Five Sources and Parallels for the Design and Construction of the Pantheon

Gene Waddell

The Pantheon is considered one of the most characteristic examples of Roman architecture, but at the time it was created, it was unusual in many respects. Its design involved a novel combination of elements from a half-dozen different building types: baths, tombs, basilicas, temples, triumphal arches, and theaters. The unprecedented span required a new approach to concrete construction. Only by virtue of its success did the Pantheon become emblematic of Roman imperial monumental architecture.

Previous studies have done much to identify likely sources for the Pantheon, but so far they have focused on individual features and placed more emphasis on design than on construction. This study explores both aspects, while special emphasis is placed on clusters of features common to the Pantheon and earlier structures. When a cluster of features occurs in a recently constructed, major monument in Rome or its environs, it can be considered a likely source. Of course, other possible sources may well not have survived, and it is also instructive to examine parallels in buildings that, while they may not have influenced the Pantheon overtly, enhance our understanding of contemporary practice.

Initially, this chapter reviews the principal buildings that have previously been identified as sources. Detailed consideration is then given to significant clusters of individual elements in these and other buildings that are either earlier or roughly contemporary with the Pantheon. A conclusion summarizes the relative importance of the influences that were combined inventively so as to create a new and influential archetype.

The Four Main Design Elements

A portico, transitional block, drum, and dome are the four main design elements making up the Pantheon. Earlier buildings provide numerous examples of each of these features, but no earlier building is known to have combined all four of them. Porticoes were initially used in Roman architecture primarily for temples, of which they were an integral part under a common roof. The Pantheon’s portico was unusual in that it was attached to the front with a separate roof.

A transitional block facilitated the attachment of a rectangular portico to a cylindrical drum; this slab-like block of masonry may be likened to the walls to which Roman architects had previously attached some temples. The Temple of Mars Ultor, for example, butts up to the firewall that separates the Forum of Augustus from an adjacent neighborhood. The transitional block of the Pantheon also provided space for stairs, as had been done previously in the back of the semicircular exedras of Trajan’s Baths (discussed inChapter Seven).

Prior to the construction of the Pantheon, the largest freestanding cylindrical structures were tombs such as the Mausoleum of Augustus (Fig. 5.1a and b). Their walls were windowless, like the Pantheon’s drum. A thick circular drum also provided the most even support for a dome, while the addition of a portico helped to distinguish this project from a tomb.

5.1. Plan of the Mausoleum of Augustus (a) and plan of the Pantheon (b) at both pavement level and at the springing of the dome, not to same scale. (Wilson Jones 2000, Figs. 4.6, 9.14)

The largest number of design features adapted for the Pantheon were used previously in Imperial Roman baths (thermae), the building type that had most fully demonstrated the potential of concrete on a large scale. A dome with an oculus had been used in baths for more than a century, but only as part of a larger complex, and not as freestanding structures.1

Major Buildings of Influence

Kjeld De Fine Licht and William L. MacDonald considered the possible sources of design for the Pantheon, and both of them identified sources of influence in major imperial buildings of the preceding decades. Agrippa’s Pantheon was nottaken into account as a probable source by these authors because at the time they wrote, it was thought to have a completely different shape (see Fig. 1.3). Today, by contrast, there is no question that the name “Pantheon” was adopted from this predecessor, and that its foundations were reused for the present portico and transitional block. It is further possible that a large circular space anticipated the present rotunda, within which a marble floor has been found beneath the existing one, though it is unclear if it was roofed over, and certainly not with a concrete dome.2 The purpose of Agrippa’s Pantheon and the question of its influence on its successor are explored in Chapter Two.3 The prospect that there was deliberate continuity between the projects is an intriguing one, yet it cannot detain us long here since the focus of this contribution is on specific physical evidence and on the details of articulation and construction.

Licht pointed out similarities with the actual Pantheon in the materials, scale, and details in the exedras of Trajan’s Baths, and he emphasized the importance of bath buildings generally, including several examples at Baiae. MacDonald pointed out numerous similarities with Trajan’s Markets. Both scholars considered the Domus Aurea a turning point in the handling of space and light. Both considered the development of individual design features, but neither considered clusters of related elements or methods of construction in detail.4 They noted important similarities, but differences also need to be taken into account in order to determine which sources were most likely to have had the most direct influence on the Pantheon.

Trajan’s Baths were constructed just northeast of the Colosseum from AD 104–109 over the top of the Domus Aurea; this immense complex set the standard for the imperial baths that were later constructed by Caracalla and Diocletian. The main space was the grandiose frigidarium with its triple cross-vaults and a span of about 85 Roman feet, while the complex also included several domes and half domes. Two domed halls were approximately 75 Roman feet in diameter (or half that of the Pantheon), but these differ in that they were inserted into square envelopes that were integral to the main building (Fig. 5.2). There also survive three exedras open to the air with half domes that are about 100 Roman feet across or nearly two-thirds as wide as the Pantheon with equivalent coffer patterns (Fig. 5.3).5

5.2. Plan of Trajan’s Baths. (Diagram Italo Gismondi, in Nash 1962, Fig. 1283)

5.3. Pair of exedras in Trajan’s Baths; the coffering only aligns with the niches on the main and diagonal axes. (Wilson Jones 2000, Fig. 9.22)

