Reverberations of Persepolis: Persianist Readings of Late Roman Wall Decoration

: Animal combats (venationes) were a popular entertainment in the Roman world. Splashy panels of inlaid marble (opus sectile) commemorate these bloody contests in several buildings in and around Rome. Among the most well-known are survivals from the 4th century CE Basilica of Junius Bassus and, several decades later, the marble-revetted hall from Porta Marina at Ostia. On the face of it, the wall decoration from these sites memorializes typical Roman activities, but the panels expose the vast geography implicated in these combat spectacles. The brilliant stones used to render them came from lands as far off as the Caspian tigers and Asiatic lions they depicted. The iconography of the panels was also foreign: the animal combat, or symplegma (intertwining), is seen on works from pre-Achaemenid sculpture to Sasanian textiles, and most recognizably, at the Achaemenid palace at Persepolis, where a lion attacks a bull in relief on the Apadana stairway. Reading these panels through a Persianist lens illuminates the ways in which the Persepolitan model animated Roman themes and visual programs. Though they recalled events in the Roman arena, they also imparted political and astrological signiﬁcation to the decoration by means of their Persian associations. By alluding to the Achaemenid empire, a great power of the past and a continuing rival in the form of the Sasanians, the Roman patron accrued to himself some measure of the veneration for this culture and showed himself able to communicate in an idiom legible to an international clientele.


Introduction
It has on its walls, and in the upper parts, some profane figures of oriental marbles, a lion that tears a deer to pieces, a leopard that kills a cow, and similar savage creatures, and so far as you can glimpse from some of the arches and windows and from many fragments in the wall, this building was all illustrated in oriental marble. 1 -From Benedetto Mellini's 17th century Descrittione di Roma.
Animal combat scenes such as the ones described in the Renaissance era testimony above were not a new invention in Late Antiquity (here, c. 300-700 CE), nor were such pairings unique to Rome. Similar combat scenes appear across the empire in numerous paintings and mosaics. Greek antecedents go back a millennium, and find far earlier precedent in Egyptian and Mesopotamian examples. Of all these earlier models, among the most salient are the bull and lion reliefs from the Apadana at Persepolis (5th century BCE). Whether by direct transmission or through some portable intermediary in silver or a less durable material, the composition leapt from stone relief to opus sectile, traversing centuries, seas, and half a continent to be reduplicated on 4th century CE Roman walls, where, in the marble revetments of civic and domestic buildings alike, tigers pounced on bulls or deer, and lions tackled wild asses. Ostensibly these scenes depicted arena contests, though their backdrops sometimes recalled the natural environments where predator and prey might have skirmished in the wild.
Though the animal pairings vary in the Roman examples, the composition is fairly consistent, comprising a "common denominator" that suggests to us the kind of stripped-down Arts 2023, 12, 102 2 of 33 schema that might have appeared in a pattern book. In this article, I raise the possibility that this common denominator might be traced not merely to the cloudiness of "Eastern origin", but might more precisely and productively be ascribed to the Achaemenid iconography of lion and bull, and even to the palatial structure at Persepolis that monumentalized this iconography. I examine the evidence for this motif's transmission and its possible mechanisms, as well as the valences of the composition for different viewing audiences. The panels take on new meanings when read against Persian antecedents, affording more layered interpretations. Yet no analysis of them or the buildings they bedecked has taken into account the parentage of the composition, or the ways its provenance would have shaped viewing, whether for Roman viewers or Persian viewers.
I follow Matthew Canepa in his Two Eyes of the Earth reading, understanding Rome and Persia in this period as opposing forces who fought bitterly but also held one another in esteem and accepted the inevitability of the other's presence and power in the region. 2 The need to compete in their coexistence gave rise to an ongoing negotiation of identity, which played out in visual practice and ritual performance. Each calculated their visual messages to make use of the vocabulary employed by the other, not because they lacked their own forms, but to ensure the claims to power they issued were commensurate with their competitor's and legible to them. I see Canepa's theory as very much in keeping with Richard White's Middle Ground theory (1991/2011), insofar as it describes a process of ongoing negotiation, invention, and interpretation-though I cannot say whether Canepa would find the comparison apt. For an application of White's theory to ancient material, I am indebted to the work of S. Rebecca Martin. 3 Martin re-examines some of the most well-known artifacts of the Hellenistic period for insight into the question of Phoenician identity. In particular, Martin's interpretation of the so-called "Slipper Slapper" statue from the Hellenistic period offers a compelling example of the negotiation of meanings between Phoenician and Greek artists, artworks, and audiences. She describes the Slipper Slapper as "a translation in multiple senses", partly because it quotes the Aphrodite of Knidos, with a twist (Martin 2017, p. 157). Martin regards such "changes to well-established types" as artistic ostentation, intended to impress and amuse both Greek and Phoenician audiences familiar with Greek prototypes in the manner of a successful bilingual pun (Martin 2017, p. 158). The Slipper Slapper incorporates a known and revered model, but subverts the viewer's expectation of it by complicating the scene and building out the grouping to capture new meanings that held special significance to viewers on Delos.
I believe it may be similarly fruitful to think of the Roman sectile panels as translations of a sort, legible to Roman audiences with no knowledge of the Apadana at Persepolis, but pleasing on multiple levels to those who understood its referent. Such an understanding would hinge on a Romano-Persian visuality, a way of seeing shaped by both Roman and Persian visual environments. Monuments such as the palace at Bishapur, with its Helleno-Sasanian mosaics, and relief sculptures such as Naqs-e-Rostam would have offered a similarly layered interpretation to viewers capable of code switching. Recognizing the affinities and slippages between Persian and Roman readings of the marble panels offers a better hope of seeing late Roman monuments in the way that their viewers and visitors saw them. This attempt at capturing a Romano-Persian visuality also offers a partial corrective to traditional narratives of Hellenization and Romanization, which characterize western influence as a centralizing force in the eastern Mediterranean.
A third framework worth bringing to bear in our examination of late Roman wall decorations here is Persianism. Persianism refers to "the ideas and associations revolving around Persia and appropriated in specific contexts for specific (socio-cultural or political) reasons". 4 Persianism does not entail imitating or acculturating, nor does it claim any universal or absolute about Persian identity. Rather, it is Persia as received, an idea of Persianness as distilled into transmissible or imitable bits and conveyed to others. Persianism is a perception formed by outsiders, and does not purport to be reflective of self-identity within the group. There is no single Persianism, because perceptions can and do change            as well, but no drawings of the building allow us to determine whether they were arrayed in the same facing arrangement as at the Ostian hall, nor is it clear how such a configuration would have been fitted around the high windows of the basilica. 6 The wall decoration from the hall at Sette Sale has not been reconstructed, so it is not clear how many animal combat panels there were, but fragments that have been pulled from the site match the shapes of individual phalanges and claws on the tiger paws seen at Ostia and the Basilica of Junius Bassus (see Figure 7). 7 Figure 7. Fragments from the parietal opus sectile at Sette Sale, 4th century CE. Image reproduced from Bianchi et al. (2000), Figure 9.
In all five surviving Roman examples, the feline's face is seen essentially head-on, so that the eye of the attacking animals confronts the eyes of the viewer, as at Persepolis. The Ostian panels all depict their prey with head turned back toward the attacker. Our instinct is to read this as looking at the aggressor, but because their eyes are in the sides of their heads, they fix their gaze on us as the viewer. The effect is one of knowingly being the object of spectatorship.

Roman Readings
Three themes are salient in the animal combat panels. First is the theme of violence as entertainment and enactment of the natural order. Second, the panels' materials and subjects map the dominion of Roman authority. The exotic animals depicted reflect the extent of the empire's reach, which enabled access to environmental resources and supported the infrastructure necessary to capture and transport them. The marbles out of which these images were formed bespeak a similar imperial mastery. At the same time, these panels demonstrate wealth and influence on the part of an individual patron, which enabled him to call on these resources for entertainment.
Despite the striking appearance of these sectile panels, few researchers have proffered an interpretation of them, or of their place in the wider decorative program of the hall (for the sites where enough of that program survives). Kiilerich's (2012) work on the visual effects of the marble walls at Ostia and Rome treats the panels' colorific effects brilliantly, but does not address figural content. 8 Several authors have discussed how animal combats in the arena functioned symbolically, whether to their live audiences, 9 as depicted in domestic decoration, 10 or on consular diptychs. 11 Some of these have suggested that the venatio was a kind of ritual offering, meant to demonstrate the natural balance of power in the world, strong against weak, and to reinforce those dynamics in the social and political world (Futrell 1997, pp. 48-51). Decorative depictions further underlined and affirmed these power structures (Brown 1992, p. 184). Such subjects were compelling because they were thrilling but not actually threatening to the onlookers (Newby 2012, p. 351). Maguire (2000) argues that later viewers would cease to see them as illustrations of literal violence and instead view them as talismanic.
The most comprehensive treatment of the panels, however, dates to the mid-20th century excavation of the Ostian hall by Giovanni Becatti. Becatti undertook the study of the Basilica of Junius Bassus and its surviving program as an adjunct to his work on the opus sectile hall at Porta Marina. 12 Becatti argues that the animal combat panels should be In all five surviving Roman examples, the feline's face is seen essentially head-on, so that the eye of the attacking animals confronts the eyes of the viewer, as at Persepolis. The Ostian panels all depict their prey with head turned back toward the attacker. Our instinct is to read this as looking at the aggressor, but because their eyes are in the sides of their heads, they fix their gaze on us as the viewer. The effect is one of knowingly being the object of spectatorship.

