Brescia Medallion

A Roman gold-glass portrait roundel of about 200-300 CE, showing three unidentified figures, fixed since the early medieval period into the face of the Cross of Desiderius, a gem-studded gold processional cross in the Museo di Santa Giulia, Brescia. Eighteenth-century antiquarians proposed it showed the empress Galla Placidia with her children; that identification stood as the standard museum label for well over a century before Joseph Breck rejected it in print in 1927.

A second theory then took its place: that the medallion, one of only two inscribed gold-glass portraits known, was made in Alexandria, argued from a supposed Alexandrian dialect feature in its Greek inscription and repeated as recently as 2007. Rachel Catherine Patt's 2024 article in the American Journal of Archaeology dismantles that argument in turn, re-reading the same letters as a common later-Roman-Imperial nickname form rather than dialect evidence.

What remains, after two rejected identifications forty years short of a century apart, is an anonymous Roman family portrait of real technical accomplishment, its sitters, its workshop and the exact date of its arrival on a Lombard king's cross all still genuinely unresolved.

A gold face fixed into a much later cross

The object this record describes is a roundel about six centimetres across, made by fusing a design in gold leaf between two layers of glass and finishing the detail in painted pigment. It shows three people from the chest up: an older woman at the centre, flanked by a younger woman and a young man. All three look outward with the fixed, wide-eyed frontality typical of the finest gold-glass portraits of the third century CE, a manner that anticipates the icon painting of the centuries that follow it far more than it recalls the sculpture that came before it.

The medallion does not stand alone in a case. It is fixed into the front face of the Cross of Desiderius, a gem-studded gold processional cross kept in the Museo di Santa Giulia in Brescia, and has been fixed there since at least the early medieval period. Its reverse and edge are concealed by the cross's own gold mounting, which means the object has effectively been examined, for as long as it has been documented, only from one side. This record treats it as its own entry because it is catalogued and discussed in the scholarly literature as an independent work with its own history, identity crisis and attribution dispute, distinct from the cross that now carries it.

How a gold-glass portrait was actually made

Gold glass, sometimes called gold sandwich glass, is a technique in which a design in gold leaf is glued to a glass disc, then worked with a fine point that scratches away the gold exactly as an engraver works a printing plate, leaving lines and details as bare glass against the gold ground. A second glass disc is then fused over the top, sealing the gold permanently between two layers of glass where it cannot tarnish, flake or be touched. On the finest examples, including this one, a painter has added colour on top of the gold outline to render faces, hair and drapery, which is why these portraits read less like glinting metalwork and more like small paintings under glass.

The technique is closely associated with the Roman glass industry of the third and fourth centuries CE. The great majority of surviving gold-glass work, several hundred pieces, comes from roundels cut from the bases of cups and bowls and cemented into the mortar of the Roman catacombs as markers for burial niches; most of these show simple devotional or domestic imagery rather than individualised portraits. The Brescia medallion belongs to a much smaller and finer group: portrait medallions of named or once-named individuals, of which only two carry an inscription that has been read as a personal name or nickname at all. This is one of them; the other is the Gennadios medallion in the Metropolitan Museum of Art.

The Cross of Desiderius

The Cross of Desiderius is a wood-cored, gold-plated processional cross of the crux gemmata type: a cross whose entire surface is set with gemstones, cameos and glass paste elements rather than left as plain metal. The Fondazione Brescia Musei, which holds and displays it, describes it as a rare example of early-Carolingian goldsmith's work of the ninth century, set with 212 gems, cameos and glass elements whose own dates of manufacture range from the Roman period to the sixteenth century. Wikipedia's account gives a close figure, 211 gemstones including around 50 ancient gems. Either count describes the same fact: this is a composite object, assembled once and added to more than once, in which a great many individually older and more valuable pieces have been set into a single later frame.

Among those older pieces, alongside a cameo of the emperor Augustus and another of Christ Pantocrator, sits the third-century gold-glass medallion this record is about. The cross is traditionally associated with Desiderius, the last king of the Lombards, and is said to have been given by him to the monastery he and his queen, Ansa, had founded in Brescia. The goldsmith's work of the cross itself, however, is dated by the museum's own scholarship to the ninth century, after the end of Desiderius's reign in 774 and, on most chronologies, after his death in Frankish captivity. That gap between a Lombard royal association and a Carolingian-period object is not resolved in the sources consulted here and is recorded as an open question below, rather than smoothed into a single tidy date.

Why a pagan-era portrait sits on a Christian reliquary

Setting a five-hundred-year-old pagan or at least religiously neutral Roman portrait into a Christian processional cross is not an anomaly to be explained away; it is the ordinary practice by which early medieval courts and monasteries built their most prestigious objects. A crux gemmata was valued precisely for the antiquity, rarity and workmanship of the individual gems it carried, not only for the gold that held them. Ancient engraved gems, cameos and, more rarely, painted glass portraits like this one were harvested from earlier Roman material, much of it already centuries old by the time it was reused, and set into new frames as visible proof of the wealth, connections and claim to Roman continuity of whoever commissioned the piece.

The Cross of Desiderius carries that practice at an unusual scale: about fifty ancient gems out of roughly two hundred and twelve pieces total, spanning from the Roman Imperial period through the sixteenth century, all mounted onto one ninth-century goldsmith's frame. The gold-glass medallion is simply the most legible example, because unlike a plain engraved gem it carries recognisable human faces and an inscription that later viewers could try, and repeatedly fail, to read correctly. Its presence on the cross says nothing at all about the identity of the people it shows; it says only that a Roman portrait of real value and beauty had survived long enough, and travelled far enough, to be judged worth including in the treasury of a Lombard royal foundation.

How the medallion became 'Galla Placidia'

The identification of the three figures as the fifth-century western Roman empress Galla Placidia with her children Valentinian III and Justa Grata Honoria is documented as an eighteenth-century antiquarian proposal. The earliest specific documentary trace located for this record is the theologian and antiquarian scholar Francesco Antonio Zaccaria, who in a work published in 1762 recalled having seen the medallion around 1725; his account is one of the threads later scholarship, including Breck's 1927 study, uses to trace the object's documented history backward. This record has not located a single named individual credited as the first to specifically propose the Galla Placidia identification, as opposed to simply describing or recording the object's presence on the cross; that gap is recorded honestly in the open questions below rather than assigned to a scholar on the strength of the pattern alone.

What is well documented is the outcome. Once attached to a real, nameable and important empress, the identification supplied exactly what an eighteenth- and nineteenth-century museum label wanted: a specific historical anchor for an otherwise anonymous ancient face, and a flattering connection between the Lombard treasury and the imperial Roman past it was already trying to claim through its cameo of Augustus. The identification held as the standard museum attribution for well over a century before it was formally and successfully challenged in print.

An interruption: was it even ancient?

Before the Galla Placidia identification was corrected, it was nearly overtaken by a different and more radical doubt: that the medallion was not an authentic ancient object at all. In the nineteenth century, the scholars Raffaele Garrucci and Hermann Vopel, both serious students of early Christian gilded glass, are recorded as having suspected the piece was a fake. Garrucci nonetheless published a careful drawing of it in his 1858 study of gilded glass from the Roman catacombs, a plate now in the public domain and reproduced in this record's gallery, which means the doubt was recorded in the same scholarly literature that eventually helped settle it.

The authenticity question and the identity question are logically separate but were resolved by the same twentieth-century turn toward technical and comparative connoisseurship: once the medallion could be set alongside a securely ancient corpus of gold-glass portraiture, both its status as a genuine third-century object and its status as a Roman face wrongly given an imperial fifth-century name became clear at more or less the same time. A reader encountering only the confident modern museum label would have no reason to know that this piece once had to survive a forgery accusation before it could even be argued over.

Breck's 1927 correction

The formal rejection of the imperial identification is credited in the modern literature to Joseph Breck, curator at the Metropolitan Museum of Art, in his 1927 Art Bulletin study 'The Ficoroni Medallion and Some Other Gilded Glasses in the Metropolitan Museum of Art.' Breck's article is principally about the Ficoroni medallion and related pieces in the Metropolitan's own collection, and it places the Brescia medallion into a comparative discussion of gilded glass portraiture rather than treating it as its sole subject; it is nonetheless the source repeatedly cited by later scholarship, including Patt (2024), as the point at which the Galla Placidia reading was set aside on scholarly grounds rather than merely doubted in passing.