Trajan’s Forum, dedicated in AD 112, was the culmination of the sequence of imperial fora. Trajan’s was the largest that had been constructed in Rome, and it included a public square, a basilica, two libraries, and Trajan’s Column (Fig. 5.4). The interior of the Basilica Ulpia had a span of 85 Roman feet, similar to that of the frigidarium of Trajan’s Baths, but in this case only the pair of side aisles were vaulted. As in earlier basilicas, the central space had a roof with wooden trusses resting on two stories of columns, and the upper level had a clerestory. The Pantheon’s dome spanned nearly twice this distance with masonry rather than wood, yet there is a dimensional similarity with the large exedras at each end of the nave that were approximately 150 Roman feet in diameter. Like the nave, these exedras were covered with a wooden roof. Depictions on coins indicate that the main entrance had a portico with a quadriga.6

5.4. Imperial Fora, Rome, including Trajan’s Forum and Trajan’s Markets. (Courtesy of James Packer)

Trajan’s Markets were built adjacent to the Forum about the same time (Fig. 5.5). The six-storied complex includes at least 170 barrel-vaulted rooms that were probably rented as shops or used as imperial offices, but its precise function is uncertain. The complex was built against the southwest end of the Quirinal, which was cut back to create Trajan’s Forum. The largest space, the Market Hall, has six cross-vaults. With its span of nearly 30 Roman feet and length of roughly 110 feet, and being well lighted by a clerestory, it is one of the most impressive surviving examples of Roman vaulting. The Markets were constructed almost wholly of brick-faced concrete with masonry vaulting throughout.7 Concrete was used in highly innovative ways, and some of the methods that were either introduced here or in other Trajanic buildings found application in the Pantheon.

5.5. Trajan’s Markets. (Waddell 2008)

The Domus Aurea, or Golden House, of Nero was designed by Severus and Celer and built from AD 64 to 68. It covered a vast area from the Oppian Hill to the Palatine that was later built over by structures that included, in addition to Trajan’s Baths, the Colosseum, Titus’s Baths, and the Temple of Venus and Rome. Part of the complex that survives beneath the platform constructed for Trajan’s Baths includes an octagonal domed room built of concrete. This room greatly impressed Nero’s contemporaries, and it is widely believed to represent a turning point in Roman architecture. For MacDonald, it “represents the first culminating stage in the creation of centralized interior architecture” (with the Pantheon as the second stage), and he noted that in its interior “shape, space, structure, and light are interlocked as one.”8 In this volume, Giangiacomo Martines has provided additional support for this building as a probable source for the design of the Pantheon. Its main space is the earliest surviving polygonal room that demonstrated the potential of concrete to create an unusual and dramatically lighted space (see Fig. 4.8). Soon afterwards, Domitian’s Palace was constructed with unusual vaulted spaces, but they are also less likely to have influenced the Pantheon than the larger and more recently constructed public buildings of Trajan’s reign.

Beyond the vicinity of Rome, Baiae has the largest concentration of major domes. Four domes have been called temples, although most if not all of them were parts of baths. They differ from the Pantheon in being constructed as relatively thin shells, and they lack coffering. Only one is unquestionably earlier.9 This, the Temple of Mercury, was probably built a century or more before the Pantheon, and it displays a quite different method of construction. It has a span of 21.55 meters or about 73 Roman feet, and even though it is partly filled with silt and water, it provides a good impression of how domes of equivalent size in Trajan’s Baths must have looked. The largest of the three domes at Baiae is the octagonal Temple of Diana, which, at 29.5 meters in diameter, is nearly two-thirds as wide as the Pantheon, but this dome is relatively pointed in profile. It has generally been dated to the second century AD or to the early third century. The octagonal Temple of Venus, which had a span of 26.3 meters, probably dates later in the second century AD than the Pantheon. Although its walls are substantially intact, little of its dome survives. Unlike the Pantheon, both the Temple of Diana and the Temple of Venus have windows in their walls.10

Licht and MacDonald noted a number of buildings at Hadrian’s Villa that are more likely to be parallels than influences depending on when they and the Pantheon were constructed (see Chapter Three in this volume). The villa was constructed near Tivoli and almost entirely during the reign of Hadrian, from AD 117 to 138. Over an extensive site, about 2 kilometers long and from 0.5 to 1.0 kilometers wide, dozens of major buildings have survived, singly or mostly arranged in scattered groups. All were constructed of concrete, but unlike the brick-faced concrete of the Pantheon, here it was typically faced with wide bands of opus reticulatum (blocks of tufa arranged in a diagonal net-like pattern) alternating with narrow bands of brick. Many structures at the villa have domes or half domes, and some domes are hemispherical and have an oculus. Other domes and half domes have vertical segments that resemble a pumpkin or gourd, shapes that are noteworthy for the sophistication of their conception and the difficulty of their execution. The “Maritime Theater” is an island villa or retreat within a cylindrical drum that has nearly the same diameter as the Pantheon’s interior (Fig. 5.6).11

5.6. So-called Maritime Theater, Hadrian’s Villa, Tivoli. (Photo author in 1982)

MacDonald mentioned the probable influence of cylindrical tombs, including that of Caecilia Metella. He also discussed the possible influence of the Greek tholos, the other main type of freestanding circular building that had been previously constructed in Rome. Examples include the Round Temple by the Tiber, the so-called Temple of Sibyl at Tivoli, and the tholos in the Largo Argentina. All three of these examples had a peristyle around a small cella with an internal diameter of about 25 Roman feet.12 Vitruvius (4.8.3) refers to domes on circular peripteral temples, but these examples are more likely to have had a wooden roof like all known Greek examples. At most, therefore, they are only indirect sources for the Pantheon.

