Two Agrippa’s Pantheon and Its Origin

Eugenio La Rocca

The Campus Martius, the Pantheon, and Dio Cassius’s Account

The vast flat area outside the Servian Wall circumscribed by the Campidoglio, the Tiber, and the slopes of the Quirinal and Pincian hills was traditionally reserved for military exercises associated with the war god Mars, from whom the name Campus Martius was derived. Initially on the fringes of the city, the area was progressively monumentalized in the course of the Republican and Augustan periods, when prominent buildings devoted to public and political functions were laid out (Fig. 2.1). Many of these were commissioned by Augustus’s chief ally and son-in-law, Marcus Agrippa, whose military successes had generated enormous wealth. Bounded by his own public bath complex to the south and an artificial pool called the stagnum to the west, and by the Saepta Iulia (a building begun by Julius Caesar housing a voting precinct, dedicated in 26 BC), the Pantheon was constructed by Agrippa and dedicated in 27 or 25 BC.1Little is known about Agrippa’s Pantheon. It apparently suffered damage in AD 80, during a fire that devastated a sector of the Campus Martius, from the Baths of Agrippa all the way to the Temple of Jupiter on the Capitoline. Domitian carried out reconstruction work of an unknown extent,2 but lightning struck during the reign of Trajan, giving rise to the reconstruction finished by Hadrian, and for that reason the Pantheon that survives to our day is referred to here as the Hadrianic building (but see the chapter by Lise Hetland).3 An understanding of that building obviously begins with its Agrippan predecessor.


2.1. Plan showing the Campus Martius in the Augustan period. (Coarelli 1997, modified by the author)

What we know of Agrippa’s monument is deduced almost exclusively from the writings of Dio Cassius, a second-century AD historian and political leader. Born in Nicaea in Bithynia, Dio had a relatively bright career under the Severan dynasty, becoming consul twice. His accounts of the much earlier Augustan period were based on firsthand information from his predecessor Livy, from the historical accounts (breviaria) of Augustus himself, and from the autobiographies of both Agrippa and Augustus’s confidante Gaius Maecenas. In addition, Dio would have had access to the senatorial archives because of his rank. The following passage is the main evidence on which to base an interpretation of the original Augustan Pantheon, a building Dio knew only in its Hadrianic redaction:

[I]n addition [Agrippa] concluded the construction of the building called the Pantheon, given this name probably because among the statues that adorn it are included images of many gods, among which are also Mars and Venus, even though in my opinion the reason can be ascribed to the domed vault, which represents the heavens. Agrippa wanted then to place there also [a statue of] Augustus and to bestow upon him the honour of having the work named after him; but since the prince did not accept either of these two honours, he had placed in the temple a statue of [Julius] Caesar pater, while in the porch he put statues of Augustus and himself.

(Dio Cassius, LIII, 27)

Dio seems to have chosen his terms in Greek cautiously, avoiding a precise definition of Augustus’s statue and distinguishing the divinities inside the building from those in the porch. Although an exhaustive study of the changing meanings of his terminology is lacking,4 he appears to have differentiated between cult statues (agalmata and theon eikones) and statues like those for Augustus and Agrippa (andriantes), which were not intended as objects of worship.5 In describing these statues, Dio accords them religious but not divine status; more significant than merely honorary, these statues bore a close affinity with those of the interior.

Dio’s text also reveals that in line with Hellenistic tradition, Agrippa had attempted to dedicate a temple to Augustus by presenting him in the immediate context of the Olympian gods. This gesture would not have been a true divinization (which neither the moment nor Roman tradition permitted), but rather an attempt to offer a statue of the princeps (first citizen) to the immortal gods. Worship rendered to the gods would thereby accrue to Augustus, who was thus allied to a god according to formulas established in the Hellenistic period by Alexander the Great and his successors.6 A location within the temple would have been interpreted as an unacceptable sign of divinization before death, yet the potential for divinization was implicit in placing the statues of Agrippa and Augustus in the porch.

Inside the cella, the statue of Julius Caesar stood in the company of gods that included Mars and Venus, progenitors of the Romans and of the gens Iulia, the family line of Caesar and his adopted son. Dio’s text makes it clear that Caesar was placed in the temple both as a god and the adoptive father of Octavian/Augustus.7 Although less blatant, the Hellenistic custom of promoting a ruler cult pervades the entire figurative program of Agrippa’s Pantheon. Had the statue of Augustus been placed in the cella in keeping with Agrippa’s initial idea, the emperor’s quasi-divine status would have been apparent; consistent with this is Dio’s remark that the Pantheon should have been called Augousteion, following the Greek custom of naming a sacred building after the divinity venerated there. In the end, the name Pantheon prevailed (at least from the Neronian period onward) indicating worship of all the gods.8 Among those gods was the divine Caesar, a mortal deified after his assassination; at its doors were statues of his living adopted son and his general, awaiting admission to the cella. These implications were encouraged by the building’s location. Wedged between the Saepta Iulia to the east and the stagnum Agrippae to the west, the Pantheon stood in the palus caprae, or Goats’ Marsh, a place predestined, it seems, for apotheosis. In fact here, according to one tradition, Romulus experienced his consecration and ascent to the heavens.9 The site was thus redolent of kindred associations that could have attached themselves to Augustus and his right-hand man.

The Physical Traces of Agrippa’s Pantheon: The Investigations of Chedanne and Armanini

Excavations conducted in the late nineteenth century by the architects Georges Chedanne and Pier Olinto Armanini and published in incomplete form by Luca Beltrami formed the basis for a modern understanding of Agrippa’s Pantheon, although their implications were not properly understood at the time.10 These excavations revealed the following:

1. The remains of a preexisting podium survive below the podium of the Hadrianic colonnade (see Plate I). Of noticeably greater width (Fig. 2.2), this earlier podium was made of concrete with chips of Monteverde tufa and occasional sections of travertine and brick, and was mostly covered by a massive cap of travertine blocks up to 3 meters deep.11 On the front, where the edges of the pre-Hadrianic and Hadrianic podia coincide, thick load-bearing travertine piers were incorporated into the concrete. By a simple upward extension, they were adapted to support the Hadrianic colonnade (Fig. 2.3a). Inside the portico, the Hadrianic columns rest on a concrete foundation constructed with a reinforcing system inserted into the preceding podium (Figs. 2.2, 2.3b, 2.4, and seePlate XI).12

2. A paved floor once existed in the entrance portico at a depth of approximately 1.50 meters below the current pavement. To judge from the remains of a bed on which impressions of paving slabs are visible, this floor inclined downward from north to south (Figs. 2.3b (arrows) and 2.5, C, D).13

3. Another pavement existed inside the rotunda, at a depth of 2.15 meters below the current one (see Plates XIX, b'–b", and XX, b"). Here again, impressions of paving slabs were preserved. The discovery of two fragments of pavonazzetto (a white marble with purple or yellow veining), along with a third, unidentified one, indicates that these slabs consisted of prized colored marbles.14 The slabs were set along an east–west and north–south alignment, with alternating intervals suggesting some sort of floor pattern (seePlate XVIII, red arrows).15

4. A cohesive mass of concrete survives 1 meter below the preceding pavement and 3.15 meters under the current level. With a thickness of around 1.20 meters, this raft rested on a stratum of alluvial clay belonging to an old bed of the Tiber River (see Plates XIX, e, and XX, g).16

Past readings of this evidence led to erroneous conclusions regarding both the chronology and form of the Agrippan Pantheon, on account of a series of flawed premises. First, it was assumed that the edifice faced or opened to the south, and had a rectangular cella arranged in transverse fashion on the site of the present portico (see Fig.1.3). Second, it was assumed that the colored marble pavement 2.15 meters below the current floor belonged to an area thatwas open to the sky.17 Third, it was assumed that this space constituted a sort of forecourt to the rectangular temple cella. Fourth, it was assumed that the layer of concrete 3.15 meters under the current pavement represented the foundation for the paving of the present edifice, which the investigators considered to be Agrippa’s Pantheon.


