Florin Curta

title.none: Hines, Nielsen and Siegmund, eds., The Pace of Change (Curta)

identifier.other: baj9928.0001.002 00.01.02

identifier.issn: 1096-746X

description.statementofresponsibility: Florin Curta, University of Florida,

publisher.none: .

date.issued: 2000

identifier.citation: Hines, John, Karen Nielsen and Frank Siegmund, eds.,. The Pace of Change: Studies in Early-Medieval Chronology. Cardiff Studies in Archaeology. Oxford: Oxbow Books, 1999. Pp. x, 194. £40.00. ISBN: 1-900-18878-3.

type.none: Review

relation.ispartof: The Medieval Review

The Medieval Review 00.01.02

Hines, John, Karen Nielsen and Frank Siegmund, eds.,. The Pace of Change: Studies in Early-Medieval Chronology. Cardiff Studies in Archaeology. Oxford: Oxbow Books, 1999. Pp. x, 194. £40.00. ISBN: 1-900-18878-3.

Reviewed by:

Florin Curta
University of Florida

This book has its origins in a symposium organized by the Danish Humanities Council at Fjordparken, Aalborg, in March 1996. The goal of this symposium was to bring together a number of archaeologists working in the early medieval period (fifth to eighth century A.D.) in western Germany, England, and Scandinavia. Their common ground was a revived interest in chronology, the sine qua non of medieval archaeology. After two or three decades of total neglect by archaeologists working within the processualist paradigm, chronological research is now growing as an important subfield of the discipline, taking advantage of two innovations of the New Archaeology era. The remarkable growth of material culture studies of the last few years is primarily based on quantitative archaeology, i.e., the use of statistical analysis for interpreting archaeological data, and, more recently, on dendrochronology. It is precisely progress along these two lines of development that explains the qualitative difference between this book and the previous collection of studies on the chronology of the Völkerwanderungszeit, edited by Georg Kossack and Joachim Reichstein and published more than twenty years ago ( Archäologische Beiträge zur Chronologie der Völkerwanderungszeit, Bonn, 1977). As John Hines shows in his introductory chapter (pp. vii-x), dendro-dates from early medieval urban sites, such as Ribe and London, "still await full integration with the artifact and burial sequences" (p. ix). In the meantime, the extraordinary potential of quantitative archaeology is fully demonstrated by studies gathered in this volume.

The basic assumption of most papers in this collection is that formulated almost two hundred years ago by the Danish archaeologist Christian Jürgensen Thomsen (1788-1865), namely that sets of artifacts found together in the same grave were buried at the same time and may thus be viewed as "closed finds." [[1]] Moreover, stylistically (or, as archaeologists would put it, "typologically") similar artifacts from various assemblages buried at the same time either within the same cemetery or on different sites, would form statistical clusters of association. By arranging these clusters in such a way that elements that are most alike will be placed close to each other, it would be possible to develop chronological sequences. The method used to obtain such sequences is called seriation. [[2]] By 1950, quantitative techniques were designed for artifact assemblage structuring, which focused on frequencies of types and increasingly relied on algorithms for ordering similarity matrices. The first computer programs for seriation were produced in the 1960s. The late 1960s and early 1970s witnessed the development of data analysis techniques and their first applications in archaeology. One of the most important data reduction techniques currently in use is the so-called correspondence analysis (CA), invented and developed by J. P. Benzécri and his team of the laboratory for Mathematical Statistics at the University of Paris VI. [[3]] Following the translation of Benzécri's book into English in 1984, the CA gradually made its appearance in Scandinavian and British archaeology. The great advantage of this technique is that the relationships between units (burials), between artifact-categories, and between artifact-categories and units may be all analyzed together and represented in the same scattergram or series of scattergrams produced by the plotting of pairs of orthogonal axes. CA works simultaneously as an analysis of relations between variables and as an analysis of relations between units. There is no need to have a Poisson distribution attached to the error structure as with previous log-linear analyses of contingency tables, nor is it necessary to know beforehand the structure of the phenomenon under study.