Parallels for Design and Sources for Structural Elements

The most significant design elements present in the Pantheon are as follows: (A) a freestanding building with an attached and pedimented portico, (B) a transitional block with stairs, (C) a cylindrical drum without windows, (D) a dome with coffers and an oculus. There are also a number of (E) other design features that had been used in Roman architecture: cross axes for a circular plan, immense monolithic shafts of granite, an internal clerestory, the use of bilateral symmetry, an orthogonal floor pattern for a circular plan, and a different elevational treatment of the attic compared with the main interior order. All parts of the design were interrelated using (F) simple proportions characteristic of Roman architecture. The most significant structural elements are (G) a dome with its base constructed of rings of concrete (step-rings) separated by layers of brick, (H) a cellular wall on multiple levels, (I) concrete with horizontal layers of aggregate, (J) “relieving arches” of brick-faced concrete, and (K) concrete vaults with brick laid flat against their intrados.

(A) Freestanding Building with an Attached and Pedimented Portico.

As already noted, a portico was usually an integral part of the overall form of temples, under the same gabled roof. Attached porticoes were less common, and they might or might not have had a pediment. Round temples ordinarily had encircling peristyles rather than projecting porticoes. In having an attached portico with a pediment, the Pantheon resembles some earlier transverse temples such as the Temple of Concord, which had a rectangular cella with a portico on one of its longer sides. During the Roman Republic, pedimented porticoes were ordinarily reserved for sacred structures, but their use became more widespread over time.13 Basilicas generally had porticoes without pediments.14 As mentioned, the Basilica Ulpia (Fig. 5.4) had an unpedimented portico in the center of a longer side, and the Basilica Nova was redesigned to have an unpedimented portico on a long side as well as at one end.15

A number of later rotundas with porticoes show the direct influence of the Pantheon. The Tor de’ Schiavi was an imperial tomb with a pedimented portico, and the “Pantheon” at Ostia was a rotunda with an unpedimented portico. However, no domed building earlier than the Pantheon is definitely known to have had an attached portico.16

(B) Transitional Block with Stairs.

The front of the transitional block of the Pantheon, despite the columns of the portico that screen it, resembles a triumphal arch, while in its plan and function it is equivalent to the stair towers built behind some of the exedras of Trajan’s Baths (Fig. 5.2). The mainfunctional purpose of the transitional block is as a container for stairs, which had been included in some Greek temples and which were used more widely in such Roman buildings as baths, triumphal arches, and triumphal columns. As Mark Wilson Jones notes in this volume, the nearest parallels for trapezoidal stairs fitted into a block that mediates between an orthogonal envelope and a circle are those associated with the exedras at Trajan’s Baths.

(C) Cylindrical Drum without Windows.

Windowless walls of concrete faced with brick were fireproof, and they have proven to be one of the most permanent types of construction ever used. In form and scale, the windowless exterior of the freestanding Pantheon most closely resembles the earlier Tomb of Caecilia Metella and the later Tomb of Hadrian (Castel Sant’Angelo), both of which were constructed primarily of concrete, but with stone rather than brick exterior facings. The former has a small corbeled dome with an oculus in the center, and the latter has a small dome in its central tower. Both of these cylinders have square bases. The Pantheon’s walls were hollowed out and given a cellular wall structure similar in plan to the outer ring of the Mausoleum of Augustus (Fig. 5.1). Adoption of the circular form of tombs facilitated support of the largest dome that had ever been created, but there needed to be a way to distinguish the Pantheon from a tomb, and as has been noted, this was accomplished primarily through the addition of a portico.

The Pantheon’s exterior has three cornices that reflect three levels of small chambers within its walls, rather in the way that cornices align with vaulted spaces on the facades of Roman theaters. This building type is unlike the Pantheon in having open arches for circulation, yet there is an affinity inasmuch as theaters were usually freestanding and constructed primarily of concrete, and they were among the few types of Greek or Roman buildings that made use of circular geometry.

(D) Dome with Coffers and an Oculus.

The Pantheon’s dome is by far its most impressive element both visually and structurally, and although there are close precedents for its design, there are also significant differences. In addition to having the same 1:1 ratio between width and height, the exedra on the east side of Trajan’s Baths have coffers that are closest in design of any known precedent to those of the Pantheon (Fig. 5.7).17 Each coffer consists of three successively smaller and more deeply recessed squares that closely resemble the uppermost row of coffers in the Pantheon; the other four rows have coffers with four successively smaller squares. In both buildings, the squares are inflected upward, which enables every square to be visible from the center of the interior.18

5.7. East exedra of Trajan’s Baths. (Photo author in 2007)

Parts of the west dome of Trajan’s Baths have survived, and its east dome is depicted on a surviving fragment of the Marble Plan of Rome, but it is not known if these domes had coffers. In the eighteenth century, Charles Cameron was able to determine that these domes were very close to half the Pantheon’s diameter.19 The Marble Plan depicts a circular room within windowless square walls, suggesting that it was illuminated by an oculus. It had large niches corresponding to the four corners.20 The arrangement of niches was similar to that of the semicircular exedras in the Pantheon, except that these occur on the cross axes rather than the diagonal axes.21

(E) Other Design Features.