2.2. Plan of excavations under portico in 1892–1893. A is edge of pre-Hadrianic podium; B is edge of Hadrianic concrete foundations; E is return to south of pre-Hadrianic foundation. (Pier Olinto Armanini in Beltrami 1898, Fig. XXXV, with new annotations)


2.3. a) Transverse section showing column bases and foundations of portico built over their pre-Hadrianic counterparts, and b) longitudinal section of trench cut in the central aisle of portico, revealed in excavations of 1892–1893; arrows indicate the pre-Hadrianic pavement. (Pier Olinto Armanini in Beltrami 1898, Figs. XIV and XII)


2.4. Section of pilaster framing east side of entrance portal and substructure, revealed in excavations of 1892–1893. a = substructure of present floor; b = substructure of pre-Hadrianic floor. (Pier Olinto Armanini in Beltrami 1898, Fig. X)


2.5. Annotated sketch of structures exposed under portico during excavations of 1892–1893. A = travertine extension of original pier foundations of facade; B, G = supplementary concrete foundations for inner portico columns; C, D = pre-Hadrianic pavement. (Pier Olinto Armanini in Beltrami 1898, Fig. XIII)

It was Rodolfo Lanciani who first posited the existence of a sort of vestibule that was circular in shape (see Plate XIV). He based this conclusion on the presence of a section of wall in reticulated masonry with a curved top – the so-called muro cordonato (encircling wall) discovered during the excavations and attributed to the Augustan building phase (Fig. 2.6, A, and Plate XX).18 According to his theory, in Domitian’s time this space constituted a forecourt which, via a flight of steps (under the intermediate block of the present building), led up to the south-facing oblong building located under the existing portico (see Fig. 1.3). Despite well-founded objections,19 Lanciani’s almost-undisputed authority enabled these ideas to persist until quite recently.20


2.6. Partial plan of 1892–1893 excavations in southeast quadrant of rotunda. The cross-hatching indicates the so-called muro cordonato or encircling wall attributed to the Augustan phase. (Pier Olinto Armanini in Beltrami 1898, Fig. XXV)

The 1996–1997 Excavations and Their Interpretation

Curiously enough, the excavations of Chedanne and Armanini did not in fact corroborate Lanciani’s reconstruction. Nor was it vindicated by the investigations of Paola Virgili and Paola Battistelli along the facade of the Pantheon in 1996–1997, which were published with the support of Giovanni Joppolo’s surveys.21 This recent activity has advanced our knowledge of the early structures in several respects.

The recent archaeological investigation brought to light the staircase at the front of Hadrian’s temple, which is now obliterated by the modern piazza, as well as the paving of the forecourt that lay in front of the building (Fig. 2.7, e).22 The stairs rose 1.30 meters above the Hadrianic paving of the forecourt to the temple’s podium (Fig. 2.7, d). The staircase was composed of seven steps and was adjoined by a fountain at each end. These fountains were composed of Proconnesian marble basins, with brick foundations to support statues from which issued the waters of the Acqua Vergine (Fig. 2.8). The fountains should be connected with the Pantheon as rebuilt after Agrippa’s time, for beneath their remains lay traces of a preexisting building. Importantly, that building had a similar configuration to the one we admire today, except that the staircase, of which traces survive under the present one, was composed of 11 steps (Figs. 2.7 and 2.8).23


2.7. Section reconstruction of two stairs belonging to the (e) present Pantheon and (f) Agrippan Pantheon; a and b denote surrounding level and podium of pre-Hadrianic building; c and d are those for Hadrianic building. (Drawing by Giovanni Joppolo)


2.8. Actual plan of portico (top) compared to Agrippa’s portico (bottom). (Drawing by Giovanni Joppolo)

The evidence recently excavated demonstrates conclusively that the pre-Hadrianic building faced north, as Chedanne had believed and as Heinrich Nissen perceptively conjectured.24 Moreover, the columns of the Hadrianic portico stand on the travertine foundation pillars of the earlier structure, and so both phases must have had comparable intercolumniations and column diameters (Fig. 2.3).25 Because the pre-Hadrianic podium was wider than the later structure, it may have had either a decastyle (10-column) facade or an octastyle facade flanked by two antae (terminal walls) (Fig.2.8).26

The pre-Hadrianic podium slopes appreciably toward the south, as Beltrami and Chedanne detected (see Plates XI, XIX, and XXI). The cause, it would seem, was the weight of the superstructures, and the slope was partially corrected during the construction of the later foundations.27 The difference in level between the pavings 1.50 meters under the portico and 2.15 meters under the rotunda is explained by this southward slope (Fig. 2.3b).28

Although literary sources document a Domitianic reconstruction, no physical trace of it has been found, nor any carbonized debris consistent with destruction by fire.29 In any case, it is unlikely that Domitian’s intervention incorporated fundamental alterations to the structure, given that the level of the Augustan podium remained unchanged until the Hadrianic renewal.

Doris and Gottfried Gruben have hypothesized that the present entrance, with its threshold in marmo africano and great bronze doors flanked by two pilasters (see Figs. 1.8 and 7.14), belonged in modified form to the Augustan Pantheon.30 This thesis suggests that elements of the original building survived successive disasters and were carefully preserved. If true, there is all the more reason to assume that Domitian’s intervention was of little significance.31 Only the fire of AD 110 definitively destroyed Agrippa’s building. Continuity between the Agrippan portico and its Hadrianic successor would also seem to be affirmed by the reprise of the Agrippan dedication carried in the inscription under the pediment. Indeed, the inscription would be much more difficult to reconcile with a total reversal of orientation and change of structure.32

We can now turn to the part where the present rotunda stands. Its exterior face coincides closely with the wall of opus reticulatum (the so-called muro cordonato, or “encircling wall,” that Lanciani attributed to the Augustan phase; Fig. 2.6). This wall is only 61 centimeters thick and its top is rounded, indicating that it could not have supported a vaulted structure. For this reason, William Loerke conjectured an open-air or hypaethral space, perhaps ringed by a barrel-vaulted annular colonnade, an idea that has been visualized in Plate XV. The exedras of the Sanctuary of Fortuna Primigenia at Palestrina offer a parallel, as does the so-called Maritime Theatre at Tivoli (whose internal diameter is fractionally less than that of the Hadrianic Pantheon).33 In Agrippa’s building, an upper order could have masked the vault and incorporated the caryatids of Diogenes of Athens extolled by Pliny.34

Yet this hypothesis has its flaws. For example, so far as we know, spaces that were completely hypaethral – open to the sky – were not paved with colored marbles because of the inevitable weathering and wear to which they would have been subjected.35The colored marble paving observed during the soundings of 1892–1893 (Plates XI, XIX, b'–b", and XX) appeared to be cut at the edges by the foundations of the rotunda. With patterns oriented east–west, this floor covered the entire interior space and not just the ring-shaped colonnade that would have been the only suitable location for a precious marble paving. Beltrami unfortunately offered limited information in this regard, but he did record slabs of varying sizes, and nothing excludes the possibility of larger slabs of white marble or travertine in the central area, with colored marbles under a perimeter portico. But by the same token, nothing excludes the possibility that the entire area was roofed with the exception of a central space crowned by an opaion, or oculus, analogous to the effect visible today. That traces of supports for such a roof have not been found cannot be considered a conclusive counterargument to this idea, given the limited archaeological surveys carried out in the rotunda.

In any event, the reticulated muro cordonato seems to represent the outermost boundary of the building.36 Its form recalls the enclosures of funerary monuments,37 leading us to wonder if the demarcation of the Augustan building was considered binding for its Hadrianic successor. If so, the foundation of the Augustan cella must have been swallowed up by that of the Hadrianic rotunda, which would explain why no trace of the former has been discovered. In conclusion, the new structure will likely have preserved the dimensions of the preexisting building.

The Relationship between the Agrippan Pantheon and the Hadrianic Pantheon

As we have seen, the Hadrianic rotunda fits neatly inside the Augustan precinct (Plates XIV and XV), while the facade columns sit exactly over their predecessors. The Hadrianic building, like so much Roman architecture, was set out according to an elemental geometry: In the circle locating the centers of the interior columns, a square with sides of about 32 meters can be inscribed. This measurement equals the width of the octastyle colonnade of the portico, enabling us to visualize a second square touching the first (Fig. 1.5 and Plate XII). Since the porticoes of the Augustan and Hadrianic plans substantially coincide, it seems that the mathematical scheme used for the Augustan plan provided the core from which the present Pantheon developed.38 The available documentation indeed suggests that, on the whole, the plan of the Augustan Pantheon – though not the elevation – resembled that of the Hadrianic Pantheon.39

Even if the reconstruction of the Augustan Pantheon produced by the Grubens turns out to be incorrect, their astute suggestion that the portal of the present building could have come from its predecessor adds a further element of continuity between the two structures.40 Such continuity must reflect an underlying conceptual motivation going beyond the symbolic value that the Grubens note was invested in the door since Homer’s day. The possible reuse of elements from the original building and the reiteration of its plan help us make sense of the ostentatious reassertion of Agrippa’s legacy by means of the bold dedicatory inscription under the pediment that was still relevant and even suggestive after a century and a half, even in political and sociocultural conditions very different from those of the Augustan period.

Precedents for Agrippa’s Pantheon and Its Constructive System

The Augustan structure cannot have had a concrete vault. As is generally accepted, the technical conditions necessary to vault such a large space did not yet exist. The Romans’ technical boldness from the time of the late Republican period using concrete for both innovative and structurally complex designs – including those with vaulted roofs and semicircular exedrae – was certainly exceptional. However, no surviving buildings from this period remotely approach the gigantic dimensions of the Pantheon. The so-called Temple of Mercury at Baiae, in reality a bathing hall or an artistically elaborated pool around a thermal spring, still conserves its concrete dome (Fig. 2.9).41 Circular in plan, with a central oculus and large windows at the haunches of the vault, its appearance bears some similarity to the Hadrianic Pantheon. Yet its diameter of approximately 21.60 meters is less than half as big.