What Breck's correction did not supply, and could not have supplied from stylistic and comparative argument alone, was a replacement identity for the three sitters. Removing an emperor's family from the object left three anonymous Roman faces where a named imperial family had stood in every guidebook for a century and a half. That vacancy is exactly what later twentieth- and twenty-first-century scholarship, including the debate this record covers next, has tried and failed to fill with certainty.

The next theory: an Alexandrian family

With the imperial identification gone, attention moved from who the sitters were to where the medallion, and its close cousin the Gennadios medallion in New York, were made. Scholarship going back nearly a century, and restated by Jas Elsner in his 2007 study Roman Eyes: Visuality and Subjectivity in Art and Text, read the endings of certain words in the Greek inscriptions on both medallions, terminating in iota, as a distinctive feature of Alexandrian Greek. On that philological basis, both medallions were proposed to be Alexandrian products, part of a broader modern habit of assigning rare, luxurious, unprovenanced ancient objects to famous ancient cities on relatively thin linguistic grounds.

This Alexandrian theory was not a fringe position. It became, in effect, the second consensus about this pair of medallions: rejected as imperial portraits, accepted as evidence of an otherwise poorly documented Alexandrian school of luxury portrait glasswork. It held that position for the better part of a century, cited in survey literature on Roman Egypt and on gold glass alike, until it was itself directly challenged.

Patt (2024) dismantles the Alexandrian dialect argument

Rachel Catherine Patt's 2024 peer-reviewed article in the American Journal of Archaeology, '"Probably Alexandria": Gold-Glass Portraiture and the Allure of Egypt,' treats the Brescia medallion and the Gennadios medallion together as the only two inscribed gold-glass portraits in the surviving corpus, and reopens exactly the linguistic argument that had supported their Alexandrian attribution. Her stated conclusion, in the article's own abstract, is that 'the grounds on which the connection rests need to be reconsidered.'

Patt's specific counter-argument is philological, not merely sceptical. The word endings in iota that earlier scholarship, repeated as recently as Elsner (2007), read as a marker of a distinctively Alexandrian dialect of Greek are shown, in her reanalysis, not to correspond to any dialect that is otherwise attested. In place of a dialect explanation, she proposes an onomastic one: these are not dialect forms at all but a specific kind of nickname, a by-name carried alongside a person's formal name, of a type that is well documented across the wider later Roman Empire, and which regularly took an ending in -i in Latin and the corresponding -i (iota) ending in Greek. If that reading holds, the entire evidentiary basis for locating the workshop in Alexandria specifically, as opposed to Rome or anywhere else Greek-speaking naming practices of this kind were current, disappears. Patt's own framing of the wider pattern is pointed: the impulse to assign rare, unprovenanced luxury objects to famous ancient metropolises such as Alexandria without independent evidence is, in her reading, closer to a habit of scholarly imagination than a documented archaeological finding.

This record follows the specification's own instruction not to flatten a nuanced philological argument into a vague statement that a scholar 'disagrees.' The disagreement here is precise: it is about whether a specific letter ending is dialect evidence or onomastic evidence, and Patt's case rests on showing that the dialect said to exist has no independent attestation, while the nickname pattern she proposes in its place does.

If not an empress, and not necessarily Alexandrian, then who

With both the Galla Placidia identification and the confident Alexandrian workshop attribution now under serious challenge, the honest state of the evidence is that the three figures on the Brescia medallion are an unidentified family or household group, portrayed with the same frontal, wide-eyed formality used across gold-glass portraiture of the period, whose names, if the inscription does record names or nicknames, cannot currently be matched to any known historical individual. Some cataloguing of comparable unidentified Roman gilded-glass portraits, including material connected to the wider Brescia civic collection, has historically used the plain Latin tag effigies ignotae, unknown likenesses, which is a more honest description of this object's status than any of the specific identities that have been proposed and withdrawn for it.

No scholarship consulted for this record proposes a specific alternative named identification for the three sitters. The object is best described, at the current state of publication, as a fine and unusually well-preserved third-century Roman family portrait in gold glass, made by an unlocated workshop, for patrons whose names are not recoverable from the surviving evidence.