Roman Readings
Three themes are salient in the animal combat panels. First is the theme of violence as entertainment and enactment of the natural order. Second, the panels' materials and subjects map the dominion of Roman authority. The exotic animals depicted reflect the extent of the empire's reach, which enabled access to environmental resources and supported the infrastructure necessary to capture and transport them. The marbles out of which these images were formed bespeak a similar imperial mastery. At the same time, these panels demonstrate wealth and influence on the part of an individual patron, which enabled him to call on these resources for entertainment.
Despite the striking appearance of these sectile panels, few researchers have proffered an interpretation of them, or of their place in the wider decorative program of the hall (for the sites where enough of that program survives). Kiilerich's (2012) work on the visual effects of the marble walls at Ostia and Rome treats the panels' colorific effects brilliantly, but does not address figural content. 8 Several authors have discussed how animal combats in the arena functioned symbolically, whether to their live audiences, 9 as depicted in domestic decoration, 10 or on consular diptychs. 11 Some of these have suggested that the venatio was a kind of ritual offering, meant to demonstrate the natural balance of power in the world, strong against weak, and to reinforce those dynamics in the social and political world (Futrell 1997, pp. 48-51). Decorative depictions further underlined and affirmed these power structures (Brown 1992, p. 184). Such subjects were compelling because they were thrilling but not actually threatening to the onlookers (Newby 2012, p. 351). Maguire (2000) argues that later viewers would cease to see them as illustrations of literal violence and instead view them as talismanic.
The most comprehensive treatment of the panels, however, dates to the mid-20th century excavation of the Ostian hall by Giovanni Becatti. Becatti undertook the study of the Basilica of Junius Bassus and its surviving program as an adjunct to his work on the opus sectile hall at Porta Marina. 12 Becatti argues that the animal combat panels should be understood as neo-Platonic allegories for the victory of the soul over death. In the context of this program, Becatti believed the herbivores depicted as the prey in these panels, a calf and baby deer (?) were symbols of helpless victims succumbing to the cruel exigencies of fate. They serve as allegories for the mortal body, which can expect no end other than destruction (Becatti 1967, p. 188). As the meeker animal falls prey to the feline, so does the body die, transcended by the soul (Becatti 1967, p. 188). Becatti understood the Basilica of Junius Bassus, the Ostian hall, and even the glass sectile hall at Kenchreai to be scholae intended for the gatherings of neo-Platonists whose esoteric discussions would be facilitated by their imagery. 13 Arts 2023, 12, 102 7 of 33 I am skeptical that all three Roman halls discussed here and potentially two other contemporary buildings were dedicated to a single current of fourth century pagan mysticism. 14 Such a proliferation would be surprising, especially given that these few survivals suggest an even greater popularity of the theme in evidence that is no longer extant. Moreover, Becatti's emphasis on the funerary examples as comparanda is problematic. We know that animal combat scenes decorated not only philosophical scholae and tombs, but also domus and imperial palaces, as well as temples, diptychs, silver plate, and textiles, where they presumably would not have been restricted to such specialized allegorical meanings. He does cite examples of similar subjects from other contexts, labeling them Dionysiac or purely ornamental, but insists on the eschatological reading as most appropriate to these fourth century halls. 15 A further objection to Becatti's interpretation is that it classes the Bassus basilica with contemporary domus, in company with most of the scholarship on the hall at that time and since. 16 My view departs from this functional designation. 17 I instead read the decoration as commemorating the euergetic gifts of the hall's patron. Bassus would have been responsible for an elaborate series of arena entertainments as part of his duties as consul-an office recorded by the basilica's dedication inscription. The sectile program-insofar as enough of one survives to make such arguments-is selected to remind the viewer of Bassus' office, his proximity to the emperor, and his generosity with the people of Rome. At the same time, the imagery works to impart associations of cosmic, universal victory, a triumph not of military type, but of civic and political achievement, conveyed through the imagery of the circus and arena (Hagan 2018a, p. 206 ff). Even if the hall were a domestic schola, in accordance with Becatti's categorization, the function of self-representation would be a high priority within, as one of the public spaces of a domus.
In the case of the Basilica of Junius Bassus, then, the venatio panels comprise part of a games-giver's self-fashioning, by commemorating his efforts as editor and manifesting the intangible resources and influence of his patronage. These are artifacts that render material and concrete his social and political relationships, even as they advertise the ongoing strength and extent of the Roman empire (Hagan 2018a, p. 318 ff). These same aims are likely to have been shared by the owners of Sette Sale and Porta Marina, whether or not they were themselves munerarii, given their participation in a similar culture of aristocratic competitive display.
In the amphitheater, pairings such as tiger vs. bovine or lion vs. ass may not have offered much in the way of suspense, but audiences nonetheless took pleasure in these combats, their predictability notwithstanding. Shelby Brown argues that this certain outcome was an intentional message and part of the panels' function (Brown 1992, pp. 184, 193). The stronger would win out over the weaker, and this was the point. This outcome was reassuring, insofar as it aligned with Roman understanding of the natural world and its order, underscoring distinctions between categories of predator and prey (Kyle 1994, p. 182). This is consistent with other practice in the arena. Romans restaged battles and built scenery that served as theatrical reenactment of hunts or other trials. In this way, the expected violence of the arena not only acclimated viewers to the violence that might be required to maintain Roman order, it also naturalized the hierarchies and structures of Roman society (Kyle 2014, p. 289;Futrell 1997, pp. 46, 51).
The use of exotic animals in the arena for display or combat conjured up associations with foreign lands. The names of various animals in ancient sources link them with their geographic origin, so each mention recalls a corner of Roman jurisdiction. 18 Tigers, for example, could be obtained from India, Hyrcania (an Achaemenid province south of the Caspian), or Armenia. 19 A diverse array of animal contestants functioned as a pageant of imperial possessions, available to Rome as living booty from their conquests. 20 The animals that graced the arena also illustrated the extent of the munerarius' influence, each representing a province on a map whence he had been able, at great risk to his investment, to summon the animals to him. If the animals displayed in the arena stood as proxies for their native lands, so too did the individual stones that made up the panels displayed in an opus sectile hall. These marbles were quarried across the Mediterranean, and, like the zoological subjects they depicted, were available for Rome to exploit because of the extent of the emperor's network of control, which could direct labor to quarries and then convey the mined marbles via the necessary transport systems (Schneider 2001). In public statuary such as that seen in the Augustan and Trajanic fora, imperial power was expressed through the selection of exotic marbles to depict defeated foes. Sometimes the stone was sourced from the very region the prisoner called home: Parthian captives robed in porphyry from Phrygia, for example. 21 The association between stone and geographic origin was maintained by literary references that labeled marbles based on their provenance, and described various regions according to the stones that abounded there. Martial called Numidia "marble-painted;" Claudian calls Phyrgia "precious for the marble painted with purple veins". 22 Thus a varied array of stones made for a "territorial display", a map of influence, whether imperial or aristocratic (Kiilerich 2012, p. 23).
These readings, through the lenses of violence and imperialism and self-fashioning, are less cabalistic than the neo-Platonic allegories proposed by Becatti. Thanks to the po-lysemic habits of Roman art, however, Becatti need not be discarded entirely, and the examples of figural decoration laid out here can sustain ongoing inquiry.

Turning to Persia
To these existing interpretive paradigms, I wish to add another that takes into account the wider socio-political context of the Basilica of Junius Bassus and its companions at Sette Sale and Porta Marina in the late empire. The previous century, the "crisis" century, had seen the resumption of war with Persia and the rise of a new dynasty over the Arsacids. In the intermittent hostilities of the following decades, the Roman emperor Valerian was captured and killed by Shapur I; the city of Dura-Europos, which had alternated between Roman and Persian control, was finally destroyed; Gordian III was killed on campaign; and Philip the Arab was cowed into an ignominious peace. Rome regained a meaningful advantage in 298 CE, when Galerius defeated Narseh, a triumph commemorated on the Arch of Galerius at Thessaloniki. Shapur II renewed attacks upon ascending to the throne, aiming to regain Mesopotamia and Armenia and restore the boundaries of the Achaemenid empire (Dignas and Winter 2007, p. 33). Rome was largely unprovoked by these efforts, perhaps preoccupied by internal affairs, but Constantine was preparing for war with Persia when he died in 337 CE. His son Constantius II took up the cause for nearly a quarter of a century, but achieved no major shift in power. Julian took the field to try to tip the scale in Rome's favor, but died after being wounded in battle. His similarly short-lived successor Jovian opted for peace, so that Rome approached the end of the fourth century under conditions more propitious to Persia, thanks in part to Shapur II's long reign.
However constant their conflict, the Persians were unique among Rome's enemies, in that they were not to be written off as entirely uncivilized or inferior. Depictions of the Persians in Roman art made their specialness clear. While opponents such as the Dacians were depicted as captives with unkempt hair in postures of defeat, Persians were not always cast as so abject. Persian costume could also signal magi, mythic hero, or cult figure in the Roman visual code (Schneider 1998). Some depictions, such as the equestrian duel on the Arch of Galerius, put the emperor and Persian king on equal footing (Dignas and Winter 2007, p. 87). Even when his army was defeated in battle, the Persian king was owed a certain respect as heir to a longstanding empire. Canepa argues that the reason for this was Rome's view of Persians as equally capable of running the world (Canepa 2009, p. 35). Official monuments address not only the victors, but also the enemy, as if begrudgingly acknowledging their interdependence in spite of their aversion.