2.9. Temple of Mercury, Baia. (Courtesy of Luciano Pedicini)

Roman architecture in the Augustan period excelled instead in the construction of trussed roofs, as confirmed by the Diribitorium, the ancient voting hall in the Campus Martius that was famed for the wide span of its roof, some 30 meters in length; similarly, the Basilica Giulia was 101 meters long and 49 meters wide, including the side aisles.42 The exedrae of the Forum of Augustus, comparable in radius to the Pantheon, had wooden roofs covered in stucco.43 Thus, large-span wooden roofs were certainly within the Romans’ technical capability. Behind it lay a Greek inheritance of buildings that were circular in plan and roofed using timber, such as the Arsinoeion of Samothrace (although this had a relatively small diameter).44 The evidence offers sufficient grounds to visualize Agrippa’s Pantheon with a cella paved in colored marble and covered by a wooden roof illuminated by a central oculus.

Such a reconstruction is consistent with the ideology and typological heritage of the monument. The conception of the vaulted space as an imitation of the starry sky and therefore of the cosmos has precedents dated well before the Augustan period.45 Cicero mentions the sphaera constructed by Posidonius “whose single rotations reproduce the motion of the sun, the moon, and the five wandering stars that occur in the sky every day and every night.”46 The frigidarium of the Stabian Baths at Pompeii is a circular structure with a truncated conical roof, central oculus, and walls frescoed with garden motifs.47 The vault was frescoed with gold stars on a blue background in imitation of a night sky.48 At the Pompeian baths, ancient visitors found themselves in an ideal setting, a garden under a starry vault with a central eye permitting a stream of natural light. Although the structure may be tiny by comparison with the Pantheon, the issue of scale does not invalidate the iconographic relationship, nor is it an isolated example. A circular aedicule with a star-studded, domed vault also appears in Pompeii in the third-style wall decoration of one of the rooms of the House of Caecilius Jucundus.49 In the ancient world, vaults often allude to this simple symbolic message: the starry sky is like a canopy, ceiling, or vault.50 In the passage introduced at the beginning of this chapter, Dio describes the dome of the Pantheon as a symbolic representation of the heavens. An echo of such associations remains in late-antique literature as well.51

A New Parallel: Chester

An additional comparison comes from an interesting structure discovered at the Roman military base of Chester, England, datable to the Domitianic period.52 There, an elliptically planned timber-roofed building (Fig. 2.10 a, b) was composed of 12 radiating chapel-like spaces that faced inward onto a colonnaded portico that supported an attic with sculptural decoration. At the heart of the plan there was a small central open-air space where a fountain was placed. Prompted by the 12-part subdivision of internal space, investigators have proposed a polyvalent religious character for the site. This was likely a celebration of the lineage of Augustus in the company of the gods – including Augustus and Caesar – but perhaps also a physical and symbolic image of the world.53 The overall length of the elliptical building, around 40 meters, is comparable to the internal diameter of the Pantheon, and one might ask if Agrippa’s project could have taken a similar form: not a simple annular ring, then, but a series of monumental “chapels” that faced a porticoed colonnade encircling a space open to the sky via an opaion. Hypothetically, where decorative roundels (clipei) may have been placed on the attic at Chester in imitation of similar features in the Forum of Augustus, the caryatids of Diogenes of Athens could have found their place in Agrippa’s Pantheon.


2.10. Plan (a) and model (b) of ancient elliptical building excavated in Chester, England. (Computer reconstructions by Julian Baum in an update of Mason 2000)

Symbolism and the ascensio ad astra of Romulus

As previously mentioned, the site of the Pantheon was associated by tradition with the apotheosis of Romulus, the first king of Rome, and as such, it served as a model for the future consecrationes of later Roman emperors. The fact that Dio Cassius interpreted the cupola of the actual building as a symbolic representation of the heavens is pertinent in this regard. Such symbolism had diverse manifestations. In one sense, it was expressed through the perfection ofmathematical relationships, in particular through the proportions of the interior. As Giangiacomo Martines explains in his chapter, these proportions correspond uncannily to Archimedes’ proposition relative to the properties of the sphere and of the cylinder since the measures of the drum and the cupola coincide (Plate XII).54Although the hypotheses that envision the rotunda as a sort of solar clock,55 or microcosmic image of the world,56 are perhaps too ambitious for the evidence, it is clear that the dome evoked the vault of the heavens. Moreover, the central opaion left open as the passage between earth and sky evokes an essential element in the process of apotheosis. Gilded bronze stars placed in the coffers of the dome may have accentuated the celestial symbolism.57 The sun’s rays enter through the opaion and, due to the effect of the rotation of the earth, illuminate the walls of the rotunda, its exedrae, and its aedicules in constantly changing ways (Plate IX). These are perhaps the fundamental elements that distinguish the Hadrianic Pantheon from related buildings.


IX. Dome and oculus. (The Bern Digital Pantheon Project)

Thus, the luminous ray of light directed toward and shining through the entrance under the sign of Torus and of the Virgin does not privilege the dates of the equinoxes (when the light falls on the zone between dome and the main order), but rather, like the orientation of the Ara Pacis, the birth date of Rome, April 21. On that date, exactly at midday, the stream of light from the opaion is centered on the entrance bay of the temple. It is as if, at precisely this moment, the emperor had entered the Pantheon illuminated by the sun as a reflection of this calculated arrangement. Like the Augustan obelisk at Campo Marzio, the Pantheon also had a strong solar association and, like the Ara Pacis, a special symbolic relationship with the date of Rome’s foundation. The movement of the sun thus exalts the connection between Romulus, first founder of the city, whose apotheosis the Pantheon celebrates, and Augustus, the new Romulus and second founder of the Urbs after decades of civil war.

Of further possible relevance to the temple’s ideological program is the recurrence of the number 7, as in the number of exedrae in the rotunda, and its multiple 28, which is the number of coffers in each ring of the dome.58Theodor Mommsen thought that the seven exedrae were occupied by statues of the seven planetary divinities.59 This theory poses some difficulty in the choice and placement of these gods, for there were other deities to take into account, including Romulus/Quirinus, the divine Caesar, and perhaps Apollo and Diana in their incarnations of Sol and Luna.60

There remains the problem of determining how much of this subtle program was understood by ancient visitors. Were they aware of the complexity of the mathematical formulas, or the symbolism embodied in the numbers 7 and 28? I don’t believe so.61Visitors instantly sensed the harmony of the space, as we do today at a distance of roughly two thousand years, having grown up with comparable images from the time of the Renaissance. The starry vault, the great hole in the dome, the apses, the aedicules, and the niches in the walls with the statues of divinities were all probably taken in without much recognition of the architect’s mathematical concerns. Visitors would no doubt have wondered about the significance of the sunbeam that moved along the walls and washed the niches and their sculptures with dramatic effect. Only subsequently, perhaps with the help of experts or local guides, would ancient visitors have sought to reconcile the number of the niches and the distribution of divinities with a familiar belief, or scheme, or programmatic logic.

It seems to me that one of the most significant elements for the ancient visitors’ cognition of the Pantheon at a semiological level is a scheme that follows the rules of a templum on a circular plan. More precisely, with its division into 16 segments (seven exedrae plus eight niches plus the entrance), the plan imitates the celestial templum according to the rules of Etruscan learning, the disciplina Etrusca, as transplanted into Roman religion, and the well-attested distribution of the gods in the sixteen regions of the celestial templum.62 Not all visitors would have appreciated the complex system that must have suggested the scheme, but certainly they would have learned some of the essential ideas of it. Perhaps the statues of the gods would have roused dormant memories and rendered comprehensible the otherwise complicated figural program. The distribution of the statues might even have reconciled the number 7 with the more or less canonical distribution of the gods in the 16 regions of the heavens.63 It is even conceivable that the luminous solar beam streaming through the opaion was intended to fall on the divine statues on their principal festival days.64 If it did not, the potential for such an effect remained as appealing for the ancients as it does for present-day visitors.

Agrippa’s Pantheon: Greek or Roman?

Dio referred to the possibility of naming the edifice after the emperor. Just as the Greek term Augousteion refers to the divinized Augustus honored in a manner consistent with Hellenistic tradition, so too Dio’s Pantheion implies an association of divinized mortals with the Olympian gods. The most accurate interpretation of the term Pantheion can be deduced from an inscription on one of the altars in the sanctuary of Demeter at Pergamon,65 which points to a translation as “temple of all the gods” rather than simply “extremely sacred.” Both the cult of the 12 gods and the cult of all the gods operated as flexible systems, subject to local variations.66 In the religious attitudes behind these cults, there is a certain liberty that contrasts with traditional cults, which typically could not be altered without loss of civic identity.

Just as a new divus, King Eumenes II, was included among the 12 Olympian gods in Pergamon, Julius Caesar would become the first mortal to be celebrated among all the gods in Rome. In both cases, these were probably only specific moments in the evolution of more complex forms of dynastic cults that extended to other members of the family. There is, in addition, a “cosmic” valence that should also be taken into account, as is evident in the figural program of the Great Altar of Zeus, where the battle between the gods and giants involves earth, sea, and sky. Symptomatic of a general trend, rulers play superhuman roles in the gods’ scheme for bringing peace and prosperity back to the world, along with an equilibrium between cosmic and terrestrial forces.

