information has been gathered from a contemporary cursive hand-written pencil note in a copy of the Heilborn 1882 catalogue kept in the Royal Library (KB) in Stockholm: “Såld till Amerika för 4500 kr. Köparens namn är Svensson.” (Sold to America for 4,500 Crowns. The buyer is named Svensson).

Kingdoms of Sweden and Norway in 1866, climbing his way into selected few of the corps consulaire. He was later decorated with the Vasa Order by King Oscar II in 1883 in recognition for continuous services to the native country.

Having transferred his main business interest from landholding in Texas to banking in New York City during the Civil War, Swenson donated the Stiernstedt Ancient Coin Collection along with other collections he had acquired in Sweden to the State of Texas in 1891. The idea was that the collections should be studied at the new University of Texas at Austin. The official acceptance of the Swenson donation to the State of Texas was signed June 27, 1891 by Texas governor James Stephen Hogg (1851–1906). Swenson’s donation was shipped from Sweden in 1892, a whole decade after the auction of the Stiernstedt Ancient Coin Collection. It finally arrived in Austin that same year. The Board of Regents of the University of Texas at Austin estimated the value of the Swenson donation to 75,000 US Dollars – no small net gain in value.

“Coins, not Finds” – Nineteenth Century Classical Education and Coin Collections

To understand how and why the Stiernstedt Ancient Coin Collection came into being, one needs to discuss the intellectual environment in which Stiernstedt existed. Despite the strong antiquarian tradition in Sweden, it is clear that coins, not finds, were considered most important even to many numismatists employed in museums. In the late nineteenth century, the study of Greek and Roman coins from the Mediterranean was associated with classical learning and education. The vast majority of these coins were decontextualized. Many towns and dioceses in Sweden and Finland that could pride themselves with museums, universities, junior colleges, seminars or senior high schools often had public coin collections. The roster includes Gothenburg, Jönköping, Helsinki (Helsingfors), Kalmar, Karlstad, Kristianstad, Lund, Örebro, Östersund, Porvoo (Borgå), Stockholm, Strängnäs, Sundsvall, Turku (Åbo), Umeå, Uppsala, Växjö and Visby (Hildebrand 1935, Hedlund 2012). In addition, there were often important libraries and book collections connected to the Greek and Roman coin collections. A case in point is Stiernstedt’s own numismatic library which was donated to the senior high school of Strängnäs along with a collection of 300 tokens. In many ways, Swenson’s acquisition of the Greek and Roman coins in the Stiernstedt Ancient Coin Collection and Palm’s collection of rare books in Latin should be seen in the light of this sentiment. The two Swedish-Texans simply wanted to reproduce in Austin what they knew of classical learning from Sweden, while also augmenting their status and self-value as generous philanthropists (Edlund-Berry 2017).

Some coin collections were associated with private societies such as masonic lodges and social clubs (Lilienberg 1910). There were also a number of private

Table 2. Coin collections with solidus finds from Sweden. CollectionFirst recordFinal transferCurrent LocationCountry Anders Bladh Collection 18261894NMF, HelsinkiFinland Borgholms fornminnesgård BorgholmSweden Borgå Gymnasium1853PorvooFinland GMVisbySweden Helsingborgs Museum1952HelsingborgSweden Hudiksvalls MuseumHudiksvallSweden Curle Gotlandic Collection1926BM, LondonUnited Kingdom KLM1871KalmarSweden LUHMLundSweden Stiernstedt Ancient Coin Collection 18821892UT Austin, TXUnited States Timmermansorden1910StockholmSweden UUM1857UppsalaSweden Visby Gymnasium GM, VisbySweden Franz Killig Collection1936KLM, KalmarSweden

collections, many of which were later donated to educational institutions or museums (Wiséhn 2009). Many of the coin collections and libraries were later subject to plunder. As a result, several book and coin collections have since been transferred to larger institutions. A case in point, the Strängnäs collections are now kept in the KB and the KMK in Stockholm (Lagerqvist 2009, Hedlund 2012, Andersson & Stenback 2018). Table 2 is a compilation of Swedish collections known to include solidi. The two county museums, Kalmar County Museum (KLM) and the Gotland County Museum (GM), and a municipal museum, Borgholms fornminnesgård, have solidi from Öland and Gotland. During her inventory 1958–1961, the American numismatist Joan M. Fagerlie (1967) visited both the KLM and GM, but the collection in the KLM has seen considerable growth since Fagerlie’s survey was published in 1967 (Herschend 1978, Westermark 1980, 1983, Kyhlberg 1986).