CA is the statistical technique behind all chronologies presented in The Pace of Change. Papers are divided into three sections on the basis of their regional focus: "the Continent" (by which the editors understand only western Germany), England, and Scandinavia, each followed by a synopsis of discussion. Elke Nieveler and Frank Siegmund open the section on Germany with a paper on the chronology of fifth- to eighth-century female and male burials in the Lower Rhine region (pp. 3-22), an excellent illustration for the degree of precision archaeologists can reach by using multivariate statistics to handle large sets of data. It is interesting to note that although making extensive use of CA, Nieveler and Siegmund relied heavily on what they call the "topo-chronological (chorological) analysis" of the Rübenach cemetery by Hermann Ament. [[4]] Ament was among the first to use a crude form of seriation in combination with the analysis of the distribution of chronologically relevant artifact-categories within a given cemetery. This analysis, also known as toposeriation, was based on the idea that interments in any cemetery started in a limited number of locations, with the earliest graves in the middle and the latest on the margins. Ament refined his Rübenach-based chronological system by applying toposeriation to other Merovingian cemeteries. Recently, extensive toposeriation was also applied to Franconian cemeteries (Iversheim, Jülich, and Lamersdorf), as well as to those of the Lower Rhine region (the large cemetery at Krefeld-Gellep, Cologne-Junkersdorf, Cologne-Müngersdorf, Duisburg-Walsum, and Düsseldorf-Stockum). The result of these incredibly detailed studies of an enormous amount of material is a regional chronology with eleven phases, in which Nieveler and Siegmund were able to define intervals of ten to forty years over a period of three centuries. More important, they included various classes of pottery (defined on the basis of both morphological attributes and decoration), which would no doubt enable correlations with settlement assemblages. In terms of absolute dates, this seems to confirm Max Martin's revised chronology of the transition between Rheinland phases 3 and 4 (somewhere between 520 and 530). [[5]]

Martin's revision is further substantiated by Claudia Theune in her paper on the chronology of Merovingian burial assemblages in southwest Germany (pp. 23-33). An updated version of an extensive, still irreplaceable study published more than ten years ago, [[6]] Theune's paper focuses, among others, on an until now neglected artifact-category: beads. Almost 8,000 beads from the Weingarten cemetery were classified and separately seriated. The result was a chronological sequence matching the five-phase chronological system developed for female burials from southwest Germany. An equivalent of Rheinland Phase 4, Theune's phase D, coin-dated to 527, is characterized by the sudden appearance of millefiori beads (p. 27). The presence of such beads in Merovingian Reihengräber was until now believed to be restricted to the second half of the sixth century, as a result of trade connections established with production centers in the Mediterranean area, especially in Italy, after the Justinianic conquest. [[7]] The revised chronology proposed by Theune seems to point to much earlier contacts, perhaps mediated by the Ostrogothic kingdom. As in the Lower Rhine region, the appearance of millefiori, reticella, and amethyst beads in female burials of southwest Germany was accompanied in phase E (ca. 550-ca. 570) by small decorative discs and pins with polyhedral heads. More interesting, beginning with phase H (ca. 610-ca. 650), bead sets typically include so-called Augenperlen, no doubt under Avar influence, as such beads were already popular in the Middle Danube region during the Early Avar period (late sixth-early seventh century). [[8]] Theune also attempts to match the chronology of female burials with that of male burials. If phases A and B (ca. 450-ca. 490) of the female series can clearly be distinguished, there are no specific male burials for this period, which is in sharp contrast with the situation in the Lower Rhine region. Theune's observation, if confirmed, would probably need an explanation, as the situation of male burial assemblages in southwest Germany during the second half of the fifth century seems to be unique on the Continent. On the other hand, the chronology of male burials seems to place several artifact-categories much earlier than it was believed until now. For example, copper-alloy buckles with triangular plates, sometimes punch-decorated, until now dated to the first half of the seventh century, appear already in phase E (ca. 550-ca. 570)(p. 31). [[9]] An interesting point of reference for a connection between male and female burial assemblages during the seventh century is the use of the animal style II for the decoration of artifacts associated with both (p. 32).