In this context, Edmund Thomas has correctly associated the Pantheon’s program with that realized by Antiochus I of Commagene (c. 86–36 BC) in the construction of the hierothesion on Mount Nemrut (Nemrud Dağ, a UNESCO World Heritage Site in modern Turkey).67 The hierothesion, a religious sanctuary, is simultaneously Antiochus’s tomb, a cult site to him and his ancestors, and also a sanctuary of all the gods.68 On the summit of the mountain, the sovereign constructed two enormous terraces, from which five divinities on thrones dominate the landscape. Four of them are syncretistic divinities of the Greco-Persian pantheon, while the fifth divinity is Antiochus himself, co-opted among the gods.69 We know that he wanted his tomb, hence his body, buried in close proximity to the gods, some of whom were planetary deities. In Rome, the Pantheon was perhaps dedicated to an analogous goal. Although the morphology of these buildings may not be comparable, it is also true that a cult of all the gods or of the 12 gods did not require a predetermined spatial configuration. Indeed the commonly accepted premise of a circular space as indicative of function as well as typology is misfounded.70 Thus, in spite of obvious differences, the important link between the ideological forms adopted by a Greek ruler and the nascent Augustan principate should not be underestimated.

Andreas Grüner has argued that Agrippa’s Pantheon was not a sacred building, but a sacred enclosure comparable less to Greek structures than to a tradition of Roman open-air (hypaethral) and centrally planned temples.71 The fundamentally Roman aspect of the project and the limited nature of comparisons with Greek buildings are not in question. Nevertheless, it cannot be a question of strictly Roman characteristics, for the very name Pantheon reveals decisive Greek connections. The case of the Tychaion of Alexandria seems to confirm those links.

The Tychaion of Alexandria

As with the Mausoleum of Augustus, parallels in the architecture and topography of Alexandria have also been cautiously advanced for the Pantheon.72 The most significant example is a structure dedicated to Tyche, goddess of fortune or prosperity, cited only in late antique documents, the earliest of which is from the fourth century AD.73 The monument apparently stood in the area of the royal palaces, by the side of the canal also called Tychaion that flowed toward the center of the city.74 The monument was described in detail in one of the ekphraseis collected under the name of Libanus of Antioch (AD 314–393), even if not by his hand.75 The text by the Pseudo-Libanus is complicated and ambiguous:

An enclosure is placed in the middle of the city, dedicated to many divinities, but on the whole it is said to be of Tyche.... The place is formed as follows: the whole is artistically elaborated from the pavement to the ceiling; the building is articulated by semicircular niches, before each of which are placed columns of various kinds. The niches were built for the exhibition of works of art, and it is possible to enumerate the niches instead of the statues; between the statues rise columns. There are not statues of all the gods, but of twelve. And a high point emphasizes the position of the founder relative to the other heights and their positions; the statue has the appearance of the Soter [royal savior] and holds that through which the city is usually nourished. And Charis stresses the nature of the earth; she is encircled by half of the statues of the gods, according to their number. Exactly at the center is placed a sculptural group depicting Tyche, who through a crown gives notice of Alexander’s victory. And Tyche crowns Gaia, who in her turn crowns the winner. Victories are at Tyche’s sides and through them the artist has emphasized well Tyche’s power, as Tyche knows all of victory. Thus the appearance of the place culminates in the laurel crown that is held by the statue. And one high on a seat discusses philosophy, another is nude and bears in the left hand an image of the heavens, but the right is held over all; nude, without clothes, he is represented as such. And bronze stelai, on which are inscribed the laws of the city, are inside on the pavement. And inside there are doors that lead to the sacred enclosure of the Muses. Kings of bronze are inside, not all those that follow one another in time, but only the most renowned of them.76

From this passage we may draw several facts and inferences: Although it survived until the seventh century,77 the Tychaion must be dated to the Hellenistic period because its center was occupied by a representation of Alexander the Great without any reference to Roman emperors. Semicircular exedrae housed statues of the 12 as-yet-unidentified gods.78 On a high point of the structure, (a high plinth, or the top of a column?), the statue of an unidentified Soter (a title associated with the Ptolemaic kings of Egypt meaning literally: savior, deliverer) held an attribute connected to fecundity.79 A statue of Charis (embodiment of grace, kindness) was surrounded by six divinities, and there was a complex sculptural group composed of Tyche, flanked by Victories, who crowned Gaia (mother earth), who in turn crowned the only mortal clearly identified, Alexander the Great.80 The location of these sculptures, like those of the unidentified kings, is a matter of conjecture. The location of bronze stelai inscribed with civic laws on the pavement of the sanctuary is equally unclear.

It is nonetheless evident that the Tychaion was a covered building of religious purpose with a variety of imagery.81 Its dedication to Tyche and the cult of the 12 gods must have been related in some way to Alexander and the unidentified Soter, both of whom are mentioned in the description. The other unamed kings mentioned were perhaps added over the course of time. Relevant to the Pantheon is the program of the Tychaion, possibly a circular building with exedras/apses containing statues of gods and of the sovereigns who bore comparison with them. The alliance of the gods and the sovereigns with political functions emerges from the presence of the civic laws inscribed on the stele of the interior.

The Pantheon as a Place of consecratio and of Imperial Veneration

Evidence for the north-facing orientation of the Pantheon of Agrippa and its similarity in plan to the present building are key elements in the interpretation of the monument and the role it played in the ideological program underlying the monumentalization of the central Campus Martius. The choice of the site can be related to the legend of the disappearance of Romulus.82 According to one version of events, the heavenly ascension of Rome’s founder occurred during a military review in a place originally called Ovile, the enclosed area for voting that was later replaced by the Saepta Iulia, not far from the palus Caprae.83 The whole area was subsequently transformed by Caesar and then Agrippa and Augustus. As a result, the marshy palus caprae was transformed into the elegant stagnum Agrippae, Agrippa’s vast artificial lake. Over time, the Saepta lost its original civic and military functions, becoming a market square and a place of diversion for idlers.84 Thus, the proximity of the Pantheon appears to hark back to mythical events associated with the very origins of Rome itself. Indeed, the reprise of the plan of the Hadrianic Pantheon in many later centrally planned mausoleums with porches, which cannot be accidental, perhaps represented a cultured reference specifically to the first Roman assumed among the gods and, more generally, the prospect of postmortem divinization.85 The right to be buried in the Campus Martius was conceded to several illustrious men by decree of the senate and at public expense for analogous symbolic reasons. Later, the northern area of the Campus Martius directly in front of the Pantheon would become sites for altars consecrated to subsequent Roman emperors.86

In topographical context, therefore, it would seem that the Pantheon operated as the focal point for an innovative religious system. It was a place of veneration of the principal Olympian divinities (probably including Romulus/Quirinus), along with the first divinized member of the gens Iulia, Julius Caesar. Judiciously, Augustus himself must have rejected the idea of presuming this honor while still alive, preferring to have his statue placed with that of Agrippa in the porch of the temple while no doubt awaiting his turn to move into the temple proper. The Pantheon was thus born with historical and dynastic functions, assimilating episodes, observances, and associative topographical values that had not previously been coherently linked: the apotheosis of Romulus; the presence of tombs of illustrious men honored by service to the Republic; the sanctification of Julius Caesar, with its echoes of the death and divinization of Romulus;87 and the future apotheosis of Augustus, the new Romulus by virtue of founding a new Rome, and thus heir to his immediate and distant forebears.

The eagle holding a crown of oak in its talons that decorated the pediment of the Pantheon, according to the interpretation of the fastening holes by Lucas Cozza, recalled the moment recounted by Suetonius (Vita Augusti 97) when a real eagle landed on the tympanum of the temple, the portent that had augured the death and divinization of Augustus (Plate XVII).88 The oculus in the dome presented that union of earth and sky that symbolized an apotheosis into the heavens. The space to the north of the Pantheon, which the pediment overlooked, was later occupied by the Temples of Matidia and Hadrian and the altars of imperial consecratio destined for the massive funerary pyres of deceased emperors. The symbolic ritual that took place for these occasions is well known: The pyre was lit and when the flames reached the summit, they consumed the knots that held an eagle imprisoned in a cage. The eagle, finally free, flew off carrying with its mighty wings the divine aspect of the emperor liberated from his mortal remains.89

Connections with the Mausoleum of Augustus

It is in this context that the relationship between the axis of the Pantheon and that of the Mausoleum of Augustus becomes significant (Plate XVI and Fig. 2.11).90 Because the two monuments were constructed at the same time – the Mausoleum was almost finished in 23 BC, at which time it received the ashes of Marcellus, while the Pantheon was dedicated by Agrippa in 27 or 25 BC – this axial correspondence cannot be accidental. It composes a unified urbanistic project whose functions may have incorporated a proportional relationship, too: The diameter of the Mausoleum is 300 feet at the base, while that of the Pantheon is 150 feet measured around a ring passing through the center of the columns.91 Previously, Augustus had built the Ara Pacis, the huge sundial (horologium), and the imperial altars. The empty space between the Mausoleum and the Pantheon must have further reinforced the link between them. Though restricted by later constructions, this axis seems to have remained as a kind of avenue.