Today, there is roughly a half-dozen nineteenth century private coin collections that still include Swedish finds of Late Roman solidi. In addition, there are two state university coin cabinets in Lund and Uppsala that keep solidi found in Sweden in their collections. Most of the solidi in these collections were probably retrieved in the late eighteenth century and the first half of the nineteenth century, before the National Board of Antiquities (RAÄ) began to acquire all prehistoric gold finds found in Sweden. By the mid-nineteenth century, the Swedish government began a policy of transferring recorded archaeological finds to Stockholm and the SHM. Under the tenure of the very able Swedish state antiquarian Bror Emil Hildebrand, the RAÄ emerged as the main buyer of solidi offered for sale on the open market. The recorded finds of solidi in the early nineteenth century from Öland and Gotland suggest that solidi could be acquired, once finds first had been reported and offered for sale to the government. In the 1930’s, the otherwise strict centralist policy of the RAÄ changed slightly. County museums were suddenly allowed to keep precious metal hoards. This meant that the counties of Kalmar and Gotland with frequent finds of solidi were allowed to keep important sections of the accumulating find horizon. At least four of the collections include solidi that are likely to come from Gotland:

the Anders Bladh Collection, LUHM, UUM, and the Stiernstedt Ancient Coin Collection. Then, there is a Christian Fredrik Herbst Collection (SHM 5307) acquired by the SHM in 1873–74 for 8,000 Riksdaler Riksmynt. It included 39 Byzantine and four West Roman gold coins, including solidi from Bornholm (Horsnaes 2009, 2013). The Byzantine coins are described in Danish in the SHM acquisition catalogue as ”en god og smuk samling”

(a good and beautiful collection). These solidi now appear to be lost or at least unaccounted for. Three of the collections were created by two brothers – Professor Carl Säve (1812–1876) and Professor Pehr Arvid Säve (1812–

1887). They were capable historians, linguists, runologists, folklorists and ethnographers born and raised on Gotland, and both were members of the KVHAA. On Gotland, the Säve brothers are known to have donated solidi to the coin collections of Visby Gymnasium (where Pehr Arvid Säve taught)

and Gotlands Fornsal (the current GM), founded by Pehr Arvid Säve). On the mainland of Sweden, Carl Säve was the keeper of the UUM from 1857 to his death (Mäkeler and Berghaus 2001).

Some of the coin collections have been removed from Sweden. Today, two collections are kept in Finland (part of Sweden until 1809), one in Britain and one in the United States. The by far largest preserved collection is the Stiernstedt Ancient Coin Collection in Texas. As a rule, these collections were not included in Fagerlie’s 1967 survey of late Roman solidi in Denmark and Sweden, although she was clearly aware of at least one specimen in the Stiernstedt Ancient Coin Collection. While Fagerlie studied solidi kept at LUHM, some lost solidi have since been recovered (Westermark 1980, 1983;

Kyhlberg 1986). Below follows a brief presentation of some of the various collections, as they offer an insight in the antiquarian environment in which Stiernstedt operated.

Professor Anders Bladh (1748–1834) was born in then Swedish Finland, served as the chief physician of the city of Stockholm and was a member of both the KVHAA and the KVA (Wiséhn 2009). He was given a reward in 1794 by the KVHAA for his lectures on Oriental coins found in Sweden.

Bladh’s collection was donated to his old alma mater Åbo Universitet (currently the University of Turku, Finland). The collection included a gold bracteate and six gold coins, three of which were solidi. After Russia had conquered Finland from Sweden, the university was later transferred in 1830 to the new capital of Helsinki (Helsingfors). In 1893, the coins were transferred anew to the newly inaugurated NMF, where coincidentally a major collection of Scandinavian Viking Age and Medieval coins that had once belonged to Stiernstedt came to be the cornerstone of the numismatic collection. The gold coins were published by Talvio (1978). Among these, Talvio noted a solidus imitation die-linked to finds from Gotland. There is also a solidus for Anthemius that is die-linked to the San Mamiliano hoard in Italy (Arslan 2015, no. 427; Fischer 2019b).