Male and female burials are also the object of Birte Brugmann's study of sixth-century assemblages in Kent (pp. 39-64). This paper opens the central section of the book dedicated to the chronology of British cemeteries. Lavishly illustrated (perhaps the best illustrated paper in the volume), Brugmann's study shows that the use of Continental artifacts in Kentish contexts is "unorthodox." Brooches appear in uncommon numbers and combinations. Unlike the situation in Germany, where they were worn only by men, belt sets, mostly buckles with a club-shaped tongue sometimes combined with kidney-shaped belt plates or shield-on-tongue buckles and shoe-shaped rivets, appear more often in female than in male burial assemblages (p. 38). Judging from the phasing of the Buckland cemetery, this coincided in time with Rheinland Phase 4 and Theune's phase D. In fact, the two best-known cemeteries in Kent (Buckland and Mill Hill) seem to have started during Rheinland Phase 3 (ca. 480/90-ca. 530). It is also interesting to note that the appearance of square-headed brooches so typical for Anglo-Saxon cemeteries [[10]] does not seem to have taken place earlier than on the Continent, an indirect indication of close contacts across the Channel. The use of Continental objects went out of fashion in Kent before buckles with plates became fashionable, i.e., before Rheinland Phase 6 (ca. 570-ca.580/90)(p. 48). One wonders why contacts with the Continent ended so abruptly in the last quarter of the sixth century, but Brugmann closes without any explanation. The disappearance of late sixth- or early seventh-century Continental brooches from Kentish burial assemblages is no doubt an interesting problem in itself, but the advised reader will also notice that the use of bow fibulae ceased on the Continent at about the same time. It is possible that the two phenomena were interrelated.

John Hines's study of sixth-century burial assemblages in Anglian England points to "a substantial change in material culture, very broadly sometime during the second half of the 6th century" (p. 76). This change took place at the same time in Anglian England, in Scandinavia, and on the Continent. On the basis of his analysis of female graves from various cemeteries in Cambridgeshire, Hines believes that in Anglian England, the transformation was "an abrupt watershed with no transitional phase" (p. 76). He attempts to extract from CA scattergrams a number of costume groups similar to those identified by Claudia Theune for southwest Germany, which may then be dated in relation to each other or to Continental parallels. He distinguishes four phases, labeled A, B, C, and D, but the division between the so-called Migration Period and the Final Phase in Anglian England is far from clear. The published evidence does not support Hines's interpretation of the scattergrams in figures 4.5 and 4.6 (pp. 72 and 74). It remains unclear how exactly a date around 590 was chosen for the transition from the Migration period to the Final Phase. In the light of what is known about the Exid Hill cemetery, namely that it has not been entirely excavated, the advised reader will hardly be persuaded that "with an earlier date than this (i.e., 590), we must infer that the rate of burial here, and thus by implication the size of the burying population (sic!), was lower in the Final Phase than earlier" (p. 66). In fact, there is no necessary or direct relation between the number of burials and the phasing and chronology of any given cemetery. Nor is it clear why Hines's interpretation of parabolas produced by his CA scattergrams is that "the chronological sequence from one end of the curve to the other is not one of early-to-late but rather early/late/early" (p. 68). Despite the use of sophisticated statistical methods, Hines seems to trust more the "established typological chronology, the validity of which is cautiously accepted here" (p. 68). The reader will no doubt wonder why, after all, was CA necessary for seriation. The lack of clear chronological markers is admittedly annoying. The scarce evidence that we have should, therefore, be treated with great caution. This is the case of the perforated coin found in burial 41 from the Kentish cemetery at Gilton, a Visigothic imitation of a coin struck for Justinian (527-565). [[11]] Because of his need to prove that the change from the Migration period to the Final Phase took place during the last quarter of the sixth century, Hines believes that the coin could not have been struck after 575. He does not find it "difficult to infer or justify a plausible date for burial circa 560-580 for this grave group, but even within that bracket we cannot be sure whether we have dated a late Migration-period grave, an early Final-phase grave, or the boundary line itself" (p. 76). The relation between the terminus post quem given by a deposited coin and the burial date is a most controversial issue. Unlike England, where before Sutton Hoo, there are no significant assemblages with coins, sixth- and seventh-century assemblages on the Continent produced a relatively large number of contemporary Byzantine coins. Because coins had various circulation rates before being deposited in graves, it is sometimes difficult to assess their chronological value. This is true, for instance, for Byzantine solidi used to date the transition from Early to Middle Avar assemblages in Hungary. [[12]] In the case of the Gilton coin, a pseudo-imperial issue of the so-called VPW type, [[13]] the problem is not so much the difficulty of assigning an exact date during (or after!) the reign of Justinian. Much more important is the fact that the coin was perforated and worn as a pendant. Whatever the date accepted for this pseudo-imperial tremissis, a perforated coin has no value for establishing any "plausible date for burial." No possibility exists at the moment to establish how long after being struck the coin was perforated, or how long after perforation, it was deposited in the Gilton burial assemblage. As a consequence, there is no evidence for Hines's proposed bracket (ca. 560-580). His conclusions about the chronology of the transition between the Migration period and the Final Phase should therefore be treated with suspicion.