2.11. Axial alignment of the Pantheon and the Mausoleum of Augustus. (Archivio della Sovraintendenza dei Beni Culturali di Roma Capitale)

The ideological and visual relationship with his family’s mausoleum would have complemented a monument celebrating Augustus the ruler and his divine associations after the fashion of Hellenistic precedents. Initially, the dynastic intent was expressed somewhat ambiguously, since Augustus’s power at this stage was still not entirely consolidated. These circumstances may explain why only faint traces of this function survive in literary sources,92 and why, by the Severan period, the precise sense of its name was no longer generally understood. According to Dio Cassius, as we have seen, it was called Pantheon “probably” because it held the images of many gods, but for him “the main reason” for the name lay in the fact that the dome fully represented the heavens.93More than a century afterward, in a visit to Rome in AD 357, Constantius II could still admire the Pantheon as a lofty vaulted space.94 Ammianus Marcellinus, in recounting Constantius II’s visit, does not mention statues of the gods, but only those of the emperors that had come to populate the rotunda. Presumably they occupied the aedicules and the niches, perhaps at the expense of some of the original, divine occupants. The Pantheon had evidently become definitively connected with the Roman emperors, its capacious vault symbolic of their dominion over a vast tract of the earth’s surface.

I would like to thank Karen Lloyd for her translation and Mark Wilson Jones for editorial suggestions. This research was completed in the summer of 2008. Since it has not been possible to review the text and notes on the basis of more recent references, I refer the reader to my article “Augustus’ Solar Meridian and the Augustan Program in the Northern Campus Martius: Attempt at a Holistic View,” Journal of Roman Archaeology, supplementary ser. 99, 2014, for updates, revisions, and bibliographic additions.

1 J. M. Roddaz, “Marcus Agrippa,” Bibliothèque des Écoles Françaises d’Athènes et de Rome 253, Rome 1984, pp. 252–277; Dio Cassius (LIII, 27) (trans. as Dio Cassius: Roman History, by Earnest Cary and Herbert Foster, Cambridge 1917). The dedicatory inscription dates the completion of the building to 27 BC (CIL [Corpus Inscriptionum Latinarum, ed. Matthaeus della Corte, Berlin 1970], vol. 6, 896), while Dio’s passage suggests 25 BC.

2 The fire of AD 80 was said to have burned the Serapeum and the Iseum, the Saepta, the Poseidonion, the Baths of Agrippa, the Pantheon, the Diribitorium, the Theatre of Balbus, the stage of the Theatre of Pompey, the Portico of Octavia, and the Temple of Jupiter Capitolinus with the surrounding buildings (Dio Cassius, 66.24, 2). Domitian assumed power in AD 81 and a reconstruction of the Pantheon is commonly attributed to him (Eusebius of Caesarea, Chronicon, 3d century BC, ed. Alfred Schoene, Berlin 1866, a. 354, 146; Hieronymus, Chronicum Eusebii ab Hieronymo retractatum ad annum Abrahae 2395, 2105 (Malcolm Drew Donalson, A Translation of Jerome’s Chronicon with Historical Commentary, Lewiston 1996), but this is doubtful, as we shall see.

3 Paulus Orosius, “Pantheum Romae fulmine concrematum,” Historiae adversum paganos, 5th century AD, ed. Zangemeister, Leipzig 1889, vol. 7, p. 12; Hieronymus, Chronicum Eusebii ab Hieronymo retractatum ad annum Abrahae 2395, 2127 (Donalson 1996). Hadrian’s restoration: Historia Augusta, Hadrianus, 19. Elsewhere in the Historia Augusta, however, the dedication is attributed to Antoninus Pius (Historia Augusta, Antonius Pius, 8). In this venue I will not go into the question of the paternity of the “new” Pantheon, namely, whether it was begun in the time of Trajan and designed by Apollodorus of Damascus, but according to custom will refer only to a “Hadrianic” Pantheon. For this debate, see Wolf-Dieter Heilmeyer, “Apollodorus von Damaskus – der Architekt des Pantheon,” Jahrbuch des Deutschen Archäologischen Instituts 90, 1975, pp. 316–347; Mark Wilson Jones, Principles of Roman Architecture, New Haven 2000, pp. 192 ff.; A. Viscogliosi, “Il Pantheon e Apollodoro di Damasco,” Tra Damasco e Roma: L’architettura di Apollodoro nella cultura classica, ed. Festa Farina et al., Rome 2001, pp. 156–161; Hetland’s chapter in this volume.

4 Translation by Cary and Foster 1917. On terminology: Hugo Hepding, “Die Arbeiten zu Pergamum 1904–1905, Die Inschriften 2,” Mitteilungen des Deutschen Archäologischen Instituts Römische Abteilung 32, 1907, pp. 241–414; Adolf Engeli, Die Oratio Variata bei Pausanias, Berlin 1907; A. D. Nock, “Synnaos Theos,”Harvard Studies in Classical Philology, 41, 1930, p. 3, note 5, p. 23; Nock, Essays on Religion and the Ancient World, vol. 1, ed. Zeph Stewert, Cambridge 1972, p. 204, note 5; p. 218 ff.; L.Robert, “Recherches épigraphiques, VII: Décret de la Confédération lycienne à Corinthe,” Revue des études anciennes 62, 1960, pp. 324–342; p. 317; L. Robert, Opera minora selecta. Epigraphie et antiquités grecques, Amsterdam 1969, vol. 2, p. 833, note 1.

5 They are also interpreted as such by Thomas Pekáry, Das römische Kaiserbildnis im Staat, Kult, und Gesellschaft, dargestellt anhand den Schriftquellen, Berlin 1985, p. 57.

6 Nock 1930; Nock 1972.

7 Lacking sons, the dictator had, in his will, adopted Gaius Octavius, son of his niece Atia and of Gaius Octavius, a figure of equestrian rank.

8 Adam Ziolkowski (“Was Agrippa’s Pantheon the Temple of Mars in Campo?” Papers of the British School at Rome 62, 1994, pp. 267–282) considers the term Pantheon to be a nickname, and that the building was originally a temple dedicated to Mars. This does not seem convincing, however, since his theories presume that Agrippa’s building was T-shaped and faced south, an idea now discredited. It is also unclear by what mechanism a temple dedicated to Mars was then rededicated to all the gods in the space of a few decades. The most reliable proof of the use of the name Pantheum comes from the Acta fratrum Arvalium of AD January 58 and January 59 (CIL, vol. VI, 2041, l. 50; John Scheid, Paola Tassini, and Jörg Rüpke, Recherches archéologiques à la Magliana. Commentarii Fratrum Arvalium qui supersunt. Les copies épigraphiques des protocoles annuels de la confrérie arvale (21 av.–304 ap. J.-C.), Rome 1998, p. 63, n. 26, line 23 (AD January 11, 58); p. 67, n. 27, line 50 (AD January 12, 59). See also the commentary of Wilhelm Henzen, Nuovi frammenti degli atti dei fratelli arvali, Rome 1867, p. 258; Henzen, “Eine Neue Arvaltafel,” Hermes 2, 1867, pp. 37–55. The same name also appears in Pliny the Elder’s Naturalis Historia (trans. H. Rackham, Cambridge 1952), 9, 121; 34, 13; 36, 38, at a date which preceded the eruption of Vesuvius in AD 79, when Pliny met his death.

9 Varro, Rerum rusticarum, III, 2; Cicero, Epistulae ad Atticum, I, 33; Livy, Ab Urbe Condita XXVI, 22, 11. See also Kjeld De Fine Licht, The Rotunda in Rome: A Study of Hadrian’s Pantheon, Copenhagen 1968, pp.191 ff.; Filippo Coarelli, “Il Pantheon, l’apoteosi di Augusto e l’apoteosi di Romolo,” Città e architettura nella Roma imperiale, Analecta Romana Instituti Danici 10, 1983, pp. 41–46 (updated in Coarelli, Il Campo Marzio: dalle origini alla fine della reppublica, Rome 1997, 195); Ferdinando Castagnoli, Topografia antica, Rome 1993, p. 251.

10 Luca Beltrami, Il Pantheon: La struttura organica della cupola e del sottostante tamburo, le fondazioni della rotonda, dell’ avancorpo, e del portico, avanzi degli edifici anteriori alle costruzioni adrianee. Relazione delle indagini eseguite dal R. Ministero della Pubblica Istruzione negli anni 1892–93, coi rilievi e disegni dell’ architetto Pier Olinto Armanini, Milan 1898; Luca Beltrami, Il Pantheon rivendicato ad Adriano 117–138 d.C., Milan 1929. For Chedanne, see William C. Loerke, “Georges Chedanne and the Pantheon: A Beaux Arts Contribution to the History of Roman Architecture,” Modulus, 1982, pp. 40–55, some of whose plans are published in Roma Antiqua: “Envois” degli architetti francesi (1786–1901). Grandi edifici pubblici, exhib. cat., Rome 1992, pp. 124 ff., nos. 71–76.