The GM was founded in 1875 as Gotlands fornsal by that ubiquitous intellectual Pehr Arvid Säve, moving into its current buildings in 1880. The nearby Visby Gymnasium was where Pehr Arvid Säve taught for many decades.

Like his brother in Uppsala, Säve had very high intellectual standards. Proof of his indefatigable spirit, the scientific collections of the Visby Gymnasium were very precious. Eventually, the school was renamed Sävegymnasiet in his memory. It was subsequently subject to considerable plunder, not least by Ragnar Engeström (1946–2008), a senior state antiquarian of the RAÄ, who served a prison sentence for stealing and selling Pehr Arvid Säve’s rare 1687 first edition of Isaac Newton’s Philosophiae Naturalis Principia Mathematica at Sotheby’s in London (Gardell & Simander 2015). Today, the combined collections from Visby kept at GM include 17 solidi.

James Curle of Melrose (1862–1944) was a British collector who traveled to Gotland on at least two occasions in 1895–96 and 1901–1902 in order to acquire and decontextualize antiquities, armed with substantial sums of money and a self-serving interpretation of Swedish law. Curle had an unorthodox

way of treating Swedish antiquarian authorities, involving varying degrees of deceit, bribery, and chauvinist self-flattery. Curle’s Gotlandic collection was acquired by the British Museum (BM) in 1921 and given the inventory number 1922,0523 the following year (Kidd & Thunmark-Nylén 1990). Some 12 solidi are known to have been in Curle’s possession after his plunder on Gotland. Yet only four solidi can be identified in the BM today. It is difficult to estimate how many solidi have been stolen from Gotland over the years by the likes of Curle and Engeström.

The diocese of Kalmar was a very learned environment in the late nineteenth century. The origins for the KLM can be traced back to 1871, when the

association for ancient monuments in Kalmar County moved into Kalmar castle. From 1927, the museum received state funding. Under the leadership of Manne Hofrén, the museum began to keep finds of solidi from Öland. There are currently some 50 solidi in KLM. The denarius hoard of Hulterstad was kept in the coin collection of Kalmar Läroverk in the nineteenth century before being transferred to the KMK. The bronze coins from the latter institution are currently kept in UUM.

The LUHM is one of the oldest museums in Sweden. It is difficult to discern the exact origin of many of the solidi kept at the LUHM. Still it is certain that parts of the Kaggeholm hoard from Uppland in central Sweden discovered in 1783 were kept in Lund already in the nineteenth century, as in the case of Borgå Lyceum in Finland (Talvio 1975). Six unprovenanced solidi in the LUHM are very likely to be Swedish finds. They would not have looked out of place in the now lost hoard of Gyllerup in Scania but could equally well come from Uppland or the Baltic islands. This brings the total of the LUHM solidus inventory to 14, eight of which were listed by Fagerlie (1967).

Timmermansorden is a Stockholm masonic order founded in 1761. Most of the coin collection in the Timmermansorden was donated in the nineteenth century by the royal chamberlain C J Roos af Hjelmsäter, an affluent noble landowner from Västergötland. The coins are likely to have a Swedish origin.

A 1910 inventory lists 3049 coins and medals (Lilienberg 1910). Today, there are two solidi in the collection, which are not listed in Lilienberg’s inventory, one is a Concordia issue for Arcadius (Fischer 2019c), the other is a RIC X 630 for Leo I with an assay mark. These were not included in the study of Fagerlie (1967), presumably because she was unfamiliar with the collection.

Uppsala University is the oldest university of Scandinavia, founded in 1477.

The earliest written record of an archaeological find coin by scholars of Uppsala University dates to 1686. The coin cabinet UUM was probably created in the eighteenth century at the bequest of Queen Lovisa Ulrika (1720-1782).

In 1857, the newly appointed director of the coin cabinet Carl Säve listed 21 Byzantine gold coins. Among the 13 solidi currently kept in the Uppsala University Coin Cabinet are 11 coins from the period 395–565, several which are pierced and assay-marked specimens typical of Gotland. For some reason, Fagerlie (1967) never included these in her study although she must have been aware of them and the fact that Carl Säve was in charge of the UUM.

The Solidi in the Stiernstedt Ancient Coin Collection.

