barbaric tremissis in Spain and Southern France

Tomasini, Wallace John, 1926-
Numismatic Notes and Monographs
American Numismatic Society
New York
Worldcat Works




Open access edition funded by the National Endowment for the Humanities/Andrew W. Mellon Foundation Humanities Open Book Program.


Table of Contents




Until the issues of Tiberius II Constantinus (578–582) there were only two reverse types used for the tremissis throughout the Mediterranean world in the sixth century. The first, which is peculiar to Byzantine Imperial currency and also found in Ostrogothic coins, is the frontal Victory holding in one hand a wreath-crown and in the other a globus cruciger, hereafter designated by the initials "VGC". The second, which has become traditionally associated with the barbarian kingdoms in the West, is the old pagan striding Victoria holding in one hand outstretched before her a wreath-crown, while with the other she holds a palm branch against her shoulder. This will be designated by the initials "VPW". These types represent two distinct versions of an old imperial personification, the Victoria Augusti. 1

The pagan Victoria tremissis (VPW) has been unanimously attributed to western mints as a product of the various barbarian Kingdoms; Ostrogothic, Visigothic, Merovingian, Burgundian, and Vandalic, because of: (1) its crude "degenerate" style; (2) its provenance in the Western Empire; (3) its ancestral relationship to later barbarian issues bearing the names or monograms of barbarian kings; (4) the pagan nature of its reverse type; (5) the alteration of its legends and (6) the poorness of its standard weight. These coins are not Byzantine imitations as they often have been erroneously classified. They are distinct issues whose consistent reverse type implies a conscious policy to separate them from the actual Byzantine tremissis. The large quantities extant and the long unbroken history of these coins from the time of Anastasius to that of Justin II, as well as the relatively good quality of their gold and weight standardization, indicate a product determined by a government-directed numis- matic policy. If this observation is correct, we are confronted with several possible answers to the VPW attribution problem. These coins are either the product of one state or of different states. If the latter assumption is correct, they must have been originated by one state and copied by the others.

The extant VGC tremisses have been attributed for the most part to the Imperial mint at Constantinople and the Ostrogothic mints at Rome and Ravenna. 2 A number of others, of a decidedly cruder style, have been assigned to the mints of the Merovingians and the Burgundians. 3 While the anonymous western mints are responsible for striking both the VPW and the VGC types, the mints of the Eastern Empire mint the VGC exclusively. It is correct therefore to refer only to the western and barbarian issues of the VGC tremissis as Byzantine imitations. Are the diverse policies of eastern and western mints significant? Only an investigation of the antecedents of these tremisses and of the incidence of the VPW type in the East and the West may amplify and clarify the VPW attribution problem. Does the striking of a distinctive VPW in the West as well as an imitative VGC corroborate the observation that one western mint may be responsible for striking the VPW?

The standard type for the tremissis in the early fifth century in both eastern and western halves of the Empire is the predominant VGC type (See Chart I). 4 After the first quarter of the century, however, the cross encircled by a wreath, or the less common chrismon encircled by a wreath, becomes the prevalent type in the West. In the East, the VGC continues to dominate the issues except for those in the name of the empresses, on which the cross-in-wreath is used exclusively. The VPW is not recorded, to my knowledge, as ever having been used on a tremissis from Constantinople in the fifth century. This situation is equally true for the West with one exception, a tremissis of the usurper John (423–425). This is recorded by Cohen with the mint mark of RX, which he considers an error on the part of Banduri from whom he is taking this specimen. 5 But from Valentinian III to the end of the reign of Romulus Augustus, the western issues are distinguished from those of the East by the use of the cross-in-wreath type. This situation continues until the reign of Anastasius when the West begins to strike two types, a VGC in imitation of the Imperial issue, and a VPW which continues the separate and distinct character of the fifth century cross-in-wreath issue.

The predominant tremissis type of the fifth century in the East continues there as a mint tradition into the sixth century. In the West, however, both a VPW and a VGC have been substituted for the customary fifth century cross or chrismon-in-wreath type. Both issues in the West are very significant, since neither has appeared on a western tremissis since the early part of the fifth century. If the unique VPW tremissis of John is discounted, the appearance of a VPW on a tremissis is last seen on the small issues of Arcadius. The more recent appearance of the VGC was with the tremisses of Valentinian III, but as a type it was rarely used in the West. Its use is similarly limited to the small issues of Honorius and Arcadius. On the other hand, the VPW had been the favored type in the West in the late fourth century. Consequently, the VPW and the VGC are reintroduced on the Anastasius tremisses in the West. The VGC issue is the result of a fiscal policy which accedes to the Byzantine Emperor the theoretical sovereignty of both halves of the Empire, while the VPW assumes the position held formerly by the cross-inwreath type, the independent and separate nature of the West.

The cross-in-wreath has been dropped for a symbol of an older tradition. If the later investigation of type significance has any validity, it should not be surprising that in the fifth century both VGC and VPW types were rarely used in the West, or that the VGC in the East was never placed on a coin issued in the name of an empress. 6 The more appropriate non-militaristic cross-in-wreath without a legend is preferred for the tremisses in either of these categories. The Victory figure, either Christian (VGC) or pagan (VPW) with her traditional legend of VICTORIA AUGUSTORUM would have been inappropriate and too unrealistic for propaganda usage.

The fifth century situation, however, should not lead us to think that the VPW is a new type for the tremissis. Even the unique example of the usurper John would discredit this. If we trace back to the earliest tremissis issue (383–387), we find that Theodosius I issued this coin with both the VGC and the VPW types. As indicated in Chart II, the tremissis was minted exclusively with the VPW at Trier (383–395), Lyon (383–392) and Aquileia (383–388). 7 The mint at Milan (383–395) struck both types, and the mint at Constantinople (383–388) issued only the VGC. 8

The analysis of the first issues of the tremisses is not complete without considering the possible dependence upon the last issues of the 1½ gold scripulum piece (1.70 gms.), their immediate predecessor (See Chart II). This coin was struck with a variety of reverse types and legends under Constantine the Great, but after 337, its reverse types become less varied. By 364, two types dominate the series, the seated Victory writing the votive inscription on a shield, and the VPW. At the mints of Trier (367–378), Milan (378–383), Aquileia (378–383) and Rome (364–367) the VPW is used exclusively. It is used at Thessalonica in 364, and at Constantinople (364–367), but is replaced by the votive shield Victory type at Constantinople (367–388) and at Thessalonica (364–375). At Antioch (364–375) the votive Victory had always been used without exception.

In the East, the tremissis descendents of the 1½ scripulum are never issued as a VPW or as a Victory with votive shield, both types being rejected for the more desirable Christian VGC, or the equally Christian cross or chrismon-in-wreath. In the West, therefore, a more continuous tradition is maintained. At Trier the VPW tremissis series (383–395) replaces the VPW 1½ scripulum issues which ceased in 378. The same is true for Aquileia where the VPW tremissis of 383 replaces a VPW 1½ scripulum of 378–383. Milan continues the VPW tradition in its tremisses but also strikes the VGC type. Rome does not issue a 1½ scripulum and does not issue a tremissis until after 395, and then only of the VGC and cross or chrismon-in-wreath types.

Although the VPW reverse for the 1½ scripulum and the early tremisses is a consistent tradition towards the end of the fourth century in the West, extant specimens indicate that it is dropped around the turn of the century. It would seem that with the defeat of Eugenius and the accession of Honorius in the West (395), the VPW is replaced by the VGC. Only three VPW tremisses are recorded after 392, one for Arcadius from Trier, a mint which would seem statistically loyal to the VPW type; another for Eugenius from Trier; and the later issue of John from Ravenna (?). 9 We can assume then that the VPW tremissis type is either dropped or issued rarely, since the above three issues are either very rare or unique. But it would seem evident, that if issued at all it would be in the West. The VPW was never issued on a tremissis from an eastern mint, nor for that matter was it ever a popular type for other eastern issues in gold, silver or bronze (See Charts I and II).

The VPW Anastasius tremissis consequently is a coin issue for which an older reverse type has been revived after an absence of almost a century. The large number of Anastasius VPW tremisses extant could not indicate, on the basis of our study, simply a larger issue of a continuing mint product or tradition. The significance of this revival and its particularly non-eastern origin and character is a most important clue by which the minter of these coins may be revealed.

An analysis of the use of the VPW type throughout the Empire from the late fourth century to the sixth will substantiate a theory of eastern antipathy or indifference to the type. Such an analysis may also further assist attempts to find a more coeval prototype for the Anastasius tremisses than the late fourth century tremissis.

A study of Chart III will quickly show that the VPW is not to be found on any gold coin in the fifth century, East or West. This makes the gold tremissis of John unique. Its rare use on a semis in the late fourth is dropped when, from Valentinian II on, the Victory with the votive shield becomes the major semis type for the next century in both the East and the West. 10

It is on the siliqua and half-siliqua that the type is noted frequently in the late fourth century and on a few specimens of the fifth century (See Chart III). It is to be noted that in the late fourth century it is the only type used on the half-siliqua pieces which Pearce considered so rare that he doubted their forming part of the regular coinage. It may be of significance that the VPW type is used exclusively on this coin, which may have been struck only as a presentation piece for distribution among certain classes of the population on festive occasions. 11 The only issue in the West in the fifth century is again a singular coin of John; 12 while in the East, in the name of Eudocia, wife of Theodosius II, three examples are found, one bearing as might be expected a cross-in-wreath and the other two the chrismon-in-wreath. 13 Tolstoi also catalogued one of the cross-in-wreath type in the name of Pulcheria, wife of Marcian. 14

On the siliqua, the VPW enjoys a long usage, particularly at Trier and Aquileia, but is non-existent in the East with the exception of one coin in the name of Arcadius listed by Sabatier. 15 Three of the four examples from the West in the fifth century are from Rome. The three coins listed by Sabatier, one of Zeno and two of Anastasius, if of eastern origin would be unique, considering the eastern antipathy to the VPW type. These would seem therefore to be the products of western mints or of Rome in particular. 16 The VPW type on the siliqua is a continuing tradition and its use is even contemporary with the use of the type on the Anastasius tremissis. This is interesting to note since the siliqua is closest in size to the tremissis and maintains a similar position in the silver currency to that of the tremissis in gold. If the silver siliqua is not the prototype for our VPW tremissis, it did help to keep alive the type tradition. The siliqua legend, VICTORIA AVGGG, might also suggest further links with the VPW tremisses, since it will be found on barbarian tremisses in the sixth century. 17

A slightly different situation presents itself at first when the bronze issues are studied, although the pattern immediately reverts to type developments already noted in the gold and silver issues (See Chart IV). Throughout the Mediterranean the VPW type is not used on either the Æ 1 or Æ 2 series, but it is one of the types used on the Æ 3 during the reigns of Valens, Valentinian I and Gratian at every mint striking bronze. This immediately changes after the accession of Theodosius from which time the East unanimously drops the VPW, never to use it in bronze again. In the West it passes naturally into the Æ 4 issues and as such is found on a coin of Romulus Augustus.

When all of the mints in the East do use this type, however, it is limited to one legend, SECURITAS REIPUBLICAE, and this consistency has all the aspects of a general imperial monetary edict (See Chart V). In the West, however, in all mints except Arles, it appears also with other legends, such as GLORIA ROMANORUM, VICTORIA AVGGG, VICTORIA DD NN AVG, and FELICITAS ROMANORUM. All such specimens, however, are rare. Trier seems to favor the type, for it is the only mint which issues it in combination with the legend GLORIA ROMANORUM at the same time as it strikes the standard GLORIA ROMANORUM issue being minted in the rest of the Empire. Only in the West is this VICTORIA AVGGG legend used on the Æ 3 issues of any type. It is this legend in combination with the VPW that is found on western issues of the Æ 4 struck at all mints down to Romulus Augustus. 18

There are limitations, therefore even when the East issues the VPW type. Pearce's analysis of the SECURITAS ROMANORUM issue in the East is most illuminating in this regard. Every strong emperor in accordance with the precedent set by Octavian Augustus maintained control of the "Aes coinage, issuing it by virtue of his 'tribunicia potestas' through the agency of the Senate." 19 Valentinian's chancellery controlled the bronze coinage and desired uniformity in order to symbolize the unity of the two parts of the Empire and the dominance of Rome. Since reverse types and legends were rigidly controlled by the chancellery as long as it was able to enforce them, the VPW appears on the Æ 3 issues throughout the Empire. 20 Pearce attributed to eastern passive resistance their reduced minting in the East after Gratian's accession (367) until that of Theodosius (379). That this type then became symbolic of Rome or at least identified with the Emperor in the West may be seen in the discontinuance of the type on the Æ 3 issues in the East in 379, under Theodosius, who by so doing may have subtly asserted eastern equality. 21 It is a coincidence in the light of later history of the type that this numismatic propagandizing occurs concomitantly with the settling of the first Visigoths within the Empire.

The significant striking of the VPW type (Æ 3) in the East with a degree of universal infrequency further supports the hypothesis that it is either unpopular or considered inappropriate in the East. Further evidence for this hypothesis can be found by recording the incidence of the VPW type from the Principate of Augustus to Flavius Victor (See Chart VI). It is only during the reign of Valentinian I that the VPW is issued frequently in the East, and as such it is unprecedented and unique.

Consequently, as in silver, the bronze issues with even greater frequency extend the life of one of the oldest pagan symbols of the Empire. They are not the only vehicles, however, which bridge the gap of the fifth century. There are three other items which bring our type into the world of Anastasius: the Invicta Roma bronzes of Theodoric, a gold medallion of Anastasius and a gold medallion of Theodoric. The Invicta Roma series attributed to Theodoric (ca. 493) may very well present the clue to the originator and first minter of the VPW Anastasius tremissis. These so-called quasi-autonomous bronzes of Rome are all non-Christian oriented in the selection of types and reveal the religious conservatism of the Roman Senatorial class. These coins also do not bear the name of any emperor, only the Invicta Roma legend and the bust of Roma on the obverse. When they are compared to those earlier bronzes bearing the name and head of Zeno on the obverse and a Victory on the reverse with the Invicta Roma legend, there is a difference in Victory types. The Zeno Roman pieces present an advancing wreath and trophy-bearing Victoria, which on the Theodoric issue has been replaced by a VPW standing on a prow before a lighted altar.

In the quasi-autonomous bronzes assigned to the reigns of Theodoric and Athalaric (494–534), distinctly Roman and pagan emblems are used for reverse types—an eagle, two eagles beneath and beside a fig tree and the wolf suckling the twins. The Roman Senate logically would hesitate to place Anastasius' name on their bronzes, since Theodoric, at the beginning of his reign, is not recognized by the emperor. The substitution of types on both obverse and reverse faces, and the nature of those types, in the time of Anastasius and Theodoric, may provide further clues in the solution of the problem of determining who is responsible for the Anastasius VPW tremissis. 22

Equally important are the two gold medallions. The VPW type has been recorded on bronze medallions from western mints during the late fourth century (See Chart IV), but its appearance on two unique and coeval gold medallions possibly may be significant. This type may have come to be regarded as appropriate to this denomination. The gold medallion of Anastasius, besides its VPW on the reverse, does have a star in the left field and a chrismon in the right field. Although there are sixth century tremisses extant with a star in the reverse field, there are none known to me with a chrismon. The legends on the medallion are similar to those found on the tremissis: DNANASTASIVS PP AVG; VICTORIA AVGVSTORVM. The similarity ends here, however, for the VPW type on the Anastasius medallion is of diverse tradition from that of the tremissis VPW. Here a VPW in three-quarter pose and profile head marches to the left in the same manner as seen on the silver siliqua of Arcadius. 23

The magnificent Theodoric three-solidi gold medallion that unquestionably was a commemorative issue or presentation piece by nature of its size, has considerable bearing upon our problem. It is an extremely fine and unique piece with a frontal bust of Theodoric on the obverse and on the reverse a Victory in girdled chiton standing right on a globe, holding a wreath in her hand and a palm branch on her left shoulder. The VPW design is similar to that found on some of the Anastasius tremisses. The globus nicephorus held by Theodoric is, in accordance with tradition, a VPW. 24 The Roman attribution of this by Wroth is reasonable permitting the selection of the design by Theodoric and any of his advisors, such as Cassiodorus. The type is as appropriate for Theodoric as it is for Rome and the Roman Senate. Speculation concerning the date of and reasons for the striking of these medallions must be left for later discussion.

Although these two medallions may have some meaningful association with regard to like use of type, they are naturally not of similar type style. Theodoric's VPW is in keeping with the most prevalent figuration of the type as found in the West from the fourth century through the sixth century. The form of Anastasius' VPW must be a less popular variation in accordance with extant examples. Its dissimilarities with the VPW of the Theodoric Medallion and the VPW of the western sixth century tremisses, suggest different prototype and tradition. It may be possible to suggest further that the singular use of this type on a coin struck in Constantinople, even though more expected on a medallion, indicates that it was somehow commemorative of an event or of persons involved in the international relations of East and West. Considering Theodoric's governmental policy, it might therefore be the justification as well as the model for his medallion.

Our study of type incidence and antecedents and possible prototypes for the Anastasius VPW tremissis has revealed several significant facts. The VPW type is more likely to be issued in the West than in the East. The VPW type is never found in the East and rarely found in the West in the fifth century. The nearest issue in time to the VPW tremissis is in the name of John (423–425). The use of the type can be traced longer in silver down to Zeno and Anastasius in what are most probably western issues. In bronze the use of the type does continue and is minted coevally with the gold coins in question. The eastern mints definitely substitute other types for the VPW used in the West. The only contemporary use of the type in gold besides the tremissis is on the two gold medallions. Finally silver and bronze issues of the type with the names of the Emperors Zeno and Anastasius are attributable to Italian mints.

We may conclude: (1) that the VPW Anastasius tremissis is an example of a new coin issue using a definitely revived reverse type, since its use on any coin and in any metal has been sporadic and rare for almost a century; (2) that the initiative for the striking of these coins is unquestionably conceded to the West; (3) that any of the western mints could have issued the coins if the workshops maintained a file of old die designs for reference; 25 (4) that the striking of a distinguishable non-imperial coin in the name of the emperor could suggest that the minter officially or publicly acknowledged his allegiance to Anastasius thereby indicating a treaty relationship; (5) that the minter is not compromised as is the minter in the East by the pagan connotations of the traditional VPW, which immediately calls to mind the long-dying pagan preferences of such cities as Rome and Arles; 26 (6) that discounting the religious issue the incidence patterns of the type might indicate a particular icono-graphical relationship between the VPW and the Western Empire or Rome understood by both the minter and the users of the coin; (7) that Theodoric may have some connection with the issue since he is already associated with a contemporary use of the type; and (8) that the western numismatic tradition responsible for the VPW Anastasius tremisses may have been already in operation with the striking of the VPW type in the names of late emperors such as Zeno and Anastasius, since the silver and bronze issues in their names seem unquestionably attributable to Italian mints.

End Notes

Vermeule, Victoria. This excellent article cannot be ignored in a study of the development of the representation of the personification Victoria. Mr. Vermeule's distinction between running and walking figures, however, has not been applied in this study. My major concern has been in the attributes which Victoria carries. In regard to stance, I have found it only necessary to distinguish "advancing" figures from standing figures.
BMC VOL, pp. 47–48, 56. The large number of VGC coins struck by the Ostrogoths in the name of Justinian may be particularly noted in the catalogue of Tolstoi.
For the study of Merovingian and Burgundian coinage see Belfort; Prou; Reinhart, Merowinger; Robert, Languedoc; Tolstoï.
The statistics for Charts I–VI have been compiled from the following sources: BMCB, BMCRE, BMCVOL, Cohen, LRBC, RIC I–IX, Sabatier, Tolstoï and Ulrich-Bansa.
Cohen, VIII, p. 208, 7. The mint mark most probably should read RV.
There is only one recorded VPW item on which the name of a Roman empress appears, an Æ 4 from Rome, but it is combined with the name of the emperor (Valentinian III and Placidia). This is recorded in LRCB.
RIC IX, pp. 28ff., 51ff.
Ibid., pp. 78ff., 232ff.
After Valentinian III in the West, the cross or chrismon-in-wreath becomes the major type. This is also seen on some imperial issues from the East. See Cohen, VIII, pp. 220 ff.
Pearce, NC 1943, pp. 97–99.
Cohen, VIII, p. 208, no. 3.
Sabatier, I, p. 120, nos. 4, 5; p. 122, no. 11. Also see Tolstoï, I, p. 86, no. 100.
Tolstoï, I, p. 107, no. 47.
Sabatier, I, p. 104, no. 26. This item is not listed by Pearce in RIC IX.
Sabatier, I, p. 140, no. 15; p. 153, no. 9; p. 154, no. 12. Also see BMCVOL, p. 44, n. 1. This illustrates a tradition in the time of Odovacar which imitates types suggested by imperial coins already struck in Italy in the name of Zeno.
Such coins can be found in my style groups JAN 11, JAN 11a, JAN 11b, e.g., nos. 396–398, 401–402, 406.
The use of the type, although of an unusual frontal kind, is noted on a unique Zeno Æ 3 VICTORIA AVGGG, however, with the mint mark of Ravenna in the exergue and what may be a rough rendition of the letters S C in the field. These have been listed but not commented on by Sabatier and Tolstoï. Cf. Sabatier, I, p. 141, no. 17, pl. VIII, 7; Tolstoï, I, p. 159, no. 69.
RIC IX, pp. xv–xvi.
RIC IX, pp. xviii, xxxiv.
Ibid., pp. xix, xxxi.
BMCVOL, p. 99, pl. XII, 20–23; pl. XIII, 1. Tolstoï agrees with the Wroth attribution. Cf., Tolstoï, I, pp. 158–159, no. 68.
Ratto, nos. 18–19.
BMCVOL, p. 54. Also see Kraus, p. 82, no. 1.
Grant, Anniversary Issues, p. 155. Also see Vermeule, Ancient Dies, p. 356.
Sidonius Apollinaris, I, pp. lxiv, cxii, cxxi; II, pp. 26–31, Bk. I, letter xi. Also see Maurice II, p. 141. As an indication of the continuing presence of paganism see the Cod. Theod. L. 25, 31 which has been quoted by Sabatier, I, p. 112, n. 1.


The incidence data compiled provide a basis for an iconographical investigation of the VPW type. The almost total disregard of the VPW in the East and its substitution by the VGC type; the diminishing use of the type in the West; the earlier conscious preference for the type during the reigns of Vespasian and Valentinian I; and the evidence of its being revived on the tremissis of Anastasius combine to suggest the use of the VPW as an intelligible symbol. The political use of numismatic types in Roman coins is common knowledge. Does the revival of a type discarded for a century on the tremissis serve the purpose of political propaganda? If it does, then it must be proved that the Victoria in conjunction with palm and wreath has a particular symbolic nature accepted by a central authority as being intelligible to the masses. Or, at the least, that the VPW is more applicable to the necessities of contemporary politics than any of the other Victory types in use: the VGC (being used in the East), the Victory with wreath and trophy as seen on the quasi-autonomous bronzes of Zeno-Odovacar, and the Brescian Victoria. Or, that the VPW symbol has an integrity of its own. However fruitful such an investigation might be and however beautiful the intricate structure of interrelating hypotheses, we cannot discount the possibility that the selection of the type may depend solely on its intimate traditional usage on Roman Imperial coins. Was it selected simply by force of habit by mints and minters who previously had a long history of use of the type and had continued to use it on small issues of silver and bronze? Was it simply the need for selecting a type distinctive from that used in Constantinople, regardless of symbolic connotation and political propaganda? Would it be so unusual for these mints which struck the type so frequently in the past to copy a type for which old die designs were possibly available?

Even a perusal of the history of Roman numismatic types will demonstrate that there was no personification, no minor deity more popular, more ever present, more imperially associated than the goddess Victoria. 1 She appears in many guises, with many attributes, with many specific connotations throughout the life of the Republic and the Empire. It is unnecessary to refer to her constant appearance on medallions, cameos, intaglios, sculptural reliefs, in paintings and in sculpture in the round, as well as on coins. She appears on gems as early as the late Etruscan period and is still found as late as the sixth century on consular diptychs. 2 This is indicative of the major role she plays in Imperial ideology, a rôle succinctly described by Graillot: "C'est la Victoire qui a fondé l'Empire; c'est par elle qu'il se perpétue." 3

A cursory review of the numismatic use of the VPW type will help our inquiries. The earliest numismatic association of Victoria with palm and wreath is to be found on a silver didrachm from Southern Italy dating between 241–222 B.C., on which Victoria holds a palm branch to which she attaches a wreath crown. 4 This may be in consequence of the tradition established ca. 293 B.C., when both the wreath and the palm of Greek tradition were awarded to the victors in games, and Victoria is alluded to as the palmaris dea. 5 Much more significant is the frequent use of the VPW on the "Victoriatus." Although the Victory on the Republican quinarius may often bear only a wreath with which to crown the trophy erected in front of her, she often bears a palm branch against a shoulder if she is to carry anything in the free hand. 6 Her first recorded solo appearance, however, is on a silver quinarius of the Marianist, L. Calpurnius Piso in 90 b.c. 7

The numismatic frequency of Victoria is accounted for by the developing importance of her personification through the territorial expansion of Rome. The Temple of Iovis Victoria, the first recorded, was dedicated on the Palatine in 294 b.c., while a century later the Aedicola of Victoria Virgo (193 b.c.) was also founded there. 8 The quinarius of Calpurnius Piso might very well commemorate these two dedications as well as the Marian cause he represents, since quinarii are so very frequently attributable to special commemorative celebrations. 9

As might be expected it is in the last century of the Republic that the reference to Victoria on the coins becomes the most popular. It is at this time that she assumes so many attributes and her depictions are so varied. Her cult, still amorphous in ritual and still lacking independent deification status, is adopted by Marius, Sulla, Pompey, Caesar, Marc Antony and Octavian. We are told that Marius placed his statue between two gold VPW's on the Capitoline and that his figure removed by Sulla was replaced by Caesar in 65 b.c. 10 This statue before the Temple of Victory is to be seen on the coins of the leading Marian, C. Marcus Censorinus. 11 Marius, if not the first to make the association, certainly revitalized or emphasized the connection of the goddess with the army. Her commemoration of military victories is always noted on the coins of moneyers of the family of Scipio and of M. Cato, 12 but the Marius association is of primary importance for her future. Caesar's allegiance to Marius, and Octavian's to Caesar, is the germinating flux necessary for transforming "Victoria" into her prime role at the core of Imperial theology.

Her frequency as a solo VPW increases in denarius and quinarius issues from 44 b.c. The personification so identified with Caesar who founded Temples of Hercules Invictus, Minerva Victrix, Venus Victrix and one at Pharsalos where he fought in the name of Venus Victrix, is used by Caesar's murderers who present on a coin a VPW with a broken wreath and a broken sceptre at her feet. 13 Marc Antony uses a VPW on a denarius from Asia Minor (31–30 b.c.) and on another from Cyrenaica (31–30 b.c.), but most important is the denarius of Octavian on the reverse of which is a VPW on a globe and the legend CAESAR DIVI F(ilius). The date of this issue is debated as either immediately preceding or following the Battle of Actium (ca. 31–29 B.C.). 14 Equally as important as its connection with the Battle of Actium is the use of a VPW in conjunction with the legend in which Octavian declares himself son of God. 15 The earliest association between Victoria and imperial inheritance is so made in a VPW, an association which contains the seeds of the type's later imperial associations with the legend, "Victoria Augusta."

This coin may also have further significance for us. The VPW on a globe would seem to be a copy of the Victory statue from Tarentum, placed on an altar in the Curia Julia by Augustus on August 28, 29 B.C., when he established the cult of Victory in commemoration of the Battle of Actium. The statue also appears on coins struck in Cyrenaica (30–27 b.c.), Corinth and on a Cistophoric issue. 16 At this altar in the Curia, Augustus prescribed that all Senators burn incense before their deliberations. 17 This tradition was to be maintained to A.D. 382, when Gratian removed the statue. This latter cause célèbre proves how even after Constantine, the Roman Senators, pagan-Christian, regarded this statue more as a symbol of State rather than just as a symbol of Imperial Victory. It was well worthy of the patriotic devotion of the Senate, since the rites before it symbolized their allegiance and obeisance to the State.

Another coin indicative of the developing imperial role of Victoria is an African bronze issue of ca. 25 b.c., of the proconsul, M. Acilius Glabrio. The obverse presents the bare head of Octavian facing an advancing VPW (ancestress of the globus nicephorus held in the hands of the Roman emperor) with the legend IMP CAESAR DIVI F AUGUST. COS IX. The reverse presents the heads of Marcellus and Julia facing each other with the name of Glabrio, the minter, on the legend. This coin implies the source of Octavian's "auctoritas" through his current consulship and therefore his imperium in Italy and his own provinces as well as through the Victoria Caesaris and his adoption by Caesar. Grant believes that the placement on the reverse of Julia and Octavian's nephew Marcellus, who were married in 25 b.c., may suggest a possible successor. 18

The VPW, which has had a traditional connection with military and athletic triumphs (its statues on the Spina in the Circus), is increasing its realm of meanings. The philosophical and morally stoic society had already associated rights and duties with Victory. It was not just "winning." 19 But she now offers guarantees of victory in life, apotheosis in death and victories for all descendants. 20 This new imperial role is clearly depicted on a sword of Tiberius in the British Museum. A Victory bearing a shield with the inscription VIC AVG stands behind the enthroned Tiberius, who offers a globus nicephorus to Germanicus, thereby signifying the victories which come from Augustus through Tiberius to Germanicus. 21 As Octavian based his rule to some extent on his descent from Divus Julius so did all the emperors from Tiberius base their "auctoritas" upon their descent from Divus Augustus through the medium of the cult of Victoria Augusta.

The final stage of the development of the cult of Victoria is attained in the Principate of Octavian, who has incorporated with the new imperial cult all the previous associations of the Venus Victrix et Felicitas of Pergamum worshiped by Sulla and Pompey, and the Victoria Genetrix of Caesar. 22 The Victoria Sullae and the Victoria Caesaris are the prototypes for the Victoria Augusta. Here the stoic philosophical and moral interpretations are combined with political and religious identifications to create a worshipful symbol of the essence of the new world, the Roman Empire. The cult will never lose its personal identification with Augustus, the founder of the Principate and even the cult itself in its wider meaning. Victoria is an heraldic device, a religious and legal symbol of the Empire. It stands for security, freedom, mildness and well-being within the borders of the Empire. It brings peace and prosperity wherever it marches ahead of the Imperial legions, whom it protects and whose right to victory it guarantees. 23 It is Roma. 24

The Victoria Augusta, always in the care of the Pontifex Maximus who is also the emperor, is inherited by each successive Augustus. In this way the legitimate rule of the Empire is handed down. 25 This accounts for the extensive use of the Victoria on coins of Vespasian who modeled his religious and dynastic program on that of Augustus in his efforts to legitimatize the right of his house to the imperial throne. Through her he identifies himself and his house with Augustus and the glories and traditions of the Empire. 26 It is through Victoria that the aegis of Roma Aeterna is passed down to the future. This concept is materialized on the reverse of a bronze issue of A.D. 71, on which a flying Victory bearing a palm presents the palladium, the symbol of the eternity and security of Rome, to Vespasian. 27

Even more indicative for us is a denarius of a.d. 70–71, which commemorates Vespasian's victories over both internal anarchy and his external enemies. The reverse depicts a VPW on a prow about which Grant presents an interesting hypothesis. Vespasian, as a conqueror of the East and as a restorer of the Republic, is identified with Augustus whom he in turn complements as a true successor of Alexander in the Hellenic non-autocratic tradition of government. A century after the Battle of Actium this coin may interpret that event in a larger historical connexus. Demetrius Poliorcetes (363–283 b.c.) placed the same type on a coin in 306 B.C., to celebrate his victories during the wars following the death of Alexander. In a naval victory off Cyprus, Demetrius had defeated Ptolemy. Thus the King of Pergamum, son of the ruler of Macedonia, is the westernmost ruler and his dynasty the most western which defeats the East, Egypt. Furthermore, one of Demetrius' successors, Antigonus Gonatus (ca. 276–239 b.c.) conceived with the advice of Stoics a theory of monarchy as a kind of "admired slavery," and this was recognized at the time as a contribution to the philosophy of the Principate. Thus Alexander, Demetrius, Augustus amd Vespasian all represent the West and the non-autocratic Hellenic tradition in their victory over the East. This was the part played certainly by Octavian at Actium in defeating Antony and Cleopatra. Although the navy was not too involved with Vespasian's victory he retains the figure of the Victory on a prow. 28

This Pergamenian association of VPW may be accounted for by other evidence: the possibility that Octavian's denarius discussed above was minted in 29 b.c. in the East from the proceeds of Egyptian spoils, since he spent some time there reorganizing affairs; that the cult of Venus Victrix et Felicitas has Pergamum origins; that Demetrius' use of a VPW may be fully in accordance with Pergamenian iconology, because a Nike with palm and wreath crowns Athena on the east frieze of the Altar of Zeus and more winged Nikes aid the gods and goddesses on the north frieze. 29

Of major importance to the Late Empire is the role that the goddess Victoria and her representation with palm and wreath plays in the time of Constantine the Great. The VPW type is, on the basis of its frequency of occurrence, a decorative and symbolic device in Constantinian workshops. The type is formularized and imitated on architectural decoration ranging from good to commercially produced pieces. Examples may be found on some consoles from the Maxentius Basilica and the Casa di Rienzo. 30 All of these date between a.d. 310–315.

The VPW type as depicted on the Arch of Constantine reveals its rôle in the early fourth century. In the Proelium Frieze she leads the emperor into battle; in the Ingressus she shares the emperor's triumph, striding beside his chariot; in the Praefactio she appears on top of a standard as a Dea Militaris. In other Constantinian reliefs on the Arch (the keystone of the main portal of the north side), she appears as a globus nicephorus held in the right hand of the enthroned Roma Aeterna. In both the western and eastern passages she is seen crowning male busts. She therefore is depicted in accordance with the representation in earlier reliefs on the Arch. In the second century Attic reliefs she appears twice on a military standard as a Dea Militaris, and on the Trajanic relief on the east wall of the main passage she crowns the emperor. It is however on the pedestal reliefs where Constantine is represented as "ubique victor," "victoriosus semper" and "victor omnium gentium" that the significance of the VPW may be inferred.

There are twenty-four of these reliefs; twelve to each of the north and south sides of the Arch; three to each of the four pedestals which support the four columns on each of the north and south faces of the Arch. Of these twenty-four reliefs, eight are the more significant not only because of the visual advantage of their physical position, but also because of the symbolic pattern they form as a unit. These are the reliefs on the front face of each of the eight pedestals. On the north side, the frontal faces of the outside pedestals depict a standing VPW with a kneeling northern captive, while the frontal faces of the inside pedestals depict a Victory with a Votive Shield and a kneeling Northerner. The side faces also fit a symmetrical pattern in which pairs complement each other in content but do not exactly duplicate each other formally. The side faces of the outside pedestals (those bearing the VPW) depict a northern family with a Trophy on one, and a Roman soldier leading a captive Northerner on the other; the side faces of the inside pedestals depict Roman soldiers with northern captives on one and the Imperial Guard with the statuettes of the Dei Militaris on the other. The north side of the Arch in both the pedestal and the frieze reliefs of the Oratio and the Liberalitas announce the "victor perpetua" through the Vota festivals as symbolized by the Votive Victory.

On the south side, the frontal faces of the outside pedestals present a standing VPW with a kneeling Oriental on one and a Northerner on the other, while the frontal faces of the inside pedestals depict a Trophy-bearing Victoria with captives. The side faces of the outside pedestals parallel those of the north side; on one, an oriental family surrounds a Trophy, and on the other a northern family surrounds a Trophy, and a Roman soldier leads an oriental captive on one and a Northerner on the other. The side faces on the inside pedestals depict Roman soldiers leading captives, and marching "signiferi" carrying the Imperial insignia. The south side commemorates through its pedestal and frieze reliefs of the Obsidio and the Proelium the "victor omnium gentium" by representing the physical attainment of the Triumph. 31 The supporting role played by the VPW on the pedestal reliefs is that of the continuous background motif — the Victoria Augusta. Its presence qualifies and characterizes the "victor perpetua" and mitigates the ravages of war. It is the palm and wreath that truly symbolize the more enduring peaceful aspects of victory. The fruits of Roman victory for even the subjugated peoples are peace, order and prosperity. 32 The conquered are welcomed into the brotherhood of the Pax Augusta, the real prize of the Victoria Augusta. The VPW appears then as the most generic form for representing the Victoria Augusta.