The paper by Christopher Scull and Alex Bayliss examines a seventh- and eighth-century cemetery at St. Stephen's Lane/Buttermarket, Ipswich (pp. 80-88). The use of coin-dated burials and radiocarbon dating made it possible to check results obtained by toposeriation. The early seventh-century dates obtained for the earliest burials suggest that no simple sequence existed of the spatial development of the cemetery. Instead of an initial, single nucleus from which the cemetery gradually grew, the evidence is consistent with "a more complex or haphazard spatial development, perhaps resulting from a polyfocal structure" (p. 85). While warning against the uncritical extrapolation from this to other cases, Scull and Bayliss point to the danger of circular argument when "horizontal stratigraphy" is used for cemetery phasing. The Buttermarket cemetery in Ipswich will probably become a pilot project for a program of high-precision radiocarbon dating of early medieval burial assemblages in England. This may prove to be a crucial decision, as it is known that regional differences disappeared during the seventh and early eighth century.

The largest section is that on Scandinavia, with five papers on the chronology of burials, hoards, and bracteates. Siv Kristoffersen's (pp. 93-114) is an excellent survey of current research on the Migration period in Norway, with an interesting discussion of the animal style II. As Kristoffersen shows, the role of the Lombards in the development of this style is possibly of less significance than believed until now. Ribbon interlace appears much earlier in Scandinavia, as indicated by the wooden handle of the weaving batten from Kvåle (p. 94). On the basis of several significant artifact-categories (cruciform and small equal-armed brooches, as well as bucket-shaped pots), Kristoffersen proposes a new chronology with three phases, broadly corresponding to the Nydam style and to the animal style I. Bente Magnus takes a fresh look at an interesting assemblage found in 1845 at Hade, Gästrikland, in Sweden (pp. 115-125). The most intriguing feature of this assemblage is the presence of equal-armed brooches without fastening pins and catches on the back. Magnus argues persuasively that the Hade assemblage is a non-funerary sacrificial deposition in a border zone of emerging polities. Since the assemblage contains accessory sets of both female and male dress, it may represent a collection belonging to a group of people, rather than to one or two individuals (p. 123). Morten Axboe's paper on the chronology of Scandinavian gold bracteates (pp. 126-147) is an abbreviated version of the chapter in the coming, concluding volume of Hauck's Die Goldbrakteaten der Völkerwanderungszeit. It is also the only paper that applies CA not to burial assemblages, but to individual artifacts, thus treating shape and ornamentation as variables. Gold bracteates occasionally appear on the Continent and in England in association with artifacts with known date (p. 140 fig. 8.12). Axboe shows that the bracteate influx from Scandinavia was restricted to short period between ca. 530 and ca. 550. This indicates that the production of bracteates could not have ceased by 520/530, as recently proposed by Lars Jørgensen, who used this as a chronological marker for the