Ansa, Desiderius and the monastery that outlasted the Lombard kingdom

The monastery traditionally connected to this cross, San Salvatore, was founded in Brescia in 753 by Desiderius, then a Lombard duke and shortly afterward the last king of the Lombards, together with his wife Ansa, using land the couple had inherited from the late King Aistulf. Their eldest daughter, Anselperga, became the monastery's first abbess. In 761 the monastery received the relics of Saint Julia, after which the complex eventually took the joint name San Salvatore-Santa Giulia by which it and the modern museum are known. An inscription recording Ansa as queen and wife of King Desiderius reportedly stood on a tomb structure in the church's southern perimeter wall until as late as the seventeenth century, which is the most direct surviving documentary trace of her personal connection to the foundation.

Desiderius became king of the Lombards in 756, three years after the monastery's founding, and was deposed by Charlemagne in 774, ending the independent Lombard kingdom. The monastery he and Ansa founded did not end with their political fall. It continued as a religious institution through the medieval and early modern periods, its own treasury, including the cross now bearing his name, surviving inside the community that guarded it.

From monastic treasury to civic museum

The Cross of Desiderius is recorded as having been kept within the monastic complex, in or near the Romanesque oratory of Santa Maria in Solario, across the medieval and early modern life of the institution. By the mid-nineteenth century the medallion and the cross were known to civic scholarship in Brescia; the medallion's earliest documented sighting, via Zaccaria's 1762 recollection of seeing it around 1725, and Garrucci's 1858 published drawing both predate any transfer to a purpose-built public museum, meaning the objects had already passed into a form of civic or ecclesiastical custody, described in period sources as the Museo Civico Cristiano, well before the twentieth century.

The monastic complex of San Salvatore-Santa Giulia itself became Brescia's city museum, the Museo di Santa Giulia, in 1998, under the Fondazione Brescia Musei. The cross has remained on display in the same oratory, Santa Maria in Solario, which is decorated with frescoes painted by Floriano Ferramola between 1513 and 1524, meaning the setting a visitor sees the medallion in today is itself a sixteenth-century addition to a much older room. The San Salvatore-Santa Giulia monastic complex, together with the Capitolium archaeological area in Brescia, forms part of a UNESCO World Heritage Site. The custodial chain, in short, runs from an early medieval monastic treasury, through nineteenth-century civic antiquarian custody, to a purpose-built modern museum, without the object ever having left the city in which it has been documented since at least the early eighteenth century.

The wider corpus: gold-glass portraits as a genre

The Brescia medallion belongs to a recognised and relatively small corpus of Roman gold-glass portrait medallions distinct from the far larger body of simple devotional roundels salvaged from the catacombs. Comparable pieces include a husband-and-wife pair in the Museo Sacro of the Vatican Museums, dated to the first half of the third century and about 4.9 centimetres in diameter; a portrait of two young men in Bologna, of the first quarter of the fourth century, about 5.1 centimetres across and inscribed with a drinking toast rather than a name; a family portrait in the Vatican Library whose date and dimensions are not established in the sources consulted; and a further husband-and-wife pair in the Vatican Museums dated broadly to the third or fourth century. Closest of all is the Gennadios medallion in the Metropolitan Museum of Art, accession 26.258, dated about 200 to 250 CE and about 4.1 centimetres in diameter, inscribed in Greek with a name, Gennadios, and a claim to musical skill; it is the subject of a separate Iconotheca record and is the only other inscribed portrait in this genre, which is precisely why Patt's 2024 article treats the two together as a matched pair of evidence rather than as two unrelated curiosities.

Set against that small comparative group, the Brescia medallion is unusual chiefly for its survival route: rather than being recovered archaeologically from a Roman catacomb, as most of its cousins were, it survived by a wholly different mechanism, incorporation as a valued antique into a much later piece of Christian goldsmith's work, which is also why its own excavation context and original findspot are simply not recoverable.

What remains unsettled

Two successive scholarly consensuses about this object, first that it showed a Roman empress, second that it was made in Alexandria, have each been formally overturned in the peer-reviewed literature, in 1927 and 2024 respectively. No third consensus has yet taken their place. The identity of the three sitters, the location of the workshop, the date and circumstances of the medallion's incorporation into the Cross of Desiderius, and even the precise translation of its two-word Greek inscription all remain open in the sources consulted for this record, and are recorded as such below rather than resolved by assertion.