As one of the Dei Militaris along with Sol Invictus, she also is most important for us, and her constant use with Roma Aeterna as a "victoriola orb" seems to make Victoria and Sol the rulers of the whole religious world of the Constantine building program. 33 From the time of the Tiberius coin previously described in which the VPW first appears as an orb in the hand of the emperor, she is a constant figure as a victoriola in imperial iconology and persists in this role even on the Theodoric medallion which presents Theodoric, like Constantine, as "victor omnium gentium." Her depiction with Roma Aeterna on the Arch may be based on the old colossal porphyry cult statue of Roma Aeterna, which was once in a Hadrianic temple dedicated to that goddess and which was in the Temple of Venus and Rome in Constantine's time. Burned down in a.d. 307, it was rebuilt by Maxentius so that the venerable cult image could be housed. That the association of VPW with the goddess may go back even into Republican times is based on the evidence that the east keystone of the Janus Arch in Rome has a similar Roma Aeterna figure. It is significant for us that Theodosius, so instrumental in removing the VPW type from coins in the East, is also responsible for replacing the globus nicephorus of Roma Aeterna with a globus cruciger. 34 For Theodosius and the Christian Roman Empire the cross is assuming all of the rôles of the VPW.

The survival of Victoria in any form throughout the fifth century is evidence of her importance, and her survival in pure pagan forms with or without associating Christian qualifying marks presents the problem faced by mint masters in the late empire. Her rôle had become so identified with the state that she was beyond the limiting bonds of religions. Besides, the fourth century population, as well as that of the fifth century, was still pagan orientated. We have evidence from both Arles and Rome to indicate that a large part of the aristocracy was still intellectually, emotionally and religiously pagan while being only politically Christian. 35 The silver bowl in Leningrad from the time of Constantinus II presents the dichotomy and incongruity of a VPW leading the way for the imperial equestrian figure, while a soldier keeping up the rear holds a shield decorated with the chrismon. 36 The victory of the emperor is thereby doubly insured.

Pearce's explanation for the Valentinian I control of the GLORIA ROMANORUM and SECURITAS REIPUBLICAE issues furthers our realization of the persisting importance of Victoria and the VPW outside the realm of religions, particularly in the West, and her association by the people of the empire with Rome. As long as the strong emperor was at Rome his chancellery was able to uniform and regularize the striking of the issues in every bronze mint of the two emperors (Valens and Valentinian I). When Gratian succeeds to the purple in a.d. 367, the East can rally only insofar as to curtail the striking of bronze coins of these issues. Once Theodosius ascends to the Eastern throne in a.d. 379, there is a dramatic and immediate change of types, and from that time the VPW is dropped in the East, although it is continued in the West. Yet, it is only here on the bronze coinage that Theodosius asserts eastern independence in the Roman world. Certainly, at this time the VPW must have been equated with Rome and the West. 37

Constantinople, the capital in the East, truly stands for a new world, a Christian Empire, and as such it more strenuously dissociates itself from all pagan references, or it must Christianize those that it cannot eradicate. Theodosius, certainly, sees his rôle in the East in this framework and with true Spanish latent iconoclastic zeal replaces the globus nicephorus with the globus cruciger. The new sister capital also needs to establish a separate identity and an iconography of its own, although tradition might have made this impossible if it had not been for the advent of a new state religion, Christianity. When Constantinople needs to effect a symbolic change to assert its own beginning and independence, its excuse is Christianity.

There seems to be no question of the traditional Roman association and that of the Senate with Victoria and with her most generic form, the VPW. It is an association maintained by the vigilance of the Senate and the ritual at the feet of Victoria on the altar in the Curia. It is an association kept alive by the legal descent from Augustus the founder of the Empire and of the cult. Rome is Victoria. Victoria is Rome. It is Augustus who consummated the marriage for the benefits of his own time and for all of his descendants.

It is not surprising considering the significance of the type that it should be revived on coins found in the West and therefore minted in the West even in a more Christian sixth century. The very nature of Victoria being a Dea Militaris would establish the worship of the cult wherever the army was stationed or wherever veterans were to be found. 38 For the entire period of the late empire in the West the army is to be found stationed in Gaul and North Italy, as would be indicated by the location of the mints. This would perhaps account for the frequent use of the VPW type particularly at the mint of Trier, and even the unusual striking of a GLORIA ROMANORUM issue of Valens and Valentinian I with a VPW type. Altars such as those erected at Lyons in 10 b.c., to the Victoria Augusta with figures of VPW, imitating the altar and cult statue in the Curia Julia, must have been common features in Gaul and other western provinces, since the altar at Lyons was specifically given by Rome. The VPW is a common cult given to Gaul by the Romans. She is also important therefore for the non-military provincials. She represents the official cult through which loyalty to the emperor is expressed. Every town has its altar and statue of Victoria, where public fetes always take place. The Victoria cult was a means by which local colonial and provincial deities could be transformed into a more Roman framework, e.g., Brigantes, an English goddess who became Dea Victoria Brigantia. 39

Should we not investigate the possibility then that the selection of the VPW type for this Anastasius tremissis is engineered purposely to distinguish it as a western mint product in the name of the only emperor then in the Empire, but at possibly a time when all of the barbarian kingdoms in the West, or the particular minter of this coin, if only one minted it, were allied to Anastasius, and for the sake of diplomatic niceties and legalities owed him allegiance? It must also most likely commemorate a military victory since the type even in its past was used most frequently in conjunction with either contemporary military campaigns and victories or Augustan ones. Whoever mints it must, in accordance with a long tradition, connect himself through the Victoria Augusta with Rome, the Roman Senate and Augustus. It would seem most unlikely that the type was used without understanding its historical and traditional significance. There is too much circumstantial evidence to give credence to the alternate hypothesis that the striking of the type is due to the accidents of necessity, expediency, familiarity and continuity. The extent of the issue in both quantity and time would indicate more satisfactorily a carefully conceived and organized numismatic policy.

End Notes

See the Synoptic Table of Allergorical Personifications in Gnecchi, Coin Types, pp. 29–35, 63. Also see Vermeule, Victoria.
Numerous references could illustrate this: Walters, Engraved Gems and Cameos, nos. 712, 1170, 1705–1709, 1717, 1718, 2228, 3058, 3059, 3789 and 3911; Gnecchi, Coin Types, pl. V; Delbrueck, Consulardiptychen, Lief I, no. 1 and Lief IV, no. 48; BMCRE, I, pp. 113, 122, 128, 146, 154, 202, and nos. 526, 551, 552, 560, 564, 641, 657, 770, and 855.
Daremberg-Saglio, V, p. 839. This is H. Graillot's article on "Victoria." There is also an excellent discussion in Gagé, RA 1930, pp. 1–3.
CCR, pp. 2–3, nos. 21 and 21a.
Daremberg-Saglio, V, p. 852. Also see Livy, X, 47, 3; Apuleius Madaurensis, Metamorphoseon, II, 4. Statues of the VPW were to be found decorating the Spina in the Circus Maximus at Rome and at various Provincial Circuses, see Daremberg-Saglio, I, figs. 1518, 1520, 1521, and 1524–1526.
CCR, p. 8, nos. 83–84, and pp. 11–12, nos. 111–121; also BMCRR, I, p. 277, nos. 213ff.
CCR, pp. 102–103, nos. 672a–672h(listed as common), nos. 673a, 673b (scarce) and no. 674 (scarce).
See Livy, X, 29, 14; X, 33, 9; XXIX, 14, 14; XXXV, 9, 6; and Ovid, Fasti, VI, 644.
Grant, RIM, p. 206.
Plutarch, Caesar, 6, 1; Velleius Paterculus, II, 43, 4; Valerius Maximus, VI, 6, 14; Suetonius, Caesar, 11; PW, Series 2, VIII, A, pt. 2, p. 2513.
CCR, pp. 111–112. The existence of a Temple of Victory on the Palatine as early as 204 b.c. is implied by a reference in Livy concerning the transference of a sacred meteorite stone of the Mater Magna to it. Livy, XXIX, 14, 3 "…… in aedam Victoria quae est in Palatio."
CCR, pp. 83, 175ff.
Ibid., p. 203, no. 1298. Also see PW, Series 2, VIII, A, pt. 2, p. 2517.
Grant, RIM, p. 13.
Julius Caesar was deified in 42 b.c.
CCR, pl. III, 60 (a denarius of L. Pinarius Scarpus). Also see BMC Corinth, pl. 15, 10 (coins struck by Duumvirs, Vatronius Labeo and Rutilius Plaucus); Daremberg-Saglio, I, p. 1213, fig. 1563; Mommsen, Monnaie Romaine, III, p. 302.
Gagé, RH 1933, p. 34; Taylor, pp. 153, 187. The ancient source is Cassius Dio Cocceianus, LI, 22, 1.
Grant, RIM, p. 27.
Ennius, Annales, 493v: "Qui vincit non est victor nisi victus fatetur;" Livy, IV, 10, 3 and XLIV, 47, 8: "…… fatentes victos se esse et imperio parere" "……eius demum animum in perpetuum vinci, cui confessit expressa sit se neque arte neque casu, sed collabis comminus viribus, iusto ac pio bello superatum." Cicero, Att., VII, 22, 1 and IX, 7c, 1; Virgil, Georgics, IV, 561; Tacitus, Germaniae, 2; Serv. Aem. IV, 618 and I, 6.
Gagé, RA 1930, pp. 1–2, 34; Gagé, RH 1933, p. 43. M. Gagé draws his conclusions from Cassius Dio Cocceianus, XLV, 17 and XLVIII, 16.
Gagé, RA 1930, p. 13; Franks, Proceedings of the Society of Antiquarians, London , III (1864–1867), p. 358; Walters, Select Bronzes, p. 867.
A temple of Venus Victrix was consecrated at Rome in 55 b.c. See Gagé, RA 1930, p. 34.
Gagé, RA 1930, pp. 2–4; Deubner, pp. 37–42; Scott, Flavians, p. 26; PW, Series 2, VIII, A, pt. 2, pp. 2509, 2510, 2520. It is stated in the latter reference that the Victoria Augusta implied the complex political and social service of Rome to all of its subjects, the VICTORIA UTI, so called by Caesar and known also as the Victoria Caesaris and Clementia Caesaris.
Victory types are one of the major personifications seen on third century medallions, and Roma is shown holding a globus nicephorus. See Toynbee, JRS 1947, pp. 135–144.
Gagé, RH 1933, p. 13; Scott, Flavians, p. 25; PW, Series 2, VIII, A, pt. 2, p. 2511.
Scott, Flavians, pp. 28–29.
Scott, Flavians, pp. 24–25; Kähler, Personificationem, pp. 31–32; BMCRE, II, p. 126, no. 586.
Grant, RIM, pp. 188–189. Grant relates Tacitus' comments concerning the plans of Vespasian to invade Africa by land and sea.
Kähler, Pergamon, pls. 3, 13, 25. In thus presenting the Nike, Pergamum, a major cultural and religious Hellenistic center, may have been equally important in giving impetus and prototypes for Roman forms as it was generally in architecture and sculpture.
Kähler, JDAI 1936, pp. 185–186, 193.
For a diagram of the numbered friezes see L'Orange, Konstantinsbogens, Abb. 16.
Deubner, pp. 37–42. Nemesis destroys the enemies of the emperor and punishes their insolence, while Victoria brings Pax.
L'Orange, Konstantinsbogens, pp. 57–58.
Ibid., p. 149.
Dill, Roman Society, pp. 27–58. This is an excellent discussion of the tenacity of paganism. Also see Alföldi, ACILRE, p. 83.
Delbrueck, Kaiserornat, pp. 1–21.
RIC IX, pp. xviii–xix, xxxi.
Daremberg-Saglio, V, p. 841.
Ibid., p. 842.


Since our compilation of evidence, factual and circumstantial, suggests Western origin and responsibility for the striking of the VPW tremissis in the name of Anastasius, it becomes necessary to consider the associations of the various barbarian kingdoms with this coin in order to discover its moneyer. By virtue of provenance and attribution none save the Suevians can be omitted. 1 Tremisses of the VPW type from Anastasius to Justin II have been in part attributed to the Vandals, the Merovingians, the Burgundians, the Ostrogoths and the Visigoths.

Vandalic attributions are the most tenuous to uphold. There is no evidence that the Vandals ever minted in gold. The very few solidi and tremisses attributed to them are done so purely on stylistic grounds which are always very questionable when not supported by external evidence. There is, however, the questionable probability of African provenance for these coins. Those in the British Museum are catalogued as Vandalic because "…the Count de Salis has pronounced (these coins) to be Vandal on grounds of style (and probably of provenance)." 2 Since de Salis did not publish any of his reasons, except for what was verbally passed on to Keary on this matter, the Wroth catalogue attributions may count for very little, particularly when we consider the total aspect of Vandalic coinage. From possibly as early as a.d. 439 to 533, there is a Vandalic tradition for the striking of only silver and bronze. The few gold pieces are anomalies in this fiscal tradition. There are no discernible mint marks on the attributed gold, since the earlier of the two groups are imitations of solidi and tremisses in the name of Valentinian III, and the latter are in the name of Anastasius. Thus if these were minted in Carthage, the early coins were under Gaiseric (428–471), and the Anastasius pieces would coincide with the reigns of Gunthamund (484–496), or Trasamund (496–523), or both. On stylistic grounds, however, the tremisses particularly are extremely close to coins found in greater abundance in southern France and will prove later to be a definite product of Gaul. Friedlaender's early doubting of Vandalic minting of gold might be an accurate hypothesis. 3

If these eight Anastasius tremisses in the British Museum were minted by the Vandals, 4 their use of the VPW type could only be accounted for by habit and familiarity, since the type is found on a large group of attributed small bronze coins of likely local unofficial issues in North Africa which bear the names of Honorius and Valentinian III. These issues bear the legends: VICTORIA AVGG and SALUS REIPUBLICAE, when both legends are not so crudely blundered as to be illegible, as IИИИ or ΛVIIII. 5 All of these may be fifth century products and therefore in accordance with VPW traditions of western bronze issues. The Vandalic attitude towards both Rome and Constantinople and the Empire seems inappropriate to any meaningful use of the VPW type for political or ideological considerations. They always conceive of themselves as separate and autonomous rather than as a part of the Empire. This is reflected in their being the first of the barbarians to place the name of their own king on their coins. 6

The possible finding of a few VPW gold coins in North Africa should not encourage the rash view of their being local mint products. Strong trade relations existed between Africa and Italy as well as southern France and Spain. Friedlaender, disbelieving their Vandalic attribution, saw them as being possibly Ostrogothic and their African provenance accounted for by the story of the Ostrogoths paying one gold coin for each "Trulle" of grain shipped by the Vandals during a famine in Italy. 7 The stylistic relationship of the gold tremissis to the Vandalic silver and bronze is less similar than to the other gold VPW tremisses that have been attributed to other barbarian mints. It is the bronzes of the Imperial Carthaginian mint after 534, however, that bear a marked resemblance to the developing stylized bust portrait on the coins of particularly Spanish provenance (See Plate B,7). The Carthaginian portrait bust even at that time presents a pectoral cross or chrismon. 8 Such connections are seen often in both artistic and other cultural and intellectual areas and may point to common mutual developments and reciprocal influences. 9 One should not be surprised to find similarities and connections between cultural products in North Africa and the Iberian Peninsula; one should expect it. This existent factor alone would leave open for consideration numismatic connections between the mint at Carthage and those in Spain, if it was not negated by the preponderance of Imperial Italian and Gallic mint products found in Spain.

The possibility that this tremissis is a product of Vandalic fiscal initiation is further obstructed by too many qualifying hypotheses. Their rulers never could have espoused the role of "Restitutor orbis Romanum" unless sardonically. 10 Nor does it seem possible that other barbarians would imitate Vandalic numismatic measures or that the Carthaginian mint would set an example for those in Gaul. It would seem an extraordinary eventuality, and there is no evidence to warrant the consideration of this as feasible. Furthermore, Gunthamund and Trasamund did not play any significant role in the affairs of Gaul and Spain to warrant any connection on their part with the Anastasius tremisses. If the Vandals minted this tremissis, therefore, it must have been in imitation of a fiscal measure already adopted at more important mints in the West such as those in Gaul or Italy.

The connections of the other barbarian kingdoms are much stronger since all occupied at some time the vicinity in which these coins must originate, southern France. The major French collections based largely on French hoard finds contain the largest number of the VPW tremisses for the Anastasius period, and their collections diminish in the number of these coins for each successive emperor, (See Charts VII and IX). Spanish collections on the other hand tend to develop in the opposite direction, with the Anastasius pieces being the smallest in number. 11 This is only one observation among many which gives increasing credence to the hypothesis that the coins originate in southern France, perhaps at Arles, or Narbonne, or Toulouse or Lyons; mints associated with the Merovingians, Burgundians, Ostrogoths and Visigoths.

The Merovingian associations, consequently, are much stronger in regard to provenance and numismatic tradition. Many VPW tremisses have been attributed to them by Tolstoï, Lenormant, Prou, Belfort and Le Gentilhomme as well as by others. 12 Reinhart further attributed all the Anastasius issues with pectoral crosses after 507, to Merovingian mints in southern France. 13 Belfort's catalogue lists many in the name of Anastasius and Justin I and attributes them to the mints of Narbonne, Toulouse, Troyes, Trier and Lyons. The Lyons group for this period would have to be Burgundian; the Narbonne group Visigothic; and the Toulousan coins only Merovingian if it can be proved that they are minted after 507; otherwise they too would fall into the Visigothic category. This would leave to the Merovingians the Troyes, Trier and Toul attributions which number in all five. Both numbers 4355 and 4356, which are in the name of Justin I and bear the letter T in the field to the right of VPW and Δ to her left, are questioned as Troyes or Trier, while number 4357 in the name of Anastasius which only has the letter T in the field left to the VPW is given unquestionably to Troyes. Extraordinarily enough, though not in the field of barbaric numis- matics, this last coin (No. 4357), given to Troyes, is very similar in style to the coin (No. 4327) which is given to Toulouse and to the coin (No. 4479) which is given to Toul (Menthe-et-Moselle). Obviously "guesswork" has played too important a role and the results often verge on the absurd. We are involved with material in which there is no strong dividing line between fact and fiction, and M. Belfort's attributions must be seriously questioned.

What might give more credence or justification to attributing the VPW anonymous issues to the Merovingians are the number, although small, which bear the names of Merovingian kings with or without the emperor's name. These occur with the name of Theodebert I (534–547) and Childebert I (511–558) in two groups, the smaller one in which VPW appears as traditionally, and the larger one in which the VPW figure is without the palm branch. 14 This latter development is a diemaker's oversight when copying the design from worn dies or coins. If all of the Lyons attributions are discounted by virtue of their descending from possible coins issued under Burgundian rule, we are left with an extremely small amount of coins (nine out of seventeen) that may descend from Merovingian mints during the reigns of Anastasius and Justin I. These coins reveal: (1) that the VPW type is very rare for most sixth century coins bearing mint marks; (2) that a small number of mints may issue it; and (3) that most of these, from which the majority of the VPW coins come, are mints that were not Merovingian in the first thirty-five years of the sixth century.

The consequent rapidity in which Merovingian coinage degenerates with excessively crude elements in design and format (See Plate C, 10–12); 15 its failure to make an organized abstract pattern from the Victory figure (certainly achieved in national Visigothic coinage of Leovigild); and its failure to produce a unified stylistic pattern universal to all mints issuing a Victory, seem to point to both a governmental disregard for the VPW type and the lack of tradition for the VPW in the Merovingian mints. There is little stylistic resemblance between many of the earlier issues and the late coins of Childebert, and there is nothing comparable to the stylistic development that is climaxed by the VPW of Leovigild the Visigoth. The late Visigothic VPW tremisses contemporary with or earlier than Childebert are unquestionably the product of a long tradition, of system, and of government-controlled mints. The later royal Merovingian VPW issues have the quality of being imitations rather than descendants from original issues. 16

When these stylistic points are coupled with the fact that the VGC is really the more popular Victory form in northern Merovingian (Austrasian) mints, so much so that Prou could say that most Merovingian Victory tremisses are of the VGC type, it is possible to question any Merovingian role in the initiation of this tremissis. 17 We see Merovingian money being executed by inexperienced local goldsmiths or moneyers with no or few traditional or professional associations with the former imperial mints, and with seemingly little central control. It is from the imperial mints almost exclusively or their vicinity that the imperial type of VPW is issued in the sixth century in Merovingian territory, imperial mints that had been Visigothic or Burgundian before they passed into the hands of the Franks. This is remarkably true even for the VGC type, which was noted by Prou as coming particularly from the areas associated with the Burgundians; mints of Lyons, Chalon, Lausanne, etc. 18 All strictly Merovingian mints begin exclusively with the several varieties of national coins with names of king and atelier and for types, crosses or monograms or both.

The type, then, survives in definite Merovingian coins only where the mint has had previous Visigothic or Burgundian association and possible late fifth century imperial relationship. The only major mint suggested for the early attributions which was in the hands of Clovis and the Franks before the fall of Burgundy in 534, was Toulouse, and this was only so since 507. This may very well be the mint of all the VPW and VW (Victory with wreath only) coins given to Trier, Troyes and Toul on the basis of the T in the field of the reverse.

If, then, the Merovingians initiated this coin, it would be first from Toulouse in 508, at the earliest. Reinhart's suggestion qualifies this. The Visigothic VPW issues of Toulouse, Reinhart said, were continued by Clovis who placed a cross on the chest of the obverse portrait. 19 Reinhart therefore considered the Visigoths as the initiators and attributed the invention of the pectoral cross of the obverse portrait to Clovis, the successful Christian crusader, on coins minted from the newly captured Visigothic treasure. This suggests that the Visigoths, so ignominiously defeated by Clovis, would adopt a device associated with him. It seems unlikely, for the stylistic grouping of the coins indicates a development toward the placement of the pectoral cross rather than an abrupt or immediate die change. 20 Just as unlikely is the idea that the Merovingians were the first to reissue the VPW tremissis. Little of the significance of the VPW and its VICTORIA AUGUSTORUM legend must have been understood by the Franks, the most recent barbarians. It would seem a rather inappropriate type to commemorate an orthodox Christian victory over an Arian king, or for Clovis' provincial orthodox ministers to suggest. The entire development of Merovingian national coinage in the sixth century would indicate the adoption and continuation of coin issues in mints recently conquered rather than the initiation of a new issue. It is not to Rome or to Ostrogothic Italy that Clovis looks. His honors come from the emperor in the East who has made him Consul after his victories over the Visigoths, 21 and his propagandists are the orthodox Christian clergy of Gaul.

The major reason for Burgundian attributions is the reading of monograms on the field of the reverse as belonging to Gundibald, Sigismund and Gundomar. 22 It has been customary to attribute all of these to the mint at Lyons, although only one coin in the entire group listed by Belfort had any mark that might associate it with that mint. 23 Nevertheless, it is natural to attribute them to the capital of the Burgundians which had a long imperial mint history. Other coins without monograms have been assigned to the capital of the Burgundians on the basis of their stylistic similarity to those coins bearing monograms. Since we have no VPW coins with the names of a Burgundian king in the legend, it is on the acceptance of the monogram interpretations and particularly of the certainty of the "Sigismund" monogram combined with their provenance that has assigned these coins to the Burgundians. The coin hoards of Alesia (1804) and of Gourdon (1846), both found in Burgundian territory and both containing over one hundred coins and buried within the reigns of Justin I and Justinian, would encourage the idea that VPW tremisses were in circulation here. 24

When we connect these hoards to the already noted role of Lyons in the minting of VPW and VGC types and the monogram identifications, it is not unreasonable to consider that the Burgundians did mint this coin in the time of Anastasius, Justin I and Justinian. Mateu y Llopis went so far as to suggest the possibility that the Burgundians were the first issuers of this type, although his dating this in the reign of Sigismund (516–524) would seem to be in error, disregarding as it does the coins bearing the monogram of Gundibald. 25 The stylistic design of types on Burgundian coinage remains fairly respectable on the whole, reflecting the influence of the old Imperial mint at Lyons. The VPW type would be in order since it was a frequently used type on Lyons issues and since Lyons itself bears strong ties with Augustus. Here was to be found the Augustan Altar of Victoria Augusta. Here also was the provincial and military center of Gaul, particularly after the abandonment of Trier in the early part of the fifth century.

Gundibald with his provincial advisors could have initiated this type after the death and defeat of Godegisel and the unification of the Burgundian kingdom under Gundibald after 500. This would coincide with establishment of the one capital at Lyons. The Burgundians do strike a large amount of the VGC type, however, and Robert's theory might apply here. For him southern Gaul (particularly modern Provence) had been long accustomed to Roman institutions with their centralized monetary policy and guaranteed prices, and they consequently maintained the imperial type and legend traditions of the few great mints such as Lyons and Arles. Rather than replace these types and legends with those of the new king, they always looked to the Roman Emperor. 26 Although the use of the VPW type might have an appropriate meaning for Gundibald's victory and the emergence of a unified Burgundian nation with its own code of law, there does not seem to be any sound reasons for the Visigoths to imitate or to do the same. Neither together nor separately do Gundibald, the Burgundians, or the mint at Lyons offer a sufficient modus vivendi for the VPW tradition which continues in Spain into the reign of Leovigild.

When the Visigoths are considered, there are a number of incontestable facts which connect them with these coins; the existence of the VPW tremissis in the name of Leovigild; the stylistic similarity of many of the anonymous tremisses with those bearing the name of Leovigild; and the large number of such VPW coins found in Spain and in Spanish collections. The Visigothic minting of this coin has been unanimously accepted. To the Visigoths have been attributed a share of the tremisses without monograms (others to Franks and Vandals and Ostrogoths), those with a monogram image erroneously interpreted as Amalaric or Narbonne, and for the most part only tremisses of the VPW type. 27 Regrettably, of all the coins of this pre-Leovigild period found in Spain, only one hoard has been kept intact and available for study. This, the Zorita de los Canes, comprises ninety pieces beginning with coins in the name of Justinian and extending to the possibly last Leovigild VPW issue. 28

There is strong stylistic evidence for a continuous and parallel development in coin issues from the time of Anastasius to that of Leovigild. In fact, the aesthetic approach and propensity for abstract design so evident on Visigothic national coinage down to ca. 714 is identical to the design principles of the pre-Leovigild anonymous issues. Visigothic art is characterized by an extremely simple, direct, austere statement. No matter how abstract, the symbol never loses its recognizable integrity. There is, as one might expect, a carefully organized sense of proportion and order. Legends never usurp their boundaries, do not invade the field, do not mix with the types. There is an overall discipline in stylistic or mint groups within which a tremendously creative sense of individualism exists. Within the rigid bounds there is freedom and variety. A freedom, however, much controlled. That dichotomy so particularly Spanish, the struggle of the real and the mystic, already is apparent here.

The undisciplined vagaries of the earliest national Merovingian issues in both legends and type and in total design are quite different from the Visigothic formulae. This distinction in contrasting the styles of the early national coinage of both nations must be traceable in the pre-national coinage. There is evident then a distinctive way in which imperial types are imitated and developed in the mints of the Franks and the Visigoths, which is not negated by a possible similar source and common origin in southern France. Circumstantially, on the basis of what the Visigothic mints do with the VPW tremissis and in the advanced degree of their total commitment to the striking of the VPW type until Leovigild's numismatic reforms, we are encouraged to accede to the Visigoths a more important rôle in the minting of the VPW Anastasius tremissis than to the Merovingians or even to the Burgundians. One feels more assured to suggest that the VPW coins were issued in southern France at Visigothic mints and were imitated by the other barbarian groups once their victories secured for them some of these same mints and the Visigothic treasure.

It is even easier to justify Visigothic adoption and revival of the type. Certainly the political meaning and association of the VPW were intelligible to the Visigoths and applicable, as suggested by D. A. Fernández Guerra in 1854 "…… Los áureos visigóticos tuvieron el objeto exclusivo de recordar victorias, beneficios de la religión, piedad o munificencia de los Reyes, duros escarmientos u otros sucesos memorables. Las medallas visigóticas, son, pues conmemorativas." 29 At the beginning of their settlement in the Empire and after the Battle of Adrianople (a.d. 379) they had been induced to become federated allies by Theodosius. They were free and semi-independent, still remaining a nation, neither provincial nor Roman citizen and paying no tribute for land but receiving government pensions. For this they were obliged to serve as federated soldiers under their own chief for the Empire. They had a legal status in the Empire beginning in 382, and although their relations with the emperor experienced many non-fraternal vicissitudes it did not affect their feeling the necessity of having some legal agreement with him. The pro-barbarian Jordanes admits to their ravaging the lands of the Empire but only when they were victimized by imperial treachery, deceit and breaking of promises and treaties. 30

Of all the late barbarian groups, the Visigoths are the first to settle legally within the Empire and from the first are legally and militarily associated with the emperor. Aside from the Ostrogoths, no other group has seen so much of the Empire from Constantinople to the Pillars of Hercules, and none has had more opportunity and time to become Romanized. Their past is glorious, even if by 500 they are living on their laurels. Until their defeat in 507, they possess the largest and one of the richest nations in the West. The prestige of the Battle of Adrianople is only seconded and surpassed by Alaric's Sack of Rome (410). They are the first to sack imperial Rome, the Roma Aeterna, the Roma Invicta. An immemorial blasphemy. Yet, at this time Alaric is a rebelling Magister Militum of Illyricum. The treaty with Honorius in 418 re-establishes their foedus status and grants them the legal and political right to occupy Roman territory in Gaul. For this imperial recognition Wallia undertook to rid Spain of all other barbarians. Of these, only the Asding Vandals and the Suevians were able to save some of their settlements, and annihilation of the Silung Vandals and the Alans gave the Visigoths Aquitania Secunda. The ambiguity of a nation within a nation did not restrain the Visigoths from enlarging their territory when expedient or of considering their lands a veritable kingdom, although Rome still considered the area as a province and governed the Roman provincial population living in it. 31

Nevertheless, we should not underestimate the Visigoths' ability to Romanize, nor their feelings about Rome, nor the continuance of Roman prestige. The Visigoths, who participate with the Romans in the defeat of Attila at Châlons, are perhaps the most important factor there. Their panegyrists can truly picture them as upholders of Empire. 32 Orosius, speaking of an even earlier King, Ataulfus, Alaric's successor, tells us that this Visigoth had admitted that only by renewing and increasing the Roman name with the arms of his Gothic followers could he achieve immortality, remembered by posterity as the restorer of Rome. 33

The name of Rome still had the power to attract and be revered. It was a symbol of civilization and the glories of past invincible and universal power. As such it would never lose the psychological allegiance of the Visigoths. It is impossible to think that the Visigoths behave any differently from the large number of other barbarians who sought for a place in the society of the Empire. The manner in which the late Roman society is Germanized reveals how unopposed to this society the barbarians were and how undesirous they were of tearing it down. Barbarians such as Merobaudes, Stilicho, Bauto and Arbogastes had risen to positions of power and were married into Imperial families. One was even grandfather of Theodosius the Great. It was not so much Roman versus barbarian as it was barbarian versus barbarian in the struggle to find a place in the sun. Even Alaric's military exploits were aimed at gaining a more satisfactory and secure position within the borders of the Empire. 34 Their vigorous blood, unfortunately, could not rejuvenate that which their desire helped to destroy. But if they were instrumental in the final killing of a dying organism they at least kept its memory and prestige alive.

The most politically intelligent barbarian leaders saw Rome and their position as did Theodoric the Ostrogoth. Theodoric, aside from his unselfish interest, set out for Rome to restore it to the Empire and to destroy Odovacar, the previous restorer. 35 But by the sixth century the behaviour of kings such as Alaric II and Gundibald and Sigismund as well as Theodoric certainly reveals "barbarians" as statesmen respectably concerned with the stability of national life, which was threatened always by the imperialism of the least Romanized Franks.

Theoretically and politically rather than realistically, the barbarian kings founded their legal authority on the foedus connection with the Roman emperor and could pose as maintainers of the Pax Augusta, the bulwark of Empire. They were agents of the emperor and, after the expulsion of Romulus Augustus, they gave their token allegiance to the Eastern emperor, because he alone had imperial dignity in the Mediterranean. He is for them a kind of proxy Roman Emperor and their allegiance is never removed from Rome. The transfer of the regalia of the western emperors to Constantinople by Odovacar and the Roman Senate best symbolizes this. 36

Without an emperor in the West and because of commercial expediency for maintaining an acceptable international coinage, the VPW coins were issued in the names of Anastasius, Justin I, Justinian and Justin II. This is also a result of the tradition of imperial monopoly of the imperial image on the gold coinage rather than of real obedience or obeisance. 37

The Visigoths under direction of the Roman emperor had acted as his agent, the agents of Victoria Augusta in Gaul and Spain. They too partook of the auspices of Victoria by this legal connection with the Roman emperor, the rightful heir of Octavian Augustus. Victoria favored them as the "Restitutor orbis Romani." 38 No barbarian kingdom, other than that of the Ostrogoths, would have had more justification for the adoption of the palm and crown Victory either as military cohorts of the Roman army or as legal political leaders of part of the Roman State, thereby partaking of some of the dignities of imperial tradition. Ironically, no one deserved it so much as those who had first shown their prowess in conquering Rome itself.

Yet, with all this circumstantial evidence, there are still problems in awarding the Visigoths the role of initiator. Their control of all of southern France west of the Rhone and along the Riviera (the major provenance area of the Anastasius groups) and the imperial and non-imperial mints of this area (Narbonne, Toulouse, and particularly Arles, that most Roman of provincial cities) would certainly only give them priority before 507 in the issuance of this coin. 39 This would imply that Alaric II is a minter of this coin and the possible initiator, a rather ironic choice for this unbellicose man. If however these tremisses are issued between 507–518, the hypothesis is confused by Visigothic disgrace and defeat; loss of the state treasure at Toulouse; 40 and the presence of Merovingians, Burgundians and Ostrogoths in former Visigothic territory. Most important is the loss of two key mints, Toulouse to the Franks (if we accept the thesis of Reinhart and others of coins being minted there) and Arles to the Ostrogoths. Narbonne which is seized jointly by Franks and Burgundians is retaken for the Visigoths by the Ostrogoth, Count Ibbas.

If the Visigothic mints (i.e., Narbonne) initiate this coin after 507, it would have to be by Amalaric after 510, since the retaking of Narbonne is not achieved until 510, and the equivocal-unstable position of Gensalic would not seem to permit a new coin issue. Nor would the issuance of the coin by Gensalic recommend it to Amalaric, Clovis, Gundibald or Theodoric. If hypothetically the coin is issued after 510, it must also have been issued before Gundibald's death in 516, if we accept the identification of his monogram on some of these coins. If we carry this further, such an issue minted between 510–516, by the Visigoths would have had to have been done with the knowledge and permission of Theodoric the Great, who is the regent of his grandson Amalaric and who ruled Spain in his own name until his death in 526. 41

There are too many qualifications yet to be established, and more circumstantial evidence to be considered before we can attempt to approach problems in dating, although they are inseparable from the problem of finding an initiator and prototype. Since the Ostrogoths are inseparable from Visigothic affairs after 510, it is necessary to consider the possible role that they may have played.

It is on the basis of style that VPW tremisses have been attributed to the Ostrogoths. Wroth catalogued two coins as being from Rome because of the "comob" inscription on one and the style of the head on both which coincides with several VGC coins attributed by him to Rome. 42 Wroth also characterizes Ostrogothic and Italian tremisses as possessing a comparatively high relief for types and lettering. 43 The fineness and the purity of the design are also considered characteristic of Italian Ostrogothic mint products. The attribution of VPW tremisses to the Ostrogoths, however, is not in accordance with the VGC majority found in Italy and Italian collections. By the coins extant, one may conclude that, if the type is minted in Italy, it is part of a limited issue, since the VGC tremisses are more preferred and more common.