transition from the Migration period to Vendel. It is more likely that this transition took place at some point during the second third of the sixth century. Anne Nørgård Jørgensen examines burials with weapons found in Scandinavia, particularly those from Bornholm and a few isolated graves in western Denmark (pp. 148-159). Unlike all other papers in this book, Jørgensen attempts to explain the results of her analysis and the chronology produced by CA. While emphasizing a sharp contrast between her phases I (ca. 520/530-ca. 560/570) and II (ca. 560/570-ca. 610/620), Jørgensen views the former as reflecting the radical changes in warfare taking place during and shortly after Clovis' reign. Indeed, the weapons found in Scandinavian burials of phase I have many parallels on the Continent (p. 158). She emphasizes the separate, somewhat later position of the English material in contrast to both Scandinavia and the Continent. Karen Høilund Nielsen's paper examines female burials from southern and eastern Scandinavia from the Migration period to Vendel (pp. 160-194). Nielsen makes the theoretically important remark that with CA and the underlying matrix, there is no need of diagnostic leading types (the so-called Leittypen), since no plot, no matter how good the parabola is, will ever display mutually exclusive phases (p. 163). On the basis of a few dress accessories (tortoise brooches, circular brooches with rim, and necklaces), Nielsen builds three separate chronologies for Bornholm, Gotland, and mainland Sweden, respectively. In Sweden, the CA plot discriminates between burial assemblages from Uppland and those from other areas, an indication that patterns identified by numerical analysis may not always represent time (p. 182). By combining all three chronologies, Nielsen eventually establishes a supra-regional chronology for a vast area in northern Europe.

Regrettably, the book abounds in misspellings (e.g., "Rhinish" for "Rhenish" on p. 4; "Theue" for Theune in the caption of fig. 2.3), grammatical errors ("has the potential to provided estimated date ranges," on p. 84; "the objects. . . , all belongs to the Early Germanic Iron Age," on p. 161; "the former model was constructed into relation to female accessories," on p. 162; "concentric circles are only found on combs associating Husby brooches," on p. 181), or simple typos (e.g., "some cemeteries seem to no longer to have grave goods," on p. 25; "club-shaped tonuge," on p. 29; "to the right are listed the the suggested new phases," on p. 190). Some papers were clearly translated by somebody else (for example, Nieveler and Siegmund's paper), others were written by scholars whose native language was not English. An attentive editorial pencil would have no doubt reduced the asperity of the original text and would have eased the access of the English-speaking reader to these very interesting papers. Occasionally, German words are also misspelled. For example, "Randdiche" for "Randdicke" is repeated no less than eight times on pages 170 and 171. There is no list of abbreviations. Those not familiar with German archaeological publications may have difficulties in understanding that "Bonner Jahrb" stands for Bonner Jahrbücher, while Studien zur Sachsenforschung is sometimes abbreviated (p. 22), sometimes not (p. 33).