In total there are 33 solidi from the period 394–565, see Table 1. Some 12 imperial personae and five official imperial mints are represented, along with five separate pseudo-imperial mints. Among the latter, two sport the legend RV, two CONOB and one COMOB. It would have sufficed for Stiernstedt just to buy two small solidus hoards, one from Öland and one from Gotland in order to have all the solidi in the catalogue. On Öland, the Hjärpestad hoard (Fagerlie 1967, hoard no. 85-86) with 15 solidi provides almost all the early types. On Gotland, the Smiss hoard (Fagerlie 1967, hoard no. 122) with 25 solidi would provide the rest. There are a number of surprises in the composition of the collection.

1. There are no solidi for Majorian or Libius Severus. That is, there are no western solidi from the period 456–467. This can be explained by the absence of such coins on the market, though, see tables 3-4.

2. There are no solidi of the type RIC X 630 for Leo I, only RIC X 605 and RIC X 617. This means that all Leo I in the collection date from the period 457–466.

3. There are no solidi from the first reign or early second reign for Zeno from Constantinople. All five solidi postdate 477. The eastern solidi for Zeno are probably of a much later date than 477 given that I have compared them to the Esquiline hoard (tpq 480). It would seem that all solidi for Zeno are probably from the late 480’s.

4. There are no Constantinopolitan or Italian solidi for either Anastasius I or Justin I. That is, there are no genuine or Ostrogothic solidi from the period 491-527.

There is no overlap between several rulers in the collection. This is remarkable as both the period 466–476 and the later period 491–527 are unusually productive. In fact, the solidus output of the years 462–476 constitutes the apex of the entire fifth century. This means that substantial parts of the most frequent solidus types are missing in the collection. It cannot be excluded that Stiernstedt could have gotten hold of separate and unrelated solidus hoards, or that he had rid himself of overlapping solidi before he died, although this latter scenario seems unlikely. The simplest explanation is that he acquired most of the solidi in Sweden and that he was happy to keep as many as possible until he could get something with a better market value for them. In this case it seems most likely that the early solidi for Leo I are from Öland and the late issues for Zeno, the two imitations for Anastasius I and the two solidi for Justinian I are from Gotland. Tables 3-4 also show that it would have been difficult for Stiernstedt to acquire many of the missing types as most such finds were reported and acquired by the Swedish government.

The Stiernstedt Ancient Coin Collection – Current State of Preservation

The preserved correspondence between Swenson and the Board of Regents of the University of Texas is kept in a registrar’s file at the Dolph Briscoe Center.

Letters exchanged between Swenson in New York and the Board of Regents of the University of Texas in Austin show that the Stiernstedt Ancient Coin Collection arrived in Austin in 1892. It seems that the collection was perhaps not as appreciated by studious Texans as Swenson had once hoped (King 1965, Kroll 1978, Castellano 2018a, 2018b). From 1902 to 1926, the coins were stored in a safe to which the key had been lost, effectively safeguarded from both study and plunder. After 1926, the coins were stored in wooden cupboards with 200 slotted drawers. An inventory was made in 1936. From 1937 until 2002, the Stiernstedt Ancient Coin Collection was kept as part of the larger Swenson Collection in many different places on the Austin campus – the Main Building, the Barker Center, the History Department, the basement of the Academic Center, and then at the Texas Memorial Museum (TMM).

Preserved records show that a second inventory of the coins was conducted at the TMM in 1975–1978 using the numbers of the Heilborn 1882 catalogue, with the addition of a new prefix number 2400. The following prefix number 2401 was reserved for some 1,915 eighteenth century medallions belonging to the larger Swenson Collection. These medallions seem to be unrelated to the Stiernstedt Ancient Coin Collection but were probably acquired in Sweden at the same time in the early 1880’s.

The 1975–78 inventory was certainly not carried out by professional numismatists. The three persons conducting the inventory listed all coins in each numbered drawer in an annotated copy of the Heilborn 1882 catalogue.

As a result, many coins were misplaced under incorrect inventory numbers.

This was partially covered up by those conducting the inventory as they must have become aware that they were gradually creating more confusion as they went along. They hence construed several new inventory numbers with letter suffixes for coins that did not fit anywhere else (e.g. 2400-1914a) and then moved on to the next drawer. These erroneous additions obviously caused further chaos down the line but luckily not beyond each drawer, because coins that had been filed under new inventory numbers were then reported as missing from the very same drawer.

The Stiernstedt Ancient Coin Collection was transferred with the rest of the Swenson Collection to the Dolph Briscoe Center in 2002. The Department of Classics initiated an inventory and new research on the coins of the Stiernstedt Ancient Coin Collection in the Swenson Digitization project 2016 (Castellano 2018a, 2018b). The first steps towards a joint Swedish-Texan research project involving Swedish numismatists and the Swedish Numismatic Society began in 2018. The following year, I could travel to Texas to begin a study of the solidi supported by a grant from the Berit Wallenberg Foundation. The preliminary

results of the research conducted at the Dolph Briscoe Center were presented at a lecture at the Department of Classics in Austin on November 1, 2019.

Scientific Method and Project Implementation

In October-November 2019, the main part of the research presented in this paper was conducted at the Dolph Briscoe Center. This research consisted of a three-stage process. Firstly, the coins and their individual file cards were compared to the registrar’s annotated copy of the Heilborn 1882 catalogue. I soon discovered that many of the new inventory numbers with letter suffixes created during the 1975–78 TMM inventory and the subsequent mixed-up inventory numbers of purportedly missing coins actually concerned the very same coins. In addition, I found that many coins had been misplaced under the wrong inventory numbers. As a rule, I was able to distinguish the misplaced coins from each other by comparing their reverse legends as indicated by the Heilborn 1882 catalogue. Coins 2400-2187, 2400-2188 and 2400-2189 (Honorius, 2 x Milan and 1 x Ravenna) could be distinguished from each other, as 2400-2187 has the legend MD in the reverse center. The bottom left seriff of the M is damaged. Thus, the Heilborn 1882 catalogue reads it as an no. Coin 2400-2188 has a regular legend MD, and coin 2400-2189 has RV (see Fig 4a-b, 5a-b. 6a-b). Coins 2400-2249 and 2400-2250 (Theodosius II) could be distinguished from each other as coin 2400-2249 is a CONCORDIA, whereas coin 2400-2250 is a VOT XXX (see Fig 8a-b, 9a-b). Coins 2400- 2255 and 2400-2256 (Marcian) could be distinguished from each other as coin 2400-2255 has the Greek letter Γ (Gamma, officina 3) at the end of the exergue, whereas coin 2400-2256 has no officina letter (see Fig 18a-b, 19a-b).

Coins 2400-2264 and 2400-2268 (Zeno) could be distinguished from each other by their officinae letters Δ (Delta, officina 4) and I (Iota, officina 10) as indicated by the Heilborn 1882 catalogue (see Fig 28a-b, 31a-b). In addition, there are errors in the Heilborn 1882 catalogue in regard to coin 2400-2266 (Zeno) where the mint mark is not IIV but RV as in Ravenna. Similarly, coin 2400-2269 (Zeno) is CONOB Iota, not just CONOB (see Fig 32a-b).

Coins 2400-2287 (a cast forgery excluded from the catalogue) and 2400-2288 (Justinian I) were misplaced but could be identified by the inventory card photographs. The solidi were then documented anew with digital equipment provided by the Department of Classics.

Secondly, the coins were determined according to the major classification systems for solidi from this period. These are the DOC (Grierson & Mays 1992), the MIB (Hahn 1973), the MIBE (Hahn & Metlich 2000), and the RIC X (Kent 1994). From an archaeological perspective, it must be emphasized that recorded finds should always take precedence over decontextualized coins regarding the empirical source value. A major problem with the classification systems is that they were not conceived by archaeologists but written by numismatists and collectors mainly working with decontextualized coins in

sales catalogues and museum collections. In the case of the DOC, the recorded finds from Scandinavia were deliberately ignored. When some coins were deemed unclassifiable by me, I asked a number of colleagues for comparative evidence.³ The coins were then inspected for other characteristics such as wear, piercings, assay marks, graffiti, and mutilations.

Thirdly, the actual dies were compared and matched to my database of some c. 7,000 individual solidi and 27,000 solidi in c. 120 hoards from Scandinavia, the European Continent and the Mediterranean. This helped identify die- linked coins, imitations and forgeries. After this initial three-stage process followed the analytical work where the solidi were attributed to specific periods that are particularly relevant for different areas of Scandinavia. Thereafter, the project was concluded by assigning the probable provenance of the various solidi following the distribution patterns of recorded solidus finds.

Wear

There are different levels of wear among the coins in the Stiernstedt Ancient Coin Collection. The general impression of the earlier solidi struck for Leo I in 457–466 is that they are all in good condition. This level of wear is typical of Öland but does not correspond to the finds from Gotland or Helgö where solidi of the same type are more worn and often pierced and/or mutilated.

Most of the coins for Zeno (a later emperor) in the Stiernstedt Ancient Coin Collection are less worn. These coins all date to Zeno’s second reign 476–491.

They are more typical of Gotland, as there are only a handful of eastern solidi for Zeno from Öland, all of which are very worn and most of which date to Zeno’s first reign in 474–476. In short, the general picture that Fagerlie (1967) discerned in regard to the correlation between wear and chronology in regard to specific solidus issues on Öland and Gotland is not contradicted but rather confirmed by the coins in the Stiernstedt Ancient Coin Collection.

Traces of Soil

A most surprising feature is that many of the solidi still have soil or clay on them – they look exactly like proper archaeological finds of solidi kept in Swedish museums. There are traces of soil not least in the obverse helmets and looped letters such as O, P, A etc. on the eastern issues for Leo I and Zeno.

There is even what looks like chip of white rock stuck in a solidus for Zeno.

The frequent presence of soil on the solidi allows for the conclusion that many of the coins did not circulate among collectors who would have cleaned them, but rather came into Stiernstedt’s possession soon after having been discovered.

3 I wish to thank Ermanno A. Arslan and Fernando López-Sánchez for responding to my queries regarding the imitations from Gaul (Coins 2400-2211, 2400-2212, 2400-2271) and Italy (Coin 2400-2289).

It would be very interesting to see if the soil could be traced to specific

locations in Scandinavia but such a research project is beyond the scope of this paper.

Mutilations

The term mutilation describes a peculiar form of piercing on eastern solidi, that of a punch mark straight through the face of the three-quarter profile cuirassed and helmeted emperor. This was obviously intentional, and the idea must have been to separate these coins from other issues. A current working hypothesis is that the process of organized mutilation was a way for the East Roman administration to deactivate coins that was meant to be recycled as new coinage in the image of the current emperor (Fischer 2019a). In this scenario, the coins would be mutilated by tax officials as soon as tax was collected, and the mutilated coins were to be dispatched back to the imperial mint. This would explain the absence of any larger quantities of such coins in the Mediterranean.

In many cases, the punch mark has subsequently been refilled with a gold pellet, suggesting that this was related to issues of weight standards. The most frequently mutilated imperial portraits are found on issues of Theodosius II, Leo I and Zeno, that is, the most common types of eastern solidus coinage circulating in the period 462–491.

The earliest mutilated issue in Scandinavia is a VOT XX for Theodosius II (issued c. 421), a type that is generally missing in the Italian material, but part of a payment clearly considered important to the early fifth century import network to Scandinavia (Fischer 2017, Fischer 2019c). Meanwhile, most of the material dates to the reign of Leo I, and is of the very common type RIC X 605, struck c. 462–466. The mutilated corpus material ends with two issues of Justinian I, both from Gotland, but these specimens have hardly anything in common with the earlier mutilated issues. It is quite rare for mutilated solidi to appear within the recorded hoards inside the Empire.

Fagerlie (1967, 145) argued that the mutilated solidi had received this treatment for political reasons somewhere in Barbaricum relatively close to the Empire, but prior to the arrival in Scandinavia given that there are two mutilated solidi in the Karsibor hoard in Pomerania (both struck for Theodosius II) and a few solidi in other European collections. This cannot be a correct general assumption given the current evidence. Once again, the important discovery of the San Mamiliano hoard complicates matters. It does contain at least one mutilated and refilled issue for Leo I and one for Basiliscus (Arslan 2015, no. 134 and no. 247). Thus, we can be certain that some form of organized mutilation of East Roman solidi struck for Leo I and Basiliscus did indeed take place inside the West Roman Empire, presumably in Italy during the turbulent years 475–477. The terminus ad quem of the San Mamiliano hoard is either 474 or 477, a solidus for Ariadne being the final coin (Arslan 2015). But the youngest pierced coin in the hoard was issued

in Constantinople for Basiliscus in 475-476. For this to be a Scandinavian piercing, the coin would first have to travel up to Scandinavia from

Constantinople, get pierced, and then be brought down to Italy and then be buried with 497 other solidi, just in time to be put aside with the Ariadne coin as the final coin – an extremely unlikely scenario. Rather, the more I look at the Mediterranean material, the more mutilations do I discover, the latest finds being in the Zeccone and Reggio-Emilia hoards. It could be argued that the mutilation process was a way for some key Continental actors to mark coinage that was about to leave their sphere of circulation, but it could simultaneously have served as a mark of quality for those about to return to Scandinavia with solidi that had been ear-marked for them. The current evidence appears inconclusive, as the key point is that the solidi were mutilated because they were about to leave the Empire for good. The Stiernstedt Ancient Coin Collection contains one mutilated solidus, coin 2400-2213 (Valentinian III).

Clippings

The clipping of a coin is a common procedure whereby excess metal outside the center of the flan or rim of the die is removed after the striking of the coin.

In periods of scarcity or relative isolation from main coin supplies, peripheral market actors often chose to clip coins further while keeping them in local circulation (Guest 2005). The Stiernstedt Ancient Coin Collection contains clipped solidi, notably coins 2400-2187 (Honorius), 2400-2216 (Julius Nepos) and 2400-2266 (Zeno). But there are also solidi that have not been clipped although they seem to have been very suited for this procedure.

Piercings

Piercings were made on solidi primarily to test gold content, not to transform the coins into pendants. Piercing a solidus does not mean that one reduces its weight, as one simply pushes the gold aside within the mass of the original flan rather than away from the coin. This can be demonstrated by looking at the pierced solidi in the Stiernstedt Ancient Coin Collection. In addition, piercings and mutilations could also be refilled with gold, increasing the weight of clipped coins. Piercings are relatively common in the Scandinavian solidus material, Fagerlie (1967, 137) listed 116 pierced solidi and 21 pierced and refilled solidi. That is to say that roughly 10% of the solidus material in Scandinavia is pierced. The regional differences are as usual quite striking.

Fagerlie listed 43 pierced solidi on Gotland, the same number on Öland, with only 18 on the Swedish mainland and a mere 11 on Bornholm. Out of the 21 pierced and refilled solidi, ten were from Gotland and six from Öland. Among the 33 solidi in the catalogue, there are two pierced solidi, coins 2400-2215

(Valentinian III) and 2400-2262 (Leo I), and one pierced and refilled solidus, coin 2400-2264 (Zeno).

Imitations

The idea to imitate a coin is only slightly younger than the idea to produce a coin. There are several imitations in the Stiernstedt Ancient Coin Collection.

Coins 2400-2211 and 2400-2212 are Gallo-Roman or Visigothic imitations in the name of Valentinian III. Both have traces of soil on them. The next two are imitations in the name of Anastasius I. Coin 2400-2271 is an early Merovingian Period imitation, probably from Gaul. Coin 2400-2270 is probably a Scandinavian imitation, possibly from Gotland. Coin 2400- 2289 is probably an Ostrogothic imitation in the name Justinian I. From an antiquarian perspective, it is clear that many of the imitations were correctly identified as such already in the nineteenth century. Thus, Stiernstedt must have been aware that he owned some imitations and may have deliberately sought to acquire imitations. In other cases, he mistakenly attributed the imitations to genuine imperial mints, presumably in good faith. The most important thing to remember, though, is that Stiernstedt’s choice of keeping certain solidus imitations may be indicative of their origin. The solidus imitations are not found everywhere. Rather, there is a distinct geographical distribution pattern that allows for some generalization.

Forgeries

Second to the imitation is the forgery. There are two kinds of forgeries in regard to nineteenth century coin collections. First, there are ancient forgeries. These may have been recognized as such by collectors in the late nineteenth century but nevertheless retained. Second, there are more recent forgeries meant to fool collectors. The latter forgeries were produced already during the Renaissance, typically pretending to be coins for extremely short-lived or even fictitious imperial personae. The Stiernstedt Ancient Coin Collection has a number of recent forgeries, but these are mostly for third century emperors and they are easy to identify (Castellano 2018a, 2018b). Among the coins I examined in the Stiernstedt Ancient Coin Collection during October-November 2019, there were a few forgeries. Only one coin had to be excluded from the catalogue below, though. Coin 2400-2287 (purportedly Justinian I, Constantinople) is a gold-plated forgery. It is cast, not struck and made of gilt bronze. It is much thicker and more orange than coins 2400-2288 and 2400-2289. It is clear that coin 2400-2288 has not been used to create a copy. A coherent study of all forgeries in the Stiernstedt Ancient Coin Collection is a considerable future challenge and beyond the scope of this paper.