Furthermore, because of the quasi-autonomous bronzes of Theodoric and their variant precursors, the Zeno-Odovacar bronzes, the attribution of this limited tremissis issue to the mint at Rome is not necessarily rash nor inconceivable. There was no mint in the Roman world with more right to the use of the type, as well as possibly no mint in the Empire more contemporarily connected with the type. The Theodoric substitution at Rome of the VPW for the trophy-bearing Victory on these bronzes might be the first indication of a prevalent attitude of the desire to distinguish coins of the new reign from those of the old. But these issues at Rome are quite out of numismatic character with Theodoric's general policies of "regnum nostrum imitatio vestri." His gold coinage is a relatively faithful reproduction and duplication of contemporary imperial money. 44 Theodoric, always careful to read his king's role as subservient to the emperor's in Byzantium and to maintain traditions, never put his own portrait on any coin nor his name except on the three solidi gold medallion. 45 Consequently, the use of the VPW type on the tremissis attributed to Rome is out of line with Theodoric's policy, a policy which faithfully imitated the VGC imperial tremissis. It would be unnatural then not to consider the scant minting of these for a specific event or purpose.

It would be even more unnatural not to attempt to find a connection between the Anastasius VPW medallion and the Theodoric medallion and the VPW tremissis issue. These may be connected with the reconciliation between Anastasius and Theodoric just before 500, and Theodoric's triumphal six month visit in Rome. These rare commemorative usages of the VPW type distinct from the imperial and Ostrogothic coinage may be significant for the VPW tremissis issues of southern France. However, the Anastasius medallion may not require any particular explanation, since small medallions such as this might have been issued annually.

If the initiative for the West comes from the mint at Rome, if in its products we have the prototypes for the other barbarian issues, the responsibility is as much Roman as it is Theodoric's. Of all the rulers in the West contemporary to Anastasius, Theodoric alone possessed the stature, prestige and connections, and the military, political and personal qualities necessary for the role of "Restitutor orbis Romani." By virtue of himself and his people and his control of Rome and Italy, he had the influence and position necessary for establishing precedent. Only Theodoric possesses the regalia of the Roman emperors. Only Theodoric was to be immortalized in both orthodox and folk history. 46 He was the prototype for Charlemagne and the Holy Roman Emperor. Theodoric, Arian though he was, first met the Pope in his church before his triumphal entrance into the city of Rome. 47

The gold medallion presents Theodoric in all his dignity as Rex and Magister Militum, the protector and guardian of the Roman Senate, the Roman People, and the Roman Tradition. Its legends propagandize this: REX THEODORICUS PIUS PRINCIS; REX THEODORICUS VICTOR GENTIUM. The prestige of Rome seems for the moment reborn, and the old Victory type with a long history of association with imperial gold and silver and senatorial bronzes again appears to declare itself as Roman. It does so on a medallion of the only Goth in the West who was a Roman citizen, who bore the title of "Patricius," and who considered himself to possess imperial blood. 48 Contemporary literary and epigraphical sources prove the high regard the native Italian population had for him and his unquestionably enlightened rule. 49 The prosperity his rule engenders and his vast building and restoration projects cause contemporaries to compare him with Trajan and Valentinian I, curiously, two emperors who made especial use of the VPW type. 50

Aside from the prestige he gained from the honors granted to him by the Roman and the Italian people, Theodoric played an important and paternal role in Western Europe. All of the evidence points to his close cooperation with all kings, even interfering in their affairs in order to maintain peace and the status quo. Here is a statesman. He seems to have conceived the idea of a system of German states in the West, playing the role of a patriarch in a family of kings to whom he was related: the father-in-law of Sigismund and Alaric II; brother-in-law of Clovis and Trasamund; uncle to the wife of the King of the Thuringians. 51 This matrimonial state system is indicative of his prestige in Western Europe and is best illustrated in his correspondence to avert the Frankish-Visigothic War in 507. 52 The moment of his greatest prestige is the victory of his troops and generals in southern France in 510, and his governing as king or regent in Italy, Provence, Narbonne and Spain. That he is the uncontested ruler, no matter in which of his two positions, is seen in his orders to Amalaric's mint workers: ". …. monetarios autem quos specialiter in usum publicum constat inventos, in privatorum didicimus transisse compendium. Qua paresumptione sublata, pro virium qualitate functionibus publicis applicentur." 53

The role of Theodoric in the wars to rid Provence and Narbonne of the Franks and Burgundians is the traditional one of Augustus. He is, as was Augustus through his generals who owed their imperium to him, the restorer of peace and the bearer of all of the fruits of the Pax, Victoria and Felicitas Augusta. Consider all of his measures to re-establish business and trade in Provence, to resuscitate the fortunes of this ravaged land, and to give tax relief to its inhabitants until civilized life had restored itself. 54 There was no other who so deserved or merited the title of "Restitutor orbis Romani," or who deserved the regalia of the precursors of Augustus and more rightly deserved to inherit the Victoria Augusta as a true descendant of Augustus.

Can anyone but Theodoric fit the hypothetical image we have created of the initiator of the VPW tremissis? We have previously determined that the coins were issued as respectable commercial currency of good weight and quality; that there was no intention to have these coins confused or mistaken for Byzantine issues. They were issued to be distinguished from imperial currency yet equally trustworthy and therefore commercially acceptable. They were minted in the West by a ruler or rulers who were allied to Anastasius and acceded, diplomatically at least, to his suzerainty, and at the same moment announced his or their independence. There is an understanding of the commemorative associations of the VPW and the Victoria Augusta with Rome, or its Senate or its emperors, or Octavian Augustus. The initiator is governed by other than religious values in selection of the type, and he may not be orthodox Christian by virtue of this type selection. He or his mint has the prestige to establish a precedent to be imitated by others. In addition the minter must be intimately involved with southern Gaul and with the Visigoths and then with the Burgundians; the coins must have been minted by 516, with the death of Gundibald. Certainly none other than Theodoric and Rome could fit the picture.

Vandalic minting can be either discounted or relegated to imitating the issue established by another ruler. Merovingian minting has all of the qualities of imitating a type of another nation or of continuing a type issued by a mint that has been recently conquered, since the type is rarely issued in Merovingian royal mints and only rarely in the south. The Burgundians who are more familiar with the VGC as are the Ostrogoths, do mint this coin with a royal monogram, a device which may have been used to distinguish their coins from those issued by another nation without monograms. Since monogrammed coins are far smaller in number, we may again see these as parts of a small issue rather than of a major one. Visigothic minting seems most definite and most centrally involved, but several circumstances prevent the acceptance of Alaric II in the rôle of prime mover: (1) his weak personality; (2) the Lex Burgundia's condemnation of his gold coins as being bad; (3) the improbability of Burgundians or Vandals or Ostrogoths imitating his coin; and (4) the difficulty of associating the issue with any specific commemoration. In regard to the Ostrogoths, although the strongest case can be made for Theodoric's participation, we have to justify the fact that Ostrogothic mints by virtue of Italian hoards and provenance predominately issue the VGC tremissis and that in all issues faithfully imitate Byzantine issues. If we accept an Ostrogothic source for these tremisses we must hypothesize either that they were not issued for use in Italy but for export or as part of a commemorative issue in small numbers. If they were issued after 510, a stronger Ostrogothic case can be made. Perhaps then they were issued for their new provinces in Gaul or perhaps Theodoric, as regent, has ordered these issued in or for his Visigothic lands.

A final decision in regard to mint initiator cannot be made until all hypotheses for dating the coins are considered. Nevertheless, the Theodoric-Visigothic association provides a basis for the strongest hypothesis and perhaps the only possible one.

End Notes

Reinhart, Swebenreiches, pp. 151–191, pls. XXXIII–XXXVI.
BMCVOL, p. XVI. Wroth puts forth the judgement of the Comte de Salis as published by Keary, pl. I, nos. 16, 17. This attribution was rejected by Tolstoï, II, p. 212, no. 152, who suggested a Merovingian or Burgundian attribution for these coins rather than a Vandalic. Le Gentilhomme, RN 1943, p. 84 does not attribute any gold coinage to the Vandals. The more recent work of Troussel also questions the Wroth attribution (pp. 153, 157) and is of the opinion that the Vandals never minted gold coins (p. 188).
Friedlaender, Vandalen, p. 6. This thesis is accepted by the majority today. Particularly note the discussion of this question in Troussel, p. 188.
BMCVOL, pp. 10–11.
Ibid., p. 20. These are attributed as products of the mint at Rome by Carson and Kent, see LRBC, p. 63.
Gunthamund (484–496) is the first to do this and he establishes a tradition which extends down to Hilderic who dies in 530. See Friedlaender, Vandalen, p. 5. The most useful text for the history of the Vandals is Ludwig Schmidt's, Geschichte der Wandalen, 2nd ed., Munich, 1942.
Friedlaender, Vandalen, p. 6. Friedlaender quotes from Olympiodori, Corpus Scriptorum Hist. Byz., pars I (Bonne, 1829), p. 46.
BMCB, I, p. IX, nos. 15–17.
Schlunk, Visigodo, pp. 227, 235.
Halphen, pp. 38–40.
See Chart VII.
Tolstoï, II, pp. 213–214, nos. 156–162. Besides the works of Belfort and Prou already cited, the studies of Charles Lenormant are largely outdated but still of interest and some importance. The best and most recent study is that of Le Gentilhomme, RN 1943.
Reinhart, Tolosanischen, p. 123.
Belfort, II, p. 187, no. 2295; III, pp. 285–286, no. 4359. Belfort attributes the Childebert coins to Childebert II, but Prou's attribution to Childebert I seems more logical.
Prou, p. 23, nos. 86–88, pl. II, 19–20; p. 124, nos. 534, 535, 537, pl. X, 1, 2.
Le Gentilhomme, RN 1943, p. 97. Le Gentilhomme describes the quantities of VGC and VPW struck by the Franks as being "d'une maigreur particuliere, dépourvue, toutefois, de la marque de cette déformation géométrique propre aux effigies des espèces frappées dans le royaume de Tolède" He also notes that the fabric differs from the Spanish items.
Prou, pp. xx-xxi. Note Reinhart's publication of the Merovingian Hoard of St. Marguerite à Monneren (Northeast of Metz in German Lotharingia) which dates from the second half of the sixth century. Of the forty tremisses in the hoard, only two are of the VPW type, the rest being VGC. Of these two VPW coins, one (Plate 4, 3) I would classify as Visigothic in Group JAN 2; the second coin (Plate 4, 4) I would classify as a crude variation of JAN 2 and most likely a Merovingian imitation of a Visigothic coin. Reinhart, Merowinger.
Ibid. Also see Le Gentilhomme, RN 1943, p. 97.
Reinhart, Tolosanischen, pp. 123–124.
See coin no. 3 in this Corpus. (Plate I)
Gregory of Tours, II, p. 78 (Bk. II, 28).
Belfort, II, pp. 173–178, 179–182. Le Gentilhomme, RN 1943, pp. 92–93.
Keary, pl. I, 10, 11. Le Gentilhomme, RN 1943, p. 93.
Descriptions of these two hoards can be found in the following: Amecourt, Châlon-sur-saöne; Blanchet; Maillard, Alise; Rossignol; Lenormant.
Mateu y Llopis, p. 129.
Robert, Maurice Tibere, p. 429.
BMCVOL, p. 56, n. 1. See also Robert, Languedoc, II, pp. 14–15. These compare with the material cited by Heiss, p. 76, and Gabriel, pp. 137–140. This monogram should read as MAR, for Gundo MAR, the Burgundian.
Zorita de los Canes. This hoard which was published by the archaeologist J. Cabré y Aguilo is now deposited in the Museo Arqueológico Nacional, Madrid.
Mateu y Llopis, p. 47. This is quoted from an unpublished manuscript generally incorporated in the Historia de España desde la invasión de los pueblos germánicos hasta la ruina de la monarquía visigoda.
Jordanes, XXV–XXVIII.
Ibid., XXX.
Orosius, VII, 43.
Bury, Invasion, p. 66.
Jordanes, LVI, 29.
Hodgkin, Theodoric, pp. 106–107.
Procopius, Bell. Goth., III, XXXIII. Also see Robert, Maurice Tibere, pp. 423–424.
Orosius, VII, 43.
Reinhart, Tolosanischen, p. 117.
Gregory of Tours, II, 27.
Bury, Invasion, pp. 205–206.
BMCVOL, p. 56, n. 1.
BMCB, I, p. 74. Also see Friedlaender, Ostgothen, p. 24, pl. III, 5.
Theodoric did not always follow exactly Byzantine patterns. In his solidi issues he ignored Anastasius' changes, the removal of the infulas of the diadem and the substitution of a chrismon for the traditional long cross. Later, when Justin I introduced the facing Victory-Angel on the solidus, Theodoric retained the Victory-Angel standing to the left. In both cases he maintained the traditions of the Italian mints. See Le Gentilhomme, RN 1943, pp. 86–87.
BMCVOL, p. XXXI. Theodoric did put his monogram on some gold and silver issues. BMCVOL, p. 46, no. 3.
Hodgkin, Italy, III, p. 341. Also see Hodgkin, Theodoric, pp. 371–372, 426; Ensslin, pp. 341–354.
Anonymi Valesiani, p. 324, no. 65.
Cassiodorus, IV, 1, 2. Also see Bury, Invasion, pp. 189–195.
Hodgkin, Theodoric, p. 242. Also see Anonymi Valesiani, p. 322, nos. 59, 60.
Anonymi Valesiani, p. 324, nos. 65–72. Also see Cassiodorus, II, no. 39.
Anonymi Valesiani, p. 322, no. 63. Also see Hodgkin, Italy, III, p. 356.
Hodgkin, Italy, p. 356.
Cassiodorus, V, 39. Also see Reinhart, AEA 1945, p. 221.
Cassiodorus, III, 32, 38, 40–44.



An understanding of Roman provincial life in both Gaul and Spain in the fifth century is necessary in order to establish the year and attendant circumstances of the initial striking of the VPW Anastasius tremissis. Increasing investigations in this "chaotic" period clarify numismatic developments. There are still too many lacunae in our numismatic information. If we relate our numismatic material to other known historical and cultural data, a more complete picture is possible. An image can be formed then of the barbarian mentality, its modus vivendi, its attitude towards the government and citizens of Rome, and its interpretation of its status within the Empire.

A cursory reading of a few sources modifies the extremist attitude that chaos is the character of this age. It would be naive to contend that the barbarian invasions are not accompanied by the usual physical and spiritual destructive consequences of war. Chaos as a state of physical reality is a momentary condition; it is more real and more lasting as a state of mind. Fifth century society moves as rapidly as it can to restore the order of daily civilized life. Its intellectuals almost immediately embroil themselves in the search for a scapegoat, pagan or Christian depending on one's affiliation, in order to preserve the core of their moral, religious and political beliefs. 1

The invasion of these tribes was, properly speaking, a resettlement. This amounted to a change in the make-up of the ruling class, with its old members retaining their class status but having to make room for these new recruits. The settlement naturally forced their relinquishing some land and wealth. What plundering occurred, and it unquestionably did, was sporadic enough for Sidonius Apollinaris to describe almost uninterrupted rococo country life for the landed gentry. 2 For him, the barbarians are a new force to contend with in the already complex life of the Late Empire. The experiences of Paulinus of Pella are quite to the contrary, but Sidonius implies that such hardships were not generally or universally suffered. 3 The tragedy of the barbarian invasions was that they compounded rather than abolished the evils of the old system, too well maintaining the conditions of provincial life.

Although such periods of confusion occur, they are short-lived when the invaders are intent on settling. Government agencies reopen and reorganize with most likely the same staff, allowing for the possible loss of some of the staff due to the exigencies of death or flight. The advisors to the Roman governors and prefects now become the advisors to the barbarian rulers, e.g., the orthodox Catholic and "jurisprudens" Leo is chief advisor to Euric; the Elder Cassiodorus serves both Odovacar and then Theodoric. The Roman provincial jurists are certainly the inspiration behind the Lex Burgundia, the Breviarum Alaricum, and Lex Visigothorum. The clergy and the provincial senatorial class still travel to Rome, maintaining the old ties. The barbarian kings must encourage rather than impede these connections. They do not come into the Empire simply to maraud; they come to live and partake of the comforts of this civilization. Although credence may be given to Keary's view of a slow undermining of society, the gradual pauperizing of old inhabitants, the sapping of all industries and the reversing of all conditions favorable to trade; 4 the barbarians cannot bear the full burden of responsibility for the economic retrogression that was already in the offing before their resettlements. 5

We have enough numismatic evidence, although still incomplete and not catalogued, to prove that at least some imperial mints remained open or reopened once peace was restored and were staffed most likely by the same hereditary class of mint workers. 6 Barbarians did not enter the ranks until long after the invasion. 7 There are large numbers of coins in the names of fifth century western and eastern emperors that by means of style and provenance cannot but be attributed to the imperial mints of Gaul, e.g., those at Lyons, Arles and Narbonne. 8 There is even evidence that new mints were organized such as those of Lugo and Braga by the Suevians as early as 420(?), in Spain 9 and that of Toulouse by the Visigoths. 10 These coins are mostly pure imitations of imperial issues distinguishable only by a lessening of imperial mint standards in regard to design and execution, which, as Keary suggested, was due to fewer controls imposed on old moneyers by less critical employers as well as an increasing number of inexperienced hands. 11 With the weakening of a central numismatic control, the increasing number of authorized and non-authorized private minters would aggravate the situation of declining standardization. We should not expect to find a different situation developing in numismatics than what has been evident in the monumental arts since the Age of Constantine. The aesthetic value system of the empire itself is changing regardless of the barbarian influx. 12 From the numismatic evidence, it appears that the greater the distance in time separating mint operation from the last imperial issue and the control of the imperial moneyer, the further the mint product strays from the aesthetic standards of the old empire. This is revealed succinctly when the stylistic development within the VPW tremissis is traced from the issues of Anastasius to those of Leovigild.

An analysis of the sixth century tremissis depends upon an understanding of the monetary circulation in the fifth century. Le Gentil- homme presents the most recent and best survey. 13 The striking of gold dominates the activity of the western mints. Silver is issued in decreasing amounts, while bronze is rarely coined, except in considerable amounts in Ostrogothic Italy. Barbarian love for gold would have demanded and encouraged the continued minting of gold and should have necessitated a continuance of silver and bronze for normal commercial needs of the provincials. The latter however does not seem to be the case as evidenced by extant coins. The probable diminishing volume of trade in the West and its trade with the East would place a greater demand on the local market and the dealing in smaller denominations. The increasing self-sufficiency of the landed estates would aggravate the commercial crisis as well as remove the necessity for large volume and long distance operations. As the century progresses and this economic situation is intensified, there is a diminishing need for the larger coin denominations. By the end of the fifth century, the solidus is gradually being superseded by the tremissis. 14 Such a market would seem to predicate the maintenance of bronze and silver issues, but such issues are of decreasing interest for the barbarian kings. Only the Ostrogoths maintain a large coinage in all three metals for the commercial needs of the urban communities of Italy, just as the silver issues of the Vandals and the numerous small bronze issues from North Africa cater to their large urban population. 15 The rest of the West manages with coins already in circulation, Ostrogothic issues, and the intermittent issues of bronze and silver by Burgundians and Franks.

The small amounts of fifth century gold extant in comparison to extant gold of the fourth and sixth centuries must be accounted for by the barbarian hoarding of gold as coin or bullion, the melting of it to fashion ornaments, or the recoining of it in order to pay taxes and tributes, and by the burying of it by provincials. 16 This scarcity is already revealed in the developing trends of the fourth century Empire where the increasing difficulty in maintaining a satisfactory gold coinage forced the imperial mint to demand a return of gold in bullion form. 17 There must have been large amounts of tribute gold in barbarian treasuries, nevertheless, not considering the amount reaped from plunder. The Visigothic treasure in literary sources attains fabulous and fictional heights. We are told that Alaric in the Sack of Rome not only carries off the Treasure of Solomon brought there by Titus, but also is conceded four thousand pounds of gold by Stilicho in the spring of 408, and five thousand pounds of gold and thirty thousand pounds of silver by Honorius in the autumn of that year, as well as being given thirty-five thousand pounds in precious metal by the Roman Senate. 18 This situation is repeated often throughout the fifth century. Besides tribute money there is ransom gold. The Ostrogoths, e.g., ca. 495, sent gold to Gundibald for the release of Italians held captive since Burgundian raids in North Italy ca. 490, during the war between Odovacar and Theodosius. 19 There would seem then to be more than enough gold available for the minting of coins for particular needs.

If it is true that the old Roman system and traditions do survive and continue through the fifth century and the evidence of earlier barbarian minting of gold is irrefutable, then we must accept the possibility that the VPW tremissis might have been struck from the very beginning of Anastasius' reign in a.d. 491. All the barbarian kings are minting coins with the exception of perhaps the Ostrogoth who does not succeed in Italy until 493. 20

Since there is evidence of Roman mints striking from the same die designs after an interval of a century or so, 21 it is not inconceivable to find the practice of restriking of old dies, or copying of old die designs increasing in the less controlled barbarian mints. With even a tendency for a new mint to imitate whatever is available to copy, 22 it is not surprising to have to contend with the problem that a mint might often strike a coin of a deceased emperor. Since there is evidence of the Ostrogoths indulging in this fiscal practice so as to avoid placing the name of their adversary, Justinian, on their coins, it has been suggested that the Visigoths in Spain did the same during the Byzantine reconquest attempts. 23 This was not the case in Visigothic Spain, as will be shown later. Here they took to jumbling the legend so that it became unintelligible. There is no stylistic or other evidence to indicate that all Anastasius coins do not date from the reign of Anastasius. A stylistic study will reveal that the coins of a given emperor in the VPW tremissis series were minted during the emperor's reign or up to the time when the news of the new emperor's accession was known.

The initial dating of this issue must be within the period of 491 to 518. Previous hypothetical conclusions encourage the connection of this issue with an event which affects the lives of the Visigoths and Theodoric. Although the initial role of the Burgundians has been questioned, the possibility should not be ignored. Both Burgundians and Visigoths could have minted this coin as soon as the news reached them of Anastasius' accession in 491, but what would have encouraged them not to continue their imperial imitation and traditional local issues which they had struck previously during the reigns of Leo and Zeno? 24 One could consider numerous dating possibilities within the reigns of Alaric II and Gundibald for a time when commercial or political expediency would necessitate a new coin issue.

If the Burgundians are considered responsible for the striking of these in their mints before 500, when the four sons of Gondovech share the kingdom, it might be attributed to a desire for a unified currency and a selection of type based on die traditions at Lyons, the major capital. When Gundibald unites the kingdom under his own rule ca. 500, the use of the monogram on the reverse field may have been inaugurated on these tremisses, or the monogrammed coins may have been issued at the same time as the non-monogrammed coins in order to distinguish those minted by Gundibald from those of his brothers. Gundibald's monogram series, however, would seem more appropriately issued after 500, when his hegemony would permit and demand the distinctive addition of the personal monogram. It is difficult to conceive of the Burgundians assuming the role of initiator after 507, of a coin also minted by Visigoths and Ostrogoths. Burgundian victories over the Visigoths would seem too short-lived to be so amply commemorated. Since there are a large number of VGC tremisses assigned to the Burgundians, it would seem more probable that the VPW tremissis was not so preferred and was a secondary or parallel issue. This might imply further that it was minted later than the VGC tremisses and therefore might not have been original with the Burgundian mints. 25

A dating for a Visigothic initiation does not seem more likely. If issued before 507, it is issued by Alaric II, but several reliable sources condemn Alaric for issuing debased coinage, and this might possibly (but not necessarily) remove him from consideration, because the VPW Anastasius tremisses are of good quality. 26 Heiss denied the possibility of Alaric's minting this tremissis on the basis of this evidence, 27 but the possibility that this condemnation might apply only to some coins rather than to all must be allowed.

Certain events or policies during the reign of Alaric might predicate the adoption of the VPW type. Possibly it might have commemorated the alliance of the Visigoths and Theodoric and thereby the Emperor Zeno against Odovacar around 490, the victorious outcome of which saw the restoration theoretically of Rome and Italy to the Empire in 493. 28 Yet Zeno's successor, Anastasius, did not look favorably on Theodoric or on Zeno's alliance with Theodoric and was not to recognize Theodoric until 498, so that such an issue for this reason in the name of Anastasius would not seem likely. 29

Besides this there is no other event commemorable in Alaric's reign except for his war with Clovis. Could this have been a propaganda issue in preparation for the war? It would not be unusual for Alaric to see his rôle of maintainer of Roman culture in Gaul in the tradition of his more stalwart predecessors whose panegyrists saw them as "Restitutores orbis Romani." The rising threat of Clovis and the Franks might have encouraged an issue, rife with military import, which announced the legal right of the Visigoths because of their alliances with the Roman emperor to settle in Gaul. The magic of the Victoria Augusta was truly on their side. Thus might Alaric also have reminded his orthodox Catholic subjects that he and he alone had a legal right regardless of his religious affiliation. Alaric could have ordered the striking of the VPW tremissis anywhere between 500 and 507, when Clovis' intentions began to be obvious. It might have been completely in character for Alaric to do this, which for all its import might be an ineffectual device. Alaric did not inherit his father's (Euric) force of character and aggressive martial attitudes. His policy towards Clovis was that of a civilized conciliator. He did not resort to force to prevent treasonable practices of the provincial population and its orthodox church, nor to stop the advance of the Franks. He sent ambassadors and suggested diplomatic conferences. Even the preparation of the Breviarum Alaricanum, an abstract of the Imperial Code, may have been a conciliatory measure for the loyalty of his Roman subjects. This was also ineffectual for his purposes and too late. The Visigoths were already too softened by civilization, too Romanized to deal successfully with a true barbarian like Clovis.

Mateu y Llopis, Reinhart and Le Gentilhomme accept the hypothesis that Alaric II minted the VPW tremisses, and therefore begin dating these coins before 507. Mateu y Llopis considers Alaric II's Anastasius issues to come from southern Gaul, from the mints of Bordeaux, Toulouse and Narbonne. 30 Reinhart agrees with this but goes further in attempting to find reasons for the distinct usage of the type. He considered that the Visigoths adopted the VPW in the time of Alaric II for religious reasons, since he thinks that the VGC would have offended their Arian faith. This statement implies that the globus cruciger was associated with the specifically orthodox creed sanctioned by the Byzantine emperor, since it was so connected with Byzantium. Reinhart suggests that the VGC type was offensive because it symbolized the victory of the orthodox creed over the Arian. 31 There is nothing to justify such a conception. If it is true then the Visigoths were alone in so thinking, for the Arian Ostrogoths and Burgundians did not find the VGC type obnoxious to their belief. This hypothesis seems to be contradictory to another of Reinhart's suggestions concerning the origin and adoption of the pectoral cross on the Visigothic obverse bust. He dated the Anastasius coins with the pectoral cross after 507, issued by Clovis who, possessing both the treasure and mint of the Visigoths at Toulouse, commemorated his victory over the Arian Visigoth. 32 Reinhart conjectures that the Visigoths adopted the pectoral cross initiated by Clovis in accordance with their typically Roman tolerant view, and we must assume that they were no longer prejudiced against the use of the symbol. In this way the pectoral cross coins were issued from Toulouse until the death of Anastasius. Reinhart attests to the Merovingian association of these on the grounds of provenance; pectoral cross coins are found more frequently in France, whereas he knew of only six in Spanish collections. 33 This is a non sequitur, since most Anastasius coins, pectoral cross or no, come from southern France, the site of the major Visigothic settlement. The move to Spain does not begin until after 507, and with the capital at Narbonne a large concentration of Visigoths would still reside in southern Gaul, Narbonne at least.

French provenance for this coin does not prove that it was any more Merovingian than Visigothic. Reinhart's hypothesis might seem a logical explanation of the manner in which the pectoral cross makes its possibly first appearance, but others would seem more appropriate. There are a few coins in which the advent of this cross is haphazard, rather than originally conceived with the rest of the design. Its presence seems due to an afterthought or correction or addition directed by a new policy. Note that No. 3 (Plate I) shows that some of the original die has been altered by the scratching away of some of the folds to make way for the little cross which appears here as a jewel. Would it not seem incongruous for the Visigoths who, according to Reinhart, did not use the VGC type because it offended them, to adopt a symbol initiated by Clovis to commemorate their own ignominious defeat? There must be a better explanation for a tradition stylistically developed and maintained by the Spanish Visigothic mints. Could it have been added to ameliorate the pagan reverse for the orthodox Christian subjects and therefore give the VPW tremisses a Christian guise? 34

There is no factual evidence or even substantial circumstantial evidence to enable us to establish an initial date before 507, and the death of Alaric at Vouillé. This is true even when we turn to consider Ostrogothic possibilities. Before 500, Theodoric's political prestige could not have been so great as to have Visigoths and Burgundians imitate his coin types. The western nations must have begun to realize the force of this statesman when he received the "ornamenta palatii" of the Roman emperor from Anastasius in 498. With the recognition of Theodoric by the emperor as his agent in Italy, Theodoric's prestige is enhanced and particularly distinguished from all other barbarian rulers of older and equally legal national positions. Of them all, Theodoric alone is the symbolic and actual inheritor of the Roman Augusti. 35 Theodoric's first visit to Rome is celebrated as an Augustan triumph and may even coincide with his Tricennalian feast of 500. 36 The three solidi medallion may very well communicate Theodoric's acceptance by Anastasius and his triumph in Rome. Thus the VPW tremissis issue might also have echoed this, minted at Rome, in what would have had to have been a limited edition, and continuing as it does one of the types found in the quasi-autonomous bronzes. The change in type between the quasi-autonomous bronzes of Zeno-Odovacar and the possible Theodoric Invicta Roma from a Victory with wreath and trophy to a VPW on a ship's prow before a lighted altar would symbolize the change in Theodoric's status from a Victor to a "proxy-Augustus." From this moment, the paternalistic rôle Theodoric plays in the West would be in character with the position of a "proxy-Augustus." It is very dubious, however, that the other barbarian rulers conceived of Theodoric in this way. His threats stopped neither the Burgundians nor the Merovingians from their attack on the Visigoths. 37 His assumed role of "proxy-Augustus" did not take on actual reality until his armies in southern Gaul so quickly proved that they could support his theoretical position. With the Ostrogothic victories in Provence and the Narbonensis, and his assumption of the Visigothic regency, Theodoric becomes the outstanding figure in the West. The Vandals are very quick to respond to his chagrin at their support of Gesalic, who is immediately abandoned by all in favor of the grandson of the new Augustus. 38 Clovis, the only possible competitor, with his armies had shown too clearly to the Romans in southern Gaul that he was not the bearer of either Pax or Felicitas Augusta for all of his being an orthodox Christian. Even though Clovis is named Roman Consul in 511 (consul suffectus not consul ordinarius) by Anastasius, and clothed in the purple tunic and mantle at St. Martin in Tours and given a triumph there, it is an ineffectual ruse on the part of Anastasius. Theodoric controls Italy, Spain, the Narbonensis and Provence. 39 What advantage this honor might have developed for Clovis was negated by his death in the same year.

It is only after 510 then, that Theodoric's prestige has the force to influence the West, numismatically or otherwise. He controls most of it. Only the Franks are a contending force, and they do not control any international commercial center. This is important in considering the impossibility of a Merovingian minted coin influencing the issues of other nations and achieving an international character. On all counts Theodoric alone is in a position to authorize the issue of the VPW tremissis soon after the establishment of peace in southern Gaul in 510. He merits this by virtue of his political, cultural, military and geographical position and his great personal prestige.

This hypothesis is supported by all the evidence accumulated so far, since only with its adoption does the evidence fall into a logical interrelating pattern. Once his generals are victorious in Gaul, Theodoric's sense for organization immediately takes over. Uppermost in his mind is the rehabilitation of these areas that have been ravaged by the recent occupation of marauding fighting armies. A large number of letters in Cassiodorus' file indicate forcibly the rapidity and efficiency of Theodoric's action. He immediately appoints Gemellus as Governor of Gaul and marks him to "… show yourself in all things such a governor as a 'Romanus Princeps' ought to send, and let the province feel such an improvement in her lot that she may rejoice to have been conquered," 40 He informs Gemellus that promptness and integrity are required in carrying out his orders 41 and is attentive himself to appeals from Gallic citizens. 42 The directives are the most telling evidence: "…… our Piety wishes that here should be order and good government everywhere in our dominions but especially in Gaul, that our new subjects there may form a good opinion of the ruler under whom they have come. Therefore by this authority we charge you to see that no violence happens in Avignon where you reside. Let our own live 'civiliter' with the Romans, and let the latter feel that our troops have come for their defense, not for their annoyance." 43 "Let other Kings desire the glory of battle won, of cities taken, of ruins made; our purpose is, God helping us, so to rule that our subjects shall grieve that they did not earlier acquire the blessing of our dominion." 44

In his proclamation to his new Gallic subjects: "Obey the Roman customs. You are now by God's blessing restored to your ancient freedom; put off the barbarian; clothe yourselves with the morals of the toga; unlearn cruelty, that you may not be unworthy to be our subjects …… Do not dislike the reign of Law because it is new to you, after the aimless seethings of the Gentilitas (barbarians) …… You may now bring out your long-hidden treasures; the rich and noble will again have a chance of suitable promotion. You may now enjoy what till now you only heard of — the triumph of Public Right, the most certain solace of human life, the help of the weak, the curb to the strong. You may now understand that men are exalted not by their bodily strength, but by reason." 45

The authority and dignity of these lines herald the coming of the Victoria Augusta. They are spoken not by a barbarian but by a veritable Augustus guided by a great, true Roman, Cassiodorus.

Theodoric's sincere intentions for reconstruction and rehabilitation cannot be questioned. He immediately releases from taxation the provincials in Gaul for the period from September 510 to August 511. 46 He concerns himself with the possibility of famine in this area, where as he says: "How can one claim taxes from the lord of a field where one knows he had not been able to cultivate it?" 47 and encourages the shipping of provisions to Gaul by preferential treatment to shipowners. 48 His policies moved to restore the commercial, industrial and agricultural activities of this province as quickly as possible, even if it demanded contributions of his own money, as in the letter to all landowners of Arles: "We wish to refresh men, but to repair cities also, that the renewed fortune of the citizens may be displayed by the splendor of their buildings. We have therefore directed that a certain sum of money be sent for the repair of the walls and old towers of Arles." In this letter he also informed the people of Arles that he had ordered grain to be shipped to them. 49 In another letter he orders that all run-away slaves in Gaul be restored to their masters and thereby sets right one more element in the confusion caused by war. 50 In his letter to Gemellus announcing that the "siliquaticum" would not be levied on corn, wine and oil, he states his reasons: "We hope thus to stimulate trade, and to benefit not only the Provincials, who are our chief care, but also the merchants." He also releases all provincials from paying the expenses of the garrison army: "We will transmit to the Duces and Praepositi sufficient money to provide 'alimonia nostris Gothis'." 51

The outlay of money must have been substantial in the first years of peace (510–512), as they must have been during the previous war years, and an order such as the following does not come as a surprise: "It is the part of true prudence to recall to the uses of Commerce the talent hidden in the earth." 52 When grave digging is authorized to reclaim gold for the state treasury, gold must be scarce and in demand. Gold is needed in the Gallic province to help revive prosperity and peace; to help maintain border garrisons and to pay the soldiery; to cover the loss from the remission of a year of taxes and to pay for the usual expenditures of government. A new monetary issue is distinctly in order, and what could be more appropriate than the VPW tremissis, a perfect commemorative issue for this situation; a most appropriate type to grace a coin that would be utilized for the military payroll. What is necessary for the Gallic province also must have been necessary for the Visigothic territory, for none of the contemporary chronicles consider anyone other than Theodoric as ruler of Spain and the Narbonensis. 53 The same attention to Visigothic lands is indicated in the one extant letter usually dated 523–526, in which he orders his officials to remedy a list of disorders that have been brought to his attention. 54 There is no aspect of government in which he is not in control, and his note to Duke Ida to re-establish certain possessions of the church of Narbonne given to it by the late Alaric II, 55 shows how he maintained control over his officers and nobles even in the Visigothic Narbonensis.

If some or all of Alaric's gold was of questionable quality, there would be even more need to begin the new peace with a new gold issue of good quality. It is interesting, if this hypothesis is correct, to question further why Theodoric would not have followed the same fiscal policy in Gaul and Spain that he had in Italy by imitating Byzantine types. Theodoric always conducted himself as King of the Goths and Regent of the Emperor in Italy, observing the limitations which had been laid down by imperial authority. 56 He was always careful to issue only edicts and not laws which were imperial prerogatives. 57 Under him the civil service in Italy remained constant with only Romans permitted in it. It is conceivable that he did the same in the new provinces. 58

Consequently, Theodoric never put his name on a coin with the exception of the gold medallion and he put his monogram only on some issues. The adoption of Theodoric's policy at Lyons as well as at Narbonne seems probable, not only because of his fiscal policy but also because of the traditions of the Italian mints, the major source of late fifth century gold. Furthermore one of the major pagan centers and imperial mints of the fifth century, Arles, was in his territory, and dies could have been sent from Italy to be copied there if older die designs of the type were no longer in their possession. 59 But if his policy at home was one of relative imitation and subservience to the Byzantine emperor, why was this not followed in the provinces? Is it because in the provinces he and he alone was the power, while in Italy he acted in the emperor's name? The provinces were responsible to him and he to the emperor.