11 Pier Olinto Armanini’s plan, published in Beltrami 1898, Fig. XXV, formed the basis for later reconstructions that erroneously give the building a south-facing facade: Rodolfo Lanciani, Rovine e scavi di Roma, Rome 1897, Fig. 185; Armin von Gerkan, “Das Pantheon im Rom,” Von antiker Architektur und Topographie 60, 1959, pp. 273–277, Fig. 1; Heinz Kähler, Der Römische Tempel, Berlin 1970, Fig. 9; Doris Gruben and Gottfried Gruben, “Die Türe des Pantheon,” Mitteilungen des Deutschen Archäologischen Instituts, Römische Abteilung 104, 1997, pp. 3–74, Fig. 29.

12 Antonio Maria Colini and Italo Gismondi, “Contributo allo studio del Pantheon: La parte frontale dell’avancorpo e la data del portico,” Bullettino della Commissione Archeologica Comunale di Roma 44, 1926, pp. 67–92. See specifically pp. 87 ff.

13 Beltrami 1898, p. 45, Figs. X–XIII; Lanciani 1897, p. 482.

14 Beltrami 1898, pp. 38 ff., Figs. VIII, XXXIV; Beltrami 1929, pp. 52–53, Plate XVI; Licht 1968, pp. 172 ff., Fig. 193; Paola Virgili and Paola Battistelli, “Indagini in piazza della Rotonda e sulla fronte del Pantheon,” Bullettino della Commissione Archeologica Comunale di Roma 100, 1999, pp. 137–154, note 33. Lanciani also mentions slabs of giallo antico, without citing evidence.

15 Beltrami 1898, p. 39. In the gallery toward the south, he states, “the impressions are present at a distance greater to those already noted (namely, in the eastern arm of the tunnel),” while in that toward the north was “found another fragment of marble still in place”; finally, in the western gallery were “discovered impressions of seams of a marble pavement at a greater distance.” Cf. Licht 1968, p. 175, Fig. 193; Virgili and Battistelli 1999, p. 140, Fig. 2.

16 The bed was made up of “stratifications of fluvial clay, of which the deepest was very compact and bluish-grey”: Beltrami 1898, pp. 37 ff., Fig. XXXIV (our Plate XX); Beltrami 1929, p. 62, Plate XVI; Gruben and Gruben 1997, p. 59, Fig. 30 a; Virgili and Battistelli 1999, Fig. 4, H–I.

17 Beltrami 1898, pp. 38, 54 note 1, 72 ff.; Lanciani 1897, p. 482. Edmund Thomas, “The Architectural History of the Pantheon in Rome from Agrippa to Septimius Severus via Hadrian,” Hephaistos 15, 1997, p. 169.

18 Lanciani 1897, pp. 482 ff., Fig. 185; Beltrami 1898, pp. 64 ff., Figs. XXV, XXVI (where the wall is marked with the letter c), XXIX, XXX (where the same wall is marked with the letter D); Beltrami 1929, pp. 62 ff., Plate XVI; Loerke 1982, p. 47, Fig. 8; E.Tortorici, “L’attività edilizia di Agrippa a Roma,” Il bimillenario di Agrippa, Genoa 1990, pp. 38, 40, Fig. 10; Gruben and Gruben 1997, pp. 60 ff., Figs. 29, 30; Thomas 1997, pp. 168 ff. For doubts about dating this wall solely on the building technique, and in the absence of photographic documentation and stratigraphical data, see Virgili and Battistelli 1999, p. 148, note 42. It may also be noted that reticulate masonry was still used in several walls of Hadrian’s Villa. It is also true, however, that the rounded top of the wall appears in many funerary enclosures of the late-Republican and Julio-Claudian periods, but is virtually unknown later, which makes an earlier, Agrippan, date more likely.

19 G. Lugli, La tecnica edilizia romana, Rome 1957, pp. 13 ff.; William C. Loerke, “A Rereading of the Interior Elevation of Hadrian’s Rotunda,” Journal of the Society of Architectural Historians 49, 1990, pp. 22–43; Loerke 1982; Tortorici 1990, pp. 28 ff.; Thomas 1997; Gruben and Gruben 1997; Wilson Jones 2000, pp. 180–182; Gerd Heene, Baustelle Pantheon: Planung, Konstruktion, Logistik, Düsseldorf 2004.

20 Lanciani 1897, pp. 480 ff., Fig. 185. Cf. Licht 1968, p. 177, note 32, Fig. 194; Coarelli 1983, pp. 41 ff.; Castagnoli 1993, pp. 248 ff.; Gruben and Gruben 1997, 59 ff., Fig. 29. Partially flawed interpretations also appear in Thomas 1997, p. 169; AndreasGrüner, “Das Pantheon und seine Vorbilder,” Mitteilungen des Deutschen Archäologischen Instituts, Römische Abteilung 111, 2004, pp. 495–512. Lanciani’s hypothesis of a circular forecourt was amended as a horseshoe-shaped plan by Doris and Gottfried Gruben (1997).

21 Eugenio La Rocca, s.v. “Pantheon (fase pre-adrianea),” in E. M. Steinby., ed., Lexicon Topographicum Urbis Romae, Rome 1995–1999, vol. 5, 1999, pp. 280–283; Paola Virgili, s.v. “Pantheon: età adrianea,” in Steinby 1995–1999, 5, Rome 1999, pp. 284–285; Virgili and Battistelli 1999. I wish to take this opportunity to pay tribute to the work of Giovanni Joppolo in the service of archaeology in the city of Rome, and his drawings of exemplary precision and expertise.

22 Virgili and Battistelli 1999, pp. 149 ff., Figs. 1, 3, 4, 6 C and D, 8.

23 Virgili and Battistelli 1999, pp. 142 ff., Figs. 1, 3, 4, 6 A and B, 7.

24 H. Nissen, Orientation. Studien zur Geschichte der Religion, vol. 3, 1910, p. 339. Even if only partially correct in his reading of the work of Chedanne and Armanini, Lugli (1957, pp. 13 ff.) also believed that the Agrippan building faced toward the north and that it had a circular covered cella. For partly similar conclusions reached by other means, see Loerke 1982, pp. 47 ff.; Tortorici 1990, pp. 28 ff.; Thomas 1997, pp. 167 ff.; Wilson Jones 2000, p. 182; Heene 2004, pp. 16 ff. Gruben and Gruben (1997, p. 72, note 217) have on the contrary rejected Loerke’s hypothesis, arguing among other things that the excavations at the end of the nineteenth century had not turned up traces of older stairs to the north. However, the 1996–1997 excavations in effect vitiated this objection.

25 Loerke 1982, pp. 48 ff. Loerke’s reading was based on drawings. Beltrami 1898, Figs. XII, XIV, XV, and XXXIV, aspects of which have been incorporated in the figures here Cf. Tortorici 1990, pp. 36, 38.

26 Loerke (1982, p. 49) proposes a decastyle portico on the basis of an elevation by Chedanne, unfortunately lost, but described by R. Phené Spiers (“Monsieur Chedanne’s Drawings of the Pantheon,” Journal of the Royal Institute of British Architects 2, 1895, p. 180), on the occasion of an exhibition held in London (on which see Loerke 1982, pp. 47, 55). Cf. Thomas 1997, p. 168; Virgili and Battistelli 1999, p. 148 and note 46. The octastyle solution with antae would be the best option according to mathematical coincidences between the pre-Hadrianic and Hadrianic Pantheon intuited by Heene (2004, pp. 17 ff., Figs. 7, 9).

27 Beltrami 1898, p. 54. The difference in level is shown in Ioppolo’s north–south section (Virgili and Battistelli 1999, 145 ff., Fig. 4), which partially revises the elevations by Armanini (Fig. 19 b, b’, b’’). The original podium related to a datum around 9 meters above sea level, which constituted the ground level for the building activity of Agrippa and Augustus in the Campus Martius generally.

28 Thus, there was no change in level to justify a hypothetical set of steps leading up to an imagined south-facing temple. See Lanciani 1897, p. 482; Beltrami 1898, p. 45, Figs. X–XIII.

29 Gruben and Gruben 1997, p. 59.

30 Gruben and Gruben 1997 passim (for the threshold, pp. 31, 54 ff.).

31 For Gruben and Gruben (1997, p. 59) the fire of AD 80 produced only limited damage, giving rise to a reconstruction that entailed an embellishment of the preexisting building: a marble podium and a pavement in precious colored marbles, approximately 1 meter above the Agrippan pavement. Cf. a note by Pier Olinto Armanini (Beltrami 1898, Plate XV), associating a possible raising of the podium with the “level of Domitian.”

32 As underlined by Gruben and Gruben (1997, p. 53): “Das stehende Bauwerk ist ohne Zweifel im engsten Sinne der Nachfolger des von Agrippa 27 v. Chr. geweihten Pantheon.” They also note (p. 55) that the symbolic ties between the two structures were not confined to the portal.

33 Loerke 1982; Loerke 1990. Cf. D. M. Jacobson, “Hadrianic Architecture and Geometry,” American Journal of Archaeology 90, no. 1, 1986, p. 84. For a schematic reconstructive drawing based on Loerke’s hypothesis, see Heene 2004, p. 16, Fig. 6.