The aim of the Fjordparken symposium was to refine the chronology of early medieval assemblages in western and northern Europe. One cannot, therefore, blame the editors of The Pace of Change for not setting this work in European context. Nonetheless, I believe many of the phenomena discussed have analogies in Central and Eastern Europe, and it would have been interesting to have had these explored as well, if only briefly. New publications point to a remarkable synchronism between various areas of the Continent [[14]] and it is worth stressing these material culture connections, as ideas of isolation and parochialism still dominate current views of the early Middle Ages. Many artifact-categories of the Rheinland Phase 4 appear not only in England and Scandinavia, but also in contemporary burial assemblages in Hungary and further to the east, in Crimea or Transcaucasia. To give but one example, burials 41 from the Bifrons cemetery, 105C from Mill Hill I, or D 3 from Finglesham, all in England, produced shoe-shaped rivets with good analogies in several cemeteries in Hungary (Berekhát, Kiszombor, Szöreg), as well as in a sixth-century cemetery in Bosnia. [[15]] Swords with damascened blades (such as that from Tamási) illustrate relations between communities in the Middle Danube and the Rhine regions, while the square-headed brooch with footplate bar from Szolnok-Szandaszöllös points to contacts with Scandinavia. The late sixth- and early seventh-century cemetery from Band, Transylvania, produced weapon burials with grave goods similar to those of Jørgensen's Phase II (square strap distributors, tongue-shaped strap ends, tweezers, wool shears, shield bosses with top spikes, and narrow saxes). Finally, a pair of Scandinavian brooches was found in a rich female burial at Gracanica, Kosovo. Similar brooches are known from Skodborgus, Denmark. [[16]] It is time to tackle the problem of this international fashion, as well as the more difficult one of how we should explain the sudden change in material culture of the early 500s, which is also visible in burial assemblages in Eastern Europe.

The Pace of Change shows the relationship between recent developments in quantitative archaeology and refined chronologies. We come to see that much of this extraordinary evidence presents itself in a new light to historians interested in material culture. The essays are not just statistical exercises, despite their noticeably limited agenda. There are no attempts at historical reconstruction or at explaining the change delineated through the minutiae of statistical analysis. Rather than simplifying the picture of sixth-century barbarian Europe, Hines, Nielsen, and Siegmund's collection makes it more complex and leaves us with more questions than answers--which is precisely what we expect from a book meant to encourage new scholarship.


1. For the concept of "closed find" and its role in modern archaeology, see Bo Gräslund, Relativ datering. Om kronologisk metod i nordisk arkeologi (Uppsala, 1974) and Bruce G. Trigger, A History of Archaeological Thought (Cambridge, 1989), pp. 76 and 78. The concept of seriation was further developed and refined by the Swedish archaeologist Gustav Oskar Montelius (1843-1921) and by the British archaeologist W. M. F. Petrie (1853-1942).

2. Andy Scott, "A parametric approach to seriation," in Computer Applications and Quantitative Methods in Archaeology CAA92, ed. by J. Andresen, M. Torsten, I. Scollar (Aarhus, 1993), pp. 317-24. See also Hartmut W. W. Tschauner, "La tipología: herramiento u obstáculo? La clasificación de artefactos en arqueología," Boletín de arqueología americana 12 (1985), 39-74. For a survey of quantitative archaeology, see François Djindjian, "Ordering and structuring in archaeology," in Mathematics and Information Science in Archaeology: A Flexible Framework, ed. by A. Voorrips (Bonn, 1990), pp. 79-92.

3. J.-P. Benzécri, L'analyse des données. II: L'analyse des correspondances (Paris, 1973). See also Erik Bølviken et al., "Correspondence analysis: An alternative to principal components," World Archaeology 14 (1982), 41-60; François Djindjian, "Seriation and toposeriation by correspondence analysis," PACT 11 (1985), 119-35; Torsten Madsen, ed., Multivariate Archaeology. Numerical Approaches in Scandinavian Archaeology (Aarhus, 1988), pp. 10-14; Torsten Madsen, "Seriation and multivariate statistics," in Computer Applications and Quantitative Methods in Archaeology 1989, ed. by S. Rahtz and J. Richards (Oxford, 1989), 205-14; T. J. Ringrose, "Bootstrapping and correspondence analysis in archaeology," Journal of Archaeological Science 19 (1992), 615-29. For a brief description of CA algebra, see Stephen Shennan, Quantifying Archaeology (Edinburgh, 1990), pp. 283-6.