Just as Octavian Augustus and his successors were responsible for the triumphs of their generals through the Victoria Augusta so was Theodoric for the victory in Gaul, achieved for him and through him by his generals. The victory of Theodoric's generals also referred back to the Byzantine emperor by virtue of his inheritance of the Victoria Augusta. For such a glorious victory accompanied by all the guarantees of the Victoria Augusta, the issue of a new coin with the distinctive VPW type was highly appropriate. It heralded the return of the authority of Rome, now an actuality rather than an illusion of a glorious past. It symbolized the hope for all of the Roma Aeterna. The minting of this coin would have been as appropriate at this time in the Gallic and Spanish mints as it had been in the western mints in the past, when commemorating past and hoped-for future victories of the armies garrisoned in the surrounding areas.

The large quantities issued and their good quality might have forced the Burgundians, for commercial reasons, to mint the same with the distinctive addition of the royal monogram. Those issued by the Ostrogothic Gallic mint would need no distinguishing mark to identify the coins knowingly issued by them. This might explain all monogrammed coins as being other than those issued by Ostrogothic mints in Gaul. The only monogrammed coin attributed to the Visigoths is the image series. Heiss, Mateu y Llopis and Reinhart read this as Amalaric and consequently a product of the mint of Narbonne. 60 Amardel considered this monogram to signify Narbonne and consequently of Visigothic origin. 61 Robert, however, remembered that the Count de Salis had considered these coins to be an Ostrogothic issue struck at the time of their domination of Provence. 62 Le Gentilhomme's thesis that these are Burgundian and the monogram is that of Gundomar seems more acceptable. 63 If the monogram does signify Amalaric, it would have to be assumed that the coins were struck at Narbonne, possibly late in the reign of Theodoric when Amalaric had attained his majority, in order to distinguish the issues of Amalaric from those issued under the regency of Theodoric. Amalaric selects a device perhaps first used by Theodoric to distinguish his bronzes from Byzantine ones. Burgundians later used this device to distinguish their coins from Theodoric's, 64 since in the mind of Theodoric the Burgundians and especially Amalaric were, relatively, in a position to him as he was to the Byzantine emperor. Amalaric was the King to Theodoric's Augustus as Theodoric the King was theoretically to Anastasius, Augustus and Emperor.

There are three possibilities for explaining the origin of the VPW tremissis. The first is that Theodoric and his advisors order a new issue of gold tremisses for the Gallic and Visigothic provinces to aid in their commercial rehabilitation as well as for the payment of the garrison troops. The first coins might have been designed in Italian mints and shipped to Gaul, 65 if, in accordance with old imperial practice, the dies were sent from Rome or another Italian mint along with the bullion necessary for striking the coins; or because of the possible temporary closing of the mints at Arles and Narbonne during the recent war. These coins were then imitated by the mints at Arles and Narbonne on orders from Theodoric, once these mints were able to function. This may account for the few extant coins of this type stylistically attributable to the Ostrogoths. The minting of these at Arles and its commercial importance for the Rhone Valley, Lyons and northern France might certainly have established the precedent for the Burgundian issues and the source for some of the coins in the Alesia and Gourdon hoards. The spasmodic and unsystematic Merovingian issues would be accounted for as random imitations of Burgundian and Visigothic-Ostrogothic issues.

A second possibility is that, when the gold issue is ordered, those to be issued for Visigothic territory and by Visigothic mints are designed to be the VPW tremisses, thereby distinguishing the coins issued by the two different categories of Ostrogothic controlled mints, the VPW for Visigothic and the VGC for Ostrogothic in Italy and Gaul. If this policy was followed, then it would imply that the Burgundians issue a coin similar to those issued by Visigoths under Ostrogothic control, thereby distinguishing their coinage from the more orthodox VGC Ostrogothic coins from Arles and North Italy. Burgundian adoption would not be imitating the Visigothic pieces or have any implication of their subservience to Theodoric. The VPW Visigothic-Ostrogothic issue could have set an admirable model and precedent for them, since they might be asserting their own victory of 507. The type was common to the mint of Lyons and therefore natural for them to issue again. The Merovingians would still be imitating these coins and continuing to strike them in former Burgundian or other southern Gallic mints that had done so prior to their occupation.

The third possibility has been verbally suggested by Mr. Philip Grierson, that there was a limited issue of the VPW tremissis struck in Rome, perhaps concurrently with the Theodoric medallion, which then was used as a deliberately non-Byzantine type in the Visigothic Kingdom. The type appears in Rome, not only on the medallion and the tremissis, but also on a few extant silver coins which bear the name of Anastasius on the obverse and the VPW and INVICTA ROMA on the reverse, with the letters SC in the field. 66 There is another issue in silver similar to this with the exception of the use of a trophy-bearing Victoria. 67

Other evidence may substantiate this hypothesis. The conception of the VPW type for the tremissis was in accordance with a developing tradition in Ostrogothic mints. This was also the conception of the Roman Senatorial aristocratic advisors of Theodoric who never lost contact with the Augustan Empire, pagan or not. The possibly commemorative circumstances surrounding the issue of the quasiautonomous bronzes, the medallion and the tremisses of Theodoric are not unlike the associations that a Roman such as Cassiodorus would again reassert on the unusual Roman bronze series of Theodahad. The king's portrait with tell-tale pectoral cross appears on the obverse and a VPW on a prow in the reverse. Theodahad's entry into Rome in 536 is associated with Theodoric's triumphal entry into Rome in 500. The legend of Victoria Principis even seems to commemorate further Theodoric with Theodahad's hopes that as Theodoric had inherited by personal force the Victoria Augusta now maybe he can inherit the right to Victory from Theodoric, the Victoria Principis. 68 The aged Cassiodorus and the Senate perhaps have resorted to an old idea, since the conception of all of these coins has the quality of an associated policy. The pectoral cross may be even more telling here. It is nowhere else to be found aside from the VPW tremisses and the Justinian Carthaginian bronzes.

We should not underestimate the importance of a symbol, especially that of the Victoria with palm and wreath. She, who was claimed by Octavian Augustus as his very own and who was thereby inherited by all of his descendants, was to grace this coin possibly conceived by Theodoric and his advisors. The Victoria Augusta, the Victoria with palm and wreath, was such an integral part of the politics and religion of the Roman empire, that the last pagan-Christian Romans could not create their new Augustus image in Theodoric without it. So much hope was there in the success of this barbarian king, who better than many before him played the rôle of Augustus, that they had to declare him a descendant of the immortal Octavian and therefore a possessor of Imperial blood. 69 The hope of a new Augustus and the new restored empire in the West was not to outlive him. With Theodoric dies the last hope of maintaining institutions of the old empire intact. The later empire of Charlemagne, the true successor to Theodoric, is a totally different world. The last appearance of the VPW in Rome is an unfulfilled wish with the hapless Theodahad. When the VPW is retired from the currency of Leovigild, it is because she is no longer a symbol of the power of Rome. The only power in Rome is the Pope, and his symbol is the cross. The VPW on the few late issues of the Merovingian Childebert II, according to the Belfort attribution, are anachronistic freaks.

End Notes

The best comprehensive discussion of this is to be found in Dill, Roman Society, pp. 285–291. Also see the writings of Sidonius Apollinaris, Paulinus of Pella, Flavius Felix Corippus, Procopius, Jordanes, Orosius and Cassiodorus. The best recent study is Courcelle's Histoire Litteraire des Grandes Invasions Germaniques.
Sidonius Apollinaris, I, pp. LXII-LXV, and LXIX; III, XII, 6; Bk. VIII, iv, I; Bk. VIII, xii.
Ibid., Bk. VII, vi, 6; Bk. V, vi, 7.
Keary, p. 71.
The best accounts of this are to be found in Dill, Roman Society, and in Dalton's Introduction to his edition of the Letters of Sidonius Apollinaris.
Keary, pp. 53–54. See also Prou, p. XIV; Robert, Maurice Tibere, pp. 423–424, 429; and Le Gentilhomme, RN 1943, pp. 56, 75, 92–93.
Theodoric's order for better treatment of barbarians implies that few if any barbarians live in the cities. See Cassiodorus, Bk. V, 39, item 13.
Reinhart, Tolosanischen, pp. 108, 117–118; Jahn, II, p. 13; Le Gentilhomme, RN 1943, p. 93.
Reinhart, Swebenreiches, pp. 154, 166–167.
Reinhart, Tolosanischen, p. 107.
Keary, p. 53.
Ibid., p. 66.
Le Gentilhomme, RN 1943, pp. 58, 89. There is an absence of tremisses in fifth century hoards, but in those of the time of Justin I and Justinian the tremissis predominates.
Ibid., p. 92.
Le Gentilhomme, RN 1943, p. 91.
Keary, pp. 71–72. Also see Kent, p. 197.
Kent, pp. 203–204.
Ulrich-Bansa, pp. 172–173. Also see Grierson, Commerce, for further ramifications of this problem.
Ennodius, Sec. 154–164, pp. 95–99; Sec. 171–173, pp. 101–103.
Reinhart, Tolosanischen, pp. 108, 110, 116–119. Also see the studies of Kraus and of Friedlaender.
Grant, Anniversary Issues, pp. 51–52. Also see Vermeule, Num.Circ.1952, pp. 356–367· An indispensible study is Vermeule, Ancient Dies.
Keary, p. 67, pl. 1 and Ulrich-Bansa, p. 170. In regard to the character of the imperial portrait see Pearce, Num. Review 1946, pp. 125ff.; RIC IX, pp. XL.
Kraus, p. 29.
Le Gentilhomme, RN 1943, pp. 74, 86, 92, and pls. I–IV.
Ibid., pp. 92–93. Also see Lenormant.
Mateu y Llopis, p. 135; Pérez-Pujol, IV, p. 473; Avitus, LXXXVII, Lex Burgundia, Constitutiones Extravagantes, XXI, 7, p. 342.
Heiss, p. 75.
Anonymi Valesiani, II, 53, p. 316; Hodgkin, Theodoric, p. 121.
Anonymi Valesiani, 64, p. 322.
Mateu y Llopis, pp. 133–140.
Reinhart, Arte, p. 54; Reinhart, Merowinger, pp. 40–41.
Reinhart, Tolosanischen, p. 128. Also see Reinhart, Merowinger, pp. 40–41.
Reinhart, Tolosanischen, p. 124. Reinhart associates the nonpectoral cross busts with the coinage of Alaric II, see Reinhart, DJN 1940, p. 75.
The presence of the pectoral cross on Visigothic currency is quite in keeping with the important position of the cross in Visigothic art. A Maltese cross form is the major religious and decorative symbol throughout the Visigothic period. It owes its dominance to the iconoclastic tastes of the early Spanish Church. Its regal associations on the votive crowns of Reccared, Reccesvinth and Swinthila suggest an old tradition. The number of pendant crosses found in the treasures of Guarrazar and of Torredonjimeno encourages one to suggest that these may be part of the Visigothic kings' regalia. It has never been satisfactorily determined whether or not these votive crowns and crosses were ever worn. The tradition of offering crowns must have begun with Leovigild, since it may be one of the Byzantine customs he introduced. However, I have never been able to find any record of a pectoral cross as being part of the Byzantine emperor's regalia. See Schlunk, Visigodo, pp. 311–313; Delbrueck, Kaiserornat.
Ensslin, p. 82. Odovacar had sent the regalia to Zeno, see Cassiodorus, Chronica, pp. 158–159; Anonymi Valesiani, no. 64, p. 322.
Kraus, p. 79.
Cassiodorus, Bk. III, Letters 1–4.
Ibid., Bk. V, Letters 43–44.
Hodgkin, Italy, p. 351.
Cassiodorus, Bk. III, Letter 16.
Ibid., Bk. IV, Letter 21.
Ibid., Letter 12. As a further illustration of this consider the appeal of St. Caesaris at Ravenna.
Ibid., Bk. III, Letter 38.
Ibid., Letter 43.
Ibid., Letter 17.
Ibid., Letters 24, 32, 42; Bk. 14, Letter 36.
Ibid., Letter 32.
Ibid., Letter 44; Bk. IV, Letters 5, 7.
Ibid., Bk. III, Letter 44.
Ibid., Letter 43.
Ibid., Letter 42.
Ibid., Bk. IV, Letter 34.
Isidore of Seville, pp. 282–283; Chron. Caesaragus , p. 223.
Cassiodorus, Bk. V, Letter 39.
Ibid., Bk. IV, Letter 17.
Bury, Invasion, p. 207.
Ibid., pp. 199, 207.
Ibid. For Theodoric's concern with the maintaining of traditions see the obversation of Le Gentilhomme, RN 1943, pp. 87, 88.
Ulrich-Bansa, pp. 325, 327–328.
Mateu y Llopis, pp. 140–141; Heiss, pp. 77–78.
Amardel, Les Monnaies à Narbonne, pp. 133–153; Armadel, Wisigoths de Narbonne, pp. 389–401.
Robert, Languedoc, p. 115.
Le Gentilhomme, RN 1943, pl. IX, 9, 10. Le Gentilhomme reads the monogram as GMA, whereas in accordance with Germanic tradition it should read MAR.
BMCVOL, p. XXXI, pl. VII, 6–13.
Kraus, p. 6. Kraus considers Arles as one of Theodoric's mints. Also see Cassiodorus, Bk. V, Letter 39 (Theodoric's orders to Spanish minters).
Tolstoï, II, p. 216, no. 165; Friedlaender, III, p. 56, no. 3.
Tolstoï, II, p. 215, nos. 163, 164, pl. 15.
BMCVOL, pp. XXXIV–XXXV, pl. IX, 13–15.


An attempt to assign a terminal date for the VPW tremissis, particularly the Visigothic issue, is aided by more concrete evidence. The few coins that may be assigned to Childebert I (511–558) or Childebert II (575–595) are too rare and too dissociated from other Merovingian coins to permit an exact dating of these last Merovingian irregular VPW issues. A combination of the monetary reforms of Leovigild, with the coins of Hermenegild, and the Zorita de los Canes Hoard, however, do provide evidence which suggests that the last VPW issue of the Visigoths occurs ca. 579–582. Determining factors may be the dynastic and political or fiscal problems of Leovigild, Hermenegild's revolt, the Byzantine occupation and the religious issue. This is based on the coins which bear the name of Leovigild and the few with the name of Hermenegild. Although the latter are so few in number, and of these some are questionable and some definite forgeries, the authenticity of the remainder necessitates their inclusion as evidence.

There are definitely three and possibly four varieties of coins which are struck in the name of Leovigild. The first is the VPW tremissis (Plate B, 1); the second is the cross-on-steps reverse issue (Plate B, 3); the third is the facing busts type (Plate B, 4); the fourth is a variation of the VPW tremissis of the INCLITUS REX design but bearing the name of Toledo instead of the customary legend (Plate B, 2). This last coin was listed and illustrated by Florez in the late eighteenth century. Although the striking of such a coin is feasible and within the developing fiscal traditions of Visigothic coinage, it has never been recorded as having been seen by anyone else and is unavailable for study. 70 Legends can so easily be misread and misinterpreted, and good forgeries can only be detected through a thorough direct investigation of the material. The use of this coin as evidence, therefore, is very limited.

Miles has suggested the following dates for these coins: the Victoria tremisses of Leovigild between 568–578; the cross-on-steps reverse between 579–584; and the facing busts between 584–586. 71 The date of the cross-on-steps reverse is based on its being an imitation of the tremissis of Tiberius II Constantinus (578–582) which was issued between September and November of 578. 72 This only proves, however, that the cross-on-steps type could not have been issued before 578. It does not necessarily follow that the striking of the VPW issue ceased immediately in 578 or 579. The date of the facing busts is based on the Cordovan issue of Leovigild, which commemorated his second capture of the city in 584, and the crush- ing of Hermenegild's revolt. 73 Miles further subdivided the VPW issues of Leovigild according to legends and assigned dates to the several groups: (1) the pseudo-imperial issues (those coins with garbled legends), ca. 568–574; (2) Types A–C (Leovigild's name on one face only), ca. 575–576; (3) Types D–G (Leovigild's name on both faces), ca. 575–576; and (4) Type H (INCLITUS REX), ca. 575–578. 74 This chronology and classification implies a progression of consecutive issues, although Miles does not exclude the possibility of simultaneous striking in different mints of several or all of the types. The final stylistic and numismatic analysis in this present study will propose a pattern combining consecutive and simultaneous progressions which, notwithstanding the diverse approach to legend, were contemporaneous products of different mints (See Chart VIII).

A study of the Zorita Hoard determined the chronology suggested by Beltrán. Considering the INCLITUS REX issues as the last form of the VPW tremissis, he dated it from 576/77–580. He placed the Florez coin in 580, and suggested that the new cross-on-steps type was issued to commemorate the Church Council of Arian Bishops held at Toledo in 580, the enactments of which it was hoped would facilitate the conversion of orthodox Catholics to Arianism. 75 The date for the destruction of the Zorita town site as well as for the burial of the hoard has been placed ca. 580. 76 Since, of the ninety coins in the hoard, only three are INCLITUS REX types (one of these is a contemporary forgery) and none are cross-on-steps types, the hoard must have been buried not too long after the INCLITUS REX type was issued and before the cross-on-steps reverse was issued or had had time to circulate freely. 77 Does this suggest that Beltrán's range of 576/77–580 is too long for the INCLITUS REX issue?

Since the Zorita Hoard contains only four or possibly six coins bearing the name of Leovigild, besides the three INCLITUS REX specimens, while it contains sixty-three with garbled legends of the Justin II types, might it be proposed that if the hoard was buried in 580, Miles Types A–G as well as Type H (INCLITUS REX) had not been in circulation for a long time? Or, if they had been, then the issues were rather sporadic and without the control of a central authority, because they do fit into the style groups of coins with garbled legends and are seemingly contemporaneous. The chronologies of Miles and Beltrán should perhaps be qualified if not discarded. The only possible verification for the dating of the VPW with the king's name as early as 575, if not even earlier, is two Zorita coins with legends that might be interpreted as Liuva and Leovigild. The legend of Zorita no. 45 is: ·VΛTIV·· ·IILLI:, VLIVV V·imageVΛ; the legends of Zorita no. 81 are: ICLIVVIƆ LOIEECIS, IVLIVVIƆ IVVCVSI. 78 I seriously question the assignment of either to Liuva. If Liuva had ordered his name placed on his coins more should be expected to have survived amongst the large number extant from this period. Although uniqueness is not an unusual characteristic of barbarian coinage, the legends of both of these coins are better explained as transition stages from legends of garbled form to legends bearing the name of Leovigild. If this is so, the Zorita Hoard contains nine coins with the name of Leovigild, including the three INCLITUS REX items. Considering the number in the hoard and the number extant, would a shorter life span be more justified for Miles' Types A-H, i.e., 577/578–580?

Miles' chronology dates Leovigild's first currency reform (the placing of his name on the tremissis) after Leovigild assumes sole rule of the kingdom in 572, on the death of Liuva; and after the successful military operations of 572–575 establish the power of his throne. At the time that he is fighting the Suevians and gradually asserting his power over the entire Iberian peninsula, he orders the placing of his name on the tremissis. Is it at this time also that his increasing power and the pressure of foreign politics also encourage him to invade further on the emperor's prerogative by adopting imperial dress, regalia and ceremony? 79 Considering Leovigild's strong dynastic instincts and his policies toward national unity, it is not incongruous that his currency should become more uniform and systematic. The INCLITUS REX issue may present the first attempt to create a uniform national coinage, since its obverse royal portrait was retained for the cross-on-steps issue which has the dignity of a royal portrait, combined with a royal name, and the name of a mint. This coin is now fully comparable to the coins of the Byzantine emperor. Is the adoption of this cross-on-steps type in accordance with Leovigild's political vision of himself as a king, free of any allegiance to the present occupant of the Byzantine throne? If this conjecture is correct, then Leovigild's currency reforms could have been determined solely on secular grounds and have nothing whatsoever to do with the Arian-Orthodox problem or the later revolt of Hermenegild. The evidence of Zorita and the contemporaneity of the adoption of the Tiberius II Constantinus cross-on-steps, however, makes impossible the discarding of Hermenegild's revolt as an encouraging factor in the adoption of the new coin type. 80

In this regard, Hermenegild's issues must be considered fully in any attempt to develop a chronology. All of Hermenegild's coins are of the same type, with the exception of three, and the authenticity of the latter should be questioned. 81 The remainder, perhaps five in all, are of INCLITUS REX style, but three of these have substituted for the INCLITUS REX legend another — REGI A DEO VITA. 82 The mint at Seville does not issue a cross-on-steps coin for Hermene- gild, or, if it did, none has survived. It has been customary to see Hermenegild's issues as dependent upon those of Leovigild. The mint at Seville, Hermenegild's capital and consequently his mint or one of them, possibly might have issued an INCLITUS REX in the name of Leovigild before the revolt took place and once the revolt commenced adapted this coin for Hermenegild's use. This would seem the perfect moment for Leovigild to adopt the cross-on-steps type in order to distinguish his coins from those of Hermenegildo. As a propaganda device this maneuver has a double thrust, both as a political and a religious weapon. The adoption of the imperial coin, as well as the adoption of the imperial regalia and ceremony, distinguishes the primacy of Leovigild by dynastic right and tradition. He assumes a role similar to that of the Byzantine emperor. If Hermenegildo conversion and the Arian-Orthodox controversy are involved in this revolt, the new cross-on-steps type asserts Leovigild's Catholic faith. This might explain in turn the adoption of the REGI A DEO VITA legend on Hermenegildo coins as a counter-propaganda thrust. Is it this legend and its reference to God that suggests to Leovigild and his advisors the CUM DEO OBTINVIT SPALI and CUM DEO ETALICA legends for the first cross-on-steps coins they issue at recaptured Seville and Italica? 83 Is this more counterpropaganda? This is not followed through, however, at Córdoba. At Córdoba, the final victory and the military glory of having twice captured this city take precedence over the protestations of Christian faith and allegiance with God in the legend: CORDOBA BIS OBTINVIT. 84 Only at one other mint, Rodas, does the CUM DEO phrase possibly occur: CVM DI ROPA. 85 If Rodas was involved in the Basque rebellion of 581, this unique legend could then be justifiably used here as it was later at Seville and Italica. 86 The cross-on-steps reverse being struck here in late 581 may indicate that Hermenegild's REGI A DEO VITA had already been issued and that the NCLITUS REX may have already been dropped. The relationship of Leovigild's CUM DEO with Hermenegild's REGI A DEO VITA may indicate a further connection between the revolt of the Basques and Hermenegild. 87 Tactically it would seem best to put down the revolt in the North first, as did Leovigild, since this frontier was flanked by Hermenegild's possible Frankish allies.

In 582, Leovigild retook Mérida, and a cross-on-steps coin is issued with the legend: EMERITA VICTORIA. 88 A non-religious legend is preferred, as it is later in 584, at Córdoba. This might disprove the CUM DEO and REGI A DEO VITA relationship. Another tradition might be at work. The epithets of VICTOR, and PIVS …………… VICTOR are found on later issues of Leovigild, Reccared, and Chindasvinth from the mint of Mérida. Add this to the importance of Mérida — AUGUSTA EMERITA — and its Roman traditions which might imply that the EMERITA VICTORIA legend is in compliance with local custom and is more suitable and meaningful to the particular circumstance. 89

If then the cross-on-steps issue owes its existence to Hermenegild's revolt as counter propaganda, it would be logical for a new type to be issued once the revolt is put down. The facing busts are issued perhaps first at Córdoba in 584, with the legend CORDOBA BIS OPTINVIT and thereby herald at one and the same time the end of the revolt and the removal of the religious question from public propaganda. The combination of the INCLITUS REX profile bust and the cross-on-steps reverse are not to be used again until 653, when Reccesvinth becomes sole ruler on the death of his father, Chindasvinth. 90 Does this not suggest Reccesvinth's identifying himself with Leovigild by a coin Leovigild issued at the time of Hermenegild's rebellion? Significant also is the re-adoption of the Leovigild profile bust by Chindasvinth in 649, when he proclaims Reccesvinth as joint ruler. 91 Is this not a direct reference to Leovigild and an attempt to associate dynastically Chindasvinth and Reccesvinth with the earlier "Pater Patriae"? Does this not encourage the acceptance of the idea that the Visigothic kings, like the Roman emperors before them, made propaganda use of the currency, and further strengthen the thesis that Hermenegild's revolt is directly responsible for the Leovigild cross-on-steps issue? The date of 580 may be suggested, therefore, as the date for Leovigild's second currency reform — the cross-on-steps reverse issue. Leovigild's first currency reform, the placing of his name on the VPW tremissis, should be dated before this — but how much before? Could this reform also depend upon Hermenegild's revolt?

In an unpublished manuscript, Grierson also rejects the hypothesis of consecutive issues and suggests a new chronology based on a simultaneous progression and on a reconsideration and re-emphasis of the coins of Hermenegild. 92 He suggests, as has been done here, that Leovigild's first currency reform could have occurred after 578, rather than before, most likely at the time of Hermenegild's rebellion in 579, thereby permitting Hermenegild's issues to be contemporaneous and parallel rather than later. Since Leovigild's and Hermenegild's issues are contemporary, Grierson feels that a better case can be made on behalf of Hermenegild for placing his name first on the coin. 93 Hermenegild's two issues are placed in both a simultaneous and consecutive progression because of the grammatical structure of the legends as well as by the variation in the formation of the letters between the coins in either series. Grierson suggests that the REGI A DEO VITA series is earlier than the REGI INCLITI because the combination of this with the dative form ERMENIGILDI is a correct rendering of the acclamation during the coronation ritual; "(Long) life to the King from God." 94 While on the other hand, "the use of the dative in the other series has no obvious justification, and must be ascribed to a mechanical blending of the dative use of the acclamation coinage with the formula REX INCLITUS copied from the coins of Leovigild." 95 He notes, furthermore, that differences in the forming of the letters R, D and E between the two legends suggests that Hermenegild's coins are issued by two different mints, with the REGI A DEO VITA most likely coming from Seville and the REGI INCLITI from Córdoba. 96 Grierson's pattern is thus that Hermenegild first strikes at Seville the REGI A DEO VITA series, which places his name on a coin that acclaims his coronation. This forces Leovigild to reciprocate, but, done in haste, his injunction to his many mints creates confusion. This accounts for the variety in the handling of the legends. 97 The hesitancy and confusion of the first appearance of Leovigild's name on his coins would signify for Grierson a policy instituted "on the spur of the moment to conteract the possible repercussions of such an action on the part of his son." Order is restored by the INCLITUS REX issue. This coin in turn is imitated by Hermenegild's mint at Córdoba.

This suggestion, although provocative, is an inadequate explanation of the numismatic evidence. It is not feasible, as both a stylistic study of the pseudo-imperial issues and an analysis of the Zorita Hoard would demonstrate. Both of Hermenegild's issues are exactly the same design as Leovigild's INCLITUS REX, which is in turn similar to one of the "curru" groups, C 3. Since the "curru" group is of garbled legends, it is presupposed that this may be earlier than Leovigild's INCLITUS REX and Hermenegild's REGI A DEO VITA. Hermenegild has replaced the legends of the "curru" coins with his name and the coronation acclamation. Leovigild takes the "curru" coin and puts the INCLITUS REX legend and his name on it. Why did the mint at Seville select the C 3 coin? Was it because the C 3 was already being struck at Seville at the time of and before the revolt? Or, if its Toledan attribution is correct, was the "curru" coin an already established national coin type standard? The latter seems more probable in the context. It would be improbable for Leovigild to adopt the coin design of the mint of Seville and of his son Hermenegild and continue to use the obverse face in his later cross-on-steps. It would not be improbable for Hermenegild to adopt a coin that his father has established as the official national coin and place his own name on it. When the legends of coins preceding the INCLITUS REX of Leovigild are studied, there is evident a tendency towards standarization, e.g., the regularity of the IVIVIVIVIVIVIV variables of JII 4, and the "curru" legends themselves. This characteristic is noticeable to a degree on all coins of the Justin II period, and it suggests that Leovigild is systematizing the coinage even before the revolt of Hermenegild. If, then, he is involved in regularizing legends and even type designs as the C 3 suggests, is it not conceivable that other measures such as the placing of his name on the coin be in the offing?

This thesis may gain in stature when Cabré's analysis of the quality of the gold in the Zorita Hoard is studied. Although Cabré does not reveal the source of his information, he has labeled some of the Zorita coins as being of "oro bajo" and "oro muy bajo." A large number of the coins placed in JII 3 are so qualified, but all are of a regular high weight standard, while a large number of Zorita coins in JII 2 are of a fine quality of gold but of a low weight standard. 98 Cabré's analysis necessitates a qualification of Reinhart's acceptance of all pre-national coinage as being of fine quality, twenty-one carats. 99 The evidence from Zorita adds to the observations made by Grierson in his article on Visigothic metrology. 100 Grierson found that the weights of Leovigild's profile busts suggested that a tremissis of 1.326 gms., or the twenty-one siliqua solidus, is the weight standard in use, a standard widely used in the former western Roman provinces. 101 He also noted that Leovigild's facing busts, as all later Visigothic national coinage, were of 18 carat gold, and that until the reign of Egica (ca. 698) the tremissis was struck with the theoretical weight of 1.516 gms. 102

The Zorita Hoard reveals that a change in the fineness of the gold and the weight standard was attempted before Leovigild's facing busts issue of 584. Grierson's evidence suggests this when he cites one Leovigild VPW and one Leovigild cross-on-steps as being of sixteen and eighteen carats respectively. 103 None of the Zorita Leovigild VPW's or INCLITUS REX items, however, are labeled as poor gold. This suggests that in the area of weight standards and gold fineness a confusion exists which parallels the confusion in legends, in the "hesitancy" in the use of the name of Leovigild and in the variety of type designs. Leovigild's first reform, the placing of his name on the VPW coin, may be the first step in the direction of a more regular currency. I would suggest that the large number of INCLITUS REX specimens is the last stage in this first reform in which other mints than Toledo strike the coin designed at the major mint, Toledo. The INCLITUS REX provides the necessary uniform type, weight standard, gold content, guaranteed by the king's name. Although the theoretical weight standard is 1.326 gms., the coin is of 21 carats. The presence of a contemporaneous counterfeit in the Zorita Hoard (Cabré no. 90) of the INCLITUS REX type may suggest the acceptance of this coin as national currency. Hermenegild's striking of it would present further evidence. The cross-on-steps issue is a second reform in which the name of the mint, lacking on the INCLITUS REX, is now supplied. The type would suggest that this has been motivated by Hermenegild's revolt as well as by his striking an INCLITUS REX coin. The facing busts is a third reform necessitated by possibly two causes, the end of the revolt and the economic necessity of legally debasing the coinage. The theoretical weight is raised to that of a Byzantine tremissis, but the gold content is reduced from twenty-one to eighteen carats. The difficulty in maintaining the twenty-one carat standard may be suggested by the existence of some cross-on-steps coins of eighteen carat gold as noted by Grierson. 104

It is more feasible to conceive of this fiscal reform developing gradually and naturally under the guidance of Leovigild. Hermenegild's reaction to Leovigild's cross-on-steps which must have stolen Hermenegild's orthodox "thunder" caused Hermenegild to continue striking the old coin but with a countering religious legend, REGI A DEO VITA. Did he need to bring the religious question further out in the open to gain supporters, particularly after the conciliatory measures adopted by the Church Council of Toledo (580)?

Can any evidence be gleaned from considering the number of the coins extant within the Justin II and Leovigild groups? Of the two hundred forty-one coins of Justin II studied, one hundred twenty-five bear garbled legends, seventy-three are "curru," twenty-two bear the name of Leovigild, and twenty-one are INCLITUS REX. Miles' catalog lists thirty-eight cross-on-steps and eighty-four facing busts. The garbled legends may extend over the thirteen years of Justin 110 reign, 565–578; the "curru" series may fall between 573/574–578, as issues of Leovigild, first struck at Toledo and then perhaps imitated by one or more mints in a preliminary move to standardize the coinage; the JII items with the name of Leovigild may date just before or just after the death of Justin II (ca. 577/578), again in response to the royal attempts to standardize the coinage.

The death of the emperor in the East may have been a perfect opportunity to permit the first great step in placing the name of Leovigild on the coinage, particularly when all the evidence points to a gradual recognition of the necessity to establish a more uniform currency. The INCLITUS REX should date soon after, ca. 578/579–580, adopting the "curru" style which may already have been promoted by the mint at the capital, Toledo, and enjoining other mints of the state to strike this in place of their local issue with the name of Leovigild (Miles' Types A–G). Hermenegild's REGI INCLITI must date from the beginning of his revolt in late 579 or early in 580, with Leovigild's cross-on-steps, and his own REGI A DEO VITA following closely behind in late 580 or early 581. Leovigild's cross-on-steps continued to be struck until 584, when it was replaced by the facing busts (584–586).

The stylistic evidence described in the following chapters will bear out this chronology which reveals Leovigild's enlarging image of himself as an independent king, possessing all of the prerogatives, the regalia, and the "uniform" of the Byzantine emperor. The death of Justin II may have permitted the final outward demonstration. This self-image also permitted the establishment of Hermenegild, the heir apparent, as DUX at Seville in late 579. Leovigild was truly an INCLITUS REX.

End Notes

Cassiodorus, Bk. IV, Letters 1, 2; Kraus, p. 79.
Florez, III, p. 175. Miles (p. 44) accepts this in his Corpus (no. 27) but is suspicious of it. Grierson in his unpublished manuscript of the Barbarian coinage in the American Numismatic Society is disposed to accept it.
Miles, pp. 44–45.
Ibid., p. 44. Also see BMCB, I, p. 105, n. 2, p. 108, n. 1.
Miles, p. 45.
Beltrán, Zorita, pp. 37, 49.
Ibid., p. 34. Also see Zorita de los Canes, pp. 52–54.
Zorita de los Canes, nos. 88, 89, 90.
The Zorita coins bearing the name of Leovigild are in my Corpus nos. 437 (Zorita no. 45), 474 (no. 83), 475 (no. 82), 477 (no. 81), 539 (no. 84), 540 (no. 85). That the legend of Zorita no. 45 does not read Liuva and is simply a coincidence in the garbled form may be seen by comparing it with the legend of the Zorita item which is no. 439 in my Corpus.
Isidore of Seville, 51 (ii, 288).
For the most recent discussions on the character of Hermenegild's revolt see Goffart, Hillgarth, Thompson.
I accept as authentic nos. 632, 633, 634 and 635. Miles (pp. 199–200) lists these as well as the Johns Hopkins (Miles 47 c) and Hispanic Society of America (Miles 46a) specimens which I consider not authentic. Miles also refers to two other specimens which I have not seen, that in the Royal Mint, London (no. 4784) and the other in the collection of Jaime Butiña of Bañolas. The two items no. 636 and 637 in the Barcelona collection I do not accept as authentic.
The authentic INCLITUS REX items are nos. 632 and 633. The authentic REGI A DEO VITA items are nos. 634 and 635, with the possible third one being the now lost item (Miles 47d) which was in the Escorial in 1580. See Miles, p. 200.
See Miles 31a, 31b and 34. Mr. Hilgarth seconds Grierson and suggests that the cross-on-steps type is in retaliation to Hermenegildo REGI A DEO VITA coin, and commemorates Leovigild's victory "….sobre el usurpador católico como un favor de Dios. Sus primitivas leyendas Cum Deo conmemoran los episodios cruciales de la guerra civil, la fortificación de Itálica, la caída de Sevilla." (pp. 45–46).
Miles 30a.–c.
Miles, pp. 85, 185, no. 18.
Thompson, p. 13. Mr. Thompson notes that "…there is no reason to think that the Basques were co-operating with Hermenegildo." John of Biclar, 581 (ii, 216). Thompson rejects the opinion of Goffart (pp. 91, 105f.) and Stroheker (p. 461), who suggest "…that Leovigild deliberately postponed his attack on Hermenegild so as to give the latter 'every chance to change his mind and repent'." Hilgarth suggests (p. 46) that the Rodas coin dates from 585 and is connected with Leovigild's war against the Catholic Franks as indication of the religious commemorative aspect of the legend. I disagree with the latter dating as being too late for the cross-on-steps type.
John of Biclar, 581 (ii, 216).
Gregory of Tours, vi, 18.
Miles, p. 119. Mérida was founded by Augustus, ca. 25 b.c. for the benefit of the 5th and 10th legionaires (Emeriti) of the Cantabrian War.
Miles, p. 51, no. 353 a.
Grierson kindly forwarded me part of a chapter on the dating of Leovigild's coinage from a volume that he is readying for publication which catalogs the barbarian coins in the ANS/HSA collections.
It must be remembered throughout this discussion that Grierson has only suggested this thesis.
This is quoted from Grierson's unpublished manuscript.
Both Goffart, p. 107, and Stroheker, pp. 460, 480, say that Córdoba and its adjacent territory was ceded to Byzantine rule at the beginning of Hermenegild's rebellion in 579. If this is so, it would seem unlikely that the mint at Córdoba would have issued an Hermenegild coin. Miles, p. 24, assigns both of these issues to Seville.
This is a condensation of Grierson's arguments in his unpublished manuscript. Miles (p. 24) gives the correct interpretation of the formula but does not connect it with the coronation laudes. Miles suggests that this formula is on Hermenegild's second coin issue and reflects his declining fortunes. This is also my opinion. There is no exact account of the Visigothic coronation ritual. Isidore of Seville (51, ii, 288) says only: "primusque inter suos regali veste opertus solio resedit; nam ante eum et habitus et consessus communis ut populo, ita et regibus erat." Grierson in a footnote in his unpublished manuscript says: "…but the phrase N…regi vita is a regular feature of all medieval rituals, which go back to the formal acclamations of late Roman times. The idea of using such a phrase on the coinage came perhaps from North Africa, where vita had appeared on the bronze of Carthage in the reign of Justin II, but the practice later spread to Constantinople itself, where multos an(nos), another of the ritual acclamations, was used on the gold of the late 7th and early 8th century (BMCB, pp. 99–101, 332, 335, 358 ff.)."
A further illustration of this is found in JII 6, composed of only Zorita items. One of the five coins (no. 529) is not labeled "oro bajo," and its weight is 1.16 gms. The "oro bajo" items (nos. 530–533) weigh respectively 1.38, 1.40, 1.35 and 1.49 gms.
Reinhart, DJN 1940, pp. 87–89; Reinhart, AEA 1945, pp. 232–233. Reinhart never suggests the change in quality of the gold although he is aware of the change in the weight standard in some of Leovigild's issues. Also see Grierson, NC 1953, pp. 83, 84 and Table II.
Grierson, NC 1953, pp. 84–85, Tables II, III.
Ibid., p. 81.
Ibid., p. 79.
Ibid., p. 84.
Ibid. Grierson suggests "…that 18 carats was the fineness which Leovigild ordered that his gold coinage should be." Grierson (p. 84, n. 17) notes that: "Hermenegild apparently reverted to the full imperial standard, for the one specimen of his coinage in the British Museum has a density of 19.08, and a fineness therefore of 98.5." This may be a correct observation, but the evidence of one coin is not sufficient.


Hypotheses have been established in regard to the provenance, the significance, the initiators, the minters, and the dating of the VPW tremisses. What remain to be studied are the coins themselves, and they must be organized into categories which will resolve specific problems of dating and attribution.

The placing of the coins into a workable order demands finding some consistent pattern that will provide or suggest a means by which groups may be collated and sequences organized. An analysis and tabulation of the legends, of the weights, and of the gold content might provide the means of analysis, yet cursory efforts in such investigations seem to complicate and confuse rather than to clarify the issue. We do find possibly more solid ground in uniting these factors with a study of the style of the coin types, the imperial portrait, and the striding Victory.

The use of the information in the obverse legend, the only definite factual material on most of these coins, may be used only in a subsidiary way, since there is no guarantee, at the outset, that the imperial name has anything to do with the date of coin. We have coins bearing the names of Anastasius, Leovigild, Theodebert I, Childebert I (or Childebert II according to Belfort), Justin and Justinian. There is a large number with consciously garbled legends and a series bearing the puzzling legends: CVRRVI CVRRVI; CVRVRVTI MVRVIVC; or variations of the same. The coins that might be read as Justin do not indicate which Justin is being referred to, Justin I or Justin II. Furthermore, it is often difficult to distinguish the legends of a Justin from those of a Justinian.

Any stylistic study of these coins must proceed first by understanding mint practices in the late Empire and then by considering three stylistic determinants. Since the coins of Anastasius establish the early terminals, it is necessary first to consider the aesthetic values upon which the imperial tremissis is based. If we accept the fact that the barbarian imitations are the products of continually producing provincial mints, late fourth century issues of Western mints must be studied. A positive, known factor, the style of Visigothic national coinage, must be considered since it represents the final stage of a long development whose origins and paths of development are conjectural.

The ascendency of the symbolic and abstract interpretation of coin types in the late Empire may be traced to the expedient measures adopted by the mints as well as the influence of both Christian and barbarian aesthetic values. The reforms of Diocletian inevitably created a confusion in the presentation of the obverse portrait bust and legend. Unavailable die models for a new emperor might necessitate the use of the deceased emperor's bust with the legend of the new emperor. 1 The consequence of the division of the provinces between Augusti and Caesari placed mints under several possible jurisdictions, with each mint continuing to issue coins in the names of its Augusti as well as his co-regents with or without the corresponding busts, so that the emperor's name could be placed around the portrait of his co-regent. 2 What develops is a symbolic rather than realistic relationship between portrait and inscription. One or more conventional portrait bust types begin to take shape by the late fourth century. The break-down in centralized controls and the consequent increasing independence of the various localities encourages this development in the fifth century. The confusion of the era and the insecurity of the imperial office might have necessitated the development and retention of a recognizable conventional effigy for commercial reasons as well as for the expediency of the minters. 3

Design variations from mint to mint and within a mint are determined by a number of factors. Mints did keep old die designs and must have kept records of previous issues. Only this could explain the issue of contorniates commemorating the tercentenary of the death of Augustus and Nero's Victory issue, with the reintroduction of Aes coinage in which Nero's coin types were used. 4 The possession of old dies may account for the surprising reappearance of coins of better and earlier imperial style after the complete breakdown of that style in the late Empire. 5 More important is the possibility of the reworking of old dies such as those noted by Alföldi in the Constantinian Solidi of Ticinium. 6 Alföldi states that the wearing out of the hubs from which the obverse dies were produced led to a less distinct portrait, and that the copy of such inferior types thus occasioned the appearance of a new type of portrait. 7 He also noted modeling lapses which gradually geometricized the portrait; the nose outlined by a contour line, the eyebrow emphasized by a row of dots, and the changes in facial proportion. This practice is most evident in fifth century coinage and may be the modus vivendi for the developing style of Visigothic coinage. The uniformity in a coin issue designated dictation from a central source, but variation in type style indicates the possibility that the local mint copied the "example" coin sent from the imperial treasury for purposes of imitation. 8

The practice of the "Monetae Publicae" in which private gold was converted into solidi, since minting of state gold was limited to the imperial residence, would not produce coins of the controlled quality of imperial mint issues. 9 The use of local goldsmiths outside the mint atelier also provided for a variety of style and less control of the purity of the design. It further increased opportunities for counterfeiting. Maurice considers this system responsible for numerous style variations and counterfeiting at both Lyons and Trier. 10 Such weakening mint practices would tend to increase rather than decrease during the crises of the fifth century. Those mints controlled by the various barbarian kingdoms in the fifth century, therefore, would make their new dies by copying imperial coins or by reworking old die designs in their own workshop. The loss of imperial centralized control, furthermore, would have necessitated the increase of private minting of gold by local goldsmiths, a situation that could have created a stylistic chaos. 11 These aspects are already evidenced in products of the mint of Lyons as early as in the Constantinian Era 12 and are certainly indicated in those late fifth century tremisses assigned by Ulrich-Bansa to the mint at Milan.

Before considering the stylistic treatment of some late fifth century western tremisses, it would be appropriate to consider the design standards of the mint of Constantinople. On an imperial tremissis of Anastasius (Plate A, 2), the treatment of the profile bust has reached that point in its stage of development, where the effect of reality is not completely evinced by plastic modeling. Line already is being used to emphasize and differentiate contours. The modeling is not of the subtle naturalism as that of the head of Augustus four centuries earlier. It is not even so finely understood as it is on the coins of Arcadius from Constantinople. By the time of Anastasius, modeling is treated almost as a simple mass which is raised slightly from the surface of the coin, and upon which line is then used to distinguish the mass as a head. This is particularly seen in the treatment of hair, which is all in line.

The die maker still had a three-dimensional sense of relationships, however, for he did not reduce the contour of the back of the head or the contour of the neck or the lower jaw to a line, but distinguished these areas by an adjustment in the surface relief. Modeling such as is used here does not seem to be sufficient for the artist. He enhances the facial features by sharper lines that give the appearance of being extra additions of gold to the surface: e.g., the sharp line of the forehead and nose; the heavy outlining of the eyebrow, the eyelids, and the tiny globules for the pupil; the globule for the nostril; and the lines which achieve the fullness of the lips. By means of such linear effects the artist attempts to distinguish between the various planes of the face and the individual features. The folds of the paludamentum also are completely flat and executed in line. There is no modeling to create the three-dimensional sense of fullness, merely the two-dimensional sense of outlines.

There is an attempt to show almost a frontal chest with the profile head instead of the three-quarter view as found in other tremisses (Plate A, 3). This may be significant in revealing the possible origin for the circular fibulae on the shoulders of the bust type, as seen in Leovigild's coinage. Where the fibula appears on one shoulder, on the other is a rough curved area that marks the upper sleeve above the pleats. This latter area treatment may be imitated later as a fibula on the left shoulder, similar to the fibula on the right, particularly when directed by an aesthetic viewpoint which is concerned with symmetrical repetitive arrangements.

The Victory figure on the reverse of this coin is achieved similarly by the simplest treatment of modeling. It is a raised mass, with patterns of line added to distinguish arms, wings, head, torso and drapery. The vertical linear folds stand out but are still contained within the contour lines of the dress. If you ignore the sense of the mass of the entire figure, you see the linear aspects which make up the later Visigothic Victory figure. For such delicate work the artist has abandoned modeling to achieve the distinction between parts of the head. An unbroken line frames the contour of the head from the top of the nose to the tresses of hair that rest on the back of the neck, and dots of gold distinguish eye, chin and lips. It readily may be seen how the misunderstanding and misinterpretation of threedimensional modeling and linear additions by barbaric imitators or perhaps more accurately, later provincial craftsmen working under barbarian governments, produce the abstract two-dimensional types found in Visigothic coinage. Perhaps the imitations were based on worn imperial coins, for the wearing down of a modeled mass does take on an increasingly linear aspect. Distinguishing naturalistic details are gradually lost.

This should not be interpreted as signifying the degeneration or barbarization of this type. The final Visigothic type must be understood in relation to the aesthetic values of these people. Their art was never expressed in three-dimensional plastic forms. They had a feeling for decorative, moving, geometric flat patterns. Their designs are purely abstract or animal and nature forms abstracted and combined in total moving patterns. This was the basis of their aesthetic, far removed from an art based on classic values and not unlike the developing aesthetic of the new Christian era. As a result of this taste, craftsmen would gradually lose the knowledge of the necessary small tricks of the trade that gave one an illusion of a three-dimensional form.

Or, perhaps more precisely, craftsmen would tend to simplify the design of the types, which does not necessarily imply a loss in skill. Certainly the evidence of the development of the coinage indicates a conscious gradual transformation from realism to abstraction.

A further factor, often forgotten, is the decidedly iconoclastic propensity of the early Spanish Church and its ties to the Levantine shore of the Mediterranean. The Council of Eliberri (Illiberis, Elvira, Granada) in the early fourth century decreed: "Placuit picturas in ecclesia esse non debere, ne quod colitur et adoratur, in parietibus depingatur." This could account for the rare appearance of the human form in early Spanish Visigothic art. The few examples that are extant date almost completely from the last third of the seventh century. 13

Tremisses of Justin I and Justinian II show the continued progression from three-dimensional plastic form to that of two-dimensional linear elements. The coins of this period present a more limited geometric cast and a greater simplification. The soft roundness of subtle classic modeling is being replaced by an angular geometry of flat patterns (Plate A, 3, 4).

The immediate sources for the style of the barbarian coins must be found however in the local mints of the Western Empire. The complex problem of separating the coins of one barbarian group from another is due largely to their similar origin. We know that from the time of the reign of Zeno there was an increasing production of tremisses in the West, which could come from the non-barbaric mints of Rome, Ravenna, Milan and Arles, or also from the barbaric mints of at least Lyons and Narbonne, if not more, e.g., Toulouse, Bordeaux and Trier. 14 Unfortunately the major tremissis type, the cross-in-wreath reverse, does not bear any mint identification marks, and, consequently, like the later barbarian issues, can only be attributed by style. Ulrich-Bansa has accredited the mint of Milan with most of the tremissis issues as well as the silver functions of siliquae because he conceives of Milan as the production center of coins for the marginal regions. 15 He is able to attribute some of the tremisses because of their stylistic relations to the silver coins which bear the mint mark of Milan. 16 The connection of these is undeniable, although Ulrich-Bansa himself comments on the variety of styles and legends within the tremissis group which he associates with these. 17 The variety might admit non-Milanese issues in the group, perhaps some from Arles.

The existence of these numerous coins in the name of Zeno and traced definitely in part to Milan, and indefinitely in part to Milan and the region of Northwest Italy or Provence, places them in the locality immediately adjacent to the provenance area of most Anastasius VPW tremisses. The stylistic connection is indubitable. It must be from coins of this type from Milan and possibly from Arles as well that the VPW tremisses of Anastasius depend. Ulrich-Bansa considers the most active period of the mint of Milan to extend from 476 to 488, when activity at Rome is negligible; that of Ravenna is very limited; and commercial needs in the West force these tremissis and siliqua issues. Ulrich-Bansa does not attribute any Anastasius tremisses to Milan although he admits the possibility. 18

Coins in the name of Zeno which have been attributed to Milan, consequently, must be analyzed. The best of the Zeno tremisses attributed to Milan or genuinely western mints still maintain a plastic three-dimensional quality in the conception of the profile bust. This plastic sense extends beyond the modeling of the head and facial features and involves the successful suggestion of the three-quarter chest pose. 19 The disintegration of this modeling refinement is quickly noticeable in the flattening chests of no. 181 (Plate C, 4) and its variables as well as nos. 176, 185 (Plate C, 1, 9). The chest is approaching the two-dimensional geometric shape of the later barbarian issues, although here no abstract pattern has been evolved. The die makers, inept when working on a small scale, have a realistic prototype in mind and so create only a symbolic confusion of parallel, curving, and perpendicular lines for the rich folds of the paludamentum. In no. 181y (Plate C, 6) the final C in the obverse legend can almost be confused with the second circular broach of the later Visigothic issues. The heads vary considerably in refinement of modeling and facial features. In the best the face, headdress and diadem all combine to form a unified whole delicately and ably conceived (nos. 177, 179; Plate C, 2, 3). In others, the geometricizing of the design is already taking shape; contours become more emphasized at the expense of the accuracy and refinement of the modeling. The nose and lips and chin are stressed, with also the glob for the eye causing the cheek and jowls of the head to protrude inordinately, e.g., nos. 182, 183, 185 (Plate C, 7, 8, 9). The tendency to emphasize and clarify details of the delicate scale of the portrait tends to distort and departmentalize, e.g., in the treatment of the diadem in nos. 181, 181a, 182 (Plate C, 4, 5, 7) where the headdress is divided into a "juliette cap" back of the head, and the frontal bangs cover ears and forehead. What we are witnessing here is the gradual breakdown in the classical value system and a change in imperial taste. This change is occuring even in an Italian mint with a long classical tradition. The conceptual difference between these and the later barbarian tremisses is the more realistic, aesthetic orientation of the former imperial mint makers. This prevents them at this stage from developing a truly abstract symbol.

The stylistic variation between these coins of Milan and earlier ones of Lyons, Arles and Trier from those of Rome and Byzantium could already be ascertained in the late fourth century issues of Valens, 20 and even earlier in those of the Constantinian Era. 21 There is a tendency in the fourth century for the coins of the Gallic mints to be closer in style and technique to each other than they are to Italian or eastern mints, and respectively they are closer to Italian issues than to those of the East. Yet in the late fifth century, with the moving of the prefecture of Gaul to Arles, the Italian mints must take a larger share in the responsibility of coining gold for the western provinces. A more local and regional style may possibly exist between North Italy and South France at this time than ever before. Ulrich-Bansa may be correct in seeing Milan as the center of western mint operations at the end of the century. On the basis of the stylistic evidence this would certainly be true. If we accept Keary's theory that as the fifth century progressed, gold from the East was used much more in northern Gaul while southern Gaul, Africa and Spain continued to use the money of the West and Italy, we may be able to justify the similarities of the Carthaginian bronzes of 534 with barbaric issues in the West and establish a stylistic basis for a western regional style. 22

The relationship between Carthaginian silver and bronze and the barbarian tremisses indicates a common heritage (Plate A, 7, 8, 9). The bronze follis shows a similar departmentalized treatment of the head as found in a type of Justinian coin, such as no. 254 (Plate XII). The frontal chest of the follis even has the central compartment which bears the cross (or in some examples the chris- mon). 23 The silver votive pieces parallel the style in which the emphasized diadem almost does away with the forelocks and bangs. Coins no. 315 (Plate XV) and Miles 374a (Reccesvinth-Emerita) may be later stages of this style.

The Ostrogothic coins which naturally descend directly from those fifth century Italian mints previously mentioned are of more consequence. Theodoric's Anastasius issues are the closest in the West to the imperial coin in excellence and refinement of technique, although the Ostrogothic versions are more conscious of line, and a line executed in a very high relief. These aspects are especially true in regard to the two VPW tremisses attributed to Rome by Wroth. 24 They have done away with representing the ear by simply extending the forelock down the entire side of the face. The only important area of modeling now is of the chin, jaw and cheek bone, which distinguishes that area from planes of the neck and the profile hollow of the meeting of the eyes with the bridge of the nose (Plates VI, 130; VII, 131). What has been said in analyzing the imperial Anastasius coin may be reiterated here. The design of these is even simpler than the imperial ones and those from Milan in the time of Zeno. Two variant types are found. The first is a typical striding Victory advancing right, and the second is a Victory posed to the left on a globe. Both are connected with Gallic and Spanish striding Victories, for the treatment of the legs of the Victory on a globe has accentuated the striding pose by hiding the taut leg under drapery and revealing the left leg advancing forward. This particular coin already shows the first stage in the development of the late Visigothic insect Victory in the treatment of the drapery folds of the chiton. Note how the garment has lost its bottom hem line, thereby permitting the drapery folds to operate freely like ribbons, ribbons which will develop later into extra legs or appendages in back and front. This is a particular design aspect seen earlier in all Victories standing on globes rather than in the striding positions. It is also seen in frontal Victories standing on a globe or striding. 25 The mints that produced these may very well have set prototype precedent in the West, particularly supported as they were by the prestige of Rome and Theodoric. Coins from local and western mints are the ancestors of the barbarian VPW tremisses and not coins or dies from the East.

Finally the style of Visigothic national coinage which is governed by certain absolute values must be considered. Since the type plays so important a role in the early coinage of Leovigild, one is hopeful that the real answer to the attribution problem lies here rather than in the few nondescript coins of Theodebert I and Childebert I or II. Visigothic coinage is a coinage in which commercial standardization supports the continuance of recognizable symbols achieved through the simplest means. Not only are letters punched, but the punches, wedge-shaped, circular, and "C" shaped, are even used to create the types, a practice particularly noticed in the time of Swinthila and Sisenand. 26 The technique and the aim are those which progressively achieve a more developed abstract reality. It must be granted that the types produced are therefore more of the character of symbols than of individual portrayals of the king. The variety of these symbolic forms in any one reign are more apropos of an elected kingship and the lack of the establishment of a strong dynasty. This is further apparent in that in certain periods one standard representation of the profile bust is regulated. This is seen with Leovigild and attests to his own strong position. It is later seen with Chindasvinth and Reccesvinth who reissue the Leovigild type. Their political gesture may be an attempt to establish a dynasty and identify it with the greatest of Visigothic kings, Leovigild, who had made a similar effort.

There is also a sense of compositional order and clarity, which continues some of the quality of the coins of the early empire. The relationship of the parts to the whole is always clearly maintained. Neither legend nor type usurps the other's territory, each retaining its identity.

There is a sense for flatness and two-dimensional emphasis on line. They prefer to see a figure as design rather than as a three-dimensional modeled form. This, as has been mentioned previously, is in keeping with all extant fragments of architectural sculpture of this period and is a constant element in Spanish art. The Spanish sense for strong flat patterns of color areas and design is rarely infringed upon by their desire for realism. 27

These, then, are the absolute values which are the result of the stylistic development of the sixth century. They are quite distinct from stylistic attitudes in Merovingian coinage, where a greater variation in technique and bust types prevails because of local rather than royal control of the mints. The individual differences in Visigothic mints are displayed within a narrower range of controlled possibilities. A much more conservative, vigorous and disciplined dependence on traditional style or type is the basis of the creative impulses of the Spanish Visigothic mint.

End Notes

Maurice, p. 4.
Ibid., pp. 5–6. Also see Pearce, NR 1946, pp. 61–62 and RIC IX, p. xi. Mattingly, Roman Coins, p. 234 gives the best and briefest picture of the change in style and the "growing barbarism."
BMCB, I, p. lxxxviii.
Grant, Anniversary Issues, p. 155. Vermeule, Num. Circ. 1952, pp. 356–357, gives further corroboration of this practice.
E.g., Coins of the JII 3 group. Also see Kent, pp. 200–201.
Alföldi, JRS 1932, pl. III.
Ibid., p. 17. Vermeule, Num. Circ. 1956, provides the following characterization of the late Constantinian style which he states continues to the time of Justinian I: (1) a constant appearance of diadem and rope-like hair arranged in neat lines over the brow and behind the ears; (2) always a flattened view of both shoulders with, as the century advances, increasingly angular drapery over the cuirass; (3) a usually beardless face which is flat, linearly defined around nose, eyes, and mouth in keeping with its relationship to the background; and (4) the portrait is so dematerialized and made so impersonal that only a trained eye can tell the actual portraits of one emperor from another. Vermeule also notes that, through the fifth century, reverse types also develop a more linear costume, a more denaturalized form while they lose the vigor of personal representation and demonstrate an increased interest in decorative pattern.
Maurice, I, p. XXVII.
Kent, pp. 200–201.
Maurice, I, p. 372; II, pp. 65–66.
Note Theodoric's edict against this, Cassiodorus, Bk. V, letter 39; Bk. VII, Formula 32.
Maurice, II, pp. 65–66.
Helmut Schlunk, Observaciones. Extant fragments of figure representation in sculpture are as follows: 1. Byzantinate sarcophagi from the late fifth century from Ecija y Alcandete, Briviesca, Itacio and Orviedo; 2. Evangelist capital at Córdoba; 3. Pilaster fragment in Almonasta; 4. Fragment of a "pila" in Toledo; 5. Remains of a capital in Córdoba. Architectural monuments in which figure relief sculpture is found are San Pedro de Nave, Santa Maria de Quintanillas de las Viñas and San Juan de Baños. All of these date most likely from the late seventh century. The iconoclastic injuction is to be found in Canon 36 of the Council of Eliberri. See Schlunk, Visigodo, pp. 212, 233–323; Garcia y Villada, I, p. 311; and Palol de Salellas, p. 104.
Ulrich-Bansa, pp. 327–328.
Ibid., p. 337.
Ibid., pp. 338–339·
Ibid., p. 340.
Ibid., p. 345, n. 37.
Ibid., pl. XV, 177, 179, 180, 183.
See the illustrations in RIC IX.
Maurice, I, pls. V, IX, X; II, pls. VII, VIII, X, XII.
Keary, p. 58.
BMCB, I, p. 65, nos. 361, 364.
BMCVOL, p. 56, nos. 71, 72, pl. VII, 3, 4.
Ibid., pl. VII, 1 and 22.
Miles, p. 151.
Consider the cubism of Picasso, the naturalism of Goya, and the flatness of Velásquez in his treatment of light. This change from a Roman to a Mediaeval aesthetic cannot be simply explained by their inability to do better. Before the inability does develop there has been an aesthetic change which dictates a form consciously away from a realistic expression. Naturally with an aesthetic based on the symbolic representation of the natural world, the techniques necessary for realistic depiction are lost. Mediaeval craftsmen create a meaningful expression within theologically established limits. Conventions and formulas govern their aesthetic discipline and aesthetic choice. They cannot create a realistic form because they do not want to nor does society want them to. It is to be noted that realistic forms in the 14th century appear quite readily once the aesthetic and cultural attitudes of artists, theologians and patrons desire a more realistic expression.


Between the poles of the attributable late fifth century and anonymous imperial issues and the late sixth century national coinage, the anonymous VPW tremisses must be arranged in a logical order to demonstrate the transition from Roman imperial coinage to that of the first mediaeval national states.

The VPW tremisses of the major collections have been organized into stylistic groups from Anastasius to Justin II. Each group has been analyzed first for consistencies of weight and legend and second for relationship with other groups. It has been found generally that most of the VPW groups have a stylistic connection with each other and particularly with the late Leovigild and so-called national Visigothic coinage. Most of the groups are associated with sub-groups, e.g., A 1 and A 1a. These sub-groups are distinguishable from their parent body for diverse reasons. Some are composed of stylistic variants which may or may not imply that they are not of the same mint, or are from an unofficial mint within the same locality, or are imitations by the mint of a different barbarian group. Others are listed as variables because, in contrast to the main body of the group, these coins bear letters or monograms in the obverse or reverse field, and this too may or may not imply a different mint product, either of the same nation or of another.

On the basis of style, the coins may be classified chronologically, since the general standard that is maintained throughout this prenational coinage is first seen in the Anastasius tremisses. The Anastasius issues also present the basis for the varied directions that the abstraction of types will follow down to Leovigild.

GROUP A 1 (Plate I, 1–20)

In the coins of this group both faces present the initial stage of later stylization. The profile bust (Plate I, 1) begins with an un- modeled chest on which only the drapery folds have been handled as line. There is an over-all sharpness of line which emphasizes the nose and eyebrow, the diadem, the bow-strings of the diadem (infulas), the pleated sleeves and the fibula on the shoulder. The mass is relegated to the head and neck alone, with the greatest density around the neck, jaw, and upper cheek. The whole effect is softened by strokes of hair falling like bangs, framing the face and covering the ears. The striding Victoria has developed much further along the lines of abstraction. The flaring folds of her light skirt now appear like fluttering ribbons; her gown no longer has a solid bottom hem to keep all the drapery folds confined within the contour of the dress. Her later insect appearance is beginning to take form. Her head and wings are still carefully modeled. The upper part of her chiton is taking the shape of the truncated pyramid.

All this is closely repeated in the coin (Plate I, 3), with the addition of a cross squeezed in beneath the folds of the emperor's paludamentum. The insertion of a cross has the look of an afterthought, of a sudden change on the die. It rests like a medallic broach against the chest. In the coins (Plate I, 9, 10) the sharpness of the styling has increased, and the pectoral cross has moved its position to the top of the head in the midst of the legend.

The portait bust (Plate I, 9) becomes more stylized in the next coins, wherein the head becomes more linearly departmentalized into the three zones, back of the head, diadem, and front framing bangs. The fibula fills the area of the shoulder above the pleated sleeve. There is a similarly increasing simplification in the Victory type, which seems to be the first stage in the insect Victory development.

The legends of these coins and the form of their inscription are similar. They read DNΛNΛSTΛ SIVSPPΛVC on the obverse, with the exception of those that include a cross in the legend at the top of the portrait bust, DNΛNΛSTΛ+SIVSPPΛVC, and those which bear the legend, DNΛNΛSTΛ SIVSPFΛVC. The reverse legends generally read VICTORIΛ ΛVGVSTORVΛ, although the difficulty in reading the last letter of the second part of the legend may result in variations such as ΛVGSTORVN or ΛVGSTOVΛI. The weights are fairly consistent, ranging from 1.465 to 1.55 gms., and only in the sub-groups are two coins found to be as low as 1.395 and 1.43 gms (See Chart X).

Their style, weight, and legend indicate that these tremisses are possibly the earliest in the Visigothic issues. The developing lines of abstraction do not interfere with the sense of order or clarity of the total design while they do reduce the classic form to a linear angular pattern.

A 1a (Plate I, 21–25; Plate II, 26)

Both legends and weight are in accordance with the main group, as is the design style in most respects. Yet, there is a slight variation in the handling of the diadem image which is close to, though not so exaggerated, in coins we identify as Burgundian and which Wroth had assigned to the Vandals. 1 The Victory figure, however, is in direct accordance with the developing six-legged type.

A 1b–A 1h (Plate II, 27–36)

These all bear stylistic affinities with the Visigothic group but with distinguishing letters in the field. Some bear closer resemblance to the group than do others in regard to style, legend and weight. All evolve around the main type of the group, however, whether or not they are Visigothic.

A 1i (Plate II, 37)

This single coin is a possible variable which is difficult to place. It bears a resemblance to some coins of Group A 2. It is not a usual combination of obverse and reverse type styles.

A 1j (Plate II, 38–41)

These coins represent a late stylization of A 1 and reveal their proximity to the style of A 3, indicating the possible lineage of A 3 from A 1.

End Notes
See the coins of Group A 7 and BMCVOL, p. 10, nos. 2–9, pl. II, 7 and 8.

GROUP A 2 (Plate II, 42–45)

The first and possibly prototype coin of this group presents a finely modeled realistic obverse portrait with a much less naturalistic Victoria. A delicate technical refinement characterizes the four coins assembled here, with an attention to facial features and headdress. This carries through into the more stylized and more linear coins in the Cabinet des Médailles, where details are delicately pinpointed.

The realism of the die cutters has made an adjustment in recreating the striding Victory, which has maintained her human two-legged form, but she has lost the flaring drapery folds in the back of her skirt by having them joined to her wings. The front drapery folds are two appendages but tightly connected to the bottom of the dress. The whole effect is that of a triangle with a concave base. The wings are expertly modeled to suggest feathers, and the nose and headdress of the little Victory are discernible. All these features have been maintained in the more stylistic Paris items but with an obviously less aesthetic substitution. The result is a more fluid design for the awkward, more realistic figure, so that Victoria does advance and stride.

A 2a–A 2c (Plates II–III, 46–64)

The solid, round, classically modeled head of A 2 again appears in yet a different and possibly later format. The chest type resembles those of A 4a and A 4b. The Victory figure presents that of A 2 in a more stylized version until in some examples (Plate III, 52) the chiton has become so small as to be almost non-existent, and only a stick figure remains. Even with this stylization there remains the same kind of design feeling, as seen in the Victories of A 2. The developing left shoulder vertical pleats are suggested in A 2, but are here well established.

The legends and weights of A 2a are fully in accord with those of A 2, and the entire group continues the standard seen in A 1. The stylistic variation of A 2b (Plate III, 56–59) maintains the weight standards as well as the legends of the group. The variation in the reverse legend of A 2c (Plate III, 60–64), VICTORIΛ ΛCUSTO, suggests the later legends seen in such four-legged Victory groups and in JAN 5. The surprising barbarization of the obverse legends in the London and New York pieces makes it difficult to consider these as of this period or as Visigothic. They do not conform in style to the Justinian or Justin II period, and the obverse legends are anomalies in any period. I suggest that A 2c is non-Visigothic. Such a possible variation developing from A 2 might indicate indirect relationships between A 2 with later Visigothic issues insofar as it presents the early stages in the development of the four-legged Victoria types. The group does bear closer resemblance, notwithstanding diverse Victory types, to A 4 than it does to A 1 or A 3.

GROUP A 3 (Plates III–IV, 65–82)

The second largest of the Anastasius groups presents a stylization of types directly dependent upon A 1 with close affinities to groups classified under Justin I. I suggest that this is a later issue of the A 1 mint. These coins present the most developed characteristics that particularly distinguish Visigothic coinage of all the Anastasius pieces. This may indicate even a later date than other groups, A 2 and A 4.

The pectoral cross has become an established motif, while the chest formation is developing towards the truncated pyramid shape. Only the fibula, the pleats of the sleeve and one fold parallel to the neckline remain of the original. Also the double contour on the left side of the chest has now appeared. The head is a simple mass, with the diadem strongly incised and the eye pronounced as a dot. The striding Victory is assuming its roughly carrot shape, although the drapery folds, front and back, are still distinct from her legs. She has even maintained her human head with discernible features.

The legends are consistent as an extension of A 1, except where obvious die copying and recutting leads to orthographic errors. There are several coins with changes in the final letters of the reverse AVGVSTORA, however, that might be meaningful, e.g., ΛVCVSTOR·T·. These coins are no different in weight and style from all others of the group. Also note the particular care given to the B of CONOB on several coins. It is included, even if it must extend above the exergue and infringe on the end of the reverse legend. This particular device develops further in group JI 1.

The weights of these coins are lower than those of A 1, with half the coins ranged from 1.44 to 1.47 gms.

A 3a (Plate IV, 83–84)

There is enough stylistic variation in the chest types of both of these to separate them from the entire group.

A 3b (Plate IV, 85–88)

Coins, particularly Nos. 85 and 86, may illustrate a natural stylistic extension of the coin types of the group. The rigid geometric structure of the design, the inverted COMOB inscription, combined with the correct reverse legend VICTORIA AUGUSTORUM and the "s" on the reverse field, however, encourages a non-Visigothic attribution. This is particularly true of the Paris coins in this subgroup which are presented here only because they also have an "s" in the reverse field.

A 3c–A 3e (Plates IV–V, 89–93)

Although associated with this group in nature of types, the image in the field of the one group and the treatment of the design in all suggest also a non-Visigothic variation of the A 3 group.

GROUP A 4 (Plate V, 94–98)

These coins have Victories similar to those of A 1 and would seem to be slightly later or contemporary even in regard to style of portrait bust, comparable weight, and more accurate transcription of the legend. There are many gaps in the structure of this group, and consequently it is more difficult to see the stages in the stylistic development. This accounts for sub-groups A 4a and A 4b which are suggested as Visigothic, either of the same mint as A 4 or of the same locality.

The rounder and more plastic obverse head relates this group to the technical workmanship and prototypes of the A 2. The chest type varies from one with repeated curving folds (Plate V, 96) to one with angular folds (Plate V, 95). Although the Victory figure (Plate V, 94–95) represents the Victory of A 1 with graceful, flowing ribbons front and back, the other two coins, (Plate V, 96–97) present a stenographic reduction of the same figure. The Victory in these has decidedly the two leg form image with less-emphasized front and back appendages. It begins to appear as if the figure was made by copying directly from a worn piece, such as No. 95, or by pencil shadowing such a coin. Legends are similar and, aside from the surprisingly low weight for one of the coins traced to the French Gourdon Hoard (1.17 gms.), the weights are standard, 1.43–1.50 gms.

A 4a (Plate V, 99–100)

The two coins resemble the first two of A 4 in obverse facial type and in Victory form but differ in the imperial chest. Instead of a single or double line on the left side framing the chest, there is a series of vertical lines balancing the vertical sleeve pleats on the right side. These coins also follow the group in maintaining the usual weight but do vary in legend. The ΛVCVSTORVΛ is replaced by ΛVCVSTORO and is preceded by a cross. The use of a cross within the legend is to be found in some coins of A 1, but here it comes at the break in the name of the emperor on the obverse DNΛNΛSTΛ+ SIVSPPΛVC.

A 4b (Plate V, 101–102)

These two coins repeat the obverse chest type of A 4 a, with a more stylized and linearly conceived profile bust. This is true in the Victory figure as well where legs and front and back appendages are becoming confused. These compare with the entire group in regard to weight and with A 4a in legends, although in the latter it is technically retrograding: VICTOIIΛ ΛΛCVSTORI. The cross, as in some coins in A 1, appears in the obverse legend.

The entire A 4 group relates to A 2, although it is the closer of the two to A 1. Furthermore, it is from A 4 that connections may be made directly to later Justin I groups.

GROUP A 5 (Plates V–VI, 103–112)

The distinctive feature is the use of a monogram in the right reverse field, either image or, as in A 5 a, image (Plate VI, 113–115). The treatment of the profile portrait and of the striding Victory is different from all previously classified Anastasius coins, although the general character of the design suggests a coeval minting with those of A 1, A 2 and A 4. Most unusual is the treatment of the Victory, with the lower hem of her chiton fully defined, preserving the nature of a skirt rather than creating flaring separate ribbons or extra legs. The legends remain fairly intact with the D of Dominus maintaining its integrity and with a preference for VICTORIΛ ΛUCUSTORUM on the reverse. The weights are likewise standard, with only three (1.35, 1.35 and 1.42 gms.) falling below 1.45 gms.

The single coin in A 5b (Plate VI, 116) is badly preserved but stylistically fits best with this major group. Instead of the monogram in the reverse field, it bears the letters and T respectively in the left and the right.

This group forms stylistic contact with only one other Anastasius group, A 6. If the monograms may be read as Gundomar and Sigismund, the Burgundian attribution of these seems highly probable. These coins may only be related to the later groups of JI 5 and Jan 11, both of which also stand apart from all the other classified groups. All may then be attributed to the Burgundians. The coins are sufficiently different from those of the other Anastasius groups to be considered as a possibly different national product.

GROUP A 6 (Plate VI, 117–124)

This group follows more closely the tradition of A 5. The Victory, with its hemmed chiton and its strong striding step, is particularly similar to that found in group A 5. The obverse portrait preserves the same design technique and approach as seen on coins of A 5. There is, however, a stronger emphasis on the legs of the Victory under her skirt. The design of both types reveals a stronger assertion of contours, a sharper and more angular line than found on coins of A 5. This is seen in the handling of the diadem where image is preferred to the image of A 5 (e.g., Plate VI, 113). These distinctive style aspects are particularly demonstrated in a more crudely executed example (Plate VI, 118).

As noted in the legends of A 5, the D of Dominus is correctly done, and the reverse legend remains VICTORIΛ ΛVCVSTORVM. The only difference is in the ending of the obverse legend, when often…. PRFΛV is preferred to…. PPΛVC of A 5. The weights in this group compare favorably with those of A 5. This group may also be Burgundian.

A 6a–A 6b–A 6c (Plate VI, 125–127)

The coins in these sub-groups are most likely of the same national mint, or of the region of the major group A 6. They are separated because of the differing letters found in the reverse field of each of the coins.

GROUP A 7 (Plates VI–VII, 128–131)

These two beautifully modeled examples must be separated from the rest of the Anastasius groups. Not only does the exquisitely detailed Victory stand on a globe on one, but also the refined craftsmanship throughout reminds us of the earlier naturalistic traditions of Roman imperial coinage. The profile bust has fully modeled hair with emphasized bangs, detracting from the diadem, so eminently stressed on all other coins studied. The fine craftsmanship extends to the legends which are carefully executed. Here, too, as seen on some other Anastasius coins, the "Pius Felix" on the obverse legend replaces the "PerPetuus Augustus".

The style of these suggests the mint of Rome.

GROUP JI 1 (Plates VII–VIII, 132–167)

This group immediately recalls Group A 3 with similar imperial facial types although with a new variation in the chest form. In the early examples, (Plate VII, 132–133), the standard right side remains the same in all cases with the sleeve pleats and fibula, but the left side presents downward and inward curving lines from the double line of the left contour as seen in coins of A 2 and A 5. These left side folds have been interrupted or erased and prevented from meeting at the bottom edge of the coin in order to provide space for the pectoral cross. These folds parallel two small folds on the lower left side of the chest. The chest type for the group develops from this, in which the central chest area occupied by the pectoral cross is framed by balancing downward curving folds in both lower corners image a device which is to be maintained into Leovigild's coinage.

The Victoria is the carrot-shaped striding type with the six legs more pronounced. The truncated pyramid shape of the bloused part of the chiton has been joined as one piece to the mass of the neck, while the skirt of the chiton has become all legs. The method of constructing the head directs us to a later development. The mass of the jaw and the cheekbone are all one piece with the upper chiton and neck; a dot provides the eye, a line the nose and forehead. The mass of the back of the head has been incised to form a diadem creating a comic hat-like impression with the rest of the raised mass of the head. This juxtaposition of mass of head and dot of the eye, intercepted by the incision of the diadem, will later create the X-shaped Victoria heads.

These coins bear legends that are similar and relate closely to those of A1 and A3. There are some that continue the … SIVSPFΛVC seen in a few coins of A 1, and there is one that uses a cross in the middle of the obverse legend. However, there is an increasing tendency for confused legends because of the failure to take care in executing the letters. The letters Ɔ and S begin to lose their character and without delicate and careful workmanship begin to be simply vertical sticks. The letter N loses its diagonal connection and becomes II. The use of triangular punches in some coins to form many of the letters will naturally increase the tendency for confusion and create unintelligible legends. This is seen in a coin in the British Museum, where the letter I is formed by a triangular punch: image and in a Paris coin, where a T is formed: image The use of geometric forms to construct letters can only create complete unintelligibility when not handled expertly. This of course does happen in the later Justinian issues that stylistically are associated with this group JI 1.

As the potentiality for confused legends rises, the standard weights of these coins become increasingly variable. The majority of the coins remain in the area of 1.43–1.46 gms., but weights below 1.40 gms. are more frequent than previously noted. The actual span is from 1.05 to 1.50 gms.

There are a few coins within the group that indicate a progression to Justinian groups, such as some in Paris from the Alesia Hoard (Plate VII, 150–152). Some of the coins in the variable groups (JI 1c, JI 1d and JI 1g), however, also indicate this Justinian connection. These may very well be examples of the group late in the reign of Justin I or early in the reign of Justinian. Their legends suggest a Justinian reading, while the unusual manufacturing of the letter R (Plate VIII, 174) with one circular and two triangular punches image is to be seen only in the time of Justin II and Leovigild (Group JII 5). This coin may very well share in the prototype rôle for the group, JII 5.

JI 1a (Plate VIII, 168)

This stylistic variable has a distinguishing reverse legend which seems to read Gloria Romanorum rather than the usual Victoria Augustorum. Its Victory reverse is closer in style to the full-fledged members of this group, but the reverse is somewhat different conceptually. Since this coin comes from the Alesia Hoard and is therefore of French provenance, it may be a Merovingian or Burgundian imitation of a Visigothic issue.

JI 1b–JI 1h (Plate VIII, 169–178)

These sub-groups are vaguely connected in style to JI 1 and may or may not be Visigothic. They may illustrate how the portrait stylization on JI 3 developed. In some (JI 1b, JI 1c, JI 1e, JI 1f) the elongated chest type is cursorily achieved so that the chest and head of JI 3 is a natural consequence. The legends reveal a less disciplined technique.

GROUP JI 2 (Plates IX–X, 179–204)

This group has its strongest ties with group A 4 in both facial type and treatment of the Victory. The pectoral cross, so much a part of the A 1–A 3– JI 1 group, is not seen on half of the group. The classical modeling traditions survive with a delicacy for detail and subtle nuances. This is best seen in the London item (Plate IX, 179) which may begin the group but is also to be seen in the Paris-Alesia Hoard coin (Plate IX, 181). The chest form retains the drapery lines that smoothly curve uninterrupted across the chest image This design later becomes angular (Plate IX, 191) image although the head and neck retain the subtle modeling. The pectoral cross bust develops naturally from these (Plate IX, 194) image It creates the same form in a long chest type (Plate X, 203) image which particularly introduces the form of a Justinian group (JAN 3). The Victory figure maintains the two-leg device, with chiton skirt folds adding the extra legs and therefore following closely the Victory of JI 1. There is a tendency, however, to maintain a sharper distinction between the blouse of the chiton and the meeting of wings and arms and neck as a shoulder point. This more human effect, added to a still careful delineation of facial features, makes this Victory more realistic. On some coins (e.g., Plate IX, 189) the left, wreath-bearing arm of the goddess is in a bent position as seen on two unique coins placed in a sub-group of JAN 3, JAN 3a (e.g., Plate XV, 313–314) which follows directly in style from JI 2.

In accordance with the greater technical skill and refinement observed in the making of the dies, the legends are more legible than those of JI 1 and are of good quality. On the reverse, the VICTORIA AVCVSTORA gradually develops into VICTORI, or VICTOR ΛVCVSTOR, or ΛVCVSI, or ΛVΛCTOS. The latter legend becomes common in what may be later groups, such as JAN 3 and JAN 5. The weight of these coins also retains a standard of 1.43–1.48 gms. with only two coins below 1.40 gms. (1.31 and 1.30 gms.). From this group naturally develops group JAN 3.

The single coin of JI 2 a (Plate X, 205) as well as the single item of JI 2b bears a questionable stylistic connection with the major group. The variable in JI 2a suggests such an early coin in the group as No. 179 (Plate IX) although the approach is more linear, and subtleties of line and mass modeling are absent. This variable also possesses a cross in the right obverse field. The Alesia item in JI 2b (Plate X, 206) has a faint relationship with possibly earlier coins in this group, but an increased dependence on line and on even less use of mass and understanding of modeling separates it from the group. It also bears a star in its left reverse field to separate it further. Its legend does resemble some of the coins, but the general conception of this coin as well as that of JI 2a suggest variations that are imitations of JI 2 coins rather than products of the same or adjacent mints.

GROUP JI 3 (Plate X, 207–217)

This group is associated with A 1–A 3–JI 1. It may be a later extension and variation of JI 1. This connection is observed when weights, legends and types are studied. The Victory remains constant with JI 1, but the obverse portrait marks a further change in style. The modeling of a JI 1 head always emphasized the transition from face to hair, so that its mass minimized the diadem; consequently, when worn, the diadem would be lost and the frontal mass of hair would substitute for it. Once the lines of hair dominate the mass of the back of the head, the normal contour of the head is lost. The effect is of a flattened head with hair standing on end.

The chest is geometrically simplified. It has assumed a square shape and no longer has the more intricate patterns that are found around the neck of the JI 1 busts. Particularly important for later developments is the joining of the first and last letters of the obverse legend to the bottom contours of the chest. This effect, combined with the fibula on the right shoulder and the natural effect of greater density at the meeting of top and left contours at the left shoulder, leads to chest types such as No. 217 (Plate X) and to those of JII 5 image The shoulder epaulettes of the Leovigild coins are evolving.

JI 3a (Plate X, 218)

This coin seems closer to the JI 3 group than to any other. The profile bust is in the JI 3 manner, although the chest is closer to Jan 2 than to JI 3. Although the reverse legend is illegible, the obverse legend unquestionably reads Justin, so that the coin must be placed within a Justin group, unless it is a rare example of the placing of an earlier emperor's name on a coin issued during the reign of a later one. However, as is so with the portrait head, the Victory on the reverse is in accordance with those of JI 3. What distinguishes this coin from the major group is the obverse chest type and the star in the left reverse field. The closeness of the chest type with those of the JAN 2 mint may suggest an influence from the JAN 2 mint and therefore may indicate that JI 3 a is a very late product of the JI 3 group, possibly continuing to be issued into the early period of Justinian.

JI 3b (Plate X, 219–220)

This variation from the Alesia Hoard also has a conglomerate character with an almost JAN 4 portrait chest, and a profile head that is a variation of the JI 3 group. The Victory figure is more within the tradition of JI 3. This coin must then be either a variation of the JI 3 theme or a single example of a new group. Its provenance might suggest that its uniqueness from the rest of the group is that it may be a Merovingian product. The Grierson item demonstrates similar characteristics.

GROUP JI 4 (Plate XI, 221–226)

This group is made up of the only Justin coins that may be associated with the earlier Anastasius groups of A 2. This is particularly obvious in the form of the Victory which has developed into a definite stick figure with two appendages in the front only. The small triangular chiton is almost nonexistent. The head does have some facial features but these, through wearing, may readily develop into the round ball as seen in No. 221 (Plate XI). This is found in later JAN 5 examples.

The chest type has changed radically from that of A 2 and may only bear some relation to No. 56 (Plate III) which is in A 2b. The head is not too unlike that which is being produced in A 3 and JI 1, but the chest is quite singular from these. The trapezoidal and square shapes are developing with contours composed of dots. The work is still in the delicate touch of the A 2 craftsmen, and a rich effect is produced. It is not a great step from No. 56 image to No. 221 image The chest type is further enriched by additions of decorative "buttons" within the border of the left side (Plate XI, 223) and in a second fibula (?) on the left-side shoulder unattached however and somewhat in the field.

Finally, the Paris coin from the Gourdon Hoard (Plate XI, 226), which may be a variation, presents a strange dwarfish Victory with better modeled body and a large head to enable its features to be delineated. This latter Victory and its accompanying legend immediately recalls the reverse of those coins of A 2c, e.g., another Gourdon treasure piece (Plate III, 60). This must be of the same mint. Three other coins of A 2c (Plate III, 62–64) reveal similar handling of delicate, rich details in the obverse faces not unlike the workmanship of the present group being discussed, and all seem to indicate being products of the same atelier. The midget-like Victoria is to be seen again in two coins bearing the name of Leovigild from the Zorita Hoard (Plate XXIV, 475; Plate XXV, 477) which are placed in JII 3a.

GROUP JI 5 (Plate XI, 227–231)

These coins present several distinguishing characteristics that disassociate them conceptually and formally from all other groups discussed so far, with the exception of groups A 5 and A 6. The reverse face evokes particular comment with its image monogram in the right field and its star in the left. It carries a solidus type legend on the reverse. The Victory figure is singular with its columnar-like chiton and the characteristic bent arm which bears the wreath-crown. The portrait bust does not include the pectoral cross but only a series of parallel folds diagonally across the chest. The diadem is reminiscent, as really is the entire obverse bust, of the types seen in the above described groups, A 5 and A 6. This may be a later product of this same mint or mints. As is true with groups A 5 and A 6, the combination of stars and monograms in the reverse fields, which are distinguishing elements from all other Anastasius groups and Justin I groups, may suggest a non-Spanish origin or provenance. A Burgundian attribution seems most probable.

GROUP JAN 1 (Plates XI–XII, 232–245)

This group follows naturally from the progression of A 1–A 3–JI 1–JI 1d. The portrait bust especially follows such coins as No. 150 (Plate VII) and No. 174 (Plate VIII). The head has been enlarged, but the same large mass of hair through which is incised the diadem reminds us of JI 1 and prepares for the head of JAN 2. The chest type of JI 1 has been further altered in the name of symmetry, although it persists in one example (Plate XI, 233). The old chest image has become image It has become increasingly geometric yet in a different line of development from JI 3. This chest begins to take on the shape of that of JAN 2 in the progression of increased simplification: image (Plate XII, 242).

The Victory is the carrot-shaped type that has developed through the period of Justin I and has the six legs and form that also belong to the A 1 progression. A further phase is noted here in comparing the Victoria of No. 232 (Plate XI) and No. 238 (Plate XI). The Victoria on the former still possesses the hat-like headdress noted before, but it has been manufactured differently. The head is now formed by a diagonal, raised line which creates the hair line. Perpendicular to it another raised line suggests the back of the head image The eye is suggested by a dot on and a little below the meeting of these two lines image A line forms the nose, image and another line the jaw image while the rest of the face is formed by the carrotshaped body image In coin No. 238 the mass delineations and distinctions have been simplified, and a rough X shape for the head evolves. As a simple symbolic design form it takes hold for the rest of its life span. A further significant aspect of some of these Victorias is the "dot" that is found in the middle of the void between the two main striding legs. It does not appear on No. 238 and No. 242, but does on No. 232 and No. 240 (Plate XI). It has a simple logical explanation and derives naturally from the "dot" ends of the ribbon drapery folds that fluttered between the central legs on the earlier coins. Simplification and further stylization have removed the folds from between the major or center pair of legs leaving behind one of the dots. That there is no further significance or interpretation of any kind to be placed on the presence of this "dot" seems unquestionable.

The legends are not so clear as in earlier coins. The name of Justinian is assumed: DNVIINIVIIVSPIИC: ƆN VSTINVIIVSPIVC; ƆNVSTIN …… VIIVSPIΛC. The reverse legends are even more unintelligible: IVTOI VOTΛVI, and VICTOΛI IIVTONΛVI. Nevertheless, these are consistent in the manner in which the inscriptions are formed. Only in coin No. 242, which must be late in the group, is there a greater degree of variation and this particularly in the obverse legend: ƆIHIT▴ HVIHIC. Its reverse legend although partially destroyed follows more closely the other coins of the group: VICT▴▴ ………IVII. The weight of the latter is also 1.39 gms. All the coins of this group that evolve between JAN 1 and JAN 2 contain unintelligible legends on the above order and, with the exception of one, all are of poor weight.

JAN 1a (Plate XII, 246)

The single coin placed in JAN 1a presents a different version of what this atelier has designed. Its chest type assumes the simplified state of the late JAN 1, yet the treatment of hair and diadem on the portrait are not of JAN 1 style. The head appears to be more of a variation on the theme of JI 3. The legend is also singular in the manner of its structure and is as close to those of JAN 1 and JAN 2 as it is to any. The Victory is within the group tradition and this combined with its general aspect places it within the A 1 progression.

GROUP JAN 2 (Plates XII–XIV, 247–277)

This, the major issue in the Justinian period, extends directly from the coins of JAN 1. In the constant development of the stylization of forms from three-dimensional plastic masses to two-dimensional linear designs, this group almost achieves the aesthetic resolution. The jaw, cheek and neck are still partially modeled, but everything else has been conceived in line. The mass of frontal bangs, the diadem, and the back of the head have been geometricized to create image a two part section of raised lines for back of head, and lines of hair filling a triangle for the frontal area, separated by the two raised lines for the diadem. The eye is formed by the usual bead of gold, but it is set in a raised circlet image The nose, a cuneiform stroke image continues from the hair line. Two raised strokes form the lips image Variation is noticed in the handling of the diadem from the angular image to the sweeping curve image or image to a straight line image from which stage the Leovigild VPW coins do not seem far away (Plate XII, 259).

The chest is completely flat and realized only in a linear pattern which continues the form already seen in the late JAN 1 image The fibula is sometimes present, and the small folds decorating the lower corners the pectoral cross zone are sometimes left out. In some the pleated sleeve persists on the right side but only as an extra line on that side image Legends vary considerably from being legible to being completely unintelligible, although there is a degree of consistency within the group. There is an increased tendency to use the repetition of straight lines and I and V combinations that are seen so frequently in the later coinage: IVIVSTIIIIIΛVIC (Plate XII, 249). The most constant form for the reverse legend is VICTOΛI VTOIΛVI, and almost all present the "conob" as image which is stylized into image In accordance with the garbling of the legends, more coins fall below the 1.40 gms. mark.

This group is of further interest because it contains the earliest coins from the Zorita de Los Canes Hoard. They account for most of the crudest examples of the group. Only two (Plate XIII, 267, 272) bear somewhat intelligible legends: IИ VƧTI NIVNV, INIVST NIΛNVV, IOTIVINIVSN and VISΛV IVΛΛO. The rest bear legends which are consistent among themselves as well as being closer to the two legible Zorita coins than the non-Zorita legible coins. The more jumbled legends, as is always the case in Visigothic coins have a more self-conscious structure than immediate impression would reveal. The Zorita legends are as follows on the obserse: ITVΛIVIC or ITΛVIITΛVII; and on the reverse IИ·IOΛVIIIV or ITTII VI. The "conob" is always image The weights of the seven coins are all below 1.40 gms. with the exception of one (1.11, 1.12, 1.23, 1.33, 1.34, 1.34. 1.41 gms.).

The distinctiveness of the Zorita coins must imply either a different mint or a different period, or both. It would seem logical for them to be part of the last issues of JAN 2 and therefore possibly after a.d. 550 and the beginnings of Byzantine occupation. A later date could be substantiated by style. The types are similar to the rest of the group but much more reduced to linear patterns. The Victoria is a stick figure and its legs are much closer to the form of the Leovigild insect variety. The profile bust is almost completely without mass.

JAN 2 is one of the most significant groups for the late stages of the anonymous coinage. Its coins (e.g., Plate XIV, 278, 280, 282) form direct ties with group JAN 8 from which the major "curru" groups and the final Leovigild INCLITUS REX must evolve. It also has an indirect relationship to or influences groups JAN 4 and JAN 5 from which stem almost all of the remaining anonymous issues of Justin II and Leovigild.

JAN 2a (Plate XIV, 278–287)

This variation is of considerable importance because it may form the link between JAN 2 and JAN 8, since it naturally forms a step in the stylistic progression between these two groups. This variant may be interpreted, consequently, as a late issue of the JAN 2 group from which the later JAN 8 develops, or it may be interpreted as a concomittant variation and development. The legends and weights, as well as design, are all in line with JAN 2.

JAN 2b (Plate XIV, 288–295)

These coins are further variations and perhaps would be more correctly grouped as "unsorted;" nevertheless, their resemblance to JAN 2 is apparent with some qualification. Such coins as Nos. 292–294 (Plate XIV) are so crudely formed that they do not seem possible as products of an official mint. They are also dissimilar to major types of coins studied so far and therefore suggest an unofficial mint. Coin No. 295 (Plate XIV) presents an interesting variation of JAN 2 because it maintains the same type style with a completely different flange composition.

JAN 2c–2d (Plate XIV, 296–297)

These coins are distinctive not only in the realm of variant designs but also because of letters in the reverse field. JAN 2c has a T in the left reverse field, while JAN 2d has an A (?) in the right obverse field.

GROUP JAN 3 (Plate XV, 298–312)

These coins are a continuation of the long chest type of JI 2, with the same unmistakably rich, delicate linear style and more realistic facial modeling. This group is comprised of coins which still adhere to classic aesthetic values, as particularly seen in the imperial portrait. The non-pectoral cross-chest type still survives, although it is obviously being replaced by the pectoral cross form. Drapery folds are never abandoned in the pectoral cross area, however, for they are squeezed in around, above, and sometimes below the cross (Plate XV, 300, 303–304). image is simplified to image

The Victory figure continues the type which can be traced through the A 4–JI 2 progression: truncated pyramid chiton with ribbon skirts; delicately slender and graceful wings; continued discernibility of facial features. The ribbon folds of the Victoria's skirt are the same in number as in the A 1 progression; however they are not so likely to be confused with the legs of the figure as on those of the A 1 progression, (e.g., Plate XV, 299). This functional awareness of the integrity of the human figure is revealed in a remarkably conceived coin (Plate XV, 308), where the skirt folds are almost nonexistent, and the figure strides boldly on two strong legs. This particular coin indicates the remarkable feeling for rhythmic design and delicacy of line characteristic of the A 4–JI 1–JAN 3 atelier. It is not unreasonable, then, to accept No. 309 (Plate XV) as a late stylization in this group. This coin presents a stick Victory with the distinctive graceful and delicate ladder wings, without details of costume, and with only two legs. The head, although still with features, very quickly transforms itself into the X or image coxcomb shape seen in the issues of JAN 5. Here, too, the obverse portrait has been reduced to an almost completely linear pattern with devices not very different from those of JAN 2, although with a different effect. The head, or more correctly, the linear strands of hair are divided into back and front sections by the almost vertical single diadem line.

The obverse legends remain very legible and the reverse legends after a few VICTORIΛ ΛVCVSTORVΛ and VICTORIΛ ΛVCVSTORO tend to abridge to VICTORI ΛVCVSI. The weights remain good with only three descending below 1.40 gms., although No. 309 (Plate XV) goes as low as 1.08 gms.

As JAN 2, JAN 3 is important for its connection with the late autonomous coinage. Its role is not so clearly and visibly defined as that of JAN 2, but an indirect relationship with JAN 5 can be acknowledged. If more coins of the group were known to us, its role might not be so problematical.

JAN 3a–3b (Plate XV, 313–315)

These variables are unique if not even bizarre. It is not difficult to establish a case for Nos. 313–314 as alterations and modifications of group JAN 3.

No. 315 bears the usual solidus legend rather than that of a tremissis.

GROUP JAN 4 (Plates XV–XVI, 316–329)

This group cannot be placed in a direct relationship with either of the A 1 or A 4 progressions but offers us a new type which may or may not suggest a different and possibly new atelier. Nevertheless, the abstraction seen here is completely within the Visigothic tradition. The chest is a simple trapezoid, outlined in two or three lines of dots with a pectoral cross in the center. The facial type reminds us of a JAN 2 in its proportions, but the treatment of the top and back of the head varies. The diadem has been made so as to cut horizontally back from the forehead and eye, thereby deleting the frontal lobe found on the other coins. Sprouting upward from the diadem are the vertical strands of hair which create again the illusion of hair standing on end. This effect was already produced in the earlier JI 3, which, however, bears no stylistic relationship to the present group.

The Victory is a long-legged, elongated type with front and back skirt folds that in some examples assume the character of legs. With their carrot-shaped bodies, they come close to the figures of JAN 2. The head is modeled very well in some examples, with a very pronounced line of the nose, image and is a prelude to the X shape head (Plate XVI, 321). This has also been previously noted in the JAN 2 group.

The distinctiveness of JAN 4 is also indicated in the consistent reverse legend: VICTORI ΛVCTINI or VICTORI ΛVCTIN. The letters are all well formed, and the obverse handling of Justinian's name is always legible and correct. There is no evidence of the usage of the triangular punches observed in JI 1, JAN 1 and JAN 2. As with the legends, the weights remain of good standard, the two lowest are 1.36 and 1.39 gms.

JAN 4a (Plate XVI, 330–331)

Attached to this group are two coins which may indicate a transitional step into the "curru" coinage. The first item bears a reverse legend which is beginning to assume the "curru" form: VTΛVOII IIOΛVRRV. Its Victory has also acquired the "rabbit-eared" head. The second coin (Plate XVI, 331), which varies slightly in chest type from the preceding one, seems definitely to lead to "curru" group C 5. Although there is no consistency in the reverse legends, particularly the reverse legend of the second coin which so much more approximates the C 5 form but does not suggest the "curru" motif, the general character of both seems to relate them to each other and to JAN 4.

JAN 4b (Plate XVI, 332–333)

One anomalous item may or may not be associated with this group. The legends are in good Justinian order, yet the portrait is on the exact form of "curru" group C 1. The Victory figure, on the other hand, is unique but it relates to this group JAN 4 in its head form and particularly in its relationship to the reverse legend. No other coin studied presents the skirt in the form it achieves here. On the obverse chest, the gold beads appear almost machine made with holes in the centers of some of them. At best this coin may be explained as a borrowing of types from different mints or a Suevian imitation. I am strongly inclined to consider it not Visigothic.

The Grierson coin, although far more crude, might be included in this group.

GROUP JAN 5 (Plates XVI–XVIII, 334–361)

The second major group of Justinian coins also consists of coins from the Zorita Hoard. The group bears a resemblance to the other major group JAN 2, which also includes Zorita coins. The stylization lines of both seem contemporaneous but in slightly variant directions. The head is longer. A characteristic diadem is formed by two almost vertical strokes; a marked hair line frames the face, and the whole head area is conceived in lines of hair strands in the hair-on-end fashion. A crescent shape identifies the ear; a round bead with a circle around it the eye. The chest is a square formed by a double line of round beads, with sometimes a third line on the right side, recalling the old pleated sleeve (Plate XVII, 344). The portrait is not without comparison to the one finally seen on Leovigild's coins.

The stick Victory is two-legged with two frontal appendages that are sometimes short and sometimes long enough to be confused with the legs. Her head is either X shaped or a rough coxcomb image and, on one Zorita coin (Plate XVII, 347), a round ball. This Victory figure recalls those of JI 4, but the proportions and execution are completely different. There are even distinctions between this and the very stylized Victory in JAN 3 (Plate XV, 309) where the skirt appendages have been entirely dropped. This distinctiveness on the part of JAN 5 coins may very well point to a new separate atelier as we have already granted to JAN 4. However, the connection with JAN 2 and particularly JAN 3 which has potentially three, four, or six legged Victories may be stronger than I am able to establish.

The legends are most consistent, ƆN IVSTINΛNVS PΛVC and VICTORΛ ΛVƧT or ΛVƧTO or ΛCVSTO or ΛΛVƧTO or ΛΛVƧTOI or VVΛƧT or ΛΛVƧTI. The reverse legend recalls both JI 2 and JAN 3. Even the manner of executing the letters and the clarity of the form is in the best traditions of the A 4–JI 2–JAN 3 progression. However, as in JAN 2, "conob" is formed as image or image The weight standard is maintained around 1.43 gms. although eight coins fall below 1.40 gms. with some going as low as 1.07 gms. This tendency has also been noted in some coins of JAN 3.

Contrary to what was indicated with Zorita coins in JAN 2, Zorita coins here relate to the rest of the group on an equal level of quality in style and legends. The importance of JAN 5 is in its being a possibly early example of a mint which also produces some major Justin II types.

GROUP JAN 6 (Plate XVIII, 362–364)

The three coins classified here are variations on the JAN 2 theme. They unquestionably relate to the A 1-JAN 2 progression. The legends are now fully jumbled, although, as is always the case, completely consistent within the orthographic customs of the mint. Punches are used in one case, although not so extensively as in the other two items. The legends: ƆIVIOIΛ IIOIVIC and ITIVIV IIIVITI are not too far afield from the jumbled types in JAN 2. The weights are standard: 1.42, 1.43, 1.45 gms.

The chest type is a simple trapezoid as in JAN 4 but richly decorated with gold bead borders and ladder-like patterns in the borders image The head is a typically late version of JAN 2 with an almost vertical diadem. The Victories with coxcomb and X shaped heads also are from JAN 2.

GROUP JAN 7 (Plate XVIII, 365–367)

These three coins are also variations of JAN 2. The obverse type is more exactly in accordance with JAN 2, while the Victory on the reverse offers a possibly later development with its "rabbit-eared" head. There is also the dot between the main legs of the figure. The reverse is in order with those found in the "curru" and INCLITUS REX issues. Triangular punches are used extensively for the letters which are again garbled but not inconsistent with those recorded for JAN 6: ƆIITΛV· ·IITA: V·, IΛTIIV ITΛVII. The weights of these coins are lower: 1.37, 1.38, 1.39 gms. The flange is large like that of JAN 6.

GROUP JAN 8 (Plates XVIII–XIX, 368–386)

This important group must extend from JAN 1 and JAN 2 because it is a transition between JAN 1 and JAN 2 types and "curru" and INCLITUS REX types. The chest is that of JAN 1 image while the head is that of JAN 2, with the addition of a vertical diadem (Plate XIV, 280–281). The more realistic modeling of JAN 1 has been maintained, although the manner of stylization is similar to JAN 5. The general appearance of the coin is more sophisticated.

The Victory is also of the six-legged variety which begins with the carefully featured head that immediately precedes the X shaped development. Some coins bear the dot, a feature of "curru" and Leovigild, between the center legs.

All the legends are achieved by a combination of punches, and it is not too difficult to see how the "curru" legend might have developed naturally, thereby withdrawing the necessity of ascribing any intelligible meaning. All of the weights are maintained between 1.40 and 1.45 gms.

JAN 8a (Plate XIX, 387–389)

The coins of this group may be non-Visigothic; No. 387 is a gross misinterpretation of the group style; and No. 389 is a variable but also has a low weight inconsistent with the rest of the group.

JAN 8b (Plate XIX, 390)

This coin has a reversed profile, is poorly executed and consequently seems too crude for the rest of the group.

GROUP JAN 9 (Plate XIX, 391–393)

These coins form a distinct unit of their own. The Alesia example seems to begin the group, with its more accurately modeled imperial head and the articulate craftsmanship in type details as well as in the letters of the legends. The Victory is formed with considerable freedom, seen infrequently in coins of JAN 3, (e.g., Plate XV, 308) and in those of JAN 4. The chest type seems to have developed from more complex examples such as seen in JAN 1 and JAN 2 and leads directly to the simple truncated form seen in the remaining two coins of the group (Plate XIX, 392–393). The character of the profile bust recalls the group JI 1, but the total style of the coin is unique in that it has no ancestral ties with any of the Anastasius, Justin I, and Justinian groups already classified.

The two other items that make up this group are from the same die and present a simplification of the Alesia examples. The chest is a truncated shape with pectoral cross. The modeling of the head is only a mass for the neck and the face. In order to achieve some of the detail of the Alesia coin, the die cutter had to enlarge the head with the resultant crowding of the head with the legend. The Victory maintains the rhythmic sweep of the Alesia coin, but even this has been reduced by combining and extending lines, e.g., one concave line forms the top of wings and arms; another forms the bottom of the wings, the chiton, and the two front appendages image

The weights are similar: 1.435, 1.43, 1.44 gms. The legends, to the contrary, have little if any connection. Since the three coins form a distinct class of their own, with unacknowledged ties to other groups, and since two of the three coins come from known hoards at Alesia and Gourdon and are not of Spanish provenance, it may be concluded that these are not of Spanish origin and therefore non-Visigothic.

GROUP JAN 10 (Plate XIX, 394–395)

The two coins of this group are also unique and difficult to relate generically to any of the groups classified. The craftsmanship is very good, and the concept of design is fully realized. There is sophistication in the stylistic resolution and no evidence of crudity. The imperial portrait seems to suggest JAN 3, (e.g., Plate XV, 303, 304, 308) and the originality of such a Victory reverse as seen in JAN 3a. The proportions of the obverse face are further reminiscent of the earlier JI 3.

The unusual Victory striding left instead of the customary right has been carefully designed to resemble more closely a human figure. The subsidiary appendages are paired with each of two legs, suggesting more accurately a "culotte-like" costume. The head of the Victory is well articulated; eye, nose, ear, mouth, cheek, forehead and hair mass. The wings show an equal concern for a deliberate, delicate touch substituting for the now customary ladder wings, a graceful herringbone pattern image The delicate workmanship, as well as a distinct rhythmic feeling for design, associates these coins with JAN 3.

The obverse legend is fairly regular and relatively accurate. The reverse legend has been reversed just as the Victory has been, which adds to the problematical character of these coins. When the reversed reverse legend is reversed it reads: ICTORИI COVTOIΛ. Both these coins may be products of the same die.

GROUP JAN 11 (Plates XIX–XX, 396–400)

This group descends directly from JI 5 and is a unique example of a mint product of exactly the same die style in both the reigns of Justin I and Justinian. It further bears the distinction of containing coins which substitute the names of the Merovingian Kings Childebert I and Theodebert I for that of Justinian. This makes the Burgundian-Merovingian attribution of the entire group unquestionable. The Childebert coins are without monograms (Plate C, 12).

The Theodebert coins are found with and without monograms, image image and also LV (Plate C, 10–11). The Victory remains fairly constant throughout, although there is some variation in the portrait bust. The pectoral cross is always absent, permitting a more elaborate system of linear folds. All the portrait busts bear resemblances to types found in Merovingian coins, (e.g., Plate XX, 399).

JAN 11a (Plate XX, 401–402)

The distinction here is that the position of the star and the monogram on the reverse field have been reversed. In addition, a new monogram or letter image is substituted for the old monogram image

JAN 11b (Plate XX, 403–406)

These coins follow exactly the design of the major group but are without stars, monograms, or letters on the reverse field.

The Gourdon piece No. 406 possesses a Victory form that suggests the Victory of JAN 9 in a more stilted and rigid manner. The stylistic relationships of the coins of this group with coins of JAN 9 and JI 5 as well as with coinage of French provenance suggest again a non-Spanish, non-Visigothic, but Merovingian attribution.

JAN 11c (Plate XX, 407)

The single item here not only is a much cruder performance but bears the letter K in the right reverse field.

GROUP JII 1 (Plate XX, 408–410)

These coins descend from JAN 3. Besides the legend tradition that is maintained, there is the handling of the letters as well as the general workmanship. The technical handling of the facial features on the obverse portrait compares exactly with that on JAN 3 coins (e.g., Plate XV, 304). The excellent modeling of the face and the subtleties achieved in the variations of cheek, jaw and mouth, as well as the conception of the eye and nose are all in JAN 3 tradition. The chest is a simplification in the manner of No. 310 (Plate XV) and retains an almost unconscious motif, the curved drapery fold under the pectoral cross, which was noted in the JAN 3 issues. The characteristics of the long chest have been retained. The head structure is not unlike those achieved in JAN 2 and JAN 8 and reveals how contemporaneous these issues may be.

The Victory figure with X shaped head also has the rhythmic sweep of wings characteristic of JAN 3. This group is a later and more successfully conceived design of No. 310. These coins may form an intermediate step between JAN 3 and JII 7.

GROUP JII 2 (Plates XX–XXII, 411–434)

This group is composed of a large number of coins from the Zorita Hoard and unquestionably descends from JAN 5. There is approxi- mately the same degree of stylization observed in the relationship of both types. The rectangular or trapezoidal chest, the vertical diadem, the carrot-stick Victoria with four legs all encourage the ancestral connection with JAN 5. This is further seen in a cursory check with the legends, for JII 2 also has a propensity for those variations of VICTOR ΛΛVƧTOI or VICTORII ΛΛVSTI, while the obverse reads: ƆIIVST VSPΛVC and never loses the Justin identification.

There is a fairly wide degree of variation in technical workmanship within the group. These coins are much cruder than JII 1. The style quickly degenerates from JAN 5's portrait to the hair-standing-on-end variety (Plate XXI, 423, 428, 430). In one example (No. 430) the Victory has the rabbit-eared headdress without the ears. The Zorita Hoard again provides many of the crude and singular specimens (e.g., Plate XXI, 429–431), as well as all of the crude and questionable variables in JII 2b.

The weights are as variable as the type styles. The standard is the lowest studied so far; out of thirty coins studied and classified in this group only seven are 1.40 and above, and only three are between 1.30–1.40 gms. There are two recorded weights as low as 0.86 and 0.98 gms. There is no direct proportion between crude and low weight coins, although there is a direct proportion between stylistic variability within a group and weight variability. Three of the five crude Zorita issues with reversed Victories weigh 1.41, 1.47, 1.48 gms. The crude and bad condition of No. 433 does suggest No. 315 (Plate XV) in JAN 3b, further encouraging the connections of JII 2 with JAN 5 and JAN 3.

JII 2a (Plate XXII, 435–437)

Grouped here are three coins of relatively similar conception, in one both types and legends are reversed, in a second this alteration is seen only on the reverse. The curious Zorita item, No. 437, bears a legend on the reverse which might be read as Liuva: VLIVV V·UVΛ. 2

End Notes
Cabré classifies this (Zorita No. 45) as a garbled legend of Justin II. I have discussed this above on pp. 65–66.

JII 2b (Plate XXII, 438–440)

These are very crude examples, two with reversed Victories, and one with both obverse and reverse types reversed. These, all Zorita pieces, have been attributed to the Visigothic Narbonnensis by Cabré, yet the design is more of an imitation of Visigothic ideas rather than an original product. These must be placed in the category of unofficial strikes.

GROUP JII 3 (Plates XXII–XXIV, 441–465)

The derivation of this distinctive style is more complex and does not stem from a single prototype. It is associated with JII 2 although not so crude, yet involved enough in similar design attitude that more stylized coins are difficult to separate into groups. Its connection with JII 2 predisposes its dependence indirectly on JAN 5 and likewise vaguely back to JAN 3 and its high standard of craftsmanship. Its Victory figures can also not help but evoke faint remembrances of the small A 2–JI 4 groups. Unquestionably we are dealing with a new group whose motivating source is outside the realm of coin progressions we have studied. The outside source is very quickly assimilated by the mint designers, and the coins with modification readjust to the stylistic level of JII 2.

This group is predominantly composed of coins from Zorita, a fact which adds to the interest of the group. Most of the series is characterized by a delicate rendition of small types. The profile portrait presents an interesting illusion of a more refined, realistic portrait in a more accurate, minute style. The square chest is worked with a double border of round beads and a large pectoral cross crowding the center. The mass of the neck and head are well modeled with the transition from neck to jaw and cheek and eye skillfully conceived in the minute scale. The diadem, almost lost in some cases because of the rich and tight treatment of the hair, is composed of round beads. It is the infulas with its two ends in round beads that is particularly emphatic. Care is given to the features with even the ear being accurately suggested. A strong contour outlines the forehead and nose, and round beads designate the chin, the lips and the nostrils. A round bead set in the modeled eye socket is topped by a strong eyebrow line. The effect of wearing enunciates the profile contour and flattens the rich headdress to a raised smooth mass (Plate XXII, 444). In some the delicate beads of the diadem are still partly visible, as in No. 469 (Plate XXIV) which is grouped in JII 3a.

The wearing of the dies necessitates their being reworked, which might explain the linear, articulated and traced-like quality of No. 449 (Plate XXIII), where the old modeling is still visible under the newly re-etched lines. The wearing also demands new dies which begin to resolve the technical problems of maintaining the craftsmanship of the first portraits. Also, the new dies begin to harmonize the first style with the general style of other contemporary mints. This is to be seen in No. 450 (Plate XXIII). Here the almost vertical diadem, not overly emphasized so as to destroy the integrity of the roundness of the head, separates the half-moon-shaped back from the spherical triangle of the frontal lobe. The accentuated eyebrow of the original substitutes for a hair line and joins with the contour of the brow and nose. Modeling of cheek and jaw and neck are still subtlely handled. The nostrils are now a straight line forming an angle with the nose profile; and the lips are also now effected by straight lines rather than round beads. The ear is still apparent, although in some cases such as No. 452 (Plate XXIII) it begins to lose its visual meaning. The excellent craftsmanship of the artisan is still apparent in the handling of the beaded diadem and the infulas image and of the center jewel at the top of the head. The chest remains as two dimensional as it was in the first pieces. The inspiration for these coins must be Imperial Byzantine issues such as those minted at Ravenna under Justin II, which they closely resemble. 3

The next step is seen in coins such as Nos. 457–460 (Plate XXIII) in which the size of the head has been enlarged to permit the execution of the facial and decorative details so consistently evident in this group. The chest is invariable, and the good quality of the head and neck persists. The size of the features is increased but handled similarly to those of No. 450 (Plate XXIII). The enlargement of the head causes the first major breakdown, because the round back of the head must be curtailed. In No. 460, the beaded diadem has taken the place of the former triangular frontal lobe, and there is only space for a few strokes for hair beyond the diadem. In later crude examples (Plate XXIV, 462–463) all parts of the head have been restored to the design at the expense of naturalism, as seen still in No. 460.

Throughout all these changes the reverse face of the coin has remained constant. The Victory figure never attained the naturalistic heights of the obverse portrait. It is always a stick figure with three or four legs, sometimes with faint back appendages, ladder wings, and coxcomb-shaped or simply rounded mass-shaped head. That the imperial silver coins from Ravenna bear only letters or crosses on the reverse, thereby providing an inspiration for only the obverse face, may explain the differences in the execution of the reverse and obverse faces.

The weights are consistently good, and only a few are found below 1.40 gms. The legends are also consistently good in noting Justin's name and in forming a reverse of ƆVICCTO VΛC with very little variation.

JII 3a (Plates XXIV–XXV, 466–477)

The coins placed here are natural, later developments of JII 3, which take the mint into the reign of Leovigild and the last moments of the VPW tremissis. The facial profiles of the obverse bust all indicate the result of the simplification of the early facial types of JII 3. The proportions of the head have returned to those seen on the early coins. The major change here is in the handling of the Victory, which has become more fluid. In some coins it has even managed facial features (Plate XXIV, 466–467), or the X shape so prevalent in the period of Justin II (Plate XXIV, 469–470). In such items as Nos. 471 and 472 (Plate XXIV) the Victory is well modeled and pleasantly conceived; the legs are modeled at the calves, although the figure is still a stick form. The facial profile is delineated, and the head is characterized as a hat-wearing type.

For the remaining issues of this mint, two types of Victory forms are used with several portrait bust types. However, the placing of them all in one mint seems logical. The Victories of Zorita Nos. 475 (Plate XXIV) and 477 (Plate XXV) are examples of the large naturalistic head types, and Nos. 473, 474 (Plate XXIV) and No. 476 (Plate XXIV) illustrate the anonymous ball-shaped head similar to the earliest Victories of JII 3. All the obverse portraits have the elongated chest type, with two in the manner of all JII 3 busts and three others influenced by developments in Leovigild's other mints image No. 477 approximates the chest of JII 5 image It is in these last coins that we begin to see the name of Leovigild, and one of these, No. 474, was classified by Miles as Type G of the Leovigild coins without mint-name. 4

One major consistency within almost all the Zorita coins in this group is that they are of a poorer quality of gold. This is true for non-Zorita coins as well. The weights, except for very few, however, are very good.

End Notes

BMCVOL, pl. XVII, 25–33.

GROUP JII 4 (Plates XXV–XXVI, 478–494)

These coins best illustrate the influence of various mints upon each other even when not within the direct progression. Associations, particularly in the treatment of the profile head, seem to link this group to JAN 4, yet the Victories, themselves unique, are closer to those of some of the JAN 3 group, such as Nos. 311–314 (Plate XV). General contemporary similarities are shared with JII 2 and JII 3.

A coin such as No. 478 (Plate XXV) may well mark an early example, since its Victory so closely follows that of No. 312. The head still maintains a few strokes for hair on the top and back, a diadem with bead infulas, and a small triangular frontal lobe. This head also recalls a JAN 3 item (No. 310) as does the elongated chest, double-bordered by round beads. The modeling of the face and treatment of nose and eye follow those coins of JAN 3.

With No. 480 (Plate XXV) the chest type of JAN 6 is immediately recalled. In place of the round beads making the square contours of the chest, a small triangular punch has been used image with the ladder pattern in the upper horizontal border and two round beads on each shoulder. The use of the triangular punch is enlarged, then, in this group, for it is not only exclusively used to form the letters, which perhaps accounts for the IVIIVIIIV inscriptions, but also to form the chest as well as the palm branch held by the Victory. Even the X head of the Victoria has been achieved by a single manufactured punch image This use of the punch may have been experimented with first in a JAN 3 variable (Plate XV, 313), on the chest and the palm branch particularly. The Victory figure with two or three legs as seen in JII 4 may develop from such as that of No. 313. Note especially the arm characteristically bent at the elbow, which is another indication of the search for a more naturalistic conception. Such awareness is still shared in the conception of the portrait in No. 479. This is even seen in a more simplified chest variation as in a Zorita item (Plate XXVI, 497) of JII 4a. It is the emphasis on the diadem, however, perhaps the almost cravass-like separation between its two strands, that eradicates the frontal lobe. A coin such as No. 481 (Plate XXV) seems almost the result of a pencil tracing of No. 479, and even the round beads on the shoulder are replaced by semi-circular outlines image

All these developed aspects remain constant for all the coins whether they read IVNVIIIVNVN, IVIIIΛIIVNVI, or IVIVIVIVIVIVIVIVIVIVIV, or DN LIVVIGILDVS REX, and regardless of whether they bear long or short chest types. Coins (JII 4a) such as No. 497 and No. 498 (Plate XXVI), although style variations, also carry such legends. Coin No. 500 (Plate XXVI), a Suevian (?) variation, bears a diverse garbled legend. A variation such as that in the Volker's collection (Plate XXVI, 501 [JII 4b]), whose legend attempts to spell LIVVIGILDVSREX and is not made by a triangular punch, is almost too crude for the group. This is the most plausible group to which to attach it, however, and it may be characterized as a very crude imitation, Suevian or otherwise.

The consistency in the system of letters evolved illustrates a conscious mint decision whether or not it is caused by the incapabilities, orthographic ignorance, or the illiteracy of the mint workers. The important point is that there is at work a conscious aesthetic concern for order, symmetry and design clarity. The mint workers, nevertheless, very quickly learn to spell Leovigild's name, and these coins are exactly the same type and style as those without his name. Coins of this group have been classed by Miles as Types B, F, and G of the group of Leovigild's coins without mint names. 5

End Notes

Miles, p. 178. Miles classified No. 475 as Type A (p. 175).

GROUP JII 5 (Plates XXVI–XXVIII, 502–524)

This series begins with coins bearing the inscription that possibly can be read as Justin (Justin II) and ends with coins in the name of Leovigild. They possess types that bear resemblances to other contemporary Justin II groups but are also distinctive in character. The portrait has an elaborate chest; its design pattern being the final stage in a long development. It is square, or sometimes an elongated rectangle with a fibula on each shoulder balanced by two fibulae at the bottom of each side. The lower fibulae were once the first and last letters of the obverse legend (the reverse Ɔ which developed from the D and the final C). The pectoral cross hangs from a triangle and seems to depict a pendant jewel. A second version of this chest replaces the inverted triangle with two curving lines suggesting breasts image The head is small, slightly modeled with a strong diagonal diadem, stressed infulas and hair that stands on end.

The Victoria, a carrot figure with six legs, bears a head of recognizable human form and like the obverse portrait seems to develop out of group JI 3. Even the legends further help to ally these groups, for the particular manner in which the letter R is formed by a circular and two triangular punches image is to be found only in one other group. That is in JI 1d which may further tie JII 5 with the A 1–A 3–JI 1– JI 3 progression. When the name of Leovigild appears, it is well handled and only here does it read VCLIVVIҀILDIREGIS rather than LIVVIGILDVSREX.

There is a unique coin of JII 5a (Plate XXVIII, 525) from the Johns Hopkins collection which bears a "curru" legend variation while also using in one place the characteristic image of JII 5. The Victory is exactly that of JII 5; the obverse chest is a variation of "curru" group C 3 image and the head is a variation of C 1. Note that even with the changes the product is still within the realm of JII 5 and seems to illustrate the mint of JII 5 attempting to adopt the type established in the mint producing "curru" coinage.

Those coins of group JII 5b (Plate XXVIII, 526–527) and JII 5c (Plate XXVIII, 528) are variations of JII 5 that seem to be imitations of it rather than products of the same mint. (Suevian?)

End Notes

Miles, pl. I, 4, 5, 6.

GROUP JII 6 (Plate XXVIII, 529–533)

This is another group that forms out of products or variations of JAN 5. The Zorita coin No. 529 (Plate XXVIII) bears a marked resemblance to JII 2, and has a characteristic chest, Victory and reverse legend of JAN 5. It seems immediately to direct the attention to later coins of garbled legend (Plate XXVIII, 530–533). The portrait heads of these remain the same as the Justinian piece; the chest, however, imitates that of the second type of JII 5 image The Victory type in two cases repeats that of the JAN 5 group. In the other two (Nos. 532, 533), a unique Victory is found who wears a triangular gown, has only two legs and a suggested human head image All these coins, like most of those of JII 3, are of "oro bajo." This coin may be the result of the mint's attempt to imitate the mint product of JII 5. The two examples with gown-clad Victoriae may result from an attempt to naturalize the appendages or insect legs into a more rational image, for note that the stick figure of Victoria is still suggested beneath the gown. This naturalistic instinct and desire is characteristic of those groups directly or indirectly associated with JAN 3 and JAN 5. It is within this stylistic area that JII 3 with its Byzantine sources also is placed.

GROUP JII 7 (Plates XXVIII–XXIX, 534–540)

This sketchy and problematic group consists of possibly similar coins rather than "exactly alikes." The legends first read Justin, break down, and then spell out LEWIҀILDIR. Two of the Leovigild items come from the Zorita Hoard and bear solidi type reverse inscriptions, VISTOI IΛVCCC. If the development is accepted, the group begins with a coin (Plate XXVIII, 534) that is reminiscent of the long-chest-without-cross type of JAN 3, and the sense for rhythm and design here and on the Victory tends to fortify this suggestion. The Victoria follows the JAN 3 type. The second coin (Plate XXVIII, 535) may very well fall within the range of JII 3, but its X head and six-legged Victory abnegate this. Note, however, that the curving fold beneath the pectoral cross continues a JAN 3 suggestion. The emperor's diadem is vertically stressed and the modeling of the face is simply done. The third item (Plate XXVIII, 536) probably follows from this, and the crude (Plate XXIX, 537) garbled legend may also connect. The culmination is in the two Zorita Leovigild pieces with trapezoidal chests and heads in which the stylization of strands of hair and almost vertical diadem have not destroyed the round, compact contour-modeling of the head. This modeling is rather subtly achieved. The Victory is four-legged and designed in an attempt to imitate the insect Victory of the "curru" types. The customary bead between the legs (placed in the center of the region bounded by the innermost legs) is also present. The head is only a round mass. This Leovigild coin is in the same category as Miles Type C. 6

"Curru" Coinage

By all means of analysis, coins with the legends bearing variations of the form "curru" written backward and forward must date within the period of Justin II. Stylistically they are contemporary with all Justin II groups and are direct forebears of the final Leovigild VPW tremisses. It is therefore through these coins that one sees the closest resemblance to the national coinage of Leovigild.

They may be classed in five groups according to their variation of what seems a common theme. The variations may or may not indicate different mint products and may or may not indicate a chronological order of progression. That is, it is questionable whether or not we are to read the groups horizontally, vertically or both.

End Notes

Ibid., pl. 1, 3.

GROUP C 1 (Plates XXIX–XXX, 541–559)

The general characteristics of all "curru" coins are established and develop naturally out of JAN 8. The six-legged, rabbit-eared Victory, with its accompanying central bead between the legs, and the developed legends, follows directly those of JAN 8. The obverse portrait is dominated by a large, almost vertical diadem which separates the frontal triangle of hair above the forehead from the curving wild hair in the back image The head is therefore assuming the Leovigild type (Miles Type H). The chest is the major variable and motif determinant of the group. It is trapezoidal, with pointed shoulders. Its contours are achieved in round beads of either two or three rows to each side. The confines of the central pectoral cross area are echoed by a pattern of repeating semi-circular lines that are in either one or two rows. This creates a rich decorative effect image which is not found on the coins of JAN 8. The contours of this pattern however are those of JAN 8.

GROUP C 2 (Plate XXX, 560–567)

These coins follow exactly those of C 1 with the exception of two aspects. The first is the development of the shoulder line of group C 1 into epaulette-like fibulae. The second is the inclusion of the cross in the middle of either the obverse or reverse legend or in both in some of the specimens. Aside from these exceptions, coins of C 1 and C 2 are similar in types, weights, fabric, and legends. It would seem most likely that they are either contemporary issues of the same mint, or earlier and later issues of the same mint.

GROUP C 3 (Plates XXX–XXXII, 568–592)

The Leovigild VPW tremissis is almost resolved in some of these coins. They follow in legends and fabric C 1 and C 2. The C 1 and C 2 types have been further developed or varied in the direction of ever-increasing proximity to Miles Type H. The diadem has been vertically placed, and the chest has been simplified. In what may be the first examples of this group (Plates XXX–XXXI, 568–573) the chest continues to have the three round beaded borders, but the epaulettes of C 2 have become round fibulae like those seen in JII 5, and the decoration of the pectoral cross area has been reduced to a curved line facing each of the four corners. The resolution is completed when the third beaded border is removed and the final Leovigild VPW tremissis bust and chest is achieved.

A development may also be seen in the Victoria, in which the final insect Victory is perfected. The six legs of the Victory in C 1, C 2 and in the early examples of C 3 (the three beaded bordered coins described above) are handled in this manner: image Within this group it is possible to trace the change to this image which almost arrives at the final step, the INCLITUS REX Victory image The only other variable aspect is the lower scale of weights in contrast to the good standards found in C 1 and C 2. In this respect as in all others, however, C3 leads directly to the INCLITUS REX issues.

GROUP C 4 (Plate XXXI, 593–595)

The variables here are important enough to suggest a different mint product in the manner of the C 3 group. The obverse portrait is the same except that the decorative curving lines have been removed. The Victory figure has almost lost its head, or it has become confused with a V in the legend. The figure is much more of the stick type and has only four legs. Furthermore, the whole scale of the coins and thereby the types is smaller. The fabric also differs as well as do the legends which read DVDVTV + ИVimageVƆ on the obverse and VISΔIVIIimageMΛI on the reverse. The three coins weigh 1.27, 1.34, and 1.40 gms., and Cabré classified the Zorita coin (Plate XXXII, 595) as made of "oro bajo."

It would seem unlikely that this comes from the same mint progression as C 1, C 2, or C 3 and may be the product of one of the mints that produced the earlier two, three, or four-legged Victories.

GROUP C 5 (Plates XXXII–XXXIII, 596–602)

These coins follow precisely the Victory and chest types of C 2. The legends, fabric and weights are all good, with the exception of the very low Zorita item (Plate XXXII, 598). The variation is in the portrait head which uses a flat-topped or horizontal diadem and hair standing vertically up from it. In two examples we get a variant that brings the head closer to the C 3 type image by describing the aspect of the frontal lobe. This seems to paraphrase the type of JAN 4 and the coins of JAN 4a, particularly No. 331 (Plate XVI). This group may then be considered a product of the JAN 4 mint imitating the "curru" standard of JAN 1-JAN 2-JAN 8.

One final "curru" coin which may or may not be connected with this group is No. 602 (Plate XXXII). This is a crude example. The chest type is a square version of C 3, and the border is composed not of round beads but of triangular punches that have been seen on coins of Group JII 4, which was also traced out of JAN 4. Besides this, the head is composed of a diadem that diagonally meets the forehead, and the hair behind this rises vertically in slight curves. Where the diadem meets the forehead contour, however, there is a second line parallel to the forehead contour which gives the impression of the beginning of the frontal lobe found in all "curru" portrait heads. The Victory figure is six-legged and roughly rabbit-eared but placed askew in relation to the exergue.

Inclitus Rex (Plates XXXIII–XXXIV, 603–630)

With this group bearing the name of Leovigild on the obverse and the legend INCLITUS REX on the reverse, the last step is reached in the development of the VPW tremisses. It follows directly from C 3 and illustrates the perfected or resolved abstract design of both obverse and reverse types. These are also the types that are found on the coin catalogued by Florez in which the name of the mint, Toledo, has replaced that of INCLITUS REX on the reverse.

This final design is totally two dimensional. All is reduced to line except for two areas of modeling which in themselves have been abstracted, flattened, and thereby stylized. The first area is the cigar-shaped or carrot-shaped body of the Victory; the second, the modeling of neck, cheek and jaw which becomes a single raised mass of this shape image All this is in order with coins of C 3. As already noted in the discussion of C 3, the only change observed in this issue is in the pattern of the six legs of the Victory which now form three lines parallel to each other at 45 degree angles to the body on each side.

The obverse portrait continues on the Leovigild cross-on-steps issue. It appears again when revived on the cross-on-steps issues of Reccesvinth (Miles 360 a, b, c) and makes its final appearance on an Egica issue from Taracona (Miles 427 a).

Hermenegild Issues

Coins bearing the name of Hermenegild are exceedingly rare. 7 The number of forgeries within this small group might make one suspicious of even the accepted authentic coins and to doubt that Hermenegild ever issued a coin. Regardless of the small number extant, nevertheless, the few pieces accepted as authentic are genuine and do fit into the stylistic pattern and fabric of Visigothic coins. Of the seven coins listed by Miles, one is lost, another has been stolen and is only known through a photograph, 8 another is a forgery (which he correctly suspects), and another which he does accept is very questionable and most likely another forgery. 9 Of the three remaining, all are of the same type and design although the legends vary. 10

H 1 (Plate XXXIV, 631)

This group contains a single coin. It bears the name of Hermenegild on the obverse and the legend INCLIT I image on the reverse. This coin which Miles has accepted as authentic has been described by Grierson as a cast and consequently a forgery. 11 Whether it is authentic or not, it does not belong to the same stylistic group as Hermenegild's two other INCLITUS REX coins as Miles suggested. It is of diverse design type. Its stylistic origins are complex, having associations with coins of JII 3, JII 6 and JII 7. The obverse chest type presents another variation of the large JII 5 group which also has been imitated in item 477 of JII 3 a and in all of the items of JII 6. The portrait head, however, is closer to the fine plastically modeled heads of JII 3 and JII 7 than to those of JII 5 or JII 6. The Victory on the reverse is developing the insect type seen on JII 5 and may mark an attempt of the style groups that relate to JAN 5 to design an insect Victory type which is not a part of their stylistic tradition. A similar attempt is seen on the reverses of JII 7. The stylistic relationships as well as even similar weight standards would most strongly connect this Hermenegild series to the already established JII 3, JII 6 and JII 7 groups and might attest the authenticity of this Hermenegild example. I must agree with Grierson, however, that "… the weight is too high for a coin that pretends to be badly worn" and that "… the lettering is unsatisfactory…". 12 The complexity of its style, furthermore, does not fit into the regular patterns of development. It is too much of a conglomeration.

End Notes
The lost item is Miles 47 (d), and the stolen one is Miles 47 (a), or, my no. 634 (Plate XXXIV) from Mateu y Llopis' catalogue of the Madrid Collection.
Miles is suspicious of the Johns Hopkins item, Miles 47 (c). However, he accepts the Hispanic Society of America item, Miles 46 (a), which I reject (Plate XXXIV, 631).
These are the following: no. 635 (LBM); no. 632 (PBN); no. 633 (VQR).
This judgment comes from Grierson's unpublished manuscript.

H 2 (Plate XXXIV, 632–633)

The two specimens placed in this group exactly duplicate Leovi-gild's INCLITUS REX series except, of course, for the different royal name on the obverse. There is no reason to question their authenticity. Grierson is correct in noting the diversity of the letter formation of the two coins of this group from those of the H 3 REGI A DEO VITA series. 13 The letters here are generally smaller and cruder. Punches are used exclusively. This is particularly seen in the letters R, D and E. In group H 3 the letters are larger, and the use of punches is combined with free-hand strokes to enhance legibility.

H 3 (Plate XXXIV, 634–635)

These represent the REGI A DEO VITA coins in the name of Hermenegild. Miles lists four, of which two have disappeared, and another, the Johns Hopkins example, is a forgery. The British Museum specimen is an exact imitation of the INCLITUS REX series of Leovigild, and there is no reason to question its legitimacy. On the basis of a photograph, the "stolen" Madrid item may safely be accepted. The Johns Hopkins example, which Miles considers as being suspicious, is unquestionably a forgery. There is not a coin in the entire VPW series designed and modeled in this way. Two aspects are immediately noted; first, the treatment of the neck and jaw region of the portrait head; second, the treatment of the lower torso of the Victory. On all of the "curru" and Leovigild INCLITUS REX coins, although the portrait head has been worked completely in line, the area of the neck and jaw still retains a remnant of the traditional modeling in the characteristic inverted "L" shape or the shape of a side-wise "T" image or image Only two graceful sweeping lines image form the neck and chin of the Johns Hopkins coin. This is the only coin of all that have been studied that introduces on the insect Victory a pattern of three short lines in the region customarily occupied by genitalia on a male figure. If genitalia are intended, then in the course of seventy-five years the sex of Victoria has changed. In addition, the design for the palm branch is unorthodox image as well as the treatment of the dot between the legs, which here is a ring or amulet punch rather than the usual bead. These charactertistics combined with the different way in which the letters of the legends are formed, cause me to believe that this is a forgery. This leaves very little for Hermenegild, but what is left is authentic.

End Notes
These quotations are taken from Grierson's unpublished manuscript.
Op. cit.

H 4 (Plate XXXIV, 636–637)

The two coins which make up this group are both in the collection of the Gabinete Numismático de Cataluña. Although they bear the REGI A DEO VITA legend of group H 3 and a similar obverse portrait chest and reverse insect Victory, there are design differences. The obverse portrait is a variation of the norm established on the INCLITUS REX issues. It possesses all the elements of the INCLITUS portrait, but the manner of execution is quite different. Here there is a greater use of modeling. The neck is a slightly raised, flat, even plane on which the head sits in higher relief; the features are evenly presented and, although wearing is indicated, the contours of nose and lips could never have been the linear forms of the INCLITUS REX type. The legible, broad, flat letters are very similar to the treatment of the modeled face and its features. The insect Victory varies slightly in the way the head is formed image rather than the image of the INCLITUS REX, and in the attachment of the wings to the body image rather than image There are also three graduated lines for the exergue image rather than the single one of all other types. Only the Johns Hopkins specimen which I have considered a forgery has this same peculiarity. A final eccentricity is the correct formation of the letter A. The distinctiveness of these two coins must be accounted for by either a different mint production than those of H 2 and H 3 or as forgeries. The latter explanation seems the most likely because these coins give the impression of having been designed by a man attuned to a more realistic aesthetic background attempting to simulate an abstract form created by an individual not realistically orientated. A final proof that one of these coins, if not both, is false is the abnormally high weight (2.247 gms.) of No. 636. This would be unique in Visigothic coinage.

End Notes

Reinhart, DJN 1940, p. 85. Reinhart knew of only three authentic items.


An analysis of the style groups and progressions and the manner in which abstractions develop should provide more material of a less circumstantial nature for the solution of problems of mint attribution and mint practices. The anonymous nature of the coins by their lack of any definite identifiable mint mark, symbol or letter, however, makes it impossible to offer conclusions that are more factual than hypothetical, even when they are classified on the basis of stylistic similarity. Considerations and judgments of style, no matter how carefully structured, are at the final moment the responsibility of individual subjectivity, if not fantasy. I do not consider myself immune to such eventualities.

When reviewing the classifications of the VPW tremisses from Anastasius to Leovigild, certain factors become apparent on the basis of which a number of hypotheses may be suggested to provide a structure for mint attribution. Such factors may be classed in the three categories of design style, inscriptions, and weights. The full interpretation of all these observations depends on the reconstruction of sixth century mint practices, which in turn is determined by the analysis of these factors.


1. The observation of primary importance, formed by a study of Chart VIII which is a visual projection of the style group relationships spelled out in the preceding chapter, is that the majority of the groups is related to definite Visigothic Leovigild coins. From this I suggest that all such major groups must be products of Visigothic mints. It must also be concluded by virtue of quantity that the Visigoths were primarily involved in the issuance of the VPW.

2. The number of groups within the Visigothic orbit increases as we proceed from Anastasius to Leovigild. There is also an increasing number of variables within each group. Simply, Visigothic coinage within this period moves in the direction of greater complexity and confusion. This is true for legends and weights as well as for types. There are, however, general consistencies among contemporary groups within each imperial reign. There are similarities in aesthetic directions and motivation. These consistencies override the boundaries of the lines of developmental progressions. All the Anastasius groups share in their predominant maintenance of naturalistic effects, while moving in the direction of geometric simplification. All the Justin I groups share in the continuation of these practices which are completed in the Justinian groups. The geometric, two-dimensional stylization of the chest is achieved in JAN 2 and JAN 5. By the time of Justinian the cross-on-chest motif becomes universal in Spanish coins. All the Justin II chest types are in order with their square or trapezoidal contours.

The consistencies within the development of groups and within and outside of the ranging stylistic progression series are extremely important in establishing the general boundaries of what must be considered Visigothic coinage. They assist in the establishment of aesthetic principles by which Visigothic issues may be distinguished from non-Visigothic. It is not possible to determine whether these consistencies in aesthetic attitude are due to centralized official policy, common sources of origin, or mutual regional influences.

The increasing number of groups and the existence of groups which do not follow directly the style progression of Justin I groups such as JAN 4, JAN 5 and JII 3 might be a manifestation of the political confusion and breakdown after 550, and the struggles against the Byzantine forces and the Suevians, as well as the orthodox Hispano-Roman population. At the same time there is an increasing possibility of Suevian imitations of Visigothic coins, unofficial mint products, as well as Byzantine issues and their influence, such as evidenced in JII 3. Concomittant with this is the transition of the Visigothic State from French to Spanish with settlements spreading increasingly further south into the peninsula and necessitating the moving of the capital to the south. From Amalaric to Leovigild there is increasing political and fiscal control of the peninsula by the Visi- goths. This expansion, as well as the moving of the capital, might have caused the establishment of a larger number of mints.

3. The increasing development towards uniformity, visible in JII issues, even notwithstanding the naturalistic innovations of JII 3, as well as the establishment of the "curru" issues, might indicate the increasing strength of the Visigoths and a stronger control of monetary practices under Leovigild. Leovigild's unification and standardization of mint products and practices as indicated in his cross-on-steps reverse and facing busts issues is seen in its early stages in the "curru" and INCLITUS REX series. This is indicated by stylistic groups in which more than one mint is involved in the issuance of "curru" coins (C 4, C 5, JII 5a are each separate from the mint of C 1, C 2, C 3). Such a policy was justified in light of the variety of coins in the groups without mint marks in the time of Leovigild.

4. Stylistic progressions may be arranged so that Leovigild types may be traced back to Anastasius issues. This stylistic development or the development of the Leovigild abstract forms out of the naturalistic ones of Anastasius results naturally from the combination of the aesthetic desire for simplification and the technical need for facility. The INCLITUS REX issues, although apparently of short duration, represent the successful resolution in which aesthetic principles are satisfied within the limitations of technical abilities. Thus really crude renderings in design types are avoided.

It must be understood that usually a stylistic program does not necessarily establish a correspondingly chronological relationship. The grading of a design on the developing scale from naturalistic to abstract form does not signify that all more abstract forms date later than non-abstract forms. It is always possible that such diverse styles might be contemporaneous, and that these progressions are to be read horizontally rather than or as well as vertically. In the material under study, nevertheless, there is a distinct chronological order in direct relationship to the process of abstraction from group to group. Evidence for this is as definite as possible in such a sub-group as JI 1d, where the transition from a Justin I legend to a Justinian legend is achieved. The chronology of developing related groups may be clarified further by understanding that when enough material is available to permit the formation of a general gradual style progression from the parent group to the descendant one, a corresponding chronological order may be imposed. When transitions are not extant between two groups such as JAN 1 and JAN 2, and a third group such as JAN 8 depends and develops from both, it is perhaps more correct to see both early groups as being issued horizontally in the scale rather than vertically.

This qualification should be extended further in consideration of mint attribution. Diverse stylistic progressions do not necessarily imply different mints; they might be explained as different moneyers of the same mint. The interpretation of the material, however, reveals a definite indication that different progressions imply different mints. The difference in designs and legends are often too consistently distinguishable to be products of the same mint. This is particularly noted in the hesitancy for the progression of A 4–JI 2– JAN 3 to adopt the pectoral cross which has been a permanent motif on the A 1– JAN 1 progression as early as A 3.

5. Stylistic correspondencies exist outside the realm of Visigothic numismatics and are observed in earlier fifth century imperial issues, and in Ostrogothic, Frankish, Burgundian and contemporary Byzantine issues. The Anastasius issues, for example, proceed naturally from late fifth century examples attributed to Italian and Gallic mints. There is a marked connection between A 1 and A 3 portrait heads and those on tremisses of Valentinian III, Avitus, Majorien, Leo, Julius Nepos, Romulus Augustus, Zeno and Basiliscus attributed by Ulrich-Bansa to Milan, 1 while coins of the same period attributed to Ravenna and Rome are connected more with A 2 and A 4. 2

The importance of the Ostrogoths in western Europe is indicated by the large quantities of their money found in France and Germany. 3 Just as the gold issues of Italy and particularly Milan must have spread over Gaul and the rest of Europe in the fifth century, so did the Ostrogothic issues in the last decade of the fifth and the early sixth. Possible proof of our earlier established theory of Ostrogothic connections with the conception of the Anastasius VPW issue is the continuing stylistic association of such groups as A 2, A 4, JI 2 and JAN 3 to Ostrogothic issues. 4 This Italian connection is continued in Byzantine Imperial issues from Ravenna and Rome and is noted between JII 3 and silver coins of Justin II. 5 The square chest types developed also in Athalaric's Ravenna Justinian issues, 6 and the many coins of Theia with broad chests of gracefully flowing folds do recall some of those of JI 2 and JAN 3. 7 Ostrogothic connections seem much more influential in the coins of the progression of A 2– JI 4 and A 4–JI 2– JAN 3.

Burgundian issues, mostly with royal monograms, are of a different type from all of the Visigothic. This is manifested most immediately in the Victory figures which never lose the modesty of their gown nor the bottom hem which prevents folds from becoming extra legs. It is further indicated by the portrait bust which cannot be stylistically grouped with any of the coins of my groups. Burgundian coins bear the closest resemblance to coins of A 2–JI 4 and A4– JI 2.

Merovingian coins become crude or resort to symbols and royal mint monograms or geometric abstract patterns so immediately that it is difficult to see any connection with coins of my groups. Since letters in the field are not in the tradition of Visigothic coinage, it may be assumed that those with such letters are Merovingian. They also are found almost exclusively on coins whose style would place them with the A 1–A 3–JI 1 group. Since these groups follow closely the fifth century issues of Milan and those of Gaul, it would not be too rash to conclude that this style descends immediately or directly from the mint products of Visigothic Toulouse before 507. Consequently, this style would have been widespread throughout Visigothic Gaul and easily continued by mints now operated or newly organized by the Merovingians. It would also be in character and traditional for this region to copy coins issued by the new Visigothic capital of Narbonne, and it would be natural for Narbonne to continue the types of Toulouse.

The majority of coins bearing Victories in the Visigothic insect manner in the name of Anastasius, Justin I, Theodebert, and Childebert I all bear a T or TΛ, or ΤΔ or TS in the field. These have been questionably attributed by Belfort and Prou to Toulouse, Toul, Troyes and Treves. 8 Yet all other VPW's in the Prou and Belfort catalogues are of the non-insect type with squarish contours to their robes; standing rather than striding figures; with a pronounced bent arm holding a wreath; and often with the palm branch deleted. Coins of this type are noted in the names of Anastasius and Justin I, bearing the monograms of Gundibald, Sigismund and Gundomar. They also appear with the names of Theodebert and Childebert I. Belfort has given these to Lyons along with some of the Theodebert issues which bear an LV in the field. 9 Some of Justin I with image or image or PR in the field have been assigned by Belfort to Narbonne; however these are the same as those I have grouped with the Burgundian issues. 10 A number of those in the name of Theodebert bear a possible NR in the field or a monogram image or an N, which Belfort has given to Narbonne. 11 The Narbonne-Merovingian attribution is out of the question, however, since Narbonne is a Visigothic city at this time.

It would seem that in the Merovingian coinage of Theodebert we have at work the influence of both Burgundian and Visigothic traditions in the distinctive representation of the Victory figure. The Visigothic Victories are found only with the letter T, alone or in combination with another letter; while the Burgundian Victories are found with letters LV, NR, N or image This factor combined with the stylistic consistencies revealed in this study would indicate: (1) Frankish usage of letters in the field to identify mints; (2) Burgundian preference for royal monograms in the field which is replaced by the mint letters under the Franks; (3) further evidence to justify the removal of all coins with image monogram variations erroneously associated with Narbonne and with Amalaric from the lists of Visigothic coinage; (4) the Visigothic Victories in association with letter T may very well indicate Toulouse. These Visigothic Victories on Merovingian coins of Theodebert with letters in the field may substantiate the inclusion of the coins bearing mint letters in the field and associated with Visigothic groups as Frankish variations and copies.

6. Stylistic correspondencies between Visigothic coins without mint names and later ones with mint names significantly reveal aesthetic continuities. The uniformity and standardization particularly of the coinage of Leovigild and Reccared, however, make it difficult to distinguish any specific mint characteristics that might have carried over from the anonymous period. My groups do relate positively and generally to the later national coinage and consequently are unquestionable parents. In order to attribute anonymous issues by comparison with later coins with mint names, a thorough investigation must be made in order to find out whether Visigothic mints maintain or possess a specific, individual character. The results of such an investigation are insufficient in themselves but combined with other material may increase the weight of circumstantial evidence.

7. A study of French and Spanish collections and their inclusion in these groups affords illuminating observations (See Chart VII). The coins in French collections decrease in number as one goes from Anastasius to Leovigild; while coins in Spanish collections increase in number from Anastasius to Leovigild. This reflects the shifting of the center of the Visigothic authority from France to Spain, specifically after 507. The largest collection of Anastasius issues is in the Cabinet des Médailles in Paris, and this is true for each of the Anastasius groups as well as for the entire period. This is equally true for coins of Justin I. It continues to be true for only one Justinian group, JAN 1. Although the Cabinet has a similar number of coins in JAN 2, it is topped by New York, London and Madrid collections. An analysis of those coins which have been so far reconstituted to the Gourdon and Alesia Hoards indicates a similar behaviour pattern to the total Paris collection in the Anastasius and Justin I periods.

The study of the collection distribution of these coins further indicates that coins with letters in the reverse or obverse field are to be found in the largest number in the Cabinet des Médailles, while one has never been recorded in the Madrid or Barcelona collections. The Zorita Hoard is void of any coins with letters in the field. Visigothic coinage from Justinian forward maintains barren fields. This seems to be a further indication that all coins bearing letters other than royal monograms are Merovingian. This custom is particularly seen in the Anastasius period. Letters in the field are less used in the period of Justin I and this then designates a Merovingian coin of Visigothic type. Its rarity after Justin I on VPW tremisses indicates the termination of the striking of Visigothic imitations by Merovingian mints.

In analysing further the character of coins of groups in the various major collections, it is found that the largest number of variables is in the PBN and forms a large part of the few coins traceable to the hoards of Gourdon and Alesia. Spanish collections have a smaller number of coins represented in variable groups even though the Zorita Hoard contains some very crude versions. The study of the variables may aid in distinguishing Visigothic from non-Visigothic imitations.

8. The Zorita Hoard provides the one opportunity to study and analyze coins which may be associated with a definite location of usage. The coins date from Justinian to Leovigild and consist of eleven tremisses of Justinian, sixty-four tremisses of Justin II, five "curru" tremisses, three INCLITUS REX coins; and six non VPW tremisses (five Merovingian [?] and one Suevian [?]). Of the five VGC coins assigned by Cabré to the Merovingians, I would suggest that Zorita no. 5 (Plate D) is close to the Hispano-Byzantine products of the mint at Cartagena, described by Grierson, and that Zorita nos. 3 and 4 (Plate D) may be unofficial Visigothic imitations of these Hispano-Byzantine products. 12 There is a stylistic relationship between these coins and those of Groups JAN 3, JAN 4, JII 3 and JII 4 which will be assigned to Andalucía. These are not Merovingian.

One of the Zorita VPW items (no. 657) which has been classified by Cabré as questionably Merovingian is an anomalous coin whose style fluctuates between JAN 3 and JAN 4 and whose inscription in exergue image may have something to do with the unique exergue image to be found on two JII 4 Leovigild coins (Plate XXV, 486–487). A Merovingian attribution seems incorrect since it does not seem to attempt an exact copy of a Visigothic issue, and its style and type variation is too much within the aesthetic orbit of all the coins studied and classified within the Visigothic series.

The Zorita Hoard is almost solely responsible for the two groups JII 2 and JII 3 and accounts for half of the extant Justin II coins catalogued. An intriguing aspect of the hoard is the preponderance of crude examples of groups strongly represented in existing collections, such as JAN 2 and JAN 5, and the large number of variations within groups. The dominance of JII 2 and JII 3 types would encourage a local mint attribution that is somewhere in the Cartha-ginensis; however, Zorita is well enough situated on commercial roads to permit coinage to be represented from all over Spain. Zorita, northeast of Toledo, is up the Tagus from Toledo on the road into the Tarraconensis, Saragossa, Tarragona, Barcelona and then Narbonne. At Toledo, that road forks south to Córdoba and Seville and southwest to Mérida and Evora. Such a group as JII 6 consists of only Zorita items, while the non-Zorita coins of JII 7 are only vaguely and questionably related to the Zorita JII 7 coins. Both groups may be even more definitely local mint products. A study of the Zorita makes possible consequent observations concerning mints in the pre-Leovigild era. Such observations cannot be so definite as one would hope.

End Notes

Ulrich-Bansa, pls. X, XIV, XV, XXI; for Group A 3, see pl. XV, 191a.
Ibid., pl. O; BMCVOL, pl. V, 17–21, pl. VI, 4.
Le Gentilhomme, RN 1943, pp. 88–90.
BMCVOL, pl. VII, 4; pl. V, 17–21; pl. VI, 4; pl. X, 17–18.
Ibid., pl. XVII, 25–31.
Ibid., pl. VII, 23.
Ibid., pl. XII, 9–12.
Belfort, nos. 2292, 2295, 4327, 4355, 4356, 4357, 4359, 4479, 4483, 4360, 4361. Also see Prou, p. 500, no. 2433; p. 124, nos. 534, 535, 537. On the last coin which Prou lists as from Nantes (?), there is a blurred legend, a VPW, and the large letters of N S in the field.
Ibid., nos. 2289, 2291, 5475–5488, 5492–5493; also nos. 2259, 2260, 2272, 2274, 2277.
Ibid., II, p. 413, nos. 3137, 3138.
Ibid., nos. 2293, 5881, 5882, 5883.
Grierson, NH 1955, pp. 313–314.


The legends bear a direct relationship to stylistic progression so that in almost every case coins may be dated in accordance with the emperor's name they bear. Justin I legends are distinguished from those of Justin II by their greater accuracy of execution. This is further substantiated by the distinctive stylistic variations in the design of the types. A study of the legends in connection with a study of type styles places the garbled legends in the late Justinian and Justin II groups, so that it is possible to see consistency within imperial reigns, as well as consistencies between groups within style progressions such as A 1 to JAN 1.

Variations between groups of different style progressions, however, are not significant enough to be sole determining factors in the cataloguing of individual coins, although the manner of inscribing the legends and its orthography do provide subsidiary circumstantial material in reconstructing a group personality. In this regard the reverse legends are perhaps more telling, particularly in the last half of the period. Those of JAN 4 and JAN 5, for example, reveal the distinctive character of each group; while the stylistic associations of JAN 2, JAN 6 and JAN 7 are paraphrased in their legends.

When the garbled legends are analyzed, two tendencies manifest themselves. On one hand there is a natural derivation of these legends from correct ones by way of technical shortcomings in reworking old dies, or in transcribing, or in copying worn coins, or in the use of punches. On the other hand, the national aesthetic takes over. It is immediately seen how quickly the D of Dominus Noster becomes a Ɔ, once the first letter of the obverse legend becomes confused with the contour line of the imperial chest; or how, in small lettering, the diagonal sections of the letter N can become confused with an H or incorrectly formed as an И, or finally II. Nor is it difficult to understand how the reverse legend becomes increasingly abbreviated because of the growing inability to place the full legend VICTORIA AVGVSTORVM on the coin. The handling of the reverse legend is further complicated by its being confused with the common reverse legend for the Byzantine solidus, VICTORIA AVCCC. It is this association which produces reverses like those of JII 7 (nos. 539 and 540) and JII 3 a (no. 467). The greater variation would naturally occur in the reverse legend where the skill of the die cutter is challenged by a serious technical problem. The reverse type and legend are far more demanding of delicate craftsmanship in order to fit on the coin. The enlargement of the Victory figure in order to include more aspects of its character impinges on the area of the legend and thereby necessitates the execution of small, more delicate letters for the legend. There seems to be an increasing inability to create these.

The increased use of punches is directly proportional to the increase in garbled legends in Justinian and Justin II issues. This is indicated by coins in groups JAN 2, JAN 7 and JII 4, e.g., nos. 271, 273, 367, 377, 380, 478, 479, 482, 483. No matter what the means, once the legends are so garbled as to become unintelligible, they are inseparable from the mintmaster's concept of total design. Almost immediately the elements of symmetry and pattern take command in the organization of the letters of the legend. This development may be seen in such legends as those of JAN 8, JII 4 and JII 5. What is apparent is that garbling cannot be accounted for by the illiteracy of the die maker or of the manager of the various mints; otherwise the minter could not have handled so easily the name of Leovigild and the mint in the national coinage. The blurring of legends may be accounted for by either technical or consciously political reasons or both. In either case, it is not a difficult or impossible transition to nonsense legends symmetrically designed or organized. "Curru" legends are explainable as the culmination of a developed, conscious pattern of letters arranged in a meaningless but aesthetic pattern. The pattern is immediately replaced once Leovigild feels strong enough to assert his independence and dynastic right.

It is interesting to note that the coins of the A 4– JI 2– JAN 3 progression and subsidiaries such as JII 1– JII 2– JII 3 provide us with the most intelligible and accurate rendering of legends. These are groups also influenced more strongly by both Ostrogothic and Byzantine issues, and for the late JII groups it might suggest their being products of mints under Byzantine control. On the other hand, garbling proceeds rapidly and reaches its final aesthetic form in the production of the "curru" legend in the A 1– JI 1– JAN 2 progression.


Crosses and stars appear in the obverse and reverse legends of some of the coins studied, but it has been impossible to interpret any significant relationship or consistency in their use. The use is scattered throughout groups in all of the several progressions. The use of the cross is more common and is found predominantly in the middle of the obverse legends (A 1, A 1c, A 1e, A 3b, A 4b, JI 1, JI 1c, JI 2, JAN 3, JII 3, C 2, C 3, C 4, IR). Its use in the reverse legend is relatively rarely found (A 4a, JI 1b, JI 2, JI 5, C 2, C 3). Stars are found much less frequently and equally avoid significant interpretation (stars in obverse legends of JII 3; and one item in the reverse legend of JAN 8).

The presence of a cross or a star in the reverse or obverse field is found infrequently in coins that might enter the Visigothic category. There is only one coin which bears a cross in a field, and this is on the obverse. It is the unique No. 205 (Plate X–JI 2a) which is very questionably included in the Visigothic orbit and seems more likely a Merovingian product. The use of stars in the obverse field is found in conjunction with the monogram image of Justin I and Justinian and with the image of Justinian. All these are Burgundian. In a coin of the Gourdon Hoard, the star alone appears in the reverse field. This and one formerly in the Reinhart collection, (Plate IX, 192) are both of JI 2, while another from the Alesia Hoard is a JI 2 variation, possibly non-Visigothic (JI 2b). A star appears one more time on the reverse field of a JI 3 variation (JI 3 a) which is stylistically unique and impossible to place in any one group with certainty. Grierson has assigned a number of VGC coins to his Hispano-Byzantine mint at Cartagena, which have stars in the right reverse field. This may account for the sporadic appearance on the few Visigothic pieces. 13 A study of the rare appearance of stars or crosses in the obverse or reverse field has failed to reveal any pattern to permit interpretation.

The pectoral cross, one of the more distinctive features of the Visigothic VPW tremissis, has never been explained adequately. Its occurrence on a coin minted in Spain should not come as a surprise, since the cross is the most recurring decorative motif and religious symbol in Spanish art of the Visigothic era. As a political and propaganda measure, Reinhart suggests that the cross, a more orthodox than Arian symbol, is a concession of the Arian kings to their orthodox subjects. 14 The Arianism of the Visigoths convinced Reinhart, as has been noted before, that this was a Frankish innovation at Toulouse, adopted by the Visigoths. Florez, however, explained that it was the Arian veneration of the cross that was behind the use of it on the coins. 15 Arian or orthodox, the devotion to the cross is an important early development in Spanish liturgy, and its place on the coin is an expression of this. Even the Lex Visigothorum prescribed the Invention of the Cross (3 May) as one of the principal feasts of the year. 16

Not discounting the religious significance, it is worth considering whether another explanation may be supplied for its presence on the tremissis bust portrait. Is it simply a realistic descriptive observation? Was the cross part of the imperial or royal regalia? Was a pectoral cross worn by the Visigothic king? The Visigothic treasures of Guarrazar and Torredonjimeno make such a suggestion strongly probable, since they are composed of sceptres and royal seals as well as diadems and crosses. 17 The pendant cross on the offering crown of Reccesvinth seems particularly to have had a separate function of its own previously as a cross hanging on a chain around the neck. 18 Stronger evidence may be the testimony of Al-Maqqarí (amongst other Arab chroniclers) who said: "…for it was a custom among the Gothic sovereigns of Andalus that the diadem worn by each of them during his life should after his death be deposited in the principal church." 19 The first appearance of the pectoral cross on the Anastasius VPW tremissis (nos. 2 and 3) is as a broach pinned to the paludamentum. That such crosses were worn and that the custom was more common in the East with both laity and clergy enhances the possibility of a similar custom being popular in Spain which so often followed Eastern usages. 20 That a pectoral cross may be a part of the regalia may be indicated by the gifts placed in St. Peter's grave by Liutprand on a visit to Gregory II (715–731). They are said to have been a mantle, vestment, sword belt, crown and silver pectoral cross. 21 At his coronation in 962, Otto I wore a pectoral cross. 22

A pectoral cross, therefore, may already have become associated with the royal apparel in the sixth century, and its presence on the coins may simply indicate this. The bronze issues of Theodahad further bear this out.

End Notes

Ibid., pp. 308–309.
Reinhart, Germania 1941, p. 193.
Florez, III, p. 173.
Lex Visigothorum, lib. XII, tit. III, leg. VI. Spaniards have attributed many of their victories over the Moors to the power of the cross, e.g., Novas de Tolosa (1212) and Oran (1509). Alfonso III gave the cross of Oviedo and Santiago in 874, as a victory offering and gave another, the "Cruz de la Victoria" in 908, bearing the inscription: HOC SIGNO TUETUR PIUS. HOC SIGNO VINCITUR INIMICUS. See Schramm, II, pp. 482–483.
Menéndez-Pidal, III, p. 215.
Ibid., p. 627.
Maqqarí, I, pp. 282–283.
Seymour, p. 251. The earliest wearing of a cross on a chain around the neck may be that worn by Pope Hilarius in 461. Gregory the Great wore a cross, and Justin I (519) gave a pectoral cross to the Pope. These may have been talismans of the True Cross, for a Theka with relics was worn by Gregory the Great, Gregory of Tours, and later Charlemagne. See Schramm, p. 310.


Evidence derived from the weight of the coins studied has limited application. Some style groups are composed of too few coins, e.g., JI 5, JAN 6, JAN 7, JAN 9, JAN 10, JII 1, JII 6, C 4. This is further compounded where such a small group contains coins whose weights have not been recorded, such as in JI 5, in which the weight of only two of the four coins is known to me. Nevertheless, a study of the Weight Charts X and XI and their accompanying graphs, Xa, Xb, XI a, XI b, does provide material for provocative and applicable observations.

Similar developments may be seen in the study of weights as in that of design. A degree of regularity in the average weight in the largest weight group may be traced from Anastasius to Justin II. As style groups become more numerous down to Leovigild, so does the variety in the weight of the individual coins in a group. Style, weight and quality of gold will be standardized by Leovigild's currency reforms. However, until the period of Justin II, the weight standards were well maintained and varied much less than did the type designs. Considering only the weights of the coins in the major groups (Chart X and Graphs X a, X b), the standard fluctuates through 1.42–1.48 gms., with the exception of groups JI 5, JAN 7, JII 1, JII 2, JII 6, C 3, C 4 and IR. JI 5, JAN 7, JII 1, JII 6 and C 4 are very small groups, and the number of coins here is insufficient to make the evidence of their weight averages valid or interpret- able. Since JI 5 can be definitely assigned to the Burgundians, only three style groups remain which are large and have a majority of coins of poor weight: JII 2 (1.096), C 3 (1.288), and IR (1.309). The widest variance occurs consequently in the period of Justin II and the early years of Leovigild. It is possible that with more specimens of known weight in groups C 2, C 4 and C 5, the weight standard maintained there might also drop to the level of C 3 and IR. The good weight of C 1 (1.413), however, is based on a large number of specimens as second largest of the "curru" issues so that it is possible that these other issues may reflect the character of this group rather than the lower standards of C 3 and IR.

The low weight standard in some of these issues that must date from the late 560's to ca. 578, presents another facet when Cabré's analysis of the quality of the gold in the Zorita Hoard is related to the coins in question. Groups JII 2 and JII 3 are known almost exclusively through the existence of the Zorita Hoard, and their average weights vary decidedly: 1.096 or 1.130 to 1.4822 or 1.4823. 23 On the basis of weight, JII 3 is an excellent issue, and JII 2 has all the characteristics of irregular currency; however, the contrary is true. Twenty-two of the twenty-four JII 2 coins are of good gold, while only seven of the twenty-eight Zorita JII 3 items are of good gold. The only poor gold coins in Zorita JII 2 weigh 1.44 and 1.47 gms. 24 The good gold specimens of Zorita JII 3 weigh 1.04, 1.49, 1.50, 1.40, 1.51, 1.44, 1.45, 25 while the good gold items of Zorita JII 2 weigh anywhere from 0.86 to 1.48 gms., with only five of the coins weighing above 1.40 gms. It would seem from this that conflicting fiscal measures are adopted at different mints to resolve the financial crisis of this unsettled moment in the Visigothic Kingdom. More important is that all the coins in the Zorita Hoard that bear the name of Leovigild are of good gold whether they fall in JII 3, JII 4, JII 7 or IR groups. This is true for the "curra" issues as well with but one exception, C 4 item no. 595 (Zorita 77). Is the small number of coins upon which this observation is based too small to conclude that Leovigild's first currency reform does not only amount to the placing of his name on the coin but also in attesting to the restored good quality of its gold and in some degree to the uniformity of its weight?

The same development is found if the weight averages of the combined major and subordinate groups are studied, except that the standard fluctuates between 1.40 to 1.51 gms. In Chart XI as in Chart X, many of the groups consist of coins whose weights range from 1.00 to 1.50 gms., e.g., JAN 5, JII 2, JII 3 and JII 5. The individual variability of coins within a group increases as they proceed from the Anastasius to the Justin II issues. It is not until the issues of Justin II, nevertheless, that there is a uniform distribution such as in JII 2, JII 4 or JII 5. While groups such as JAN 5 have coins at the bottom of the weight chart, the bulk of the specimens are found in one range, 1.41–1.45 gms. This is to some extent true in JII 3 (1.45–1.50 gms.). Such patterns are determined by the number of coins which have been assigned to a group, and a larger number of coins assigned to any of the Justin II groups might also show a concentration in one weight bracket as is seen in the earlier groups. Nevertheless, a general working assumption may be suggested for predicting the weights of specimens in any of the large groups; that the weights of such coins would most probably fall into the same area as that of the majority group. A case in point is the collection of the Gabinete Numismático de Cataluña which is not included in the statistical compilations for the weight charts, since the charts were tabulated before this collection was studied. Note the concordance on page 151.

Of the twenty items, ten do not fall into the weight range of the largest number of the coins in the group; however, a closer inspection of these ten reveals that more of them than appear to do fit the majority pattern. Those coins that do match exactly are specimens of JI 3, JAN 5, JII 2, JII 4a, JII 7, C 2 and C 3. Of these JI 3, JII 7 and C 2 are small groups, and a full pattern is not evident. JAN 5, JII 2 and JII 4 a are large groups, but the coins of the group are evenly distributed throughout the weight chart. Perhaps then it may be concluded that coins which are assigned to these latter large groups tend to even out the distribution in the weight range. To

No. 2 A 3 1.448 1.41–1.45 1.44
8 JI 1 1.42 1.41–1.45 1.43
4 JI 3 1.37 1.45–1.50 1.48
5 JAN 1 1.45 1.41–1.45 1.42
10 JAN 2a 1.42 1.41–1.45 1.429
6 JAN 5 1.389 1.41–1.45 1.43
7 JAN 5 1.436 1.41–1.45 1.43
13 JAN 8 1.437 1.41–1.45 1.43
9 JII 2 1.16 1.06–1.10 1.096
18 JII 2 1.374 1.06–1.10 1.096
19 JII 4a 1.203 1.41–1.45 1.43
11 JII 5 1.463 1.45–1.50 1.485
12 JII 5 1.496 1.45–1.50 1.485
21 JII 7 1.376 141–1.45 1.485
17 C 2 1.222 1.41–1.45 1.45
14 C 3 1.334 1.26–1.30 1.2888
15 C 3 1.416 1.26–1.30 1.2888
16 C 3 1.028 1.26–1.30 1.2888
20 IR 1.32 1.31–1.35 1.309
22 IR 1.33 1.31–1.35 1.309
project further, coins of lower weight might have once existed or are still to be assigned to such groups of a wide range of weight distribution such as JAN 5 and JII 2. Another qualifying aspect is that the weights of the JAN 5, JII 2 (Barcelona No. 9), and the JII 7 items are very close to the weights of the majority group. This leaves only one of the JII 2 and some of the "curru" items as being out of order with the average weight of the majority coins in each of the style groups.

If the hypotheses developed from an interpretation of the weight charts are valid, is there any correspondence between the weight patterns of the Leovigild coins struck by specific mints and earlier mintless issues of Leovigild and his predecessors? In both of Leovi- gild's series with mint identification, the lowest weights and the widest range of weight distribution are to be found at Córdoba (1.03), Seville (1.08), Evora (1.09), and Mérida (1.06). 26 In the pre-Leovigild issues only in JAN 5, JII 2 and JII 3 are such poor weights found combined with a pattern of wide range distribution. Could we then work on the assumption that an Andalucían provenance for these style groups is a probability? Note also that this situation is true again in the time of Reccared for the mint of Mérida. The mint of Narbonne is characterized from Leovigild through Witteric by a narrow range of weight variability as well as a tendency to maintain traditional weight standards. 27 Can we suggest such early groups as A 1, JI 1, JI 2 and JAN 2 as being products of Narbonne? Might this be said also for such groups as C 1 and C 2? Naturally before such a conclusion may be accepted other evidence is necessary.

No substantial information can be gleaned by combining the information of the weight charts with the stylistic progressions as seen in Charts X c through h. Does the uniformity of the weight standard in Charts X d, X h, X i indicate that the progressions in each chart signify the products of a single mint; while the variability in X c and X g indicates that these progressions are formed from the product of different mints? This evidence is not sufficient to accept these conclusions.

What is apparent in a study of the weight charts is that from the time of Anastasius to that of Leovigild the Spanish mints strike a tremissis of a theoretical weight standard of 1.516 gms. This standard is altered only in some of the issues of Justin II and Leovigild, JII 1, JII 2, C 3, C 4 and IR in which the tremissis is struck at a theoretical weight standard of 1.326 gms. as in France. 28 This light tremissis of twenty-one carat gold is not retained but is replaced by Leovigild's heavy tremissis (facing busts) of eighteen carats.

End Notes

Duchesne, Lib. pontif., I, p. 408.
Schramm, p. 311.
See Chart X.
The Zorita JII 2 poor gold items are nos. 428, 431.
The Zorita JII 3 poor gold items are nos. 441–447, 454, 459, 460, 461–463, 466–471, 472 and 476.


When all the evidence, numismatic and otherwise, is analyzed, it is impossible not to conclude that the Visigoths are primarily involved in the striking of the VPW tremisses, not only because of the persistent issuance of the coin into the Leovigild period, but also because of the aesthetic formularization of the types. The latter indicates the maturation of a "home grown" product, rather than a cultural borrowing or imitation. A study of the political and cultural aspects of the period encourages the acceptance of the qualifying hypothesis that the initiator and prime mover is Theodoric, regent of the Visigothic Kingdom, and his advisor Cassiodorus. The Merovingian and Burgundian minting of this VPW coin would otherwise remain inexplicable if the Visigoths were solely responsible for reintroducing this coin. The acceptance of this hypothesis must also force acceptance of a date of ca. 510–511 as the possible initial date of the Anastasius VPW issue. This in turn determines attributions, since it makes the assignation of any of the VPW issues to Alaric II and his Kingdom of Toulouse impossible. This is the possible flaw in the hypothesis, however, for if the VPW Anastasius issues are removed from Alaric II's coinage, what is left to be assigned to him? Considering that Alaric II would continue the policy of his predecessor, he would have struck gold solidi as well as a tremissis. This tremissis could be a VGC or a cross-in-wreath, although the former seems a more correct type for the time. There are no Anastasius cross-inwreath coins known to me. The VGC tremisses should then be studied for possible Visigothic (Alaric II) issues. The numerous accusations that Alaric's gold currency is not trustworthy might encourage the acceptance of the thesis that because this poor gold was melted down and recoined in a pure state, most of Alaric's gold, if not all of it, is not extant. 29

The attribution of Anastasius issues, consequently, should include Visigothic issues of Narbonne, possible Burgundian and Merovingian issues, and Ostrogothic issues which provided the prototype. Our evidence further suggests that most, if not all, the Anastasius groups come from southern France, their major provenance area. The classification and attribution of the Anastasius series will differ from Reinhart's, who traced the VPW back into the reign of Alaric II and the mint of his capital, Toulouse. If the VPW was struck by Alaric II as Reinhart and everyone else has determined, the Anasta- sius Visigothic issues would have to be divided between Toulouse and Narbonne before and after 507.

Before the various groups can be assigned to mint cities two further observations, suggested first by Reinhart, must be considered. First, that on the basis of earlier fourth and fifth century mint practices, it may be assumed that coins would be issued at least at cities of royal residence, which would provide the location of the major mint for that period, if not the only one. Second, that it would seem plausible that mints which issued Leovigild coins with their mint name on them also issued the earlier Leovigild coins without mint names. Reinhart organized his coins partially on this basis.

In regard to this "capital theory" the possibilities would be as follows:

Anastasius (491–518) Toulouse (to 507) (Alaric II) Narbonne (511–518) (Amalaric)
Justin I (518–527) Narbonne (Amalaric)
Justinian (527–565) Narbonne (527–531) (Amalaric)
Barcelona (531–540's) (Theudis, who dies in Seville 548)
Seville (540's–549) (Theudis and Theudigils, who dies here in 549)
Mérida (ca. 550–554) (Agila, who dies here in 554)
Toledo (ca. 554–565) (Athanagild)
Justin II (565–578) Toledo (565–567) (Athanagild)
Narbonne (567–572/3) (Liuva)
Toledo (573–) (Leovigild)

This list must then be enlarged by an analysis of the Leovigild mints. 30 These number fifteen in all, although only four issue a large number of coins if we accept the thesis that the extant coins are directly proportionate to the original mint product: Toledo (44), Mérida (22), Evora (21), and Seville (11). These mints are also responsible for the largest number of coins issued in the reign of Reccared: Toledo (69), Mérida (102), Evora (35), and Seville (42). 31 Other important mints for both reigns are Saragossa (7,23) and Córdoba (6,20). Both Barcelona and Tarragona are poorly represented for Leovigild (2 and 1 respectively) but have larger extant Reccared pieces (26,28). In the last period of the anonymous coinage, therefore, Toledo, Mérida, Evora and Seville most likely issued coins. These might be the major mints while Barcelona, Tarragona, Saragossa and Córdoba would be of secondary importance. Cordoba might be more important than extant Leovigild coins suggest since the Hermenegild revolt interfered with mint production in the 580's and with the adoption of Leovigild mint name coins. The exultant boasting in Leovigild's commemorative CORDOBA BIS OPTINUIT series of 584, might indicate the importance of this center which had been under Byzantine rule from 567 to 572.

The city of Narbonne for the Leovigild and Reccared period is much less important, sharing seventh place with Rodas (5,9). Since it was the capital of Liuva from 567–573, a larger number of anonymous coins might be associated with it. Narbonne should also share in the anonymous issues. The final list of possible mint locations could be Toledo, Mérida, Evora, Seville, Narbonne, Córdoba, Barcelona, Tarragona and Saragossa.

The capital hypothesis presents another problem. Did the mint move with the capital and thereby establish the mint workers as part of the royal entourage, or once a mint was established did it continue to exist even after the city lost its royal status? Since the numismatic evidence suggests different forms of progression, the latter assumption would seem preferable. There are too many styles existing contemporaneously to allow the attribution to one mint. Reinhart's suggestion of provenance determining the mint location, with all the attendant difficulties of such a method, would also dispel the idea of a single monetary source for Visigothic coinage at any given period.

What is revealed when the capital theory is related to the stylistic groupings and to the scant provenance information is a pattern of large, extant stylistic groups with both contemporaneous dependent and completely independent groups. The large groups might logically be assumed to be those of the main mint which at that time would be a capital city, and minor groups might be located in former capitals or other important cities. The measures taken by Theodoric as noted in Cassiodorus' letters to control and prohibit non-official minters would seem to indicate the possibility of many local mints. 32 This suggestion would be in accordance with traditional mint practices of the late empire in the fourth and fifth centuries. 33

Reinhart attributed the Visigothic VPW coinage and dated it as follows: 34

Anastasius Alaric II Toulouse to 507
Amalaric Narbonne to 518
Justin I Amalaric Narbonne to 527
Justinian A Amalaric Narbonne to 531
A Theudis Barcelona to 548
B Theudigils Seville to 549
C Agila Mérida to 554
D Athanagild Toledo to 565
Justin II D Athanagild Toledo to 567
D Leovigild Toledo

He expanded this capital theory to permit the issuance of coins on a continuum basis at former capitals, that is, the Narbonne- Barcelona Justinian coins are issued through Agila; the Justinian Seville group by Agila, Athanagild, Liuva, and Leovigild; the Méridan issues by Agila, through Leovigild; and the Toledo group by Agila, Athanagild and Leovigild. He further expanded his Seville group to include a suggestion of possible Córdoban attribution for these and a continuation under Byzantine rule. He classified his coins, therefore, according to the capital pattern with one exception; the possible coins of Liuva from Narbonne. Reinhart attributed no Justin II group to Narbonne and Barcelona. For him the mints moved south, along with the spreading Visigothic population. This is the major weakness with Reinhart's groupings. Reinhart further substantiated his attributions by assuming that local collections were built predominantly upon locally found specimens. Coins of Group A, therefore, which are found mostly in southeastern France and northeastern Spain are of Narbonne and Barcelona; coins of Group B found mostly in Spain and in Andalucía are of Seville and Córdoba; those of Group C from western Spain and Portugal are from Mérida; and those of Group D found primarily in Spain and so closely associated with the only VPW coin bearing a mint mark of Toledo and with the money of Leovigild, non-extant but described by Florez, must be Toledo.

My conclusions will indicate that I do not deviate greatly from the attributions brilliantly conceived by Reinhart. It is regretted that he did not publish his statistical findings or a complete catalog of coins studied. Such an endeavor was naturally beyond the scope of an article. It must be realized, however, that Reinhart's attributions were presented only as considered suggestions, and the changes in the make-up of his groups from his early and late articles reflect his own questioning and reorganizing attitudes. The chart on page 158 compares the composition of Reinhart groups with my stylistic groups.

The major difficulties in comparing both groups are due to the more general amorphic quality of Reinhart's. The groups are not distinct or specific enough. His classifications overlap many of mine. The less specific nature of his groups is responsible for the major disagreement between Reinhart's Merovingian attributions and those presented here. He would assign all of Group A 3, most of JI 1, and JAN 3 to

Kgd. of Toulouse until 507 A 1, A 2, A 2a, A 2c, A 1c, A 1j, A4
Narbonne continuing Toulouse Types 511–527 JI 2, JI 3
Group A (Narbonne-Barcelona) JI 1. JAN 1, JAN 2, JAN 7, JAN 8
Group B (Seville-Córdoba) JAN 3, JAN 5, JII 5, JII 1, JII 2, JII 3
Group C (Mérida) JII 4, JAN 4
Group D (Toledo) JAN 4, C 1, C 2, C 3, IR
the Merovingians. All of these I classify as Visigothic. 35 Coins of JI 1 and JAN 3 groups are also placed by him in Visigothic Groups A and B respectively. 36 If the coins illustrating his Merovingian article are compared with my lists, only sixteen coins of the forty-two depicted agree with my non-Visigothic attribution, but of these I consider half as possibly Burgundian rather than Merovingian. 37 In that same article, nevertheless, I would agree with his Burgundian-Merovingian attribution of the seven coins illustrated in Abb. 1, and his southern Gaul Merovingian copies of Visigothic issues in Abb. 3. 38

In order to attribute the style groups to particular mints I have been governed particularly by the stylistic progression analysis as seen in Chart VIII, which has encouraged the separation of the groups into Visigothic, Merovingian, Burgundian and Ostrogothic. These groups have been separated further as to major and minor by both quantitative and qualitative considerations. A major group would be defined as one which consists of large extant quantities and is important for stylistic influence. To the arrangement of these the hypothesis of the capitals as mints and the mints of Leovigild's national coinage have been applied. Substantiation for some of the attributions has been found by relating anonymous pre-Leovigild and Leovigild groups to Visigothic national coins of known mints; and to Merovingian, Burgundian, Ostrogothic, and Italian Byzantine coins.

A study of stylistic Chart VIII has resulted in some general attribution determinants. The progression beginning with A 1 and ending in the INCLITUS REX group is the major purely Visigothic channel. It traces uninterruptedly the development of the six-legged insect Victory type and the Leovigild portrait bust. The second major progression A 4 becomes diffused after JAN 3 but strongly influences groups such as JAN 4 and JAN 5 which are also influenced by JAN 2. Because of the closer ties of the A 1 progression with national coinage of Toledo and the mints of the Tarraconensis, this progression is attributed wholly to the Narbonne-Barcelona-Toledo area. The intricate interrelationships of JAN 4, JAN 5 and JAN 3 and all of their descendent Justin II groups, and their relationships to the coinage of the national mints of Seville, Mérida, Evora, Córdoba and Italica encourage an attribution for this progression (at least from JAN 3 on) to southern Spain. These Andalucían groups have been assigned also because of their strong connections with Ostrogoth, Italian Byzantine and Hispano-Byzantine coins.

End Notes