The revolving door of conflict, migrations (many of them forced), diplomacy, war, and commerce of the 3rd and 4th centuries would have shaped the non-official realm of visual culture as much as it did the imperial register. Non-imperial individuals might also Arts 2023, 12, 102 9 of 33 use foreign iconography to convey messages about their status and taste, as, for example, through the collecting of imported luxury goods or the use of decorative styles. 23

The Apadana at Persepolis
The building I wish to put in dialogue with the opus sectile panels is the Apadana, or audience hall, from the Achaemenid capital of Persepolis. The Apadana was begun under Darius near the end of the 6th century BCE and completed by his son Xerxes. Located on the terrace of the city with other significant public buildings, including royal palaces and the treasury, the square hypostyle hall sat on a podium 2.6 m high, which was accessed by staircases on the north and east sides. Its 72 columns were topped with addorsed animal head capitals, many of which were bulls and lions. 24 Sculptural reliefs adorned the outward-facing surfaces of the stairways leading to the Apadana. The exterior parapet was a trapezoidal form, at the center of which was a scene of Darius and Xerxes receiving an official (see Figure 8). On either side of this vignette was a triangular passage containing a lion attacking a bull. The lion at the left faced right, toward the audience scene, while the lion at the right faced left, so that the two were mirror reflections. The lion is arched onto hind legs, and the bull in front of him rears, creating an upward-slanting diagonal composition that fills the height of the triangle naturally. Behind the animal pair, stalks of water lilies at descending heights fill the slender point of the triangle (see Figures 9 and 10). The lion-bull pairing, or symplegma, appears not only on both the north and east sides of the Apadana, but is transposed/reduplicated on every major staircase on the terrace, including those of the council hall and the palaces. 25 Behind this parapet, on the inner stairway's wall, were registers of figures arrayed in procession. Some represent various peoples and provinces under Persian rule, distinguished by their costume and the gifts they bring as tribute. Other figures form the royal cortege (Root 1979, p. 88). Both lines lead to the center of the portico, as if they are ascending the stairs to the Apadana or waiting their turn to do so. These processions were long believed to depict a Persian New Year's festival (Schmidt 1953, p. 82), and while that interpretation has not sustained scholarly favor, the lion-bull pairing has been ascribed astrological meanings, whether as surrogates for the sun and moon or for Leo and Taurus. 26     The tribute-bearing figures have been given far more extensive treatment than the lion-bull symplegma; perhaps the latter were dismissed as ornament or filler. It seems to me, though, that they are given a hugely important role in this structure and its surrounds. They are given a central place, next to that of the king, flanking him like guardians. Their repetition at the site signals not rote pattern replication, but rather talismanic reiteration, the potency of which arises from the significance of the animals in a long material history.
Bulls and lions both have great resonance attached to kingship in Near Eastern material traditions. Lions were creatures of super strength and kingly bearing, endowed with melam, a kind of aura of power and splendor (Ulanowski 2015, p. 255; Watanabe 2002, The tribute-bearing figures have been given far more extensive treatment than the lion-bull symplegma; perhaps the latter were dismissed as ornament or filler. It seems to me, though, that they are given a hugely important role in this structure and its surrounds. They are given a central place, next to that of the king, flanking him like guardians. Their Arts 2023, 12, 102 11 of 33 repetition at the site signals not rote pattern replication, but rather talismanic reiteration, the potency of which arises from the significance of the animals in a long material history. Bulls and lions both have great resonance attached to kingship in Near Eastern material traditions. Lions were creatures of super strength and kingly bearing, endowed with melam, a kind of aura of power and splendor (Ulanowski 2015, p. 255;Watanabe 2002, p. 48). In Near Eastern art they could symbolize destructive chaos and, as such, they represented a threat to civilization (Ulanowski 2015, p. 256). The figure of the king was often seen vanquishing or having vanquished a lion, as in Assyrian palace reliefs, and in the "master of animals" motif, which appears on Assyrian seals. 27 Because the powers of the lion and the king were closely matched, this conferred upon the lion kingly associations, so that by about the end of the Iron Age, the lion appeared not in opposition to the king but instead as a symbol of his royalty, and as guardians to thrones and entrances (Ulanowski 2015, pp. 260, 262). Some texts equated "lion" and "king", using the terms interchangeably (Ulanowski 2015, p. 259). Lions also adorn royal garments and thrones depicted in relief. Reclining lions with crossed paws were found at Persepolis, but find precedent in court works from Susa as early as the third millennium BCE. 28 Bulls were also symbols of royal power. Reverenced for their purity, they were believed to be the first animal to live on the earth, and became significant in Zoroastrian ritual (Root 2001, p. 194). Sculpted bulls guarded Persepolis from the late 6th century BCE, and their heads topped palace capitals (Root 2001, p. 202). Stylized horns protruded from the western wall of the citadel (Root 2001, pp. 197, 202).
Given their extensive royal associations, it is unsurprising to see both animals in royal contexts, such as the throne furniture at Persepolis (Root 2001, p. 202). For all their profusion as individual animals, though, the lion-bull attack is not typical of Persian art, apart from the Apadana reliefs. 29 When lions are depicted with prey, the quarry is typically a deer (Root 2001, p. 202). This is one indication that the lion-bull symplegma may not be a straightforward hunting scene. However meek the domesticated animal might seem, Root notes, a wild bull made a formidable foe, and was "not necessarily always conceived of as the victims in these emblems" (Root 2001, p. 201). This means the predictable order of nature cannot have been the intended takeaway here as it might have been for the Roman panels. If this were a hunting attack, moreover, we would be more likely to see the female lion at work. Instead, we have a pairing more stylized than violent, in which a lion lunges at the back end of the bull. The bull turns back to look at the attacker, rearing up onto its hind legs. Root observes that this looks less like a predatory hunt than the prelude to mating. She suggests that the visual nod to copulation implies a union of these two animalsnot literally, but in symbolic terms, a tangle of well-matched powers, signifying opposing natural forces balanced under the Achaemenid empire's rule (Root 2001, pp. 201-2). The astrological meanings of the lion and bull, the sun (represented by the sign Leo) and moon (represented by the sign Taurus), respectively, support a reading of opposition and balance. This pairing of sun and moon is displayed in other royal contexts, such as above the fire altar at Achaemenid royal tombs (Root 2001, p. 202).

Looking across Monuments
The compositional similarities between the marble panels and the Persepolitan relief sculpture are notable. The attacking animal, whether tiger or lion, is balanced on its hind legs, its front paws clutching its quarry from behind. At Persepolis, the lion grasps the hindquarters of the bull; in the Roman examples, the feline's purchase is higher on the body, on the chest or at the shoulder point. In the sectile examples, as in the relief sculpture, the cat's front paws are both in evidence, one pressed flat into the facing flank of the herbivore, and one wrapped around its body from behind. The cat's claws slash at the flesh of the animals, tracing champlevée incisions onto their marble hides and even drawing drops of glass-paste blood (see again Figures 4 and 5). The cat holds its prey close, teeth sinking into the caught animal. 30 From this position the feline in every example looks out at the viewer head-on. The bull in the Persian reliefs turns back over its shoulder, facing its opponent. The ruminant (gazelle?) and onager in the Ostian panels twist back in the same way, showing the viewer their heads in profile, and thus meeting the eye of the viewer in the same way that the predator does. 31 Though the cows in the Bassus panels do not achieve this same degree of flexion, still the eye glares out, at once terrified and confrontational.
In these examples nearly a millennium apart, finer features of the faces also find parallels. In the Persepolis reliefs, the lion's eyes are outlined in a teardrop shape that points toward the ear. There is another nodule on the bottom of the eyelid, near the side of the nose. These nodules are imitated in the Ostian lion, whose eyes are outlined in black (possibly a dark serpentine), as if the animal were wearing eyeliner. The dark outline recalls the deep recess of the lion's eye in the relief, the modeling of which brings shadows into the depths of its features. The same winged eyelids appear on the Bassus tiger panels. 32 The Ostian lions have a small oval highlight over each eye, paler than the surrounding marble, which articulates the ridge of the brow. The lion in relief displays the same articulation: two circles between the eye, positioned one on either side of a muzzle crease (see Figure 11). The affinity between the composition and the graphic choices we see in images 800 years and 3000 miles apart is paralleled-which is not to say explained-by the analogy between opus sectile and relief sculpture as media. Both the Persepolis reliefs and the sectile panels are made of stone, a material for sculpting, and thus associated with working in three dimensions. Yet both are also surface decoration, which renders them superficial by definition. Limited to two dimensions, opus sectile is sometimes referred to as "painting in stone," an allusion to a graphic medium whose effects it approximates (Barry 2011). Yet in these examples, depth is suggested only by high color contrast. Marble joins may touch, but never layer or soften into sfumato; though there is variegation within individual pieces of stone, still each joins the adjacent piece at a defined edge. The effect is schematic, as joining puzzle pieces whose interlocking seams are easily traced. This would not be called painterly in Wölfflinian terms, but linear.
This linearity is met in parts of the Persepolis reliefs, such as the lower part of the lion's hind legs, where the job of delineating joints or bone is done by tracing sinew (see Figure 12; notice how this matches the right-facing cow from Bassus hall, with its round knees articulated). The Persepolis reliefs are not wholly without modeling, but there are places such as the grooved ridges of the lion's muzzle that correlate to the edge-fired pieces of marble that make up the concentric pear-shaped pieces of the lion's yellow nose (see again Figure 11). Incised patterns create visual texture in the lion's mane and on the neck and torso of the bull. 33 The affinity between the composition and the graphic choices we see in images 800 years and 3000 miles apart is paralleled-which is not to say explained-by the analogy between opus sectile and relief sculpture as media. Both the Persepolis reliefs and the sectile panels are made of stone, a material for sculpting, and thus associated with working in three dimensions. Yet both are also surface decoration, which renders them superficial by definition. Limited to two dimensions, opus sectile is sometimes referred to as "painting in stone", an allusion to a graphic medium whose effects it approximates (Barry 2011). Yet in these examples, depth is suggested only by high color contrast. Marble joins may touch, but never layer or soften into sfumato; though there is variegation within individual pieces of stone, still each joins the adjacent piece at a defined edge. The effect is schematic, as joining puzzle pieces whose interlocking seams are easily traced. This would not be called painterly in Wölfflinian terms, but linear.