Many Roman buildings orchestrate space by means of axes subordinate to the main one, and the Pantheon is a particularly rich example of this principle. The main axis of the interior is made immediately apparent by the arched entrance and apse. The cross axis is emphasized by exedras that align with the square pattern of the floor and coffers.22 A closely similar floor pattern had been used for the apsidal ends of the Basilica Ulpia. Indeed, the patterns of both the floor and walls consisted of basic geometric forms; this design feature and the restraint with which the forms were used are characteristic of Trajanic architecture in Rome. The lavish use of color and of exotic materials characterized imperial architecture generally, but color was usually subordinated to form.

Monolithic shafts of granite were more difficult to quarry and transport than drums of marble, but a granite shaft in one piece was considered worth the extra effort and expense. Granite was one of many types of exotic colored stones that were imported from newly acquired provinces to embellish Rome in the imperial period. Large granite monoliths had been used as early as Vespasian’s Forum of Peace, which was dedicated in AD 75.23 They had been used extensively in the Forum of Trajan for the main floor of the interior of the Basilica Ulpia, which had columns with shafts 30 Roman feet tall, or three-quarters as tall as the 40-foot granite shafts of the Pantheon’s portico.24 Here, the shafts are of two different colors (gray and red), and the monolithic shafts of marble fronting the exedras in the interior are also of two colors (although this is not obvious today due to staining). The use of different colors of marble for the same elements was not unusual, and it was done symmetrically. What is unusual is for a portico with columns to have square column-like piers, or antae, that are different in material (marble rather than granite), in color (white rather than gray or red), and in construction (drums rather than monoliths).

The attic of the Pantheon’s interior resembles the clerestories of basilicas and other buildings, except that its six framed openings or “windows” allow the light passing through the oculus to illuminate the exedras. This arrangement is the reverse of the usual way that a clerestory admits light into a space. Blind windows centered above the adedicules and eight structural piers maintain the pattern and rhythm of the attic. The orders of the main floor and attic story have similar patterns, each with a wider central intercolumniation.25 Although these orders differ in scale and alignment, their stylistic affinity, the similar intercolumniations, and the same materials ensure compositional unity.

The architect Apollodorus had recently used clerestories for the Basilica Ulpia and for Trajan’s Baths. The hemicycle of Trajan’s Markets has alternating triangular and segmental pediments, which prefigure those over the aedicules of the Pantheon. The windows of the Markets are located on an upper floor that resembles a clerestory, and they admit light to the spaces behind them. The pilasters supporting these pediments rest on a running pedestal in an arrangement that is similar to the attic of the Pantheon. As with Roman theaters generally, the uppermost level of the Colosseum exterior has a similar running pedestal, with windows and pilasters rather than an arcade, and its pedestal serves as a facing for barrel vaults in a way that is equivalent to the attic of the Pantheon’s interior.26

The main story of the Pantheon’s interior has exedras fronted by two columns with flanking pilasters, much like the in antis (between the walls) arrangement that had often been used in Greek architecture. The wider intercolumniation in the center was characteristic of Greek Ionic and of Roman architecture generally, and it was used in both the portico and the interior of the Pantheon. Indeed, the A-B-A compositional principle was in fact used throughout the project. Other examples include the way that the portico is flanked by the drum and the way that niches within the portico flank the wider arch of the entrance, a proportional arrangement that closely resembles the openings of a triumphal arch with a similar A-B-A rhythm as on the interior. (A triumphal arch was located opposite the entrance of the Pantheon in the large open space in front.)

The orthogonal floor pattern of the Pantheon most closely resembles the pattern of the Basilica Ulpia, which had circles and squares alternately within a grid of squares. In addition, the way in which this rectilinear pattern was cut through abruptly on the perimeter of its semicircular exedras compares directly with the edge condition of the Pantheon’s floor. In turn, the floor pattern of the Basilica Ulpia is similar to that of the adjacent Forum of Augustus, which has squares of similar size separated by broad bands of marble.

The pattern of pilasters on the attic story of the Pantheon’s interior was cut off above the apse and entrance. Since the floor pattern is cut through around its perimeter, cutting through the pilasters can be considered a solution rather than a solecism, as was presumed by Renaissance critics. Despite the relative independence of the patterns of its orders, the interior was given unity through axial coordination and the use of the same materials and colors for the walls and floor, as well as through the use of similar shapes for the walls, floor, and dome.

(F) Simple Proportions.

As Martines discusses in Chapter Four, the proportions of the Pantheon are straightforward and characteristic of Roman architecture generally. All three surviving exedras of Trajan’s Baths have overall heights that are approximately equal to their widths, just as in the Pantheon’s interior.27 The 1:1 proportion seems also to have been used as a basis for the design of the transitional block, which is nearly square like some triumphal arches with triple openings (see Fig. 7.13).28

The Basilica Ulpia had semicircular ends or exedras about 150 Roman feet in diameter, or approximately the same diameter as the Pantheon, as the Maritime Theater at Hadrian’s Villa, and as the hemicycles of Augustus’s Forum.29 This was also the height of the Mausoleum of Augustus, which was 150 feet in radius. The use of 150 feet as a standard measurement was thus well established before it was used for all three dimensions of the Pantheon’s interior.30 The Pantheon’s width on both the main and cross axis as well as its height relate to one another as 1:1:1, and since all three dimensions are so close to 150 Roman feet, the use of this dimension was undoubtedly intended.31