34 Pliny the Elder, Naturalis Historia, 36.38. Cf. Gruben and Gruben 1997, pp. 58 ff. and note 155.

35 Similar observations were put forward by Lugli 1957, p. 14, and by Wilson Jones 2000, p. 182.

36 Tortorici 1990, pp. 38, 40.

37 Such walls tend to have a double-sloped or curved top. For those of the tombs of the necropolis of Porta Nocera at Pompeii, see Antonio D’Ambrosio and Stefano De Caro, Un impegno per Pompeii: Fotopiano e documentazione della necropoli di Porta Nocera, Milan 1983.

38 Heene 2004, p. 18, Fig. 9; p. 20, Fig. 10. For the mathematical scheme of the Hadrianic Pantheon, along with associated formal and conceptual implications, see Mark Wilson Jones 1989b (“Principles of Design in Roman Architecture: The Setting Out of Centralised Buildings,” Papers of the British School at Rome 57, 1989, pp. 108, 118, 127, Fig. 5, Table 1); Wilson Jones 2000, pp. 184 ff., Fig. 9.11.

39 For similar conclusions, see Tortorici 1990, pp. 40, 42; Loerke 1990; C. J. Simpson, “The Northern Orientation of Agrippa’s Pantheon: Additional Considerations,” L’antiquité classique 66, 1997; Thomas 1997; Wilson Jones 2000, p. 182.

40 Gruben and Gruben 1997.

41 A. Maiuri, “Restauro di una sala termale a Baia,” Bolletino d’arte 36, 1930, pp. 359–364; Licht 1968, pp. 205 ff., 214, 216, Figs. 206–207; Friedrich Rakob, “Römische Kuppelbauten in Baiae. Die Gewölbeprofile,” Mitteilungen des Deutschen Archäologischen Instituts, Römische Abteilung 95, 1988, pp. 257–301, Figs. 1–9, Plates 102, 1, 4; 103–107. Friedrich Rakob, “The Vaults of Baia,” in Civiltà dei Campi Flegrei, Atti del Convegno Internazionale, ed. Marcello Gigante, Naples 1992, pp. 229–258; pp. 237 ff., Plates 5–10. The rotunda at the Sanctuary of Fortuna at Palestrina, undoubtedly older (end of the second century BC), was much smaller, although it did have a dome decorated with coffers and a central oculus. See Friedrich Rakob, “La rotunda a Palestrina,” Urbanistica ed architettura dell’antica Praeneste: Atti del convegno di studi archeologici, Palestrina 1989, pp. 87–113; Friedrich Rakob, “Die Rotunde in Palestrine: mit einer Bauaufnahme und Rekonstruktion von Mertin Kleibrink,” Mitteilungen des Deutschen Archäologischen Instituts, Römische Abteilung 97, 1990, pp. 61–92.

42 The roof of the Diribitorium, a building that measured ca. 43 meters in width externally, and which was ca. 120 meters long, was constructed with larch beams 100 feet long and 1.5 feet thick. An evident marvel, one of the unused beams was placed in the Saepta. See M. Torelli, s.v. “Diribitorium” in Steinby 1995–1999, vol. 3, Rome 1997, p. 18; Tortorici 1990, p. 40; M. Pia Muzzioli, “I lavori per la via Nazionale e il Diribitorium,” Rivista dell’Istituto Nazionale di Archeologia e Storia dell’Arte 18, 1995, pp. 139–167; Coarelli 1997, pp. 155 ff. The central hall of the Basilica Giulia, 30 meters high, measured 75 x 16 meters in the plan and carried a roof made of wooden trusses. See H. Lauter, “Zwei Bemerkungen zur Basilica Julia,” Mitteilungen des Deutschen Archäologischen Instituts, Römische Abteilung 89, 1982, pp. 447 ff.; C. F. Giuliani and P. Verduchi, s.v. “Basilica Julia,” in Steinby 1995–1999, vol. 1, Rome 1995, pp. 177–179.

43 Valentin Kockel, s.v. “Forum Augustum,” in Steinby 1995–1999, vol. 1, Rome 1995, pp. 289–295.

44 The outer diameter of the circular foundation is ca. 20.30 meters, and the width of the foundation walls is 2.44 meters. See J. R. McCredie, G. Roux, S. M. Shaw, and J. Kurtich, Samothrace 7: The Rotunda of Arsinoe, Princeton 1992; Wolfram Hoepfner, “Zum Arsinoeion auf Samothrake,” Archäologische Anzeiger, 2001, pp. 467–480. On the wooden roofs of Greek tholoi, see H. Pomtow, Die grosse Tholos zu Delphi und die Bestimmung der delphischen Rundbauten: eine architekturgeschichtliche Studie, Leipzig 1912, pp. 216 ff.; H. Thiersch, “Antike Bauten für Musik,” Zeitschrift für Geschichte der Architektur 2, 1909, pp. 33 ff.; J. Charbonneaux, Fouilles de Delphes II: Topographie et architecture: le sanctuaire d’Athèna Pronaia, Paris 1925; J. Charbonneaux, “Tholos et prytanée,” Bulletin de Correspondance Hellénique 49, 1925; K. Lehmann, “The Dome of Heaven,” Art Bulletin 27, 1945, pp. 1–27; p. 20.

45 Lehmann 1945.

46 Cicero, De Natura Deorum, 88 (ed. and trans. H. Rackham, Cambridge 1933).

47 August Mau, Pompeji in Leben und Kunst, Liepzig 1908, p. 196, Fig. 96; Licht 1968, p. 212, Fig. 214; Hans Eschebach, Die stabianer Thermen in Pompeji, Berlin 1979, pp. 11, 58 ff., Plates 7 b, 8 a, 41; E. La Rocca and M. and A. de Vos, Pompeii, Milan 1994, pp. 308–310 and Fig. a.

48 F. Niccolini, Le case e i monumenti di Pompeii disegnati e descritti, Naples 1854–1896, Plates VI–VII; Lehmann 1945, p. 21, Fig. 59; Ida Baldassare, Pompeii. Pitture e mosaici: la documentazione nell’opera di disegnatori e pittori dei secoli XVIII e XIX, Rome 1995, p. 418, Fig. 245; M. and A. de Vos, in Eschebach 1979, pp. 85 ff., Plates 66 a–c, 67 a.

49 Lehmann 1945, p. 20, Fig. 58.

50 Robert Eisler, Weltmantel und Himmelszelt. Religionsgeschichtliche Untersuchungen zur Urgeschichte des antiken Weltbildes, Munich 1910, passim; Lehmann 1945; E. Baldwin Smith, The Dome: A Study in the History of Ideas, Princeton 1950, especially pp. 79 ff.

51 Servius, Commentarii in Vergilii Aeneidos libros, I, 505; Eisler 1910, p. 614.

52 David J. P. Mason, Excavations at Chester, The Elliptical Building: An Image of the Roman World? Excavations in 1939 and 1963–1969, Chester 2000.

53 Mason 2000, pp. 76 ff.

54 Giangiacomo Martines, “Argomenti di geometria antica a proposito della cupola del Pantheon,” Quaderni dell’Istituto di Storia dell’Architettura 13, 1989, pp. 3–10; Thomas 1997, pp. 178 ff., Fig. 8; Gerd Sperling, Das Pantheon in Rom, Neuried 1999, pp. 25 ff.

55 Frank Granger, “Julius Africanus and the Library of the Pantheon,” Journal of Theological Studies 34, 1933, pp. 157–161.

56 William Lloyd MacDonald, The Architecture of the Roman Empire, vol. 1: An Introductory Study, London 1965 (2nd ed. rev. New Haven 1982, p. 120). It was MacDonald who first suggested a reference to heliocentric theories of the universe in the structural conception of the Pantheon (p. 118: “... as the earth rotates, Hadrian’s sun-show spins on”), a theory expanded by Sperling 1999, pp. 169 ff. For skepticism, however, see Thomas 1997, pp. 181 ff.

57 Baldwin Smith 1950, p. 91 and note 139.

58 Guglielmo De Angelis d’Ossat, “La forma e la costruzione delle cupole nell’architettura romana,” Rome 1938, repr. Realtà dell’architettura. Apporti alla sua storia 1933–1978, ed. L. Marcucci et al., Rome 1982, Plate XIX. For more on this theme, seeChapter Four by Martines.

59 The hypothesis, on which Mommsen never wrote anything beyond a citation (Archäologischen Zeitung, Berlin 1867, p. 55), was referred to by Heinrich Nissen (Antiquarische Untersuchungen, Berlin 1869, p. 224), and H. Jordan (Topographie der Stadt Rom im Alterthum, Berlin 1907, pp. 581 ff., note 61). Cf. S. B. Platner, A Topographical Dictionary of Ancient Rome, ed. Thomas Ashby, Oxford 1929, pp. 382 ff.

60 Nissen 1869, p. 224.

61 The choice of 28 for the coffers may also have a compositional logic, as pointed out by Wilson Jones 2000, pp. 191 ff., in order to set up a dynamic interaction with the 16-part radial partition of the ground plan.

62 Nissen’s interpretation is stimulating, despite its inaccuracy: “the last and most complete form of the templum is the circle, so also the name urbs is strictly related to orbis and so the building, that visibly represents the order of the world, the Roman Pantheon, will be built as a centrally-planned temple and constructed with a cupola.” See Nissen 1869, pp. 150, 219 ff., and 181 ff. for the subdivision of the templum in 16 regions.

63 Nissen 1869, pp. 223–226; Nissen 1910, vol. 3, pp. 339 ff.

64 Thomas 1997, p. 174.

65 Georg Wissowa, Religion und Kultus der Römer, Munich 1912, pp. 77 ff., note 7. The inscription is in Friedrich Jacobi, Pantes Theoi, Halle 1930, p. 48.

66 The distinction between the cult of the 12 gods and the cult of all the gods was not clear-cut. See G. Ziegler, “Pantheon,” Realencyclopädie der Klassischen Altertumswissenschaft 18, no. 3, 1949; E. Will, “Dodekathéon et Panthéon,” Bulletin de Correspondance Hellénique 75, 1951, pp. 233–246. On the cult of all the gods at Pergamum, see Jacobi 1930, pp. 18, 31, 36, 48 ff., 66 ff., 96, 98, 103, 105, 108, 110, 117; E. Ohlemutz, Die Kulte und Heiligtumer der Gotter in Pergamum, Darmstadt 1940, pp. 219 ff., 281 ff. On the inscriptions in the sanctuary of Demeter: Jacobi 1930, p. 36, 9 a, b (“basi”); pp. 48 ff., e, f (“are”). Altars dedicated to all of the gods and all of the goddesses were found also in other places in the city, while an important precedent is offered by the case of Philip II and Alexander the Great, each of which was honored as a thirteenth god: Charlotte R. Long, The Twelve Gods of Greece and Rome, New York 1987, pp. 207 ff. In a nocturnal procession to Aigai [present-day Vergina], the statue of the living Philip II, similar to that of a god, was carried together with that of the 12 gods (Diodorus Siculus, Bibliotheca historica XVI, 92, 5).

67 Edmund Thomas, “From the Pantheon of the Gods to the Pantheon of Rome,” Pantheons: Transformations of a Monumental Idea, ed. Richard Wrigley, Aldershot 2004. Cf. Donald H. Sanders, ed., Nemrud Dagi: The Hierothesion of Antiochus I of Commagene: Results of the American Excavations Directed by Theresa B. Goell, Winona Lake, Ind., 1996.

68 Sanders 1996, pp. 208–214.

69 Sanders 1996, pp. 133 ff.

70 Will (1951) has also shown to be a fallacy the hypothesis that the cult of the 12 gods (and, by extension, of all the gods) was celebrated nearly exclusively in circular buildings (cf. Ziegler 1949, col. 741 ff.).

71 Grüner 2004, esp. pp. 506 ff.

72 La Rocca 1999, p. 283.

73 Giacomo Lumbroso, “Cenni sull’antica Alessandria tratti dal Pseudo-Callistene,” Annali dell’Istituto di Corrispondenza Archeologica 47, 1875, pp. 5 ff.; G. Botti, Plan de la ville d’Alexandrie à l’époque ptolémaïque, Alexandria 1898, pp. 37 ff.; A. Ausfeld, “Zur Topographie von Alexandria und Pseudokallisthenes,” Rheinisches Museum für Philologie 55, 1900, p. 367; Aristide Calderini, Dizionario dei nomi geografici e topografici dell’Egitto greco-romano, Cairo 1935, p. 155, s.v. “Tychaion”; Achille Adriani, s.v. “Tychaion,” Repertorio d’arte dell’Egitto greco-romano, Palermo 1966, Serie C, 1–2, pp. 258 ff.; P. Goukowsky, Essai sur les origins du mythe d’Alexandre (336–270 av. J.C.), Nancy 1978, p. 150; Barbara Tkaczow, “Remarques sur la topographie et l’architecture de l’ancienne Alexandrie dans les texts antiques,” Archeologia 35, 1984, p. 15; P. M. Fraser, Ptolemaic Alexandria, Oxford 1972, vol. 1, p. 242, 2, p. 392, note 417; Long 1987, pp. 84 ff., T 24. A., pp. 212 ff., 307 ff.; Andrew Stewart, Faces of Power: Alexander’s Image and Hellenistic Politics, Berkeley and Oxford 1993, pp. 243 ff., 383 ff.; Gunter Grimm, Alexandria: Die erste Königsstadt der hellenistischen Welt, Mainz 1998, p. 70; Elena Ghisellini, Atene e la corte tolemaica. L’ara con dodekatheon nel Museo Greco-Romano di Alessandria, Alexandria 1999, pp. 97 ff.

74 Pseudo-Callisthenes I, 31, 4. However, Thomas (2004) opts instead for a location in Antioch, erroneously in my view.

75 Libanius, Progymnasmata 12, Ekphraseis 25, in R. Foerster, ed., Libanii Opera VIII, Leipzig 1915, pp. 438 ff., 529 ff.; Bernhard Hebert, Spätantike Beschreibung von Kunstwerken. Archäologischer Kommentar zu den Ekphraseis des Libanios und Nikolaus(Diss. Universität Graz), Graz 1983, pp. 8 ff.

76 My thanks to Emanuele Dettori for his help in the translation and explanation of the text. The building is also mentioned apropos the dramatic events that led to the killing of the Byzantine Emperor Maurice in AD 602. In his account, Simocatta relates that a famous calligrapher, “reaching the zone of the city called Tychaion ... saw the most famous images sliding down from their bases.” See Theophylactus Simocatta, Historia, ed. Carolus de Boor, Leipzig 1887, vol. 8, 13, 342 B.

77 Besides the event of 602 described by Simocatta, a sixth-century text could refer to the sanctuary: T. D. Néroutsos-Bey, “Inscriptions grecques et latines recueilles dans la ville d’Alexandrie et aux ses environs,” Revue archéologique 3, no. 9, 1887, p. 203, n. 8. See also Christopher Haas, Alexandria in Late Antiquity: Topography and Social Conflict, Baltimore and London 1997, p. 167.

78 On the 12 gods in Alexandria, see E. Ghisellini 1999, esp. pp. 100 ff. for the sculptures of the Tychaion.

79 As to the precise identity of this ruler, the progenitor of the Ptolemaic dynasty, Ptolemy I, seems more likely than the lackluster Ptolemy X Philometor Soter II.

80 This is the preferred configuration as described by Libanus, and not that offered by Thomas (2004), who thinks that Alexander is crowned by Tyche, who in her turn was crowned by Gaia.

81 Only Stewart 1993, pp. 243 ff., 383 ff., T 95, proposes a building on a square plan. It has also been argued that the structure was an open square with two opposing hemicycles and niches in the walls. Simocatta also refers to a site called Tychaion, a term that could indicate either a building or an open-air sacred precinct (just as at Samos, the term Heraion could denote the whole sanctuary of Hera or her temple).

82 Coarelli 1983, pp. 41 ff., esp. p. 45. Cf. Roddaz 1984, pp. 275 ff.; Thomas 1997, pp. 163 ff.; Thomas 2004.

83 For sources see n. 9; Filippo Coarelli, s.v. “Caprae palus,” in Steinby 1995–1999, vol. 1, 1993, p. 234. For objections, see Ziolkowski 1994.

84 Statius, Silvae 4.5.2; Martialis (Martial) 2.14.5; 57.2; 9.59.1; 10.80.4.

85 On formal and typological affinities between the Pantheon and later mausolea see Wilson Jones 1989b, 108 ff.

86 Coarelli 1997, pp. 591 ff.

87 Angelo Brelich, “Quirinus: una divinità romana alla luce della comparazione storica,” Studi e materiali di storia delle religioni 31, 1960, pp. 63–119; Andrea Carandini, Remo e Romolo. Dai rioni dei Quiriti alla città dei Romani (775/750–700/675 a.C.), Turin 2006, pp. 299 ff., 467 ff.

88 Licht 1968, pp. 45 ff. (based on the observations of Lucos Cozza); Roddaz 1984, p. 274.

89 Penelope Davies, Death and the Emperors: Roman Imperial Funerary Monuments from Augustus to Marcus Aurelius, Cambridge 2000.

90 Nissen 1869, 226. Cf. Ferdinando Castagnoli, “Il Campo Marzio nell’antichità,” Atti dell’Accademia Nazionale dei Lincei, 1948, pp. 148 ff.; Loerke 1982, p. 51; Loerke 1990, p. 42 and note 47; Thomas 1997, pp. 174 ff., Fig. 6.

91 On the general significance of round dimensions in Roman architecture, see Wilson Jones 1989b.

92 One clue is offered by the information that the college of the fratres Arvales assembled in the Pantheon on AD January 11, 58 and January 12, 59 (see note 8). As this priesthood was concerned with the celebration of the imperial family, it can be deduced that the Pantheon had found service in this connection, at least in the Neronian period.

93 Dio Cassius, 53.27. 2–3.

94 Ammianus Marcellinus, Rerum gestarum libri, 4th century AD, ed. J. C. Rolfe, Cambridge 1956, pp. 16, 10, 14: “... velut regionem teretem speciosa celsitudinem fornicatam; elatosque vertices qui scansili suggestu consurgunt, priorum principum imitamenta portantes.” The text is, unfortunately, not without interpretive difficulty.

If you find an error please notify us in the comments. Thank you!