This linearity is met in parts of the Persepolis reliefs, such as the lower part of the lion's hind legs, where the job of delineating joints or bone is done by tracing sinew (see Figure 12; notice how this matches the right-facing cow from Bassus hall, with its round knees articulated). The Persepolis reliefs are not wholly without modeling, but there are places such as the grooved ridges of the lion's muzzle that correlate to the edge-fired pieces of marble that make up the concentric pear-shaped pieces of the lion's yellow nose (see again Figure 11). Incised patterns create visual texture in the lion's mane and on the neck and torso of the bull. 33 This linearity is met in parts of the Persepolis reliefs, such as the lower part of the lion's hind legs, where the job of delineating joints or bone is done by tracing sinew (see Figure 12; notice how this matches the right-facing cow from Bassus hall, with its round knees articulated). The Persepolis reliefs are not wholly without modeling, but there are places such as the grooved ridges of the lion's muzzle that correlate to the edge-fired pieces of marble that make up the concentric pear-shaped pieces of the lion's yellow nose (see again Figure 11). Incised patterns create visual texture in the lion's mane and on the neck and torso of the bull. 33

Installing the Wall
While the finer details of joins and textural treatment bear close looking, the Persepolis symplegmata and the 4 th century opus sectile panels should also be compared as modules within their respective installations. Each animal combat scene is a self-

Installing the Wall
While the finer details of joins and textural treatment bear close looking, the Persepolis symplegmata and the 4th century opus sectile panels should also be compared as modules within their respective installations. Each animal combat scene is a self-contained horizontal unit in the plane of the wall, which does not participate in the same imagined space or decorative demesne as the surrounding imagery. That is: each time the lion-bull combat appears at Persepolis, it conforms to the shape of the available surface, so that the animals might bound upward to follow the ascent of a staircase or crouch lower to fit themselves into a more regular rectangular space, as called for by the surface area in each instance. The lion-bull motif complies with the limits of its borders, and does not inhabit the same imagined space as neighboring images, such as the processions. Cordoned off by rosettedotted bands, the combats do not share the same scale or pretend to the same setting, but are instead entirely separate entities that happen to be placed in close proximity.
Similarly, the Ostian panels (the only ones for which their placement on the wall is supported by archaeological evidence) bear no relationship to the content or scale of the panels they are fitted next to. The lion panel on the long wall follows after a pitched roof structure at the beginning of the row, but neither it nor its mirror image attempts to match the scale of that building or participate in an illusion of depth that would place them inside it. The Ostian hall-and arguably the Bassus hall as well, as far as we can tell-lacks any unity of surface; rather, the wall is treated as a board on which discrete pictures are arranged. They are punctuated by sham architecture, such as the column of scrollwork, but they do not participate in the feigned space of that architecture.
While it was typical for a Roman wall to have divided sections of decoration, for example, a dado of imitation stone, while the upper wall depicted a narrative scene or a trompe l'oeil architectural fantasy, the proportions and greater number of subdivisions we see at Ostia (and in the Bassus hall, as far as we trust the post-antique documentation) are irregular for a Roman model. 34 At Ostia, there are at least three zones of decoration from floor to ceiling, divided just above head height by a band of scroll. No zone's vertical divisions align with those below. The uppermost panels are varying widths, not regular divisions of the wall. The refusal of vertical unity is similar to the divided-strip coverage of the synagogue walls at Dura Europos. What we see does not fit the expected scheme of a painted wall, nor is its boxiness consistent with the all-over approach of a mosaic "carpet". It is instead closer to the Arch of Constantine, with its assorted reliefs inserted in different registers, self-contained and independent of any other module's internal visual logic.
We see a similar treatment in the mosaics at Bishapur, where the decoration around the perimeter of the iwan comprises rectangular boxes of varying heights to fit into corresponding niches in the wall. 35 The tallest ones contain figural scenes, many with dancers or musicians (see Figure 13). The white ground indicates no depth in the image or indeed in any of the companion scenes. These figural scenes are flanked by shorter rectangles, which are filled with a geometric pattern or swath of imitation marble. The next rectangle in the descending height sequence is a row of Dionysiac masks. Each rectangle is heavily outlined, as if the border were meant to emphasize the separateness of each panel. The repetition of this stepped crenellation (reminiscent of the Persepolis Apadana's own parapets, and of the stepped band of rosettes that traces the stairways' upward or downward direction there) is evident in plan. 36 The framed panels are like mosaic emblemata, except that an emblema typically functioned as a centerpiece, surpassing the surrounds in detail and serving as the focal point of a surface, aided by the elaborate ornamentation of its framing. Here, as in the Ostian hall, each panel is a work in its own right, but its relation to the whole is unclear. All together, the graphic decoration at these sites put one in mind of the arrangement of fitted blocks in a Tetris game, or the imposition of separate motifs in a textile according to the size and shape of the negative space. The placement of each "box" or panel is somewhat indiscriminate, as if any two modular units of the same dimensions were fungible. 37 At the same time, their seeming indifference to the rest of the mural arrangement notwithstanding, the panels' orientation makes clear that there was some deliberate systematization in their disposition. The combat panels are arrayed such that they face one another, with a non-figural panel between them. Similarly, the Persepolis emblems appear in facing pairs, butterflied. The aesthetic at work on these varied surfaces, whether in relief, sectile, or mosaic, is characterized by a bringing together of somewhat disparate and even disjointed parts into an irregular fabric of surface decoration.

Lineage and Transmission of the Animal Combat Theme
There is general acceptance that the animal combat scenes we find in the Basilica of Junius Bassus, the hall at Porta Marina, and Sette Sale have their origins in material from the Near East and Egypt. This point is made summarily in most publications that discuss these depictions, but only leanly explicated, as to assign credit for the earliest appearances of the motif. 38 This is of minimal interest outside an art history that seeks to describe motif transmission as a stepping stone in art's forward progress toward a Greco-Roman pinnacle.
As Greco-Roman legacy expands to fill the imagination, the acknowledged role of Persians in the construction of the cultural landscape becomes narrower and narrower (Gates 2002, p. 119). Persia as an entity is regarded as passive, an instrument in service of higher attainment. This stance puts cultural patrimony at risk of erasure, a peril that extends beyond the academic. By laying out possible mechanisms for transmission of the feline combat motif generally, and demonstrating the close compositional parallels between late Roman wall decoration and the Apadana reliefs in particular, I hope to credit Achaemenid art in a way that redresses that narrowing.
I would like to flesh out the spindly link between Persian forms and their Greco-Roman adoption by illustrating a selection of animal combat scenes that date to the period between Persepolis' foundation and the making of the opus sectile panels. Given that some eight centuries passed between those two points, it may seem a far leap to suggest a direct line between the Apadana at Persepolis and the marble-clad halls in and around All together, the graphic decoration at these sites put one in mind of the arrangement of fitted blocks in a Tetris game, or the imposition of separate motifs in a textile according to the size and shape of the negative space. The placement of each "box" or panel is somewhat indiscriminate, as if any two modular units of the same dimensions were fungible. 37 At the same time, their seeming indifference to the rest of the mural arrangement notwithstanding, the panels' orientation makes clear that there was some deliberate systematization in their disposition. The combat panels are arrayed such that they face one another, with a nonfigural panel between them. Similarly, the Persepolis emblems appear in facing pairs, butterflied. The aesthetic at work on these varied surfaces, whether in relief, sectile, or mosaic, is characterized by a bringing together of somewhat disparate and even disjointed parts into an irregular fabric of surface decoration.

Lineage and Transmission of the Animal Combat Theme
There is general acceptance that the animal combat scenes we find in the Basilica of Junius Bassus, the hall at Porta Marina, and Sette Sale have their origins in material from the Near East and Egypt. This point is made summarily in most publications that discuss these depictions, but only leanly explicated, as to assign credit for the earliest appearances of the motif. 38 This is of minimal interest outside an art history that seeks to describe motif transmission as a stepping stone in art's forward progress toward a Greco-Roman pinnacle.
As Greco-Roman legacy expands to fill the imagination, the acknowledged role of Persians in the construction of the cultural landscape becomes narrower and narrower (Gates 2002, p. 119). Persia as an entity is regarded as passive, an instrument in service of higher attainment. This stance puts cultural patrimony at risk of erasure, a peril that extends beyond the academic. By laying out possible mechanisms for transmission of the feline combat motif generally, and demonstrating the close compositional parallels between late Roman wall decoration and the Apadana reliefs in particular, I hope to credit Achaemenid art in a way that redresses that narrowing.
I would like to flesh out the spindly link between Persian forms and their Greco-Roman adoption by illustrating a selection of animal combat scenes that date to the period between Persepolis' foundation and the making of the opus sectile panels. Given that some eight centuries passed between those two points, it may seem a far leap to suggest a direct line between the Apadana at Persepolis and the marble-clad halls in and around Rome. The examples adduced below, many of which belong to a robust Greco-Roman tradition from the Hellenistic period through the late empire, may seem likelier kin to the opus sectile depictions. While there are indeed some details that bear a close resemblance to the marble panels, this set includes multiple types, with variations that depart from the animal-on-animal pairing. My selection does not aim at a comprehensive catalog, but at demonstrating the range of artistic interpretations visible in this corpus. The differences between them demonstrate the ways the theme morphed to serve the aesthetic and social needs of the patron. They also make clear that there was a wide range of options for depicting this motif, such that the opus sectile panels did not end up looking the way they did merely by chance. My hope is that a closer look at how the subject was transposed into the art of the Hellenistic period will better equip us to see how closely attuned the sectile panels are with the Persepolis examples. I argue that the panels-however well they suited Roman contexts and mirrored arena combats-remained recognizably eastern (and arguably, explicitly Persian, even pointedly Achaemenid).
I elide here the pre-Persepolis history of the motif, leaving that origin story to specialized historians of Mesopotamia. 39 The subject is popular on various types of Archaic and Classical period Greek pottery. 40 Roughly contemporary with the building of the Apadana are several lion combats in Greek temple sculpture (Markoe 1989, pp. 96-102). Athens also took up the lion and bull on their coinage, which Markoe regards as having been useful as a symbol of power legible both within Greece and outside of it (Markoe 1989, pp. 107-9).
Cylinder seals offer further instances of the official use of the theme. An impression of PFS 857s from not long after the Apadana was built carries a scene of two lions attacking a stag, one at the underside of the sprawled prey and one at its back. 41 The seal is known as the Seal of Gobryas, for the high-ranking Persian official who carried its impression on a sealed document from the king. 42 The seal could not be more incontrovertibly linked to the administrative functions of the court at Persepolis. This bureaucratic use endowed the imagery on the seal with the capacity to peregrinate far and wide, such that seals comprised a highly portable mechanism for image transmission. 43 Every recipient of a seal-impressed document who viewed the motif and responded to its bearer further ratified the symbolic meaning intended by the sender. Sealings also served to miniaturize the iconography, and would undoubtedly have played a part in allowing viewers to contemplate the transfer of its meaning to any number of different contexts. 44 Coins, too, bore images of the lion-bull combat. A coin minted by Mazaeus, a member of the Achaemenid aristocracy and contemporary of Alexander the Great, features this motif on its reverse (see Figure 14). 45 This object employs the lion-bull pairing as an indication not merely of Mazaeus' authority, but of the royal sovereign in whose name he ruled. Minted at Tarsus, in what is today south-central Turkey, the coin would have circulated across the Mediterranean, its replication further proliferating the animal combat imagery.
By the Hellenistic period, the evidence we have for the transmission of the feline hunt shows it to have moved outside the official or administrative sphere and into the domestic realm. In this era, when we find two-dimensional parallels in mosaic, we approach more closely the composition of the sectile panels. The animals vary-both attacker and attacked-but the theme persists in art from the Hellenistic period, and is used frequently as both centerpiece of narrative scenes and component of decorative material in floor mosaics. From c. 370 BCE comes a mosaic at the House of the Mosaics in Eritrea, Greece, in which a lion attacks a horse from behind. This duo is depicted twice, on opposite sides of a square, alternating with a griffin-warrior-griffin sequence to form a figural border (see Figure 15). A house in Motya (Mozia) in Sicily has a black and white pebble mosaic from the same century that depicts several different animals, some paired in combat, while others are shown alone, rendered in white within the confines of their respective black boxes, facing another single animal (see Figure 16) (Belis 2016, cat. 5 note 4). One of the pairs shows a cat attacking a bull, but the predator is seen as if standing atop the prey, perhaps because the black and white pebble motif makes it difficult to depict overlapping bodies at this scale. In the panel most similar to the Persepolis composition (not pictured here), a griffin attacks a horse or onager, the creature's head turning 180 degrees towards the attacker to show a single eye glaring out from the profile. The griffin, though, is also seen in profile, rather than face-on like the lions at Persepolis or the felines in the opus sectile panels. The boxy frame around each single animal or pair yields the impression that each is a unit that can be detached, slid along one axis or another, and reassembled in a different configuration. In this respect they resemble the Ostian panels, which also stand alone as rectangular framings abutting one another in the grille configuration of the wall. Their undifferentiated backgrounds call to mind that of the Persepolis emblems. By the Hellenistic period, the evidence we have for the transmission of the feline hunt shows it to have moved outside the official or administrative sphere and into the domestic realm. In this era, when we find two-dimensional parallels in mosaic, we approach more closely the composition of the sectile panels. The animals vary-both attacker and attacked-but the theme persists in art from the Hellenistic period, and is used frequently as both centerpiece of narrative scenes and component of decorative material in floor mosaics. From c. 370 BCE comes a mosaic at the House of the Mosaics in Eritrea, Greece, in which a lion attacks a horse from behind. This duo is depicted twice, on opposite sides of a square, alternating with a griffin-warrior-griffin sequence to form a figural border (see Figure 15). A house in Motya (Mozia) in Sicily has a black and white pebble mosaic from the same century that depicts several different animals, some paired in combat, while others are shown alone, rendered in white within the confines of their respective black boxes, facing another single animal (see Figure 16) (Belis 2016, cat. 5 note 4). One of the pairs shows a cat attacking a bull, but the predator is seen as if standing atop the prey, perhaps because the black and white pebble motif makes it difficult to depict overlapping bodies at this scale. In the panel most similar to the Persepolis composition (not pictured here), a griffin attacks a horse or onager, the creature's head turning 180 degrees towards the attacker to show a single eye glaring out from the profile. The griffin, though, is also seen in profile, rather than face-on like the lions at Persepolis or the felines in the opus sectile panels. The boxy frame around each single animal or pair yields the impression that each is a unit that can be detached, slid along one axis or another, and reassembled in a different configuration. In this respect they resemble the Ostian panels, which also stand alone as rectangular framings abutting one another in the grille configuration of the wall. Their undifferentiated backgrounds call to mind that of the Persepolis emblems.   The subject passed into Roman use, where it is often depicted in a lushly rendered perspectival landscape, as at Hadrian's villa, where we see a lion attack a bull (see Figure  17). In the triclinium, multiple species of exotic cats fight centaurs. These depictions are very painterly, with a good deal of color variation and modeling and some depth of field. The 2nd century CE Tunisian example from a villa in Hadrumentum pares down the landscape, detailing it in the foreground but filling in the area above the horizon line with only a field of white tesserae against which the brawling animals stand out (see Figure  18). This example bears some resemblance to the Bassus panels, in that the head of the cat is frontal, while that of the onager is seen in profile, and we see both of the tiger's front paws; but this onager is twisted, nearly pinned to its side so that its belly is exposed to the The subject passed into Roman use, where it is often depicted in a lushly rendered perspectival landscape, as at Hadrian's villa, where we see a lion attack a bull (see Figure 17). In the triclinium, multiple species of exotic cats fight centaurs. These depictions are very painterly, with a good deal of color variation and modeling and some depth of field. The 2nd century CE Tunisian example from a villa in Hadrumentum pares down the landscape, detailing it in the foreground but filling in the area above the horizon line with only a field of white tesserae against which the brawling animals stand out (see Figure 18). This example bears some resemblance to the Bassus panels, in that the head of the cat is frontal, while that of the onager is seen in profile, and we see both of the tiger's front paws; but this onager is twisted, nearly pinned to its side so that its belly is exposed to the viewer, even though the animal's head and neck fight to come upright. Blood drips down the onager's flank and onto the soil of the riverbank. Later Roman images show less depth of perspective, and the landscape, if it is depicted at all, tends to become a more perfunctory gesture toward implying the scene takes place in natural space, rather than trying to create an immersive natural scene (Belis 2016, cat. 5). The Carthage mosaic in the Smithsonian illustrates this category nicely: the ground is mostly white, but some tufts of greenery indicate to the viewer that we are seeing a scene unfold in a green space. 46 In terms of landscape fullness, the opus sectile panels fit in this category.
Other examples are not located in landscape at all, and instead sit at the center of a blank ground, surrounded by a circle or polygon that frames the image. Examples of this type are the 3rd century CE Lod mosaic, where animal combat pairs and animals on display fill the hexagonal compartments of an all-over mosaic, like so many shiftable tiles (see Figure 19). Some have groundlines or shadows that imply a location in space; others simply float. Outline is used to demarcate the contours of the animals' anatomy. The quality of the mosaic is distinguished by its color and variety, not by unreserved commitment to naturalism.
One variant on this type seen in painting and mosaic shows an exotic cat with the dismembered body of a victim. The Terrace Houses at Ephesus have a mosaic of a lion standing over the head of a bull, one paw atop its head while it gazes menacingly out (see Figure 20). In this example, the flick of the lion's tail up and over its back rhymes nicely with the same figure in the opus sectile examples, ending in a generous brush-like tuft. As with other types, the animals can be swapped for others. At Verulamium, a mosaic lion holds in its mouth the bloodied head of a stag (see Figure 21). 47 At Dura Europos, the dado of the densely decorated synagogue is populated by leopards, panthers, lions, and tigers, many of which are harnessed in the manner of the Ostian lions (though less richly). Some of the carnivores stand in the apparent aftermath of a venatio with one paw resting on the head of a deer, ram, or bull in the manner seen at Ephesus. 48 The location of these paintings in a border city where Roman-Persian interchange was guaranteed by the city's changing governance makes them especially germane to this study. They are one of the few examples that permit us to compare a vertical wall, instead of a floor, to the wall decorations from Rome. The Dura decoration is also particularly interesting for the gap between the medium used and the medium referenced: the painted animal pairs are at the center of imitation marble roundels, which are set into rectangular slabs of contrasting faux marble. 49 The variegation and veining of the faux marble shows an interest in the visual effects of opus sectile, but the composition of the figures does not resemble the actual sectile panels made in the century that followed.
Arts 2023, 12, x FOR PEER REVIEW 19 of 35 viewer, even though the animal's head and neck fight to come upright. Blood drips down the onager's flank and onto the soil of the riverbank. Later Roman images show less depth of perspective, and the landscape, if it is depicted at all, tends to become a more perfunctory gesture toward implying the scene takes place in natural space, rather than trying to create an immersive natural scene (Belis 2016, cat. 5). The Carthage mosaic in the Smithsonian illustrates this category nicely: the ground is mostly white, but some tufts of greenery indicate to the viewer that we are seeing a scene unfold in a green space. 46 In terms of landscape fullness, the opus sectile panels fit in this category.    viewer, even though the animal's head and neck fight to come upright. Blood drips down the onager's flank and onto the soil of the riverbank. Later Roman images show less depth of perspective, and the landscape, if it is depicted at all, tends to become a more perfunctory gesture toward implying the scene takes place in natural space, rather than trying to create an immersive natural scene (Belis 2016, cat. 5). The Carthage mosaic in the Smithsonian illustrates this category nicely: the ground is mostly white, but some tufts of greenery indicate to the viewer that we are seeing a scene unfold in a green space. 46 In terms of landscape fullness, the opus sectile panels fit in this category.   Other examples are not located in landscape at all, and instead sit at the center of a blank ground, surrounded by a circle or polygon that frames the image. Examples of this type are the 3rd century CE Lod mosaic, where animal combat pairs and animals on display fill the hexagonal compartments of an all-over mosaic, like so many shiftable tiles (see Figure 19). Some have groundlines or shadows that imply a location in space; others simply float. Outline is used to demarcate the contours of the animals' anatomy. The quality of the mosaic is distinguished by its color and variety, not by unreserved commitment to naturalism.
One variant on this type seen in painting and mosaic shows an exotic cat with the dismembered body of a victim. The Terrace Houses at Ephesus have a mosaic of a lion standing over the head of a bull, one paw atop its head while it gazes menacingly out (see Figure 20). In this example, the flick of the lion's tail up and over its back rhymes nicely with the same figure in the opus sectile examples, ending in a generous brush-like tuft. As with other types, the animals can be swapped for others. At Verulamium, a mosaic lion holds in its mouth the bloodied head of a stag (see Figure 21). 47 At Dura Europos, the dado of the densely decorated synagogue is populated by leopards, panthers, lions, and tigers, many of which are harnessed in the manner of the Ostian lions (though less richly). Some of the carnivores stand in the apparent aftermath of a venatio with one paw resting on the head of a deer, ram, or bull in the manner seen at Ephesus. 48 The location of these paintings in a border city where Roman-Persian interchange was guaranteed by the city's changing governance makes them especially germane to this study. They are one of the few examples that permit us to compare a vertical wall, instead of a floor, to the wall decorations from Rome. The Dura decoration is also particularly interesting for the gap between the medium used and the medium referenced: the painted animal pairs are at the center of imitation marble roundels, which are set into rectangular slabs of contrasting faux marble. 49 The variegation and veining of the faux marble shows an interest in the visual effects of opus sectile, but the composition of the figures does not resemble the actual sectile panels made in the century that followed. The atrium (17) of the House of the Laberii in Oudna is sprinkled with animal combats, which receive less attention than the more well-known hunt mosaic in the same space (Parrish 1987, p . 119 and note 25). The elongated rectangular panels show a cat dashing after a stag or other quarry. The panels face inward, which would likely not have been obvious to a viewer standing on the floor, since a column stood between them (see The atrium (17) of the House of the Laberii in Oudna is sprinkled with animal combats, which receive less attention than the more well-known hunt mosaic in the same space (Parrish 1987, p . 119 and note 25). The elongated rectangular panels show a cat The atrium (17) of the House of the Laberii in Oudna is sprinkled with animal combats, which receive less attention than the more well-known hunt mosaic in the same space (Parrish 1987, p. 119 and note 25). The elongated rectangular panels show a cat dashing after a stag or other quarry. The panels face inward, which would likely not have been obvious to a viewer standing on the floor, since a column stood between them (see Figure 22). Laid out in bird's eye view, though, that band of the floor is reminiscent of either a wall dado like that at Dura or the paratactic arrangement of opus sectile panels on a wall, as opposed to the carpet decoration of the sea it runs alongside. Here, as elsewhere (at Piazza Armerina, for example), the animal combat theme appears in conjunction with the theme of the hunt, calling to mind not only the aristocratic and kingly activity of hunting, but also the capture of animals for arena entertainments. In other Roman examples, the arena setting is made evident because the animal combat is depicted alongside other entertainments, such as gladiatorial combats. Many more examples could be offered from the Roman repertoire. Contemporary with them are examples from Persia or its surrounds that depict the subject as handled by artisans working in a Persian mode. The Miho Museum holds one such example, a pre-Achaemenid silver vessel of a lion attacking a bull (see Figure 23). The lion is fully on top of the bull's back, rather than behind it, which collapses the upward-springing diagonal of the composition from Persepolis. In many fine details, though, the miniature silver version matches the monumental relief that was to come. The bull's front legs are splayed in opposite directions, one jutting out from the main body of the vessel, just as the Persepolis bull's leg extends before the animal in an unlikely, dancerly position. The artist has finely incised in the surface of the metal some of the same indications of volume and texture as we see at Persepolis: the scale-like forms that denote the texture of the lion's fur; the articulated knuckle joints of the paw slung over the bull's side; the articulated semi-circular brows above the bull's eyes; and even the grooves that indicate whisker lines on the lion's muzzle. All seem to adhere to the same guidebook for fleshing out the combatants in surface detail, difference in medium and scale notwithstanding.
Many more examples could be offered from the Roman repertoire. Contemporary with them are examples from Persia or its surrounds that depict the subject as handled by artisans working in a Persian mode. The Miho Museum holds one such example, a pre-Achaemenid silver vessel of a lion attacking a bull (see Figure 23). The lion is fully on top of the bull's back, rather than behind it, which collapses the upward-springing diagonal of the composition from Persepolis. In many fine details, though, the miniature silver version matches the monumental relief that was to come. The bull's front legs are splayed in opposite directions, one jutting out from the main body of the vessel, just as the Persepolis bull's leg extends before the animal in an unlikely, dancerly position. The artist has finely incised in the surface of the metal some of the same indications of volume and texture as we see at Persepolis: the scale-like forms that denote the texture of the lion's fur; the articulated knuckle joints of the paw slung over the bull's side; the articulated semi-circular brows above the bull's eyes; and even the grooves that indicate whisker lines on the lion's muzzle. All seem to adhere to the same guidebook for fleshing out the combatants in surface detail, difference in medium and scale notwithstanding.
A set of Parthian phalera (horse trappings) offers another portable take on the lionprey theme, again in metal (see Figure 24). The quarry this time is a stag, however, which, as Root says, is more typical of hunting scenarios, and therefore suited to the purpose. The stag in the example on the right used to have an inlaid gem in its eye, such that it stared out at the viewer in the same confrontational way as the Roman examples. The lion's tail flicks back into a curve, as we see in some of the opus sectile panels.
Persepolis bull's leg extends before the animal in an unlikely, dancerly position. The artist has finely incised in the surface of the metal some of the same indications of volume and texture as we see at Persepolis: the scale-like forms that denote the texture of the lion's fur; the articulated knuckle joints of the paw slung over the bull's side; the articulated semi-circular brows above the bull's eyes; and even the grooves that indicate whisker lines on the lion's muzzle. All seem to adhere to the same guidebook for fleshing out the combatants in surface detail, difference in medium and scale notwithstanding. A set of Parthian phalera (horse trappings) offers another portable take on the lionprey theme, again in metal (see Figure 24). The quarry this time is a stag, however, which, as Root says, is more typical of hunting scenarios, and therefore suited to the purpose. The stag in the example on the right used to have an inlaid gem in its eye, such that it stared out at the viewer in the same confrontational way as the Roman examples. The lion's tail flicks back into a curve, as we see in some of the opus sectile panels. In an 8th century CE textile of Persian or Sogdian manufacture, figures hunt with bows and arrows on horseback, simultaneous to felines hunting their own prey, which they mount from behind (see Figure 25). As at Persepolis, the victim's head swivels back 180 degrees so that its eye looks out. The textile also shows us the repetition of figural groups in a way reminiscent of the reiteration of the lion-bull combat at Persepolis. The animal pair, victim wounded by prey and prey wounded by an arrow, is woven just under the mounted equestrian figure. In an 8th century CE textile of Persian or Sogdian manufacture, figures hunt with bows and arrows on horseback, simultaneous to felines hunting their own prey, which they mount from behind (see Figure 25). As at Persepolis, the victim's head swivels back 180 degrees so that its eye looks out. The textile also shows us the repetition of figural groups in a way reminiscent of the reiteration of the lion-bull combat at Persepolis. The animal pair, victim wounded by prey and prey wounded by an arrow, is woven just under the mounted equestrian figure.
All of these examples were portable luxury goods, which served as one vector through which new eyes could become acquainted with unfamiliar visual material, and which might then be adapted or imitated. These few illustrations allow us to see the treatment of the lion combat motif across multiple media in Persian art of several distinct periods. None is identical to the Persepolis reliefs or to the Roman panels, but all show the lion attacking its prey from behind, the victim's neck twisted back toward the predator.
In an 8th century CE textile of Persian or Sogdian manufacture, figures hunt with bows and arrows on horseback, simultaneous to felines hunting their own prey, which they mount from behind (see Figure 25). As at Persepolis, the victim's head swivels back 180 degrees so that its eye looks out. The textile also shows us the repetition of figural groups in a way reminiscent of the reiteration of the lion-bull combat at Persepolis. The animal pair, victim wounded by prey and prey wounded by an arrow, is woven just under the mounted equestrian figure. All of these examples were portable luxury goods, which served as one vector through which new eyes could become acquainted with unfamiliar visual material, and which might then be adapted or imitated. These few illustrations allow us to see the treatment of the lion combat motif across multiple media in Persian art of several distinct periods. None is identical to the Persepolis reliefs or to the Roman panels, but all show

Between Rome and Persia
Given the resemblance between the opus sectile panels from Rome and its surrounds and the Apadana at Persepolis, we might wonder whether Romans could have encountered this monument in the fourth century CE. We know that the third and fourth centuries were a time of high contact between Rome and Persia, comprised of both formal interactions at the imperial level and more mundane associations on the part of soldiers, prisoners, artisans, and other private individuals. The wars, treaties, and trade between the two empires in these two centuries facilitated the exchange of visual material through several mechanisms.
It may seem overly conjectural to wonder whether the artisans who produced opus sectile panels for fourth century CE Roman patrons might have seen the Apadana firsthand, but this possibility is not to be discounted. Though Alexander burned Persepolis down, there were many remaining ruins, "a continual resource for Iranian artists long before they were ever excavated by the Oriental Institute" (Root 1994, pp. 36-37). The north facade of the Apadana was among these, and arguably, if Roman artisans used these ruins to inform their work, they would not have been the first: Root believes the frieze of the Parthenon-a monument that could not possibly be more emblematic of Greekness-emulated Persian forms and meaning from the Apadana. 50 "Nevertheless", writes Root, "in most discussions of artistic transmission relating to the ancient Near East, the notion of direct encounter with visible, non-portable remains is ignored or dismissed" (Root 1994, p. 36). She attributes this to the desire to identify Greek art as purely Greek, because the influence of Persian art on Greek forms would be regarded as a kind of contamination (Root 1985, p. 118).
Root's work suggests that Roman artisans could easily have had, if not firsthand witness to the Apadana, then certainly access to drawings or copies that sufficed as well to make the monument known to them. She notes the conservativism of artistic training processes, which demanded direct copying, and thus preserved very ancient works as models (Root 1994, p. 25ff). Artists studied by working from replicas, which they might collect as part of a personal archive, or which hung on studio walls for the benefit of the atelier. Those replicas were sometimes many centuries old, as Roman casts made of works from the Greek Archaic period would indicate (Root 1994, p. 28). Temple archives preserved records that provided artists a means of learning artistic conventions or researching their predecessors' approach to artistic problems (Root 1994, p. 30). All of these mechanisms might have given artists access to even very ancient models. Root suggests, moreover, that Arts 2023, 12, 102 23 of 33 as some works are known to have been replicated for other cities in Persia, parts of the Apadana might also have been reproduced at other locations, such as Susa. 51 This would have offered another possible avenue of contact.
The above scenarios all explain how forms might have traveled, but so too did people (including artists) and goods. At the uppermost levels of governmental administration, officials and their diplomats traveled to Persia from Rome and vice versa. Three of the tetrarchs were themselves in Persia, and architectural and decorative aspects of palaces such as Diocletian's palace at Split and Shapur I's at Bishapur betray the involvement or influence of the other party. 52 Luxury goods gifted by diplomatic embassies served as surrogates for foreign taste. 53 Shapur II sent munera to Constantius II by way of his envoy in 357/8 CE, according to Ammianus (Cutler 2009b, p. 14;Amm. Marc. Res Gestae, XVII.5.2). We also know Constantius sent gifts to Arsaces of Armenia and Meribanes of Hiberia, hoping these border regions would side with Rome over Persia. 54 Other goods, less freely given, were taken as loot by the enemy in the course of war. 55 War drove the forced migration and resettlement of populations over these centuries. Inscriptions tell us that Shapur I founded new cities in Persian lands, including Bishapur and Djund-i-Shapur/Vez-az-Andiyok-Shapur to house those he had deported from conquered cities such as Antioch. 56 Khosroes I did the same three centuries later, founding a city of exiles near Ctesiphon called Veh-Antiokh-Khusro. 57 Romans also deported prisoners of war, capturing Sasanian cities and moving their populations to Thrace (Dignas and Winter 2007, p. 259). This would have enabled close contact of peoples with very different visual vocabularies, resulting in exposure to different artistic repertories and conventions. Some of those deported were builders and artisans who would have continued to ply their trades, now working alongside Persian artists for Persian patrons-or for Roman patrons, but in Persian lands. Théophylacte Simocatta discusses one such foundation where mosaics were "probably executed by the Greek artists and artisans that Justinian had sent him [Khosroes I]" (Balty 2006, p. 40).
Aside from textual sources that tell us about mass displacement of city populations, we see material evidence in the archaeological record in places such as Bishapur, where we can be fairly certain that Roman and Persian craftsmen were working alongside one another, or at the very least, that the work done there was an attempt to integrate imported subjects with local form. The excavator of Bishapur, Ghirshman, supposed that the cartoons for the mosaics were western, and that the artists who executed them were from the region, probably Syrian (Ghirshman 1956, pp. 8-9). Balty takes the opposite view, arguing that the designs were done by Persian artists familiar with local taste, while the placing of mosaic tiles was carried out by Roman or Roman-trained workers who followed the prescribed designs even though they thought them subpar (Balty 2006, pp. 36, 39). The fact that these scholars, working five decades apart, both recognized in this mosaic program the fusion of Persian and Roman elements, while arriving at opposite conclusions about which party performed which role, is itself illustrative of the blending that is evident here. Techniques and media also suggest Roman involvement at the Persian king's palace: the orthogonal plan may suggest an affinity with Roman camp design, while plaster walls, ashlar masonry, and the very use of mosaics would have been customary for construction undertaken by Roman planners (Hunnell Chen 2016, p. 235).
Outside the hardships of war, goods were bought and sold. Both empires played significant roles in the ancient silk trade, whether as supplier of raw materials, producer of textiles, or end buyer. 58 Silver was another luxury good traded across Eurasia (Cutler 2009a). These media, as well as small objects such as coins, cameos, and medallions, carried with them not only foreign materials, but also stylistic conventions, popular subject matter, and other visual codes that could be transmitted and imitated (Canepa 2009, p. 67).
This discussion on Persian-Roman exchange has centered on portability, inasmuch as we are looking at ways that subjects "traveled" to other cultures through goods and people. Of course, both Rome and Persia erected monuments that were very much immovable, but nonetheless addressed the enemy from wherever they stood. Canepa has written extensively about the way that triumphal monuments produced by each empire incorporated the visual language of the other, sometimes inverting familiar tropes in order to direct the monument's message at their enemies as much as at their own people. A relief at Naqs-e-Rostam depicting the Triumph of Shapur over Philip and Valerian, for example, shows the genuflection of Philip the Arab to Shapur I. This kneeling gesture had not been used previously in Sasanian (or earlier Persian) art, and may instead have represented a transfer of Roman iconography that depicted defeated barbarians or submissive Persians as kneeling. 59 Another version of the same subject at Bishapur includes the dead body of Gordian, incorporating the defeated emperor to draw a parallel between his subjugation by Shapur and the trampling of Ahreman at the end of time by Ohrmazd. 60 So too did Romans incorporate Sasanian imagery into their own victory monuments. The so-called "Paris Cameo", a carved sardonyx, casts Valerian opposite Shapur in an equestrian duel, signifying Valerian's defeat at Edessa. Roman artists adopted this format but reworked it to show instead the emperor Galerius triumphant against the Sasanian king Narses. 61 In neither case was the intent to borrow forms from the other culture and import them whole cloth, meaning and all, to a new image with different protagonists. Instead, the objective was to use multiple formal languages to express messages of victory, such that they were legible to friends and foes, and insiders and interlopers on the international stage (Canepa 2009, p. 75).
When we look more closely at the mechanisms of exchange, it becomes clear that one culture does not merely copy while the other exerts its artistic influence. Instead, a theme is volleyed back and forth, each time re-rendered. Neither does an ethnos have an innate or absolute identity or style; nor do they exist in a vacuum. Instead, culture and material culture are continually reshaped by vicissitudes of context and reception, coexisting and interconnected (Gates 2002, pp. 105, 119). In antiquity, this reality called for a fluidity in material expression that the east-west binaries prevalent in Classical scholarship fail to capture, and that I believe is better elucidated by re-readings from a kind of "middle ground".

Persianist Re-Readings: The Middle Ground
Given the visual affinities between the opus sectile panels and the Persepolis reliefs, as well as the many channels of exchange that might explain their concordance, the Achaemenid iconography of the lion-bull combat may have served as a model still consciously in mind for the fourth century makers and viewers of the Roman combat panels. The notion adds some precision and toothiness lacking from the usual incidental attributions of the motif to Near Eastern traditions, and invites us to look to the meanings of the Persepolitan prototypes to shade in our understanding of the later Roman renditions of the subject.
We must imagine a viewer equipped to read the panels as a Roman would, but also to read into them the allusion to Persepolis. This suggests a late antique "Romano-Persian visuality", one informed by Persian culture as well as Roman. Such a sensibility would be shared by Romans who had spent time in Persia, who married Persian women, or who lived as forced migrants in new and unfamiliar cities. So, too, would it be accessible to Persian dignitaries who visited Rome, collectors of luxury goods, and those who read the witness of writers such as Ammianus Marcellinus (Ammianus 1963). These viewers could flexibly bear in mind multiple image canons, effectively code-switching depending on the context of their associations.
Among such viewers, the panels may have imparted associations with related Persian themes such as hunting, the lion hunt in particular being a kingly sport and a visual shorthand for the kingly maintenance of order over chaos. These would have had resonance to Roman eyes as well, thanks to the Roman inheritance of varied eastern traditions of royal hunting, though their royal appurtenance was excised in Roman use. In addition, the monumental precedent of the Apadana as a royal structure where the King of Kings received supplicants lent a share in that royal glory, without Roman patrons making explicit claims to a throne and risking the great displeasure of the emperor. The animal combats allowed him to underscore his role as patron and munerarius, while also invoking a distant tie to elite and kingly virtus. The language of Persepolis elevated this further: by reference to a royal monument where Xerxes and Darius had held court, the decoration cast the dominus alongside those kings and situated him as if at the Achaemenid capital. However dazzling a hall might have been when it was bedecked in varieties of gleaming marble, its reference to the Apadana invoked the scale and grandeur of that many-columned hall, the antiquity of its foundation, its fame in the ancient world, and its identification as a place of supplication and tribute-bearing.
If we bring to bear Margaret Cool Root's reading of the symplegma as an emblem of balance and mutuality, we have another lens on the animal combat theme: not only as a depiction of the natural world and its appointed order, or of the constructed world of the arena, but of a contemporary world in which Romans and Persians remained enemies locked in a cycle of combat over borders and dominance, each making occasional gains but not overturning the balanced scale. This accords well with Canepa's paradigm of visual negotiation between the two empires. Each celebrated individual victories as if they were total, but also begrudgingly regarded the opponent as worthy rival, even sibling.
The decoration may also signal the Roman framing of Persian identity as universal or non-specific. Though Roman conception conceded Persians were distinguished among barbarians, it did not admit of the idea that Persians were to be distinguished from one another. It seems neither artistic nor literary depictions differentiated between Achaemenids and Parthians or Arsacids and Sasanians, viewing each empire as a continuation of one larger category subsumed under the designation "Persian" (Sommer 2017, pp. 349-51). Romans felt themselves to be up against a long-abiding empire, fighting a perennial foe. This made the Sasanians not merely the rival of the moment, but a fulfillment of the archetypal eastern foe. While Greek and Roman writers cast them as if they strove to continue their predecessors' legacy, there is some disagreement over whether the Sasanians consciously viewed themselves as heirs of the Achaemenids-or indeed, whether they were conscious of the Achaemenids at all (Daryaee 2006). Canepa believes the Sasanians not only were aware of Achaemenid history, but that they drew from this past for inspiration, even if they chose to underscore an Avestan past over a Persian one (Canepa 2009, p. 44;Canepa 2010a). The monumental landscape of the Achaemenids would still have been very present to the Sasanians (Canepa 2009, p. 43;Daryaee 2006, pp. 500-3). Dignas and Winter (2007, pp. 56-60) argue that the foreign policy of the Sasanians was shaped by a desire to return to the extent of Achaemenid territorial claims, and that monuments such as trilingual inscriptions consciously recalled the epigraphic habits of their Achaemenid predecessors.
If the Sasanians did have some conception of themselves as the Achaemenids returned, the use of themes from the Achaemenid capital at Persepolis by their enemy would have had a particular potency. For a Roman patron to take up a markedly Persian composition like this one-and in some instances even to depict animals of Persian origin in that composition-was to arrogate it and all that it contained to Roman culture. It served to insinuate superiority, to flaunt fluency in Persian style, and to suggest that however great Persia might be, it could be subsumed under a Roman umbrella and contained within the frame of Roman decoration. The specificity of the model made the emblem not just amorphously Near Eastern, but recognizably Achaemenid. This possibility expands Canepa's proposition of negotiation through manipulation of contemporary visual material: not only were the Romans appropriating or repurposing contemporary material from their enemy, they were pillaging the ancient visual canon of that enemy's ancestors and grafting their own universalizing claims onto it. Whether received as homage or aggression, this was spoliation-by-transposition, a retrojection of Roman authority onto the Achaemenid past.
Perhaps it is significant that the revetments recall the ancient capital of Persepolis at a moment when Rome had lost its grip on its position as capital of the empire. Bringing an allusion to Persepolis in may have been one way to maintain a tie to an illustrious capital, and to fold its prestige into a universalizing concept of Rome as everywhere and nowhere at once. 62 Even if the Roman patron or viewer did not mark the distinction between empires, his Persianist viewpoint allowed him to see Persepolis as sufficiently metonymic to speak to contemporary Sasanian realities even though it came out of a culture almost a millennium older.
Balty would see these late antique monuments as part of the trend in decoration "à la perse", alongside mosaics at Antioch and Daphne that incorporate animals with royal associations in Sasanian culture. These animals are cited as "loans" of iconography and motifs from the east, increasingly more popular in fifth and sixth century Romano-Byzantine mosaics (Balty 2006, p. 40). While Balty is to be credited for granting the east some agency in the east-west exchange, rather than casting its peoples as passive recipients of Hellenization, this characterization may not be sufficient to capture the scope of meanings communicated by the Roman wall revetments or other applications of Persian-originating material. Richard White applied the term "middle ground" when theorizing the space of interaction and competition between Native American tribes of the Great Lakes region and the French settlers who occupied their land beginning in the 17th century. European settlement forced Ottawa, Miami, Potawatomi, and other native communities into a multidimensional relationship with the French as neighbors, traders, hosts, and, in some cases, kin. Sometimes in conflict and at other times in cooperation, Native Americans and Europeans "created an elaborate network of economic, political, cultural, and social ties to meet the demands of a particular historical situation" (Deloria 2006, p. 18). Arising from this relationship was not only interchange but also "new cultural production" (Deloria 2006, p. 22). The same processes occurred in the alchemical interaction of Persians and Romans in the 3rd and 4th centuries-described not as middle ground, but with terms such as reciprocity, negotiation, appropriation, and "agonistic exchange" (Canepa 2009, p. 21).
Ettinghausen uses instead the language of transfer, adoption, and integration when speaking of such "loan" material (Ettinghausen 1972, p. 1). 63 While transfer is rare and short-lived, since it meant putting an untranslatable motif into a new context with no change, adoption allows for more input and negotiation (Ettinghausen 1972, p. 1). More flexible still is integration, "a felicitous co-equal intermingling [that] could occur in an off-beat marginal region" (Ettinghausen 1972, p. 2). Ettinghausen observes that under these circumstances the once-foreign material takes on its own meanings, sometimes derived from "the selection of secondary or unusual features that suddenly took on a new significance in their new historical setting" (Ettinghausen 1972, p. 2).
Ettinghausen's "secondary features" calls to mind Richard White's "misunderstandings", referring to the misinterpretations and distortions one group makes of another's values or customs (White 2011, p. xxvi). For example, the Roman animal combat panels make substitutions that would not be compatible with the Persepolis program: lion and bull are replaced by other predators and prey. In the Roman examples, an emblem that points toward royal preeminence is embraced for a setting that is without a monarch, and yet there it is arranged with respect to other cues that signal imperial majesty and cosmic triumph . It is not identical to the Persepolitan prototype; it is not, in Ettinghausen's terms, a "transfer:" "It is well known that when iconographies cross cultural boundaries, they seldom maintain their original appearance and meaning". 64 Rather, there is enough recognizable content to identify the model, and enough changed to make it appertain to its Roman setting. In the interplay of the two, a new thing is created.
White's theory is useful not least because it permits viewers some degree of agency, rather than restricting that agency to the artist. Key to Roman reception, a multiplicity of meanings is built in: "An image or object could have as many 'authors' as it had patrons, artists, or viewers, who reinterpreted it, or contexts that provided new and unexpected meanings" (Canepa 2010b, p. 19). There could be different viewers or audiences; viewers whose perceptions were made different (transformed) by virtue of their experiences and exposures; and the objects could change, particularly in relation to their surrounds.
White's middle ground, a shared space of creation, communication, and conflict, is premised upon a world of interconnection not unlike our own. This vantage point allows us to consider a shifting spectrum of meanings to which both Greco-Roman and Persian images and audiences contributed. When we consider the co-construction of the meanings (Root 2001, p. 202). See also (Root 2003, pp. 21-22), which suggests that further valences of cyclical time and seasonal abundance are imparted by the rosettes that outline sections of the Apadana parapet frieze. Each rosette has 12 petals, matching the 12 months of the year. 27 (Ulanowski 2015, pp. 263, 265;Watanabe 2002, p. 54). On the lion hunt in Near Eastern art, see (Almagor 2021, p. 15). 28 (Root 2001, p. 200). The Achaemenids used it too: (Markoe 1989, p. 103). 29 (Root 2001, p. 202) notes some exceptions on seals. 30 Among the extant panels, the exception to this composition is the Ostian tigress, who lifts her chest and exposes it to the viewer, front legs extended as if holding its prey at a slight distance. 31 The onager, or wild ass, is easy to mistake for a horse, but can be identified by black stripes. See (Parrish 1987, p. 114 and note 4). 32 The left-facing tiger at the Basilica of Junius Bassus (MC1226) has only a single exaggerated nodule, rather than three, and incised pupils instead of inlaid ones, making the eye look more like a skillet with an egg in it. This panel may have undergone restoration when it was removed from the wall of the basilica and turned into a tabletop at Sant'Antonio, some time during the second half of the 17th century according to the report of (Ciampini 1690, p. 56). See (Hagan 2018a, pp. 95-96), and for a discussion of conservation history see (Cima and Rubolino 2000, p. 81). It is unclear whether the tigers' eyes were part of the modern intervention. This tigress may have had her front right paw wrapped around the neck of the bull, so that it was visible like in the other panels. Given the clear alteration of the head and neck of the bull, it is likely that this paw was edited out during the restoration.