Since the Pantheon’s exedras are the same width as the piers in between them, they relate to one another as 1:1. Its apse was made somewhat wider and the entrance somewhat narrower for visual and structural reasons, respectively. The ratio of the attic story, or upper part of the interior wall, to the main or lower part is 2:3, another simple proportion that was widely employed.32

The Corinthian columns of the Pantheon, both exterior and interior, conform to the orthodox pattern of imperial practice that took hold around Augustus’s time.33 In accordance with these principles, the heights of shafts are multiples of 5 feet (40 feet in the portico, 30 feet in the interior), while the heights of the complete columns (base, shaft, and capital) are 6/5 greater. A ratio was also probably used to determine the thickness of the walls of the rotunda in relation to the distance that needed to be spanned, since proportion represented a convenient way to record what previously had proven to provide adequate structural support. The Pantheon’s wall thickness in relation to its span is 1:7.3, somewhere between the extremes of one-eighth to one-fifth encountered in other examples.34

In many cases, the same buildings that could have influenced the design of the Pantheon also present similar structural elements. The combination of design and structural elements increases the probability of a direct source.

(G) Dome with Its Base Constructed of Rings of Concrete (Step-Rings) Separated by Layers of Brick.

Through-courses of brick (sometimes called bonding courses) are layers of brick a single brick thick that subdivide sections of concrete at intervals. A through-course of bipedales (bricks two feet square) served the purpose of covering the top of each section of concrete and prevented it from drying too fast while it cured (Fig. 5.8). This shows that buildings of the Trajanic and Hadrianic periods were usually constructed one horizontal section at a time. On the exterior and in the staircases of the Pantheon, layers of bipedales are conspicuous at intervals of about 1.2 meters, being thicker as well as wider than the regular facing of bricks that toothed into the concrete and provided a permanent formwork (see Plate XXIII).35

5.8. Typical Roman brick-faced concrete wall with triangular pieces of bessales separated by layers of bipedales. (Middleton 1892, vol. 1, p. 57, Fig. 8)

A concrete half dome at the north end of the hemicycle of Trajan’s Markets resembles the Pantheon’s dome, but was constructed differently. As Lancaster pointed out, its relatively thin shell was constructed first, and a thick layer of opus signinum (a highly waterproof hydraulic mortar made with ground potsherds) was applied on top of the upper surface before a surcharge of concrete was added.36 This surcharge takes a stepped profile, but was not created as an integral part of the dome and does not contain through-courses of bipedales, as is the case with the step-rings of the Pantheon. A thick layer of opus signinum was also used to coat the dome of the Pantheon, but this was added after the step-rings had been created. This exceptionally waterproof layer contributed to the permanence of the building, particularly in the period between the theft of the bronze roof tiles and their replacement by lead roofing.

The Pantheon’s dome was constructed one step-ring at a time, and as each step-ring was added, it was cantilevered inward (see Fig. 1.12).37 Roman domes were usually constructed as monoliths, but up to the top of the step-rings, the Pantheon’s dome was constructed in much the same way as its walls. Martines and Wilson Jones note that in part for this reason, the need for centering was minimized. After setting up, the top step-ring provided a firm base on which to cast a saucer dome, and the step-rings contained the thrust of the saucer dome while it cured. This part of the Pantheon’s dome is a true dome, and it spans about one-third of the width of the interior (not counting the 30-foot oculus).

Step-rings occur in the Serapeum at Hadrian’s Villa, a building similar to the Pantheon in having curving walls with a sequence of piers that radiates like spokes and with a cellular wall on three levels (Fig. 5.9).38 The Serapeum’s step-rings are conspicuous when viewed from the back (Fig. 5.10). Since the building contains bricks that are datable to AD 123, it must be later than the Pantheon, and possibly influenced by it. However, mutual influence cannot be ruled out, nor a lost precedent that influenced both buildings.39

5.9. Plan of Serapeum, Hadrian’s Villa, Tivoli. (Aurigemma 1961, Fig. 84)

5.10. View of east side of Serapeum, Hadrian’s Villa, Tivoli, showing step-rings, concrete voussoirs, and openings to two of three levels of empty chambers in the walls. (Photo author in 2005)

The Roccabruna at Hadrian’s Villa had two domes, one above the other. Its lower dome survives intact, but is less comparable to the Pantheon by being part of a square in plan. The upper part of the Roccabruna, which has been variously reconstructed, had a dome with step-rings, as indicated by large fragments located nearby.

Turning to later examples, the dome of the nymphaeum called the “Temple of Minerva Medica” (Fig. 5.11) is closest to that of the Pantheon in having both step-rings and through-courses that align with one another. Unlike the Pantheon, it has brick ribs or lattices that could easily have misled FrancescoPiranesi into thinking that a similar method was used for the Pantheon.40 The half dome of the northeast exedra of Diocletian’s Baths also has through-courses.

5.11. Section of Temple of Minerva Medica, Rome. (Julien Guadet, Eléments et théorie de l’architecture cours professé à l’école nationale et special des beaux-arts, Paris 1915, Fig. 943)

(H) Cellular Wall on Multiple Levels.