4. C. Neuffer-Müller and H. Ament, Das fränkische Gräberfeld von Rübenach (Berlin, 1973). For toposeriation see Djindjian, "Seriation and toposeriation."

5. Max Martin, "Bemerkungen zur chronologischen Gliederung der frühen Merowingerzeit," Germania 67 (1989), 121-41.

6. Helmut Roth and Claudia Theune, SW I-V: Zur Chronologie merowingerzeitlicher Frauengräber in Südwestdeutschland (Stuttgart, 1988).

7. Ursula Koch, "Mediterrane und fränkische Glasperlen des 6. und 7. Jahrhunderts aus Finnland," in Studien zur vor- und frühgeschichtlichen Archäologie. Festschrift für Joachim Werner zum 65. Geburtstag, vol. 2 (Munich, 1974), pp. 497-503.

8. Eye-shaped beads appear in association with a tremissis struck for Justin II in burial no. 2 at Szentendre and with a solidus minted for Phocas in burial no. 116 at Jutas. See Eva Garam, "Die münzdatierten Gräber der Awarenzeit," in Awarenforschungen, ed. by F. Daim (Vienna, 1992), p. 151; Attila Kiss, Das awarenzeitliche gepidische Gräberfeld von Kölked-Feketekapu A (Innsbruck, 1996), p. 197.

9. Roksanda Swoboda, "Zu spätantiken Bronzeschnallen mit festem, dreieckigem Beschlag," Germania 64 (1986), 91-103; Attila Kiss, "Germanen im awarenzeitlichen Karpatenbecken," in Awarenforschungen, p. 56.

10. See, more recently, John Hines, A New Corpus of Anglo-Saxon Great Square-Headed Brooches (Woodbridge/Rochester, 1997).

11. B. Faussett, Inventorium Sepulchrale (London, 1856), p. 16 pl. 11/2. The coin is also reproduced on the cover of The Pace of Change.

12. See István Bóna, "Studien zum frühawarischen Reitergrab von Szegvár," Acta Archaeologica Academiae Scientiarum Hungaricae 32 (1980), 31-95; Csanád Bálint, "Probleme der archäologischen Forschung zur awarischen Landnahme," in Ausgewählte Probleme europäischer Landnahmen des Früh- und Hochmittelalters, ed. by M. Müller-Wille and R. Schneider, vol. 1 (Sigmaringen, 1993), 195-273.

13. See Philip Grierson and Mark Blackburn, Medieval European Coinage (Cambridge, 1986), pp. 48-9.

14. See, for example, Radu Harhoiu, Die frühe Völkerwanderungszeit in Rumänien (Bucharest, 1997).

15. Dezsö Csallány, Archäologische Denkmäler der Gepiden im Mitteldonaubecken (Budapest, 1961), pls. LXIII/3, 9; LXX/15; CXLVII/5; CXLVIII/9, 10; CL/5, 10; Nada Miletic, "Das frühmittelalterliche Gräberfeld in Rakovcani bei Prijedor, " Wissenschaftliche Mitteilungen des bosnisch-herzegovinischen Landesmuseums 5 (1975), pl. IV/33.

16. Tamási: István Bóna, "Neue Langobardenfunde in Ungarn," in Problemi seobe naroda u Karpatskoj kotlini, ed. by D. Dimitrijevic, J. Kovacevic, and Z. Vinski (Belgrade, 1978), p. 112. Szolnok-Szandaszöllös: Csallán y, Archäologische Denkmäler, pl. CXCII/2. Band: István Kovács, "A mezöbándi ásatások," Dolgozátok az Erdélyi Nemzéti Múzeum érem- és régiség tárából 4 (1913), 265-329. Gracanica: Zdenko Vinski, "Krstoliki nakit epohe seobe naroda u Jugoslaviji," Vjesnik Arheolo^Úkog Muzeja u Zagrebu 3 (1968), 106 and pls. II-III.