1The re-attribution of the late Roman frieze on Rome’s Arch of Constantine to earlier periods – much like the reassignment of the arch itself to earlier emperors – has a long history. As this matter has been raised once again in a recent article, a review of the relevant evidence, the history of scholarship, and the outstanding questions concerning the arch and its frieze is warranted. The present essay offers this review while adducing new evidence in confirmation of the traditional date.

2In modern times, there have been three basic interpretations of the arch and its decoration:

3 i) The arch (Fig. 1) was built, probably hurriedly, almost entirely of spoliated material, for Constantine following his defeat of Maxentius in 312[1]. Reliefs together with architectural fragments were adapted freely for reuse to build and decorate the new arch in celebration of Constantine’s victory; to these were added others, newly carved, prominent among which are those of the frieze.

4 ii) The arch was built at an earlier point on its present site, and then refashioned for Constantine, employing spolia for its decoration – including the long historical frieze that wraps around the entire structure[2].

5 iii) A lost earlier arch, unreported in the sources, and built elsewhere (though presumably in Rome), was disassembled and then rebuilt as a new arch for Constantine[3].

6In all three cases, the dating of the arch’s late Roman frieze and the elucidation of its imagery plays a significant role. What follows attempts a sketch of the various problems that bear on the frieze’s interpretation, noting the strengths and weaknesses of the various analyses. Rather than follow the course of the scenes as they encircle the monument, we treat a variety of issues in a sequence that, we hope, builds progressively toward a coherent interpretation. As shall become clear, we believe that the arguments for the frieze’s Constantinian date and its topical significance remain the most compelling.

7An insistent claim about the arch’s historical frieze is that it was taken from a Diocletianic monument and rededicated to Constantine. Four scholars have attempted to make the case. Wace argued that the arch’s creator must have been Diocletian because (1) the original imperial portraits had been removed and replaced and feet were missing or poorly carved on two reliefs; (2) the style of the reliefs signaled a date of ca. 300; and (3) the frieze’s imagery seemed to require the new dedicatee to have celebrated a triumph. While conceding that two of the six scenes that comprise the frieze were likely to have been Constantinian (Obsidio and Proelium apud pontem), he believed that the remaining four (Profectio, Ingressus, Oratio, Liberalitas) were of Diocletianic origin[4]. Two decades later, Byvanck rightly noted that the frieze was carved directly onto the monument but felt, on the basis of comparison with the Tetrarchic column base in the forum, that it had been sculpted in situ in the reign of Diocletian and Maximian[5]. Sixty years later still, Knudsen made the claim that all six scenes were originally representations of Diocletian[6]. Recently, Rose resurrected these arguments in favor of a Diocletianic date[7]. Not only does his new treatment reiterate a series of long-recognized problems (the recarved heads, the missing feet, the presence of bearded soldiers, the question of a triumph), but it also offers a re-interpretation for the long-held significance of the river battle – now claimed to be a representation of an event poorly acknowledged in the historical record and otherwise unknown in the visual tradition as a victory of Diocletian’s. All of these claims, along with several other issues, shall be treated below.

8The relief that has traditionally been interpreted as marking the first in the narrative cycle constituted by the six friezes in the central band of the arch is that on the west side depicting an army setting out on campaign – a profectio (Fig. 2). Taken in conjunction with the other five reliefs, it immediately raises two questions crucial to the interpretation of the frieze as either Diocletianic or Constantinian. If the arch’s frieze was originally dedicated to a member of the Tetrarchy, why does no section of the entire frieze feature more than one emperor? And, does every scene, including this one, actually include the emperor?

9The Profectio scene has, at times, been held to display no emperor whatsoever. A seated figure rides in the wagon at the far left of the scene, and his smaller stature (Fig. 3) has been presumed sufficient grounds for denying him imperial status, on the assumption that difference in scale was primarily intended to be understood hierarchically. Yet the relief’s designers were equally concerned with two other distinctive aspects of composition. First, the appearance of the wagon at the far left of the Profectio scene was clearly intended to mirror the form of the composition on the east end of the arch where the Ingressus (Fig. 4) depicts an emperor (whose identity, despite the missing head, is unambiguous) similarly riding at the rear of the procession. Given that the visual form of one procession clearly parallels that of the other, should we not understand the elevated figure on a wheeled conveyance at the back of each parade to represent the emperor? And second, both of the processional sections of the frieze, Profectio and Ingressus, are distinguished by their conspicuous isocephalic arrangements: a single file in the former, and staggered, two abreast, in the latter. In the Profectio, all the figures stand nearly the full height of the frieze, their heads aligned, while the horses and camels accompanying them are rendered shorter than the human figures and smaller than lifesize, suggesting their lesser significance as aspects of staffage. As a result of the isocephalic design, the relative size of the prominently elevated figure in the Profectio, seated atop a cushion set in the bed of his carriage, was necessarily diminished, rendering him almost comically small in comparison to the next figure to the right and all others on the frieze save for the wagon’s driver[8]. Indeed, as we shall see (§ 13 below), there are grounds for arguing that this was the first part of the frieze to be carved and presented problems that would later be resolved in the remaining sections.

10Nevertheless, might we understand this seated figure to represent the emperor? There is hardly a consensus view[9] The recent claim that the emperor’s presence should be excluded because it would be unique among the “encircling” processional friezes amid the small corpus of surviving arches can be excluded as inaccurate and inapposite. The comparanda adduced in support of this argument present fundamentally generic images lacking in precisely the specific historical imagery found on Constantine’s arch[10].

11New grounds for any argument concerning the identity of the single, unambiguously seated protagonist in the wagon are essential. Some basis is provided by the earliest surviving depictions of the arch’s reliefs, the prints produced by Pietro Santi Bartoli (Fig. 5. 6) and Luigi Rossini (Fig. 7)[11]. Both artists took care to depict the reliefs in condition they found them, using hatchings to signal the most obviously damaged parts of the frieze, although Bartoli’s display what may be a penchant to reinstate certain elements for aesthetic reasons (see below § 12).

12In the case of the Profectio, the figure in the wagon still bears its head to this day, albeit badly damaged. It is depicted by both Bartoli and Rossini intact, with the extended right arm bearing a globe. If these renderings were accurate,[12] there can be no doubt that this figure was intended to represent the emperor – for no other mortal would have been depicted on a public monument with this long-standing symbol of power, so central to the imperial image (cf. Fig. 9. 46. 47). And if this is the emperor, he was represented here without a tetrarchic colleague (cf. Fig. 8. 9). Given that the soldiers who lead this march bear standards surmounted by images of Victoria and Sol Invictus (whose connections to Constantine are discussed below, § 6), this figure must be meant to represent Constantine.

13Why in all the other scenes does the emperor appear, unambiguously, alone?

14Diocletian had come to power in November 284 with one other emperor ruling in the West, Carinus, whom he then eliminated in a series of conflicts in upper Moesia (see below, § 14). By late 285, however, Diocletian had named Maximian as his co-ruler, and in 293 he expanded the imperial college with the addition of two further emperors to form what modern scholars now term the Tetrarchy[13].

15This system of joint rule between two senior Augusti and two junior Caesares formed the bedrock not only of imperial policy but also of imperial iconography in the years between 293 and Diocletian’s and Maximian’s joint abdication in May 305. Joint rule was the hallmark of laws, inscriptions, and coins, and when imperial images were produced, these followed the imperative of emphasizing that imperial power was a collegial affair, among two graded pairs of equals. It is thus that we regularly find the four rulers grouped together on their monuments: the porphyry tetrarchs now on the corner of St. Mark’s Basilica in Venice, as well as those now in the Vatican collections[14]; in the wall paintings of the Temple of Ammon at Luxor [15]; on the sole surviving base of the Five-Column Monument in the Forum[16]; on the coinage, notably with the four rulers standing side-by-side around a sacrificial brazier before a military camp gate.[17] In other contexts we sometimes see only the two senior rulers – the Augusti (Fig. 8. 9), a pairing also found on the newly discovered reliefs from Nicomedia (Izmit), which date to the period of Diocletian’s and Maximian’s diarchy (before 293)[18].

16Any divergence from this well-established pattern should give pause, for when exceptions occur, it is for a clear reason. Thus, on his arch in Thessalonica, Galerius appears alone in two scenes with the intent of glorifying his personal victory over the Persian Shanashah Narses in 298. Yet the same monument is careful to display all four tetrarchs grouped side-by-side on a neighboring panel[19]. By analogy, the isolated image of Galerius might be comparable to the lone emperor in the Roman arch’s two battle scenes, but what of the other scenes on the frieze? As we have seen, the Profectio and Ingressus display a single emperor; so too the Oratio and Liberalitas, the latter long a subject used to showcase joint rulership (see below, § 9). One is hard-pressed to understand why on a frieze said to be dated soon after Diocletian’s joint visit to Rome with Maximian to celebrate their Vicennalia in 303, the other tetrarchs – or at least Maximian, who equaled Diocletian in rank – were omitted[20].

17It cannot be denied that, when the arch was dedicated in 315, Constantine also had a co-emperor in the person of Licinius. Yet the arch’s iconography, with its stinting acknowledgement of the existence of any other imperial colleague, matches the circumstances of Constantine’s and Licinius’ shared rulership much better than it could any construal of the first tetrarchy. When Constantine defeated Maxentius in 312, he was still heavily reliant on his colleague to secure the east against their more aggressive and less cooperative third remaining co-ruler, Maximinus Daia. He thus arranged for Licinius to meet him in Milan in February 313 and to marry his sister, Constantia. There they also jointly issued the decree now known as the Edict of Milan. That same summer, Licinius defeated Maximinus, who committed suicide, thus setting the scene for what could have been a secure new diarchy. But Constantine had other plans. He had already made clear his desire for domination among his co-rulers in 312 by having the Roman senate proclaim him maximvs avgvstvs, a title he advertised regularly on coins and inscriptions, including the arch’s epigraph[21]. In 315, when the arch was completed for Constantine’s celebration of his decennalia in Rome, he also attempted to promote as a colleague a distant relative named Bassianus against Licinius’ wishes, an episode which provoked a civil war that soon resulted in Licinius’ defeat at the Battle of Cibalae and his loss of control over Illyricum to his western “colleague.” Thus, by the time the arch was erected, Constantine’s relationship with his one remaining co-ruler was anything but collegial, which explains why Licinius would not have been equally celebrated on the monument. In this environment of imperial conflict, the completion of the arch in such a way as to showcase the city of Rome’s de facto singular ruler should hardly surprise, despite the institutional reality of co-rule.

18Indeed, two heads on separate roundels in the arch’s north façade originally depicting Hadrian were recarved for the arch as a different person, whose status is confirmed by the nimbus (Fig. 10; cf. Fig. 17). L’Orange had identified these as representations of Licinius in acknowledgement of the political reality, despite a tense and muted co-rulership[22]. Yet the imperial figure on these roundels does not display the physiognomy of Licinius as seen on either coins or marble portraits, even granting the tendency of Tetrarchic iconography to efface individual appearances. In light of this, a growing consensus has emerged that these Hadrianic roundels were refashioned with portraits of Constantine’s father, Constantius Chlorus[23]. Whoever this figure may represent, (as his features are not typologically related to either Constantine’s co-ruler or father), his subordination in the overall imagery of the monument is patent, and suggests, again, that there is no representation of either Tetrarchic or Dyarchic collegiality anywhere on the arch. This again points to the Constantinian origin of the frieze[24].

19The circumstances of Diocletian’s brief sojourn in Rome and the record of his monumental construction in the Urbs tell against the assumption that there was a “lost” arch from which the frieze was removed. We are fortunate to have a fourth-century source, local to Rome, that catalogs in some detail the architectural patronage of Maximian and Diocletian, the Chronicon Calendar of 354. This compendium reports a list of emperors and their benefactions to the city, including an enumeration of the building programs for these two co-rulers, who are still paired a half-century later:

“While these [men] were emperors, many public works were constructed: the senate house, the forum of Caesar, the Basilica Julia, the Theater of Pompey, II porticoes, III nymphea, II temples [one] to Isis and [one to] Serapis, the new arch (arcum novum), the Baths of Diocletian[25].”

20This list, detailed and extensive, corresponds with buildings generally known from other sources to have been built or rebuilt under Diocletian. The mention of a “new arch” (or perhaps “the New Arch”), also reported in the late compendia that inventoried the regions of the city, the Curiosum and Notitia, has long been associated with another monument whose remains were discovered in the northern part of the city (Region VII) on the site of S. Maria in Via Lata, dated to 293/294 by a fragmentary inscription recording Diocletian’s Decennalia[26]. This tetrarchic monument, destroyed 1488/1492, was apparently decorated with spolia of the Claudian period, but these are not the only known decorations to have survived. Its Diocletianic elements, two pedestal reliefs now in the Boboli Gardens, are more carefully rendered and classicizing in style than the reliefs on the Five-Column Monument in the Forum – or the frieze on the Constantinian arch and clearly belong to the late Roman era[27]. If the Boboli pedestals do indeed derive from the Arcus Novus, they indicate that it, like the Arch of Constantine, made use of a mix of old spolia and new carvings[28].

21But if the Arcus Novus was not the source of the Constantinian frieze, and Diocletian built a second arch, why was this not enumerated in the Chronicon Calendar of 354 or in the Curiosum or Notitia lists? Furthermore, it is arguable that Diocletian would not have been keen on allowing the construction of an arch at Rome in the years following his joint adventus with Maximian in 303. To be sure, the Five Column Monument in the forum was completed in that year, but since it appears in the background of one of the frieze segments (see below § 6. 8), these latter must postdate the Five Column structure[29]. But, as Lactantius reports, Diocletian and Maximian had anything but an irenic visit to the city when they came in November of this year. In the face of civic unrest, Diocletian departed Rome for Ravenna on December 20, less than two weeks before he was to assume his ninth consulship. Leaving as he did in wintry weather, he contracted a serious illness on his journey northward that would dog him until his abdication and hasten its approach[30]. Diocletian was therefore not happy with the Romans of Rome – nor, apparently, they with him, making it hard to imagine that they would have celebrated him with a second arch in the city following his angry departure in late 303.

22The Obsidio scene (Fig. 11) is universally acknowledged as Constantinian. Chief among the reasons is the partially preserved head of the emperor. Even in its present state, having lost nearly all its surface, likely due to natural deterioration of the stone, it still retains enough of the carved marble to show the recognizable profile of a clean-shaven Constantine (Fig. 12). The visage, with its distinctive long and aquiline nose, jutting jawline, and prominent chin, corresponds compellingly with that of Constantine known from his portraits (Fig. 13), his numismatic iconography (Fig. 14), and even some of the recarved heads in the Hadrianic roundels.

23A panegyric to Constantine, delivered probably at Trier in 313 (Pan. Lat. 12 [9]), recounts his war with Maxentius in elaborate detail, beginning with his crossing of the Alps and moving through his early victories at Segusio (Susa), Augusta Taurinorum (Torino), Mediolanum (Milan), Verona, and Aquileia before turning to the denouement at the Milvian Bridge [31]. The panegyrist describes each of these early victories but devotes particular attention to Verona, which proved the most difficult and most important of the battles on Constantine’s march to Rome. After relating how the emperor succeeded in bottling up the Maxentian forces inside the city to which he then laid siege, the orator proceeds to reprimand Constantine, in a classic display of parrhesia, for having personally become involved in the battle: “Why did you enter the fray yourself, why did you thrust yourself into the densest throng of the enemy, why did you send the state’s salvation into such great danger?”[32] This corresponds precisely to the scene depicted on the frieze: the emperor stands, taller than any other figure, in the thick of combat, calmly directing the fight, while Victoria crowns him from above. In the panegyric, the parallel invocation of the emperor’s unusual presence at the heart of the fighting suggests that it had rapidly become legendary and would soon become a rhetorical trope. In his Panegyric of 321, Nazarius also reports the same sequence of battles before dwelling on the siege of Verona, where he reports how the city was defended by “a great force of men along the entire circuit of the walls (per tota admodum moenia magna vis hominum)” and notes, again, that Constantine fought in person[33]. In both instances, we can witness Constantinian verbal and visual rhetoric operating in tandem, conveying, in a highly coordinated manner, the same imagery in both media[34].

24Another argument for Constantinian dating can be built on the horse prominently featured in the leftmost segment of the relief. It is marked, despite its small scale, by its carefully rendered bridle, its gaping mouth champing at the bit, and the distinctive locks of its unusually bushy mane (Fig. 15). These features distinguish this horse from all the others in the entire frieze cycle, which have closely cropped manes that stand stiff and upright. This majestic charger, caparisoned with jeweled tack, can only be the emperor’s mount. The significance of the emperor’s horse in the scene is two-fold. First, by showing Constantine dismounted, it emphasizes his brazen entry into the fray to join his infantry on foot, as the panegyrics report. And second, the appearance of this horse’s head closely resembles that of a horse that became a standard feature of Constantine’s coin obverses in the aftermath of his victory over Maxentius. This same horse appears on numerous coin types produced at five western mints beginning in 315 – the same year the arch was erected – and most notably, on Constantine’s famous Ticinum medallion, now dated to 321 (Fig. 16). Both of these aspects played a role in the increasing emphasis on the military in the coinage beginning in 306. These various factors constitute yet another argument for the frieze’s date[35].

25One of the frieze’s curious features is that some of the figures are bearded while others are not. Can the presence of some bearded figures have evidentiary value for the frieze’s dating?

26Rose has argued that the high number of bearded figures on the frieze indicates a Diocletianic date, taking the Profectio scene on the arch’s west end (Fig. 2) as a prime case in point[36]. But a closer look reveals that the army it depicts comprises 21 figures, 17 of which have enough of their faces preserved to determine whether or not they wear beards: 9 do, while 8 are clean-shaven. There is a clear distinction between the figures at the front of the parade, footsoldiers who wear helmets and are beardless, and those at the back (behind the camel), officers who wear beards along with Pannonian hats. In fact, since more of the figures at the front have damaged faces, it is quite probable that the relief originally had more beardless than bearded figures.

27Styles of personal appearance have long played a role in the study of Roman portrait sculpture, significantly in that aspect of visible ideological conformity known as the Zeitgesicht. Yet there are numerous mostly unidentifiable figures, datable on contextual or archaeological grounds to the reign of a particular emperor, whose preferences for facial hair do not conform to the style of that ruler and his time. Thus the presence or absence of beards, in any period, hardly offers consistent or definitive grounds for attribution to a particular era: some men wore beards in the time of Trajan and Constantine, while some in the time of Hadrian and Antoninus did not. And in every era, soldiers might wear the short stubble of a “campaign beard” – akin to the tetrarchic style – that was one highly visible consequence of life on the march and at the front[37].

28The wearing of beards continued into the Constantinian era, and the fashion is only probative of identity in this period for the emperor himself, once he broke with tetrarchic convention and began having himself represented as clean shaven[38]. But this did not commence at the start of his reign, for Constantine is regularly depicted with a tetrarchic beard on coin obverses for the first six years of his rule[39]. A beard also appears on the head recarved to represent Constantius Chlorus (or Licinius) in the Hadrianic roundels (Fig. 17), where the figure is depicted sporting a typically tetrarchic stubble beard. But this beard is carved using the same technique employed in the beards of the soldiers at the rear of the parade on the west frieze (Fig. 18): thick stubble rendered schematically with stippling inside a raised field demarcated from smooth skin by a curvilinear tracing running high above the jaw line. Given the similarities of style and technique, it is likely that the same team of sculptors – or at least sculptors of the same era – were responsible for both the refashioning of the portraits on the roundels and the carving of the Profectio on the frieze. Thus, while the presence of bearded men offers no compelling indication of the frieze’s date, the rendering of beards on the arch’s various late Roman reliefs might nevertheless indicate that they were all carved as part of a single campaign.

29The Constantinian arch is bristling with the pairing of Victoria and Sol Invictus, especially on military standards. Both divinities were lionized by Constantine, for the Sun God became paramount in his iconography after his so-called “pagan vision” of Apollo in 310. But is Sol Invictus just as likely to point to Diocletian, as Rose contends?[40] It is true that Sol had appeared on Diocletian’s coinage and on one of the reliefs from the surviving base of the tetrarchic Five-Column Monument in the Forum. The view of Bardill, who focused intensively on the question, was that Sol Invictus had been dropped from imperial iconography by Diocletian after his first year in power and only reappeared late in the tetrarchy in association not with Diocletian but with his Caesar, Galerius[41]. Marlowe had pointed out that the base from the Five-Column Monument with Sol Invictus must be associated not with Diocletian but with either Constantius Chlorus or Galerius since its inscription reads caesarvm decennalia feliciter[42]. Two of the monument’s other bases have also been documented, and although now lost, their inscriptions were recorded: avgvstorvm vicennalia feliciter and vicennalia imperatorvm. L’Orange, followed by Kähler, made the plausible argument that the former of these two lost bases bore an image of one of the Augusti and the latter the figure of Jupiter, as we see on the rendering in the arch’s Oratio scene. The third of the bases – the surviving one mentioned above – was for a column and statue to a Caesar, and there would have been two more, one for each remaining member of the Tetrarchy[43].

30If the lone extant base of the Five Column Monument honored not Diocletian but one of his Caesars, the presence of Sol Invictus on one of its sides thus offers no strong evidence of Diocletian’s interest in or association with this deity. Indeed, the side of the base with Sol also features Mars, and on the opposite side a parade of senators is interspersed with three military standards featuring an Aquila, a Victoria, and an Aquila above the Genius Populi Romani[44]. Had military standards pairing Victoria and Sol Invictus of the sort featured on the Profectio seemed important to the Tetrarchy when the Five-Column Monument was erected in 303, why does the pair not feature in this relief?

31Numismatic evidence confirms the view that Sol Invictus was not characteristically Diocletianic. In the years between the rise of the dyarchy in 284 and the death of Maximinus Daia in 313, attested coin types with legends dedicated to Sol Invictus number 126 for coins with Constantine’s obverse and only 1 for those with Diocletian’s. Indeed, of the 174 type-legend combinations with Sol Invictus in RIC VI, only 7 (4%) were even minted before Diocletian’s abdication in 305 (see chart: Fig. 19). This is a clear indication that Sol Invictus was not only of scant importance to Diocletian but also to other members of the “first tetrarchy” before its dissolution in 305. Of the other tetrarchs, only Maximinus Daia – who came to power upon Diocletian’s abdication – shows any marked inclination to favor Sol Invictus (on 27 issues), but he does so only after Constantine begins showing a preference for this deity[45].

32An abundance of evidence suggests that not only the presence of Sol Invictus, but also Sol’s pairing with Victoria, as well as the arch’s “triumphal” character, were fundamental to any understanding of the monument’s Constantinian iconographic program[46]. This conclusion is supported by three interrelated arguments. First, Constantine’s obsession with solar imagery and the sun god are widely attested in the sources, both pagan and Christian – even Eusebius’ Life of Constantine is filled with solar imagery, not to mention Constantine’s own abundant writings[47]. Moreover, in the years between 310 and 315, precisely the period when the arch was constructed, Constantine began proclaiming his attachment to the late antique incarnation of the sun god by adopting the title invictvs as part of his official nomenclature, a designation that becomes nearly universal by the year of the arch’s dedication[48]. At this same time, his most common numismatic reverse is dedicated to soli invicto comiti and depicts the Sun God nude but for a cape draped over his shoulders and billowing, wearing a radiate crown, and bearing a globe in his left hand and a whip in his right[49].

33Secondly, on the arch’s west frieze, Sol Invictus is depicted alongside Victoria atop the only standards in the parade (Fig. 20). This same pairing appears in a crucial passage from the anonymous panegyric delivered to Constantine in 310:

“For you saw, I believe, O Constantine, your Apollo accompanied by Victoria, offering you laurel wreaths, each one of which carries a portent of thirty years”[50].

34The broader context in which this comment appears describes the famous “pagan vision” of Constantine when the emperor first claimed to have been visited by the appearance of a divine light near the Temple of Apollo at Grand (Vosges). In his recreation of the event for the emperor, the anonymous orator uses the classicizing Apollo rather than the late imperial Sol Invictus to name the deity, but there has never been any doubt that this is our first attestation of what would become a standard pairing of the Sun God with Victoria. Constantine eventually came to reinterpret the vision as a sign deriving not from the traditional pagan Sun God but rather from the Christian deity – perhaps as the result of a dream he is reported to have had in the days leading up to the battle of the Milvian Bridge[51].

35The Sun God – Sol Invictus – was not only a defining figure in Constantinian verbal and visual propaganda in the middle years of his reign, but, together with the personification of Victoria, he represents a persistent leitmotif in the arch’s iconography, which constitutes a third argument. We have already seen that, on the west end Profectio (Fig. 20), the soldiers hold aloft the standards crowned by the figures of the twin divinities. Similarly, on two of the socle reliefs carved in place below the columns on the arch’s north façade, soldiers carry the same paired military standards (Fig. 21, for one of these). Although badly eroded, the pair also seems to appear in wreathed roundels below the aquilae on the standards of the southeast socle flanking the south side entry as well. Higher up on the arch, in one of the Antonine reliefs in the attic, the Rex Datus, the Constantinian sculptors have also recarved the two leftmost figures on the four standards being displayed in the archways (Fig. 22) so as to transform them into Victoria and Sol Invictus as well[52]. In the arch’s eastern passageway Victoria and Sol Invictus (Fig. 23) are found yet again, this time facing each other on opposite walls, but with Victoria now crowning the emperor – as she is said to do in the Panegyric of 310 and as she does in the Obsidio relief[53].

36The message is thus clear and consistent: Sol Invictus and Victoria are the tutelary deities who guaranteed the emperor’s success. Thus it seems most likely that, since the west end Profectio, in which they prominently figure, is central to the arch’s broader Constantinian iconographic program, it can hardly be a randomly spoliated relief originally intended for Diocletian.

37The troops depicted are not only to be identified by the signa they carry. Constantine’s forces include a row of Mauretanian archers[54], recognizable by their hair and headgear and by the bow-pulling gesture they make (although their bows have largely eroded away). Mauretanians might seem odd in the forces of an emperor attacking Italy from the northwestern provinces, but it should be remembered that Constantine also had control of Spain, with which Roman Mauretania was always more closely connected than with the rest of North Africa[55]. These archers might then have been auxiliary forces recruited throughout the westernmost provinces under Constantine’s authority. Indeed, as Zosimus reports, many of the forces Constantine used to march on Maxentius fit a similar description:

“[Constantine] recruited forces among barbarian war captives whom he had in his power and among Germans (Γερμανῶν) and the other Celtic peoples, as well as men assembled in Britain, to the number of about 90,000 infantry and 8,000 cavalry and crossed the Alps into Italy[56].”

38Constantine seems to have devised a strategy for his attack on Italy that relied heavily on a mobile army assembled from non-Roman auxiliaries consisting especially of barbarians. Scholars of Roman military history agree that this constitutes the first firm evidence for the creation of the late Roman comitatensian army[57]. The effects of this mobile field army’s creation over the course of the century are recorded in the Notitia Dignitatum, a rich source for details of Roman military and administrative organization[58]. The Notitia’s official list of military units includes a number of auxiliary troops that are connected with Constantine’s initial mobilization of the comitatenses – comprising, according to Zosimus, a broad group recruited in the West – among which was a unit called the Cornuti, of Germanic origin, which came to play a major role in military history in the decades to follow[59]. We are fortunate to have not just this name cataloged in the Notitia but also an image of their distinctive shield ensign in the illustrations accompanying the Notitia’s lists (Fig. 24). The symbol, a peculiar pair of heads curling toward one another in a crescent shape, appears on the folio alongside the shields of units that accompanied the Magister Militum Praesentalis of the Western field army[60]. Notwithstanding the vagaries of the Notitia’s manuscript transmission, this represented the 15th-century copyist’s attempt to replicate what he understood to be the original. Fortunately, the arch itself preserves what has been taken to be that original fourth-century form on one of the socle reliefs at the base of the north façade (Fig. 25). This Constantinian-era image features a shield with a winged Victoria above the boss, and below it the same crescent-shaped emblem discernible on the shield in the Notitia, although here one can clearly make out the facing heads of goats, with long spindly horns projecting from their temples: this is clearly what was intended with the name Cornuti.

39These same spindly horns, formed into a crescent, also appear on the front of the helmets worn by a number of the infantrymen in the Obsidio relief (Fig. 26)[61]. The helmets are of themselves odd relative to earlier Roman military headgear, with their bowl-like shape and curling flange at the rear; but the frontal horns they bear are nothing short of bizarre in the context of standard Roman military headgear. They are clearly depicted in this way to draw attention to a new, distinctively foreign kind of auxiliary fighter. In the Obsidio relief, these Cornuti are not strangers to the fight but constitute the advance guard of Constantine’s siege lines. Moreover, it should be noted that they appear not just in the Obsidio relief but also in the scenes of the Profectio and the Proelium apud pontem (Fig. 27. 28), making it clear that the arch’s designers wished to draw attention to these soldiers as crucial to every military operation they depict.

40If these auxiliary soldiers can be connected with Zosimus’ report concerning Constantine’s mobilization of barbarian forces in his expeditionary army of 312, this too would suggest a Constantinian origin for these reliefs. And even if this connection is not conceded, one would need to explain why this peculiar feature appears on all three reliefs – Obsidio, Profectio, Proelium apud pontem – and on the shield on the Constantinian socle if they were not part of the same iconographic program. If, as seen above (at § 4), the Obsidio relief is almost certainly Constantinian, should not these unusual helmets on the other two frieze reliefs suggest that they were all carved together as parts of the same – Constantinian – sculptural project?

41The Oratio scene is set in the Forum Romanum (Fig. 29), where the emperor stands on the Rostra and, as his gesture indicates, holds forth before a great throng that gathers about the platform. His addressees include members of the senate, wearing their voluminous togae contabulatae, and standing just beyond these, the plebs Romana, clad in tunics and sometimes also paenulae. This specificity of his audience’s social standing is hardly unique on Roman public monuments[62] and is matched by that of the setting, which is clearly established by the series of monuments that fill the background: from left to right, the arcaded Basilica Iulia, the single-bayed Arch of Tiberius, the columns of the so-called Five Column Monument projecting behind the Rostra, and the three-bayed Arch of Septimius Severus.

42The Five Column Monument was the most recent among these, a tetrarchic structure set up in 303 on the occasion of the Vicennalia of the Augusti and Decennalia of the Caesares to celebrate Diocletian’s and Maximian’s joint visit to Rome[63]. In the years between its construction and dedication of the arch (303–315), imperial itineraries reveal that only four emperors spent time in Rome: Diocletian, Maximian, Maxentius, and Constantine. Maximian and Diocletian’s 303 appearance was a joint affair, and this should lead one to expect their co-presence in public scenes like this one. Unless we are to imagine there was a similar panel on which the absent colleague (Maximian or Diocletian) would have appeared analogously alone but that is now lost, the event depicted in the Oratio relief on the arch cannot be connected to their conjoined Vicennalian festivities in 303. With regard to Maxentius, too little is known of his public oratorical performances to argue that he is to be taken as this lone speaker[64].

43In the case of Constantine, by contrast, we have direct testimony that he made more than one public oration following his victory in 312. This is reported in the Panegyric of 313:

“Now why should I mention your decisions and acts in the curia, by which you restored to the Senate its previous authority, refrained from boasting of the salvation which they had received through you, and promised that its memory would rest eternally in your breast. I would say more about your imperial speeches (de divinis orationibus tuis), about your kindness (benignitas) voluntarily extended rather than procured by entreaty, if I did not prefer to let your words (dicta) remain untold while I am in haste to praise your deeds. However little we know of the substance of what you said in the Senate (verba tua in senatu habita), yet the fame of your clemency (clementia) proclaims its spirit[65].”

44Constantine thus clearly held one or more speeches before the senate as well as the city’s broader public[66]. These largely concerned the restoration of senatorial privileges and, more broadly, the return of civic institutions. Yet there is also a bellicose undertone. The celebration of Constantine’s clementia implicitly compares him favorably with the infamous vengeance of former rulers – Cinna, Marius, Sulla – and is confirmed by the apostrophe to Roma that follows: “fortunate at last in a civil victory!” (O tandem felix civili, Roma, victoria!)[67]. This tone is echoed in the Oratio relief as well, where the emperor appears on the rostra dressed in tunic and paludamentum – military, not civic, attire. The scene represents a proclamation to the citizenry by a military leader who had, as the arch’s dedicatory inscription declares, freed the Romans from a tyrant and whose rule effectively undermined the long-standing ideology that distinguished the spheres of domi militaeque [68]. This was a visible declaration that the Roman aristocracy recognized the new state of political affairs and acknowledged the emperor’s powers. Constantine, Optimus (since 312?), and now Maximus (as of 313), was in reality sole ruler despite the nominal continuation of the diarchy: the singular presence on the rostra can only have been Constantine[69].

45The Liberalitas relief (Fig. 30) is even more striking for its representation of a single emperor, seated at the scene’s center and pouring down coins from a tabula into the folds of a senator’s toga. Numismatic evidence indicates that, when multiple emperors were ruling, they were regularly shown together distributing coins on the occasions of such largitiones, which were then billed as joint affairs by the plural legend liberalitas avgvstorvm[70]. Shared benefactions were central to tetrarchic ideology, and there is no evidence, visual or literary, that might lead us to believe that Diocletian would have seen fit to leave his co-Augustus entirely out of this scene.

46This is all the more pointed if we consider that a joint congiarium is attested in connection with Diocletian and Maximian’s vicennalia of 303 in the Chronicon Calendar of 354: “Diocletian and Maximian were emperors for 21 years 11 months and 12 days. They gave (dederunt) a congiarium of 1.550. While they were emperors many public works were constructed…” [the catalog that follows is given above at n. 25][71].

47Previous studies of congiaria have established that they generally occurred in imperial fora or basilicae, enclosed spaces with two-storied porticoes whose second floors could be commandeered by imperial personnel so that they could manage these massive cash distributions well out of reach of the crowds and thereby prevent theft or unrest. In these spaces, the individual recipients would be allowed up the stairways in orderly fashion to receive their allotment on the upper floor. We find just such a scene in the arch’s Liberalitas relief. At its center, we see the emperor situated on what must be the ground floor of an interior structure, as suggested by the candles born by the guards at his two sides. He sits on a suggestus, well above the crowd of senators, and offers one of them twelve coins from a tabula. Above him, to either side, treasury officials are shown on a balconied upper story, seated in four groups of three each, while the torso of a fourth figure rises from the lower frame, an indication that he is climbing the stairs between floors to receive his allotment – in this case just six coins – also from a tabula[72]. It is clear that order and record-keeping were at a premium, for each citizen was entitled to only one donation, and the size of the crowds and amount of treasure on hand would have formed an explosive mix if such a tight system of controls were not imposed.

48The Chronicon Calendar’s account continues by reporting a second distribution in the year 303, a sparsio in the circus:

“They scattered gold and silver coins in the circus (sparserunt in circo aureos et argenteos). The footing of the covering tent (partectorum podius) fell and crushed 13 persons.”

49This second distribution will surely have been smaller than the congiarium, conducted as it was in an open air space, yet the memory of it – and its attendant tragedy – were clearly still vivid in the eyes of the Romans four decades on. This memory included the joint participation of both emperors, much as the Chronicon Calendar reports for the congiarium[73]. Given that the arch’s Liberalitas so obviously focuses on a singular emperor, centrally placed as a focal point in the relief, it should not be taken as a representation of either event.

50Instead, it is surely preferable to connect the arch’s single-ruler Liberalitas with a Constantinian benefaction, especially because such a benefaction is indeed found in the sources[74]. A distribution of coin in Rome in 312/313 may be alluded to in the passage of the Panegyric of 313 quoted above (“your kindness [benignitas] voluntarily extended rather than procured by entreaty”). And the Panegyric of Nazarius from 321 suggests the same (“It would take long to enumerate our leader’s benefactions [beneficia] from that point, which glisten forth returning to the world without measure interwoven with his goodwill [benignitate]”)[75]. These generalizing allusions are reinforced by the evidence of the coins. A contemporary solidus (Fig. 31) from the mint of Ticinum, well known for its closeness with Constantine’s consistory, shows on its reverse a personification of Liberalitas, holding a tabula similar to those found on the arch’s relief. The coin is datable to the precise period of the arch’s dedication by its reverse legend, which reports Constantine’s eleventh anniversary as emperor and fourth consulship (xi imp iiii cos), i.e. July 25 – December 31, 315[76]. And a recent study of coin types from the mint for Rome has confirmed that there were four periods of greatly increased monetary output from that mint which coincided with public benefactions in the city under Constantine: 312–313, 315, 326, 335–336[77]. The last three correspond with Constantine’s Decennalia, Vicennalia, and Tricennalia, and the first, obviously, with his entry to Rome in the aftermath of the Milvian Bridge Victory. Thus, with the Liberalitas scene, we have another instance in which the preponderance of the evidence points to Constantine.

51On the arch’s eastern end, a military procession celebrates an entrance into a city (Fig. 4). As it turns the corner of the monument the scene continues on the north façade, and the parade passes through an archway that clearly features on its summit elephants with projecting trunks (Fig. 32). As L’Orange recognized, this is Rome’s Porta Triumphalis over the Via Flaminia, which Domitian had reconstructed and adorned with a quadriga pulled by elephants[78]. The Flaminia was the road which crossed over the Tiber just north of Rome via the Milvian Bridge, and the frieze’s designers surely intended that its beholders would comprehend the topographical implications. They reinforced the connection by the placement of the Antonine Profectio relief in the attic above the frieze in which the same aspects of the city’s built landscape appear, with the elephant arch above a personification of the Via Flaminia[79].

52It has long been recognized that the scene on the arch’s east end is not a triumph, especially because the emperor sits in a four-wheeled parade wagon rather than standing in a chariot[80]. But is this comprehensible as Diocletian’s arrival for his Vicennalia? That was, of course, the event celebrated with the erection of the Five-Column monument depicted in the Oratio scene (Fig. 29), where the whole point was to emphasize the shared nature of Tetrarchic rule.

53But the emperor parading in the carruca on the Ingressus frieze is clearly alone, with no space for a co-ruler[81]. If this relief had been intended, like the Five-Column Monument, to represent Diocletian’s visit to the city with Maximian on the occasion of their Vicennalia, we would expect to see the co-emperors depicted together. Maximian’s absence suggests either that there was another section of the frieze on the putative Diocletianic monument which duplicated the imagery for the other Augustus, or that this is not the relief’s subject – and Diocletian not its protagonist. Such a “doubling” is unlikely, however, since similar scenes enacted by these co-rulers earlier in their joint reign had featured both of them together. Diocletian and Maximian are represented on a gold medallion of 287 (Fig. 8) jointly celebrating their consular procession in an elephant quadriga, and they appear standing side-by-side holding globes and scepters crowned by Victoria on an aureus of 293 (Fig. 9). Both rulers also appear on a unique mold for a series of terracotta reliefs depicting their joint triumph of 303, again in an elephant quadriga[82].

54The scene on the arch is better interpreted as an image of Constantine’s adventus in 312, especially since it corresponds to a description preserved in Nazarius’ Panegyric of 321. This recounts Constantine’s entry into Rome following the battle at the Milvian Bridge:

“The emperor’s entrance into the City must be described (dicendus in Urbem ingressus est imperatoris), and in expressing the greatest rejoicing of the senate and people of Rome no oration is likely to please unless it is exuberant itself. No day since the founding of the City (post Urbem conditam) has shone upon the Roman Empire for which there either was or ought to have been a thanksgiving so lavish and distinguished; no triumphs which the antiquity of annals has immortalized in literature were so joyful (nulli tam laeti triumphi). Leaders in chains were not driven before a chariot (ante currum), but the nobility marched along, freed at last. Barbarians were not cast into prison, but ex-consuls were led out of it. Captive foreigners did not make his entrance (introitus) honorable but rather Rome itself now freed (Roma iam libera)[83].”

55Nazarius took pains to point out that Constantine’s was a ceremonial introitus or ingressus, not a triumph – even if it was modeled on the celebratory vernacular of that most exorbitant Roman event. And this is precisely what one sees in the parade on the arch[84]. Indeed, the glorious entry of Constantine into Rome on the day following the Milvian Bridge is celebrated not just in Nazarius but also in multiple contemporary accounts from as early as 313, and the momentous event continued to be celebrated annually as the adventus divi (Constantini) on October 29 deep into the fourth century, as we learn in the Chronicon Calendar of 354 [85].

56Constantine’s armed troops are shown on the arch’s east frieze as they cross the pomerium and enter the city. They appear clean-shaven and accompany their lone ruler with the pomp characteristic of a triumph but the restraint necessary for the celebration of a civil war victory. These are Constantine’s troops, and the scene is that emperor’s ingressus into Rome.

57It is widely agreed among those who have studied the arch’s construction that all its marble blocks were re-used. These re-usages are of differing kinds and employed differing kinds of marble. One should distinguish between, on the one hand, spoliated works of sculpture as well as finished architectural elements re-employed in their original forms, and on the other, the structural re-use of reclaimed marble blocks either at their earlier size and shape or recut for new purposes. The re-used stones in the fabric of the Constantinian arch were structural, not decorative, and thus differ from both the re-deployed traditional decorative moldings and also from the well-known reliefs of the Trajanic, Hadrianic, and Antonine periods, which were recarved so as to transform the emperor’s heads to represent and honor Constantine[86].

58In contrast with the Trajanic panels affixed in the passageway and on the attic’s ends as well as the Antonine reliefs and Hadrianic roundels that adorn both the north and south façades, the frieze is carved from the very blocks that form the surface and architectural fabric of the arch,[87] and thus any reference to the frieze’s “slabs” is misconceived and misleading[88]. The blocks onto which the frieze was carved on the main façades form horizontal fields that are all roughly uniform in height, while those on the arch’s east and west ends are visibly taller (on which, see below). This change in the vertical dimensions of the frieze is clearly visible where the east and west friezes continue on the north and south returns as far as the pilasters of the façades[89]. Moreover, the length of these frieze blocks varies considerably, and they are laid in courses, two of which compose the major portion of the frieze and its upper moldings, while a third course below was at times used to fashion the lower moldings – under the Oratio, Liberalitas, and the left end of the Ingressus[90].

59Appropriate comparisons to this construction technique are not the other re-used sculptural panels on the arch, but the fabric of the two Anaglypha, now in the Curia Romana, which are similarly constructed of blocks of varying size and shape[91]. This manner of construction is typical of the relationship between a building’s fabric and the reliefs carved into it, as seen on similarly adorned arches and other monuments (e.g., the Arch of Titus, the Arch at Orange, the Glanum cenotaph)[92]. It is echoed in the large blocks of the pedestals into which the Victories and Captives are carved, whose technique corresponds to that used on the Arch of Septimius Severus[93], and is particularly evident in certain unfinished aspects of the Arch at Orange[94]. As a carving technique, this was neither unusual nor reserved for figural representations, and it was also employed, as is well known, for architectural forms such as the integral half columns and moldings on the Colosseum and the Theater of Marcellus, or the decorative elements of the voussoirs on the Constantinian arch itself.

60It is in this context – that is, of the arch’s marble fabric – that the “missing” feet of the Oratio (Fig. 33) and Liberalitas sections of the frieze are readily explained[95]. The bottom part of these reliefs has not been “trimmed to reduce its height, thereby cutting off the lower legs and feet of the onlookers”[96]. Nor was there a “break” running across the lower part of the scene.[97] Rather, this was the seam between the two bottom blocks of the frieze, on the lowest of which both the molding and the feet were intended to be carved – even if the latter were often left unfinished. This had been pointed out long ago by von Gerkan in the case of the Oratio:

“Stellt man sich eine sorgfältige Ausarbeitung vor, so genügen die verfügbaren 4 cm durchaus, um die Figuren in ihrer Proportion zu vervollständigen, und hier, wie im NW [viz., Liberalitas], liegt in der Tat nichts anderes vor, als eine Flüchtigkeit der Ausführung an eine Stelle, die das entschuldigt: die perspektivische Überschneidung des Gesimses verdeckt sie in der Nähe, während man aus größerer Entfernung überhaupt nichts mehr erkennen kann[98].“

61This placement of the feet on the lowest block is readily apparent in von Gerkan’s measured diagrams (Fig. 34. 35), and his conclusions were confirmed by Koeppel – well as by Pensabene and Panella – but ignored by Rose[99]. Comparison suggests that this was a favored location for making unobtrusive joins on the monuments. Something very similar is found on the Arch of Titus passageway reliefs, where the feet were also carved on the lowest blocks out of the massive supporting element beneath them[100].

62The revival of old arguments about the “missing feet” requires further discussion about the fabric of the frieze and about the arch more broadly. Firstly, the blocks from which the Constantinian frieze is carved do not form a coherently articulated unit. Unlike the Antonine reliefs on the attic or the passageway segments from the Great Trajanic Frieze, they comprise neither a single panel nor a conjoined series of upright slabs. As part of the arch’s fabric, the frieze is not only carved integrally with its moldings (see below, § 13), but it is also fashioned out of oddly shaped blocks that extend into the spandrels below and the field housing the Hadrianic roundels above[101]. Given the degree to which so many irregularly shaped blocks were refashioned for the arch’s construction and the manner in which some blocks included not only figural but also architectural and structural elements, it is highly unlikely that these blocks were disassembled from another monument and reassembled in the same order on the Constantinian arch[102]. While it is true that such a disassembly, movement, and reassembly of a major monument is known – the famous example being the Temple of Ares in Athens – it is, for technical reasons, highly unlikely to have been the case here (see below, § 13). For this would assume not only the undocumented existence of a prior arch of the same basic dimensions, but also that the Constantinian arch was built expressly to house these elements according to their highly irregular forms and dimensions.

63Secondly, both the siting of the frieze on the arch and its dimensions were, in turn, dictated by several factors which suggest that its reuse, as an integral figural component, is hardly possible. It is apparent that the arch’s overall design was based on the size of its colossal columnar order, which established the proportional relationships between the width of the façade, the depth of the arch, the height of the entablature and the podia, and the interaxial dimension of the paired columns – as well as other features[103]. These proportional calculations also established the height of the frieze panels on the façade, and this calculation of the frieze’s placement is confirmed by the geometrical scheme of the façade, based on a grid of three-by-three squares determined by the columns’ height[104].

64Despite the regularity that the arch’s modular design implies, certain, perhaps inevitable, variations were introduced in the building process; a clear tolerance for limited discrepancies prevailed[105]. Among these is the fact that the overall lengths of the various sections of the frieze vary significantly. The façade’s pilasters, responding to the columns, were clearly planned (if not carved) before the frieze, and thus dictated the length of its segments on the north and south faces, which display a close correspondence (Oratio = 541 cm; Liberalitas = 538 cm; Proelium apud pontem = 547 cm; Obsidio = 539 cm). Placement of the columns and pilasters also necessitated the inclusion of the short returns at the façade ends in order to produce the effect of a continuous circuit around the arch. The scenes on the east and west ends, however, are substantially longer than those on the façades – by more than 1 meter (Ingressus = 648 cm; Profectio = 682 cm), and this would appear to have been the result rather than the cause of the arch’s depth, which, as was observed by Wilson Jones, was determined by the module of the columns’ height.

65Thirdly, as was noted above, the overall height of the frieze sections on the ends differs noticeably from that of the sections on the two façades[106]. This might be construed as evidence that the two end sections of the frieze, the Profectio and Ingressus, were carved for another monument as a pair – for the coupling of Profectio and Adventus (or, Ingressus) had a long tradition as a subject in Roman historical relief sculpture[107]. But this is unlikely, for several reasons. It is hard to imagine that the length of two supposedly “pre-existing” friezes would have set the module for height of the columns (and so much else) on the Constantinian monument. This is clear from a comparison of the Constantinian arch with its model, that of Septimius Severus, where that exemplar was adapted to suit a new decorative scheme[108]. Even more decisive is the awkward and inconsistent treatment of the framing of the east and west end sections of the frieze. These two sections display a clearly different profile from that employed on the other parts of the frieze (Fig. 36. 37). An ornate 12 cm molding was carved on blocks inserted between the frieze and the tondi and must originally have extended the full width of the roughly square fields in which the medallions were set[109]. The huge slabs with the Constantinian tondi were then placed atop these intermediate blocks, while the dressed blocks of the attic zone to both sides sat directly on the frieze itself (Fig. 4).

66Moreover, on the east and west ends of the arch there is a small (5 cm) upper cavetto molding to the left of the tondo’s field, which is missing to the right. This was never carved: at the right-hand ends of both the Profectio and Ingressus, the masons carved the frieze up to the join with the next course of blocks, without introducing the usual set-back and the upper concave molding. This was surely a mistake, probably one of many that resulted from hurried work, but also a distinctive feature that demonstrates that the frieze was carved in situ[110]. All of the factors that have been described here demonstrate that the blocks of the Constantinian frieze were not removed from an earlier monument, but carved on the arch – as was the case with all of our surviving comparanda[111].

67The frieze displays spaces hollowed-out for the insertion of heads for the figures of the emperor in four scenes (Fig. 38. 39. 40. 45). When and why this was done has long been related to the problem of the frieze’s date, subject matter, and technique. A cogent solution to the conundrum is complicated by the fact that the material fabric of the frieze does not, in its present state, offer sufficient clues for a definitive answer.

68Various explanations have been proposed. Our early evidence for the state of the arch – Renaissance and Baroque drawings and prints as well as accounts of documented restorations – is of limited probative value, as it is demonstrable that later artists did not always record the monument faithfully, and because the record of known restorations and related interventions is patchy and often insufficiently detailed.

69The earliest recorded restoration of the arch mentions a payment on 3 November 1498 for “(re)placing certain marble heads,” although what these were is hardly clear[112]. The early drawings of the arch present equally ambiguous evidence, and some are outright fanciful[113]. Even those that appear to be scrupulous renderings, such as those attributed to Giuliano da Sangallo and dated c. 1465 (Fig. 41) or the sketch in the Codex Escurialensis of c. 1500 (Fig. 42), offer divergent impressions[114]. These early drawings offer even less evidence about the heads on the Constantinian frieze and the passageway reliefs. In most instances these are rendered, if at all, with little concern for accuracy, or at a scale too small for certainty. The sole exception is Claudio Duchetti’s 1583 engraving of the north façade for the great series of prints known as the Speculum Romanae Magnificentiae, which displays intact not only the Dacian heads, but also in summary form, those of the passageway busts and of the emperor in the Liberalitas scene of the frieze.

70It is only in the seventeenth and eighteenth centuries (see above, § 1) that arguably more scrupulous representations appear. Bartoli’s rendering of the south façade (Fig. 43) presents not only headless Dacians, but, in summary fashion, the portrait busts in the side passageways. And in his separate, detailed renderings of the frieze (Fig. 5. 6) he indicates the most heavily damaged and/or entirely missing figures, such as that of Constantine in the Proelium apud pontem section of the frieze, with dark hatchings. The question thus arises as to why his engravings depict the emperor’s head undamaged in the Ingressus, Oratio, Liberalitas, and Obsidio scenes: were these in place on the monument in the seventeenth century, or did he restore them imaginatively in his engravings? In Rossini’s 1836 publication (Fig. 7), the first three of these heads are missing and were most certainly lost, while the fourth was damaged (much as we see them today). Bartoli had asserted his fidelity to the monuments. And the publisher of the volume with his engravings of the arches, Jacopo de’ Rossi, claimed that Bartoli had compared the present state of the monument with the drawings done by Giuliano da Sangallo in the Codex Barberini (ca. 1465), “when they were not so mutilated and destroyed as they are now” (Fig. 41). The comparison does not instill confidence, and it is thus hard to know how accurate Bartoli’s renderings of the Constantinian frieze are[115].

71Given the absence of definitive visual evidence about the survival of the imperial portrait heads on the frieze, and about whether their disappearance occurred in ancient or early modern times, there are at least four different accounts that have been offered as possible explanations for why the four imperial heads are missing[116]:

72 i) The frieze was carved for a monument dedicated to an earlier emperor whose portraits it bore (Diocletian? Maxentius?)—and these were subsequently redone to represent Constantine. By this interpretation, the frieze was another of the arch’s spolia, and its refashioned portraits corresponded to the other materials reused for the arch’s decoration. However, in all other instances of spolia on the arch, the heads were recarved, not replaced, a distinction which argues against the likelihood of the frieze’s similar reuse. In any event, the variation in technique itself demands an interpretation (to which we shall return below, § 13).

73The preceding analysis of the “missing feet” has shown (§ 11, above) that the re-use of a frieze carved for another monument is an argument almost certainly to be rejected. But even if those explanations are dismissed, there is at least one other seemingly intractable – and arguably more notable – problem that remains concerning the missing heads. For if one imagines that the arch itself was built earlier than the Constantinian period, or that the frieze was a spolium that had originally depicted a previous emperor, the integrally carved portrait of Constantine on the Obsidio section (Fig. 12) would need to be explained: this part of the frieze was unambiguously Constantinian in date, and thus in significance[117].

74 ii) The frieze was carved for a Constantinian arch, but at some point still in antiquity the emperor’s portraits were deliberately destroyed as an act of iconoclasm; this is not so dissimilar to the arguments for the transformation of a Diocletianic original. The most notable example of this view is Wilpert’s vision of an anti-Constantinian “profanatio”, but this is unlikely given the lack of comparanda and similarly fails to take account of the original imperial portrait in the scene of the Obsidio (and, if our argument at § 1 is correct, the Profectio), which would seem to have survived late antiquity intact (above at § 4).

75 iii) It is not impossible that the torsos were hollowed out much later for the restoration of damaged heads. It would seem that some heads – notably those of the emperor in the Obsidio and Proelium scenes – sloughed away over the course of time due to flaws in the stone. Yet it is well-documented that certain heads on the arch were mutilated by Lorenzino de’ Medici in 1534, although contemporary testimony to this act of iconoclasm is not sufficiently detailed to reveal precisely which heads were damaged[118]. If these were the heads of Constantine in the frieze, we might imagine an unrecorded attempt to repair the damage in which the torsos on the protagonists in four of the six frieze scenes along with two of the bust-length figures in the side passageways were hollowed-out by the carving of sockets for new heads. While documentary evidence for restorations to various damaged heads on the arch survives[119], we have no record of a specific attempt to restore the heads that were vandalized by Lorenzino – whichever they may have been.

76 iv) A last possibility is that all of the missing imperial portraits were Constantinian in origin, despite the difference in their manner of carving. The imperial figure of the Obsidio scene displays an originally carved head bearing a recognizable Constantinian profile, and this, along with the seated imperial figure of the Profectio, suggests that some of the figures of the emperor were carved with integral portraits. What must be explained is a presumed change in the carving technique, from integral portraits to separately-made heads designed for insertion in the hollowed torsos of the other imperial images in the remaining four scenes (and in the passageways). The iconography of at least one of these, the Proelium, is almost certainly Constantinian (see below, § 14–16). Despite the missing body of one of the chief protagonists – surely the emperor – the roughly-worked cavity above its shoulder line confirms that it had been prepared for the insertion of a separately-made head prior to the natural fragmentation of the stone that sloughed off the rest of the body. Granting the subject of this section of the frieze, the change in carving technique from the use of integral heads (Profectio, Obsidio) must have occurred while the arch was under construction in the Constantinian period.

77A coherent hypothesis presents itself: the frieze was planned and the carving begun according to a design that called for integral portrait heads. These were completed for the Obsidio and the adjacent Profectio – making them the earliest in the series of the six sections; then the sculptors shifted to having separately-worked heads produced by specialized portraitists to be set in prepared sockets[120]. Such a hypothesis fits the evidence provided by the frieze and once again suggests its Constantinian date – but it requires further amplification.

78How many teams of Constantinian sculptors worked simultaneously on the arch? It is generally assumed that such grand projects required numerous teams of artisans, working simultaneously, and that their presence is betrayed by observable differences in style, despite the fact that the extent and significance of these distinctions have been debated. For example, numerous groups of sculptors are imagined for the carving of the Column of Trajan[121]; four have been assumed for the Great Trajanic frieze[122]; and two different teams are considered responsible for the Antonine panels on the arch[123]. The near certainty that the arch was constructed in great haste – in the two and a half years between late 312 and mid-315 – makes a large workforce all but certain, and in this regard the construction of the Constantinian arch can have been little different from that of other major Roman monuments.

79The use of two different techniques for the carving of the imperial portrait heads, integrally or worked separately, was recognized early on[124]. And the idea of distinct teams of sculptors, employing slightly different techniques, is also suggested by other aspects of the work. For the treatments of the moldings, both above and below the frieze, signal, if not different teams of sculptors, at the very least a rethinking of their design and its execution as work was completed on the other sections of the frieze. The six parts and subjects are distinguished as two groups of three contiguous sections. Profectio, Obsidio, and Proelium apud pontem all are divorced from a separate lower molding, while the Ingressus, Oratio, and Liberalitas are carved integrally with the prominent moldings beneath, whose blocks are variously extended below to form parts of the voussoirs of the two side arches, as we have already noted; on the Oratio and Liberalitas, the lower blocks of the frieze include the figures’ feet. All six sections, however, include an integral upper molding, in each case a thin cavetto, although in the Proelium apud pontem and the Obsidio (at the left and right ends, respectively) an upper block of the frieze extends beyond the cavetto into the field above[125]. And, in the Profectio and Ingressus, on the right side of each section (Fig. 2. 4), the upper cavetto was omitted, and the figures were carved to the top of their blocks at the join with those above the frieze[126]. This varied treatment of the upper border suggests different teams of sculptors working on the scaffolding at the two ends of the same frieze section. The presumed rapid construction of the arch would have necessitated these multiple teams of sculptors, and it is clear with regard to the details reported here that they worked somewhat independently.

80Can a sequence of the work be envisioned? We might imagine that work on the frieze began on the south façade, at the west end (Obsidio), and also on the Profectio adjacent to it on the west. These are the two sections that originally had integral emperor’s heads, and it is conceivable that they were sculpted before the allocation of work was changed in favor of separately worked portraits to be inserted in the now-requisite torso sockets. One consequence of this proposed sequence of work on the frieze is that the detachment and disappearance of the originally inset heads – whenever that occurred – was all the more readily effected.

81Apart from the Obsidio with its surviving portrait, the most difficult portion of the frieze to re-attribute to Diocletian rather than Constantine is that depicting the river battle over the arch’s southeastern bay (Fig. 44). The scene is focused entirely on a military conflict that occurred on a bridge over a river at the edges of a walled city. Given that the arch’s inscription (dedicated in 315) makes it clear that the entire monument was designed to glorify Constantine’s defeat of the “tyrant” Maxentius and his “faction,” and that Constantine widely celebrated his victory at the Milvian Bridge (in 312) as the singular most important event in the series of battles that led to Maxentius’ demise, the scene has traditionally been interpreted as a depiction of this battle – even by those who believe that many of the other sections of the frieze are spolia from monuments erected for other emperors[127].

82Yet Rose has argued that the relief was originally intended to depict Diocletian’s defeat of his rival Carinus at the “Battle of the River Margus” in 285[128]. The problem is that our sources indicate that there were multiple conflicts between Diocletian and Carinus, that among these was a major battle near a place called Margus (or Margum) which may have been won by Carinus, and – most importantly for our purpose – that a river or a bridge are not known to have played any role in these conflicts.

83The sources can be broken into five groups: those which follow Enmann’s Kaisergeschichte, the Chronicon Calendar of 354, the Itinerarium Burdigalense, a Greek tradition reflected in Zosimus, and the Eusebian chronicle tradition.

84The Kaisergeschichte, though written in the mid-fourth century, is now lost, but its contents are widely agreed to constitute the common font for several fourth-century epitomes of imperial history[129]. This is the only branch of the tradition to use the name Margus/um, making it worthwhile to quote their brief reports in full:

Aurelius Victor, liber de caesaribus, 39, 11 (ed. Pichlmayr): “But when Carinus reached Moesia, he immediately confronted Diocletian Marcum [sic] iuxta, and when he bore down hard against the defeated forces, he perished by a blow from his own men, for being out of control with desire, he had lustfully pursued many of his soldiers’ women, whose angry husbands displaced their wrath and upset onto the outcome of the battle.”

Eutropius, 9, 20, 2 (ed. Santini): “Afterwards, he defeated Carinus, who was living with the hatred and detestation of all, in a huge battle apud Margum, and Carinus was betrayed by his own army, although he had the more powerful force, and was actually deserted by it, between Viminacium and Aureus Mons (inter Viminacium atque Aureum montem).”

Historia Augusta, Carus, Carinus et Numerianus, 18, 1–2 (ed. Hohl): “He [Carinus] learned that his father had been killed by lightning, his brother by a comrade, and that Diocletian had been proclaimed Augustus, and he then committed even greater vices and disgraces, as if he had been set free from the reins of his homely piety by the death of his relatives. But he was nonetheless missing the strength of mind to reclaim the empire for himself. For he engaged in conflict against Diocletian in many battles but died after being defeated in a final battle fought apud Margum”[130].

85We have left Margum in all three sources untranslated because, grammatically, it could refer to one of two toponyms: the River Margus (Great Morava) or the similarly named town of Margum (Požarevac, near Dubravica, Serbia)[131]. The prepositions apud and iuxta that accompany it mean simply “near” in conjunction with both rivers and towns in Latin, such that, in this instance, we are frankly unable to say grammatically or lexically which of the two is intended[132]. Neither preposition implies, however, that the battle was fought hard on the river or the town, for either could imply a very broad geographical range from feet to miles[133]. Thus, these sources leave it unclear whether they are referring to the town or the river, and none says anything about a river playing a significant role in the battle.

86More importantly, a second source tradition confirms that the battle apud Margum was not fought on the river but on a neighboring plain: the Chronicon Calendar of 354, not related to the Kaisergeschichte, reports that Carinus “was killed on the Margian field”[134]. This does not contradict the use of the preposition apud in the Kaisergeschichte and speaks against the river having been a factor.

87A third source tradition, independent of the previous two, is represented by the Itinerarium Burdigalense, a travellogue by an early fourth-century Holy Land pilgrim who visited the region on his way east. This offers a list of the towns through which its author passed including: “The city (civitas) of Aureus Mons, VI miles (i.e. from the previous station); the changing station (mutatio) of Vingeius, VI miles; the city of Margum, VIIII miles; the city of Viminacium, X miles, where Diocletian killed Carinus”[135]. The information is compressed but would seem to indicate that, in the early fourth century, local knowledge presumed that Carinus had been killed ten miles east of the city of Margum near the legionary fortress of Viminacium, and thus far removed from the River Margus.

88The question thus arises, how are we to connect the battle apud Margum with the demise of Carinus? A careful reading of Aurelius Victor, our earliest source from the Kaisergeschichte tradition, suggests that Carinus actually won the initial field battle at Margus/um but was then betrayed and murdered by his own men: dum victos avide premeret, suorum ictu interiit [136]. And Eutropius implies awareness of something related when he reports that Carinus’ execution occurred not apud Margum but between Viminacium and Aureus Mons, two towns which, like Margum, sit along the military road that runs beside the middle Danube in Moesia. The Kaisergeschichte tradition thus indicates that Carinus met his end following a victory at the big battle apud Margum at a locale further downriver, much as the Itinerarium Burdigalense reports.

89It seems, then, that Carinus’ forces may have defeated Diocletian at Margus/um but then assassinated their leader as they pursued Diocletian on his retreat eastward, probably because they had been suborned. This may lie behind our fourth source tradition, the Greek reports in the early sixth-century historians Zosimus and John of Antioch, who also relate that Carinus was killed by one of his own soldiers whose wife Carinus had violated[137].

90A fifth and final tradition, originally transmitted in the Greek Chronicon of Eusebius, reports instead that Carinus was killed at Cornacum (modern Sotin, Croatia), a town some 164 miles west of Margum. Although Péter Kovács has argued in favor of this location for Carinus’ death, it appears only in the later Armenian and Syriac translations of the Chronicon and may represent yet another in the complicated series of conflicts that eventuated in Carinus’s assassination[138].

91Our sources for the ‘Battle apud Marguum’ are thus anything but uniform. Close examination points to at least four major problems with any effort to associate the arch’s Proelium apud pontem with the River Margus. First and foremost, no ancient source speaks of this conflict as a river battle nor definitively identifies the River Margus. Instead, some sources report only that a battle occurred “near” either the river or city that share the name Margus/um, and one states explicitly that the battle occurred on a plain called the campus Margensis. Second, our earliest witness in the Kaisergeschichte tradition, Aurelius Victor, indicates that Carinus in fact won this battle, and the same may be implied by the Greek tradition transmitted through Zosimus. If the battle apud Margum represented a defeat for Diocletian, it was hardly an event to be publicly celebrated in monumental reliefs. Third, the sources indicate that Carinus was killed not near the River Margus nor the city of Margum, but 10 miles eastward near Viminacium, or alternatively 164 miles to the west near Cornacum, and that he died at the hands of his own troops. A river thus played no role in Carinus’ demise, which the sources present as a sort of anti-climax to the main battle apud Margum. Finally, however one parses the reports of our ancient authorities, their prevalent confusion demonstrates that this was not an event celebrated widely in imperial propaganda – and thus no canonical version of the events ever emerged.

92Indeed, we have no source – textual, numismatic, or iconographic – that indicates Diocletian ever drew attention to this conflict. Nor is this silence attributable to a lacuna in our source material for Tetrarchic propaganda, since we have three extant Latin panegyrics dedicated to fellow tetrarchs in which we might expect the glorification of Diocletian’s victory had this been the order of the day[139]. In all of this material, we find only one allusion to the circumstances of Carinus’s demise, a brief mention of Diocletian’s elimination of “savage despotism” from the world early in his reign at Panegyrici Latini 11(3)[140]. This would appear to pick up a strain of tetrarchic court propaganda that emphasized Carinus’ cruelty and lack of restraint, qualities that also figure in many of the ancient narratives reported above[141]. Reconfirming this, the sixth-century historian Petrus Patricius, who often used contemporary sources from the third and fourth centuries, reports that Diocletian actually “said that he did not eliminate Carinus because he was desirous of empire but because he felt pity for the republic”[142]. By combining this with the allusion to “savage despotism” in the panegyric and the reports of perversity and cruelty in the Greek and Latin historiographic traditions, we get the sense that the only thing Diocletian and his fellow tetrarchs wanted known of Carinus’ elimination was that it freed the world from a depraved despot.

93The sparse and ambiguous account of the battle apud Margum stands in marked contrast with Constantine’s Battle of the Milvian Bridge. It would be superfluous here to reproduce the evidence for Constantine’s broader campaign of glorification of his civil war victory over Maxentius. But a narrower look at the textual evidence that was produced by Constantine’s chancery reveals the remarkable degree to which these sources correspond with the scene presented on the frieze (Fig. 44).

94Ancient accounts of the battle begin to appear immediately, and it is striking how abundant they are. These include the panegyric delivered to the emperor at Trier in 313 (Pan. Lat. 12[9]); the text titled On the Deaths of the Persecutors, completed by Lactantius, the court scholar and teacher of Constantine’s sons, in 315; the second edition of Eusebius’ Ecclesiastical History, published c. 315/316; and the Panegyric of Nazarius (Pan. Lat. 4[10]), delivered in Rome for the decennalian celebration of Constantine’s sons in 321. All four, written very much within Constantine’s inner circle, tell a remarkably consistent story that leaves no room for doubt that both the River Tiber and the Milvian Bridge itself were believed to have played a crucial role in the battle. The earliest among them is also the most detailed and is worth quoting at length:

“Then at the first sight of your majesty and at the first attack of your army so often victorious, the enemy was terrified, routed, hindered by the narrowness of the Milvian Bridge, and with the exception of the first instigators of that usurpation who in despair of pardon covered with their bodies the place which they had chosen for combat, all the rest went headlong into the river, so that there was at last some abridgement of the slaughter for the weary right hands of your men. After the Tiber had swallowed the impious, the same Tiber also snatched up their leader himself in its whirlpool and devoured him, when he attempted in vain to escape with his horse and distinctive armor (cum equo et armis insignibus) by ascending the opposite bank… The swirling river rolled along the bodies and arms of other enemies and carried them away; that one, however, it held in the same place where it had killed him, lest the Roman people should long be in doubt whether it was to be believed that the man, the confirmation of whose death was sought, had actually escaped. Sacred Tiber (Sancte Thybri), once adviser to the sojourner Aeneas, next savior of the exposed Romulus, you allowed neither the false Romulus to live for long[143] nor the City’s murderer to swim away. You who nourished Rome by conveying provisions, you who promoted her by encircling the walls, rightly wished to partake of Constantine’s victory, to have him drive the enemy to you, and you to slay him (merito Constantini victoriae particeps voluisti, ut hostem in te propelleret, tu necares)[144].”

95The apostrophe invoking the River Tiber then continues, making it abundantly clear that the orator wished to emphasize that it was Rome’s river and its tutelary deity, Sanctus Tiberis, who had worked in tandem with Constantine to defeat and destroy Maxentius.

96The oration’s description, so consistent with the scene on the frieze, was written and disseminated even as the arch was being erected. It is conceivable that the arch’s designers were aware of this very testimony – certainly its senatorial patrons would have been. And those senators would have undoubtedly had knowledge of the imperial communiques that informed both the panegyric’s content and the arch’s imagery[145].

97Writing just two years later and thus coincident with the dedication of the arch, Lactantius tells much the same story, emphasizing the fighting at the Milvian Bridge, the terrified flight of Maxentius and his army, and his death in the River Tiber, further adding the detail that the bridge had been cut by Maxentius’ own forces[146]. So too Eusebius, writing perhaps a year later from the other side of the empire in Palestine, had already received news of the battle at the Milvian Bridge, which he names along with the River Tiber as crucial protagonists in the battle’s outcome[147]. Eusebius also mentions the breaking of the bridge and even indicates that a bridge of boats had been built alongside it by Maxentius but had given way under the rush of his fleeing cavalry[148]. Furthermore, Eusebius elaborates a motif drawn from the Panegyric of 313 by emphasizing the heavy armor of Maxentius’ soldiers, which proved especially disadvantageous when they entered the river and “sank like a stone”[149]. Finally, Nazarius in his Panegyric from nine years after the battle also describes the conflict as a river battle and emphasizes the role played by the Tiber, repeating the rhetoric of the 313 panegyrist in describing the river’s waters churning with the bodies of dead soldiers. His account is briefer than that of 313, but this is only because – as he reports – he had delivered an entire oration on the subject of the Battle of the Milvian Bridge on the previous day[150]. The battle is also described in numerous other sources extant from the period immediately following Constantine’s death, especially Eusebius’ Life of Constantine and the anonymous Origo Constantini [151]. The river battle at the Milvian Bridge was thus the central military accomplishment being trumpeted by the Constantinian court in the decade between 312 and 321. Indeed, it could rightly be claimed that Constantine treated it as his most important military achievement until the day he died.

98These official pronouncements coincide precisely with the images we find on this frieze. This is true on at least five counts.

99 i) On the arch, the river is an active agent in the destruction of the enemy forces, as multiple enemy fighters are shown falling into or drowning in the water. This is exactly as the scene is portrayed in the Panegyric of 313, in Eusebius’ Church History, and in the Panegyric of Nazarius.

100 ii) The winning and losing forces are clearly distinguished by the amount of heavy armor each wears, with the defeated soldiers uniformly clad in weighty scale-mail, while their opponents, both infantry and cavalry, wear only helmets and loose-fitting tunics. Here too, we find echoes of a theme emphasized in the Panegyric of 313 and Eusebius.

101 iii) The frieze depicts not just a river but very clearly features a bridge under whose arches the river roars in full spate. It was thus important to the arch’s designers to depict not just a river battle but a bridge battle – a specific reference found in most contemporary as well as later textual and epigraphic accounts.

102 iv) Beneath the feet of the striding goddess on the bridge and the (now missing) emperor, the figure of a river god reclines on his right elbow (Fig. 44. 45). Although the relief is now badly eroded, enough of its lineaments are preserved to show the naked chest, flowing hair, bushy beard, and upturned amphora – all attributes of the god Tiberis[152]. This is precisely what Bartoli saw when he identified the deity as “Thyberis” in his engraving in 1690 (Fig. 5).

103 v) Alongside the bridge, we can also still make out the prow of a boat curving upward into the arc of the right leg of the goddess Victoria. The bridge of boats effected by Maxentius, “an engine of destruction for himself,” was a key theme emphasized in the account of Eusebius and in other sources[153].

104For all these reasons, it is clear that the River Battle is set at the city of Rome on the Milvian Bridge[154].

105The three imposing figures at the left end of the river battle scene are all damaged (Fig. 44. 45). The one in the middle is almost completely destroyed, but given its size, imposing stride, and the still evident paludamentum fluttering below the knees, it must be the emperor; at the right is Victoria, whose wings are unmistakable. Yet the third figure, at the left, poses a problem much debated in studies of Roman iconography – one that, in this context, requires elaboration.

106Is this the apparition of the Dea Roma or the personification of Virtus? Both were regularly portrayed as Amazonian females in a high-crested helmet, buskins, and a short chiton that revealed the right shoulder and breast, with the lower legs exposed. Examples are known from the republican period through the late Roman era, and their iconographies are virtually indistinguishable[155]. Either identification would be appropriate at a scene of military significance, and such figures have been variously construed on the coinage and on various Roman state monuments (e.g., the Cancelleria relief A, the Great Trajanic frieze panel in the passageway of the Constantinian arch, the Antonine panels on its attic, or the Jupiter Column in Mainz). Art historians have regularly been at odds over which of the two might figure in any given instance[156].

107But who appears in the bridge scene? It has been asked, if it is the Dea Roma with Constantine, why does she appear “on the wrong side of the river?”[157] This assumes that the goddess Roma appears primarily as a topographical symbol, as does the figure of the River Tiber below her.

108But this need not be the case, and two arguments favor the interpretation of Dea Roma. First, topography. The apparition of the Goddess Roma sometimes occurs on monuments and the coinage with no explicit topographical significance – as when the goddess offers the globe to the emperor (Fig. 46. 47; cf. Fig. 48)[158]. By contrast, she might also serve as a divine presence and a topographical symbol at one and the same time: so she appears elsewhere on the arch, with her familiar costume and attributes, on the Trajanic adventus relief in the main passageway (Fig. 49), where it is she, not Virtus, who welcomes the triumphant emperor to the city as he is crowned by Victoria. Indeed, there the goddess’ topographical role is reinforced by the Constantinian inscription facing the relief: liberatori vrbis. Roma plays the same role on the arch’s attic, in the Antonine Adventus panel (Fig. 50), where she greets the emperor on his arrival to her city[159]. And it is not without significance that in these two other instances on the arch, the Dea Roma appears outside the city’s gates. As the coinage makes plain, Roma might well appear with the emperor anywhere – as she does when accompanying Constantine’s entry to Milan in February of 313 (Fig. 51) or when she presents the Dea Constantinopolis to the emperor in 335 (Fig. 52)[160].

109In the case of the river battle on the frieze, the broad significance of Roma’s conventional usage allowed the arch’s designers to allude to the paradox of how it was that Maxentius – who had so openly and vigorously proclaimed himself to be Rome’s champion (conservator vrbis svae) and the beneficiary of her favor (avctrici avg n: Fig. 47) – found himself deserted by the Dea Roma in his final battle before her gates[161]. Rome’s senate, when it dedicated the Arch, seems to have intended a claim that the Dea Roma had abandoned Maxentius and left her own walls to battle alongside the rightful emperor. This is precisely what the panegyrist of 313 says to Constantine when describing Maxentius’ hesitation to abandon the safety of Rome’s walls:

“Meanwhile even in public he [Maxentius] kept wishing that an advance would be made all the way to the gates. He did not realize that the City’s very Majesty (maiestatem illam Urbis), which had once tempted advancing armies, had gone over to support you now that it was disgraced by his crimes and driven from its seat...”[162].

110The artists of the Proelium apud pontem scene visualized the same legend, and their imagery drove home the point that Constantine had not just Victoria but also the Dea Roma at his side when he defeated Maxentius outside the walls of Rome. Underscoring Roma’s preference for Constantine, the Trier mint issued a coin (Fig. 46) pointedly alluding to the coin of Maxentius (Fig. 47), even as it mocked him by styling Constantine the restitvtor libertatis, a title also asserted in the arch’s central bay[163].

111Numismatic comparanda provide a second argument for an identification of Dea Roma rather than Virtus. While Virtus as an abstract quality represented by its female personification was common on coins in the second century, such depictions had tapered off by the late third, and the personification appears to be entirely absent from tetrarchic coinage in the period 284–305[164]. Nevertheless, the idea of “Virtus” as an imperial ideal continues to appear on tetrarchic coins with great regularity in legends such as virtvs avgg, virtvs militvm, and virtvs exercitvs. On these coins, however, Virtus no longer appears as a traditional female personification, but now as an exemplification of “manly power,” depicting the individual emperors themselves, or the gods of the tetrarchy (Jupiter, Hercules, Mars [Fig. 53]), soldiers (Fig. 54), or all four tetrarchs sacrificing before the gates of a military camp[165]. There is thus nothing in our surviving iconography to support the identification of Virtus on a frieze presumed to be Diocletianic in date.

112Nevertheless, in 315 the figure on the bridge might have been construed as Virtus since Constantine had reintroduced the traditionally female personification of an Amazonian Virtus on numismatic reverses in the aftermath of Diocletian’s abdication (Fig. 55)[166]. But as the sources on the Battle of the Milvian bridge that we have examined demonstrate, the senators and the moneyers of the Constantinian era recognized – and this we take to be decisive – that the Dea Roma had deserted Maxentius in her city and gone over to the side of Constantine to lead him to victory.

113The tradition of reattributing the frieze on the Arch of Constantine to earlier emperors goes back to the early twentieth century, as indeed does the reattribution of the monument itself. When L’Orange and von Gerkan published their magisterial study of the arch’s fabric and its imagery, it was in part to reinforce the dating of the decorative program’s varied elements, including the assignment of the frieze and other obviously late antique components, to Constantine’s reign. They did so not only by employing technical and iconographic arguments, but also by demonstrating that the frieze’s six reliefs work in unison to relate a continuous narrative that must have been conceived as a coherent whole[167].

114The frieze reliefs were not a collage of spolia with Constantinian additions, but rather a carefully designed Constantinian sequence. The six scenes display a historical progression of Constantine’s march on Rome and successful routing of the tyrant Maxentius. They thus offer a distinctive complement to the narrative of the Trajanic battle scenes, and a contrast to the Antonine panels’ representations of a host of traditional imperial virtues depicted according to conventionalized abstract formulae that suggest the fruits of victory. The two representational modes – historical narrative and its transformation into time-honored symbolizations – were skillfully united by the Constantinian designers and successfully convey the era’s characteristic conjunction of virtus exercised by the emperor when Rome was at war and when Rome was at peace. And the monument’s entire repertory of images was, in turn, presented as if eternal – framed as it is on the arch by the cosmic cycle of the setting of the moon on the western side and the rising of the sun on the east. All of the themes evoked by the arch’s decorative ensemble, whether new sculpture or old, found their counterparts in the broad array of Constantinian imagery, monumental, textual, and numismatic.

115Thus we close by reviewing the conclusions reached in the foregoing study of the frieze. We have argued that the evidence of the arch points again and again to a Constantinian origin for all six of its sections:

116 i) The sequence begins with the Profectio scene, which must feature an emperor seated on the wagon at the far left (just as on the arch’s opposite end in the Ingressus), a realization confirmed by the early modern engravings showing this figure with raised arm and globe. This must be Constantine, given that his army marches under the standards of Sol Invictus and Victoria, a pairing omnipresent on the arch and explicitly attested as fundamental to Constantine’s public presentation but absent from Diocletian’s.

117 ii) The Obsidio retains enough of the emperor’s profile to leave no question but that Constantine was its original subject, an identification corroborated by the presence of his distinctive war horse and of troops wearing the helmet of the Cornuti.

118 iii) The Proelium itself is highly unlikely to have depicted Diocletian’s battle apud Margum. The circumstances, outcome, and import of this battle are never directly attested, and it receives no attention in the propagandistic program of Diocletian. It must instead represent the Battle of the Milvian Bridge, for it depicts the River Tiber, the bridge itself, and even the temporary boat-bridge alongside it, all of which together with Roma and Victory, played a key role in the combat, much as they do in the contemporary rhetorical and historiographic literature.

119 iv) None of the scenes displays more than one emperor; certain subjects, under the Tetrarchy, would have demanded otherwise. In the Liberalitas, a lone emperor appears in a scene that ruling colleagues customarily celebrated (and portrayed) as a joint benefaction – as did Diocletian and Maximian in Rome in 303. The Ingressus, with the parade entering Rome at the Porta Triumphalis, features a sole figure, much as Constantine’s ingressus in 312 is described in contemporary panegyric.

120These narrative and iconographic arguments are reinforced by technical observations which obviate the recently made case for the spoliation of the frieze from an earlier monument.

121 v) The frieze was fashioned not on slabs but from multiple irregularly shaped blocks which had been built into the fabric of the arch before sculptural carving was undertaken. What appear to some as “missing” feet from the Oratio and Liberalitas reliefs are best understood as signs of hasty and incomplete carving on a monument erected within a two-and-a-half year timeframe.

122 vi) And, finally, the “missing” heads suggest a change in carving technique from integral heads to separately carved portraits as the artists of the arch progressed with their work. While this solution to the problem cannot be proven, an acknowledgment that the friezes were carved in situ eliminates the theory that the heads needed replacing because they had originally derived from a different monument and depicted another emperor, whether Diocletian or Maxentius.

123At nearly every turn the evidence points to Constantine as the dedicatee, the Roman Senate as the patron, and that era’s artists as the creators of the arch’s frieze. To be sure, it seemed a great boon to some to have discovered a new set of sculptures from the first Tetrarchy hiding in plain sight in the imperial capital for 1700 years[168]. But too much is lost in the effort to throw off the immense weight of the evidence to make it worth constructing a Diocletianic theory on such shaky foundations. The dating of the frieze matters greatly – for the history of the Tetrarchy, for our assessment of Constantinian art, and for our understanding of the development of public architecture and the use of spolia in late antiquity. An earlier dating for the frieze would throw into question much of what we claim to know about these issues – and more. Revisionism is always welcome when it helps us unlearn what was wrong to begin with, but the weakness of the arguments for a redating of the frieze, and the continued strength of the case for Constantine, make the resurrection of previously rejected arguments for an alternative chronology an impediment to scholarly progress.

124The authors would like to thank Bert Smith, who made a number of very helpful suggestions on a late draft. All translations are by the authors unless otherwise specified.

This article reassesses recent efforts to redate the frieze on the Arch of Constantine to the end of the reign of Diocletian. It offers a series of arguments to demonstrate that these reliefs are, as has been long thought, Constantinian. These focus on iconography (the use of known Constantinian imagery), narrative (allusions to events in Constantine’s military and political career known from written sources), historiography (the interpretation of the source tradition, especially for the arch’s “River Battle” scene), and technical matters (the fact that the frieze reliefs were carved in situ and the likelihood that the arch’s sculptors fashioned inset portrait heads of the emperor). Numismatic material and early modern prints provide new iconographic parallels that reinforce a conclusion that the frieze cycle was produced in the Constantinian era.