Like the Pantheon, the Serapeum has thick walls that include three levels of wall cells disposed radially and with openings to the outside (Figs. 5.9; 5.10).41 The earlier building most similar in structure was the Mausoleum of Augustus (Fig. 5.1, A).42 This cellular or diaphragm structure confers the benefit of great rigidity that could resist, in the case of the Mausoleum, the thrust of a mound of earth and, in the case of the Pantheon and the Serapeum, the thrust of their domes. The wall cells of the Serapeum and Pantheon were too small and poorly lighted to be useful for much other than storage, and most were inaccessible. Being open to the outside, they enabled the escape of the heat generated chemically as the concrete cured. The largely hollow walls that resulted are also less likely than solid walls to crack while curing. Like the Serapeum, the Pantheon’s three levels of wall cells include an upper set at the level of the base of its dome (Fig. 5.10, and see Plate IV).

(I) Concrete with Horizontal Layers of Graded Aggregate and Through-Courses.

We have already seen that from the Trajanic period, concrete was often laid in layers separated by through-courses of bipedales (Fig. 5.8). As an initially independent development, as early as the Colosseum, heavier aggregates had been used in lower layers of concrete and lighter aggregates in higher layers.43 At Trajan’s Baths, both techniques come together, though not systematically, and here half domes have heavier tufa at lower levels and lighter tufa in the upper part. The Pantheon followed a similar principle, but with more gradations of density in between successive through-courses. Alberto Terenzio established that the heaviest aggregate (travertine) was used for the foundation and the lightest (tufa and pumice) for the saucer dome (see Fig. 1.12).44

The damaged front of the Serapeum at Hadrian’s Villa provides a good illustration of how the layers of concrete were constructed for most Roman domes. The exposed surfaces of its intrados show that the concrete was built up with fist-sized aggregate set into place individually in horizontal layers. Horizontal layers of aggregate show clearly in photographs of the Pantheon’s intrados when stucco was removed in the 1930s.

As mentioned earlier, Hadrian’s Villa has numerous other domes and half domes. The vestibule of the Piazza d’Oro has curved walls that support a “pumpkin” dome with single and double curving elements alternating much like the dome of the Serapeum. Its plan resembles the Pantheon’s in having alternately rectangular and curved niches, and it, too, had an oculus (as did the dome of the Large Baths at the Villa). This combination of features is more striking than any one feature, and they occur in what is essentially a freestanding building. The method of construction used for the vestibule’s dome was similar to the Pantheon’s in that a ring of concrete was created first by cantilevering inward to decrease the opening that needed to be spanned. Elsewhere at Hadrian’s Villa, cantilevered balconies can be seen on the nearby Barracks of the Vigili and the long building parallel to the pool of the Canopus. Corbeling had been used to similar effect previously, for example in the main hall of Trajan’s Markets, but prior to the Pantheon this structural principle is not known to have been used to create a dome or half dome.45

(J) “Relieving Arches.”

Concrete had largely replaced stone for the construction of arches and vaults before the end of the first century AD. The designation “relieving arches” applies not to open arches but to ones filled with masonry underneath; their purpose was often, it seems, to reduce or relieve pressure over an opening lower down. As Martines also notes, this type of arch was used widely in concrete buildings dating from the reigns of Trajan and Hadrian, including Trajan’s Baths, Trajan’s Forum, Trajan’s Markets, and those at Hadrian’s Villa. It is generally assumed that relieving arches were made of brick and mortar rather than concrete;46 however, examples in the exedras of Trajan’s Baths can readily be seen to be made not just of brick but of brick and concrete in combination. Equivalent arches exist in numerous earlier Roman buildings and in the construction of aqueducts; where the internal structure of the arch can be seen, it seems invariably to have been constructed of bricks alternating with portions of concrete shaped like voussoirs.47 Examples of concrete voussoirs that were used before and after the Pantheon can be seen together in a section of the Arcus Neroniana (Caelemontana) adjacent to the Scala Sancta (Fig. 5.12).48 Examples constructed during the Hadrianic period include the Serapeum (Fig. 5.10) and the Small Baths at Hadrian’s Villa. There concrete voussoirs alternated with portions of brick in the same way that the walls of the Pantheon were no doubt constructed, in line with the typical configuration of relieving arches in this period (Fig. 5.13). When the level of the arch was reached, bipedales were placed radially on top of the centering and were separated from one another by a facing made up of half or quarter bipedales, leaving spaces later to be filled by concrete.

5.12. Detail of brick-faced concrete with concrete voussoirs in the Arcus Neroniana (Caelemontana), built under Nero (AD 54–68) and reinforced during the reign of Septimius Severus (AD 193–211), who inserted a similar arch underneath. (Photo author in 2007)

5.13. Diagram of brick-faced concrete showing the construction of concrete voussoirs. (Courtesy of Nikolaos Karidis)

In conformity with the general perception that relieving arches were made of solid brick, the Pantheon’s great arches are usually presumed to have been made of brick and mortar extending entirely through its walls (as in Fig. 5.14).49 The main evidence for this view is the brick barrel vaulting that projects from the center of the south side of the Pantheon (see Fig. 7.5), but this served a different purpose in acting like the flanking buttresses of the main hall in Trajan’s Markets or the buttresses of the Basilica Nova.50 It may be noted how, as a consequence of its radial plan, all of the relieving arches of the rotunda have to taper; this means that the great arches that show on the exterior become considerably smaller on the interior (see Fig. 7.1). Such tapering or conical vaults had long been used in the substructures of theaters, and they were easily constructed of concrete. To have constructed them of brick alone would have required that some bricks be made wedge shaped both vertically and horizontally and/or that mortar joints be wedge shaped.

5.14. Pantheon, cutaway view. (Josef Dell, “Das Pantheon in Rom,” Zeitschrift für bildenden Kunste, ns. 4, 1892, p. 273)

Evidence that the great arches of the Pantheon may actually be brick-faced concrete was found in 1882 when a building attached to the back of the rotunda was removed, and a damaged wall surface revealed what was behind its brick facing. A specialist in Roman construction methods, J. Henry Middleton, examined these sections of wall before they were restored, and he stated that “the brick facing, including the arches, only tails into the wall to an average depth of 5 or 6 inches, so that in reality these apparently relieving arches are of little more use (as regards the pressure) than if they were painted on the surface.”51 The illustration he provided makes the same point for general practice. In this volume, Wilson Jones discusses a brick-faced arch he found in the west staircase of the Pantheon that can be seen to be toothed into the concrete. However, only an examination of the concrete core of the building itself could reveal whether its relieving arches were constructed of solid brick and mortar or brick-faced concrete.

(K) Concrete Vaults with Brick Linings.

The soffits of vaults in the mid-imperial period were often made of brick linings. These were laid like tiles (not on edge) on a centering to form the intrados of a concrete vault, and they remained attached when the concrete cured. The earliest-known example of this construction method appears in the barrel vaults at Trajan’s Markets and other Trajanic structures, and it had been used experimentally in several different forms. In all known cases where the structure is visible, brick linings face vaults made of concrete, not of brick.52 The intrados of the great arches of the Pantheon’s exedras have a lining of bipedales; even if they could have been attached using mortar – they weigh about 25 kilograms each – they would not have been needed on the intrados of a solid brick arch or vault.


The round plan of the Pantheon provided a secure and permanent support for its dome. The almost exclusive use of masonry enabled a fireproof building to be constructed so as to avoid the fate of its predecessor, and the use of concrete enabled an unprecedented span. A rectangular portico was married to the cylinder by means of a transitional block in between, housing the stairs. Although previous buildings provided the elements for its design, none proved to be equally permanent, and no larger dome has ever been made of unreinforced masonry. A new means of construction was required to produce the desired effect.

While taking its cue from Agrippa’s project, the Pantheon was created by combining design and structural features that had been used for at least six different building types. Baths provided precedents for domes or half domes with coffers and an oculus and an equivalent arrangement for the incorporation of trapezoidal stairs. Tombs provided a precedent for a windowless cylinder that was freestanding. Both tombs and theaters provided an example of a cellular structure and a geometrical method for laying out circular plans. Theaters provided models for conical barrel vaults and multiple levels of cellular structure. Basilicas offered examples of a forecourt, an attached portico, and a clerestory arrangement. Triumphal arches provided exemplars for the form and proportions for the transitional block. Temples provided precedents for a pedimented portico.

Of the precedents erected in the preceding decades, the octagonal hall of the Domus Aurea was the most innovative, but it differs in many respects from the Pantheon. Its plan is more complex, and its use of light is more sophisticated, but its structure and method of construction are simpler. Compared with the Pantheon, there are other significant differences: the shape (with an octagonal plan and largely octagonal vault rather than a round plan and hemispherical dome), the lighting (with openings at a high level around the perimeter aside from the oculus), the decoration (with a plain intrados rather than coffers), and the structure (part of a larger building rather than freestanding). In addition, its relatively small and thin dome is monolithic. At Baiae, the even earlier Temple of Mercury is likewise relatively small and has a thin shell, and it too lacks coffers. This is also true of the more imposing Temple of Diana and the Temple of Venus, although these have windows in their walls, and it is uncertain if either was constructed earlier than the Pantheon. Regardless of their exact dates, none of the three domes at Baiae provides a close parallel with the Pantheon in terms of its design or construction, and all formed part of complexes rather than being freestanding. Tholoi are still less likely to have been directly influential in that none is definitely known to have had a dome, although they had circular peristyles rather than projecting porticoes.

The domes and half domes most nearly equivalent in scale and materials to the Pantheon had recently been constructed for Trajan’s Baths and Trajan’s Markets, one certainly and the other possibly designed by Apollodorus. Even though only the lower parts of their domes survive, enough remains of the coffering and other details to show just how important an influence the Baths undoubtedly were. As a more complete geometrical form, a dome has inherently more strength than a half dome, and since half domes had been successfully built that were 100 feet across, about two-thirds as wide as the Pantheon interior, these examples could well have suggested the practicality of constructing a still larger dome with the same proportions, as well as suggesting a similar approach to the incorporation of the stairs. There are some similarities in constructional technique, but this apparently did not extend to layers divided by through-courses of brick, of which only occasional use was made in the walls. Like the Pantheon, Trajan’s Markets includes numerous concrete vaults with brick linings, as well as walls with concrete cores separated by layers of bipedales. Its half domes, however, were not constructed this way. Such qualifications aside, the largest number of design elements and constructional features of the Pantheon occur in buildings constructed during Trajan’s reign.

At Hadrian’s Villa several buildings have features in common with the Pantheon, but without more exact dating, they must be considered parallels rather than sources. The Serapeum of Hadrian’s Villa provides the closest parallel for the way in which the Pantheon’s walls were constructed. Both buildings have cellular walls thick enough to provide internal buttressing. The Roccabruna has step-rings, and the dome of the vestibule of the Piazza d’Oro was cantilevered inward to reduce the span that had to be fully supported by centering during construction.

Regardless of whether constructional refinements such as step-rings with through-courses were invented specifically to help cope with the enormous span of the Pantheon, they represent the culmination of centuries of development in the creation of space through the use of concrete. As impressive as its technical accomplishments are, though, the principal achievement of the Pantheon lies in its captivating and memorable space. The visual impact of the interior was greatly enhanced thanks to the dramatic chiaroscuro of the exedras, the articulation of the coffering, and the illumination that floods in through the ample oculus. These features and many others were derived from a wide variety of sources that can be identified directly or indirectly, but despite the extent of its multifaceted indebtedness to earlier architectural achievements, the design of the Pantheon exceeded the achievements of its sources and set a new standard for the conception of interior space.

Mark Wilson Jones provided detailed advice that has improved every aspect of this chapter. When our conclusions differ, mine were reached reluctantly.

1 Domes with an oculus were used in bath buildings as early as the first century BC (Vitruvius, 5.11.5; see Vitruvius: Ten Books on Architecture, trans. Ingrid D. Rowland, commentary and illustrations by Thomas Noble Howe with additional commentary by Ingrid D. Rowland and Michael J. Dewar, New York 1999. Early examples are discussed in Fikret Yegül, Baths and Bathing in Classical Antiquity, Cambridge 1992, pp. 37–38, where the author provides persuasive evidence that domes “started with bath buildings” (p. 3).

2 The architect Georges Chedanne established that the existing Pantheon was constructed more than a century after Agrippa’s death; he exhibited a drawing of the key brickstamps that he discovered and publicized his findings, but he never published them. They were, however, summarized by R. Phené Spiers, “Monsieur Chedanne’s Drawings of the Pantheon,” Journal of the Royal Institute of British Architects 2, 1895, pp. 180–182. The results of the excavations in 1892–1893 were published by 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, and Il Pantheon rivendicato ad Adriano 117–138 d.C., Milan 1929); Beltrami reconstructed Agrippa’s Pantheon as a south-facing T-shaped structure. The evidence was limited, however, to a small piece of foundation; see Kjeld De Fine Licht, The Rotunda in Rome: A Study of Hadrian’s Pantheon, Copenhagen 1968, p. 219, and Mark Wilson Jones, Principles of Roman Architecture, New Haven 2000, Fig. 3.3). For the recent excavations in front of the portico, see 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. 377–394.

3 Most of the ancient texts relating to Agrippa’s Pantheon and the existing Pantheon can be conveniently compared in Licht 1968, pp. 180–183. Paul Godfrey and David Hemsoll argued that none of the three Pantheons was a temple. Noting that the present building is only known to have been used as a tribunal, they added that “despite their many uses, the imperial administration of justice was not conducted in temple buildings” (Godfrey and Hemsoll, “The Pantheon: Temple or Rotunda?” Pagan Gods and Shrines of the Roman Empire, ed. Henig et. al., Oxford 1986, p. 202).

4 Licht (1968, pp. 203–225) mentioned the importance of considering groups of features, but he discussed separately each building that was a possible source, rather than seeking clusters of features. MacDonald’s book on the Pantheon contained his fullest treatment of its sources (William L. MacDonald, The Pantheon: Design, Meaning, and Progeny, London 1976, pp. 44–75); cf. MacDonald, The Architecture of the Roman Empire, vol. 1:An Introductory Study, London 1965; 2nd ed. rev., New Haven 1982, for additional details. For a more detailed discussion of the sources for the Pantheon’s design and construction, see Gene Waddell, Creating the Pantheon: Design, Materials, and Construction, Rome 2008.

5 Kjeld De Fine Licht, Untersuchungen an den Trajansthermen zu Rom, Copenhagen 1974. Dimensions for the exedras of the Baths of Trajan are given in meters and Roman feet in Wilson Jones 2000, p. 218.

6 The span of the central space of the Basilica Ulpia was 23.43 meters or 79.15 Roman feet (James Packer, The Forum of Trajan in Rome: A Study of the Monuments, Berkeley 1997, folio 25). Some early reconstructions reproduced by Packer show the exedras of the Basilica Ulpia with half domes like the exedras of Trajan’s Baths, but subsequent archaeological evidence has established that their roof structures were not vaulted. Recent excavations of Trajan’s Forum have been interpreted differently; see RobertoMeneghini, “Il foro Traiano. Ricostruzione architettonica e analisi strutturale,” Mitteilungen des Deutschen Archäologischen Instituts, Römische Abteilung 108, 2001, pp. 245–268, and James Packer, “Templum Divi Traiani Partici et Plotinae: a Debate with R. Meneghini,”Journal of Roman Archaeology 16, 2003, pp. 109–136; some uncertainty remains about how its east and west ends were configured. The dates for Trajanic buildings are from Amanda Claridge, Rome: An Oxford Archaeological Guide, Oxford 1998, pp. 161–164, 170–172, and 288–290. The diameter of the Maritime Theater or “Island Enclosure” is 44.20 meters or 148.5 Roman feet (William L. MacDonald and John Pinto, Hadrian’s Villa and Its Legacy, New Haven 1995, p. 82).

7 For the construction of Trajan’s Markets, see Lynne Lancaster, “Building Trajan’s Markets,” American Journal of Archaeology 102, 1998, 283–308, and Lynne Lancaster, “Building Trajan’s Markets 2: The Construction Process,” American Journal of Archaeology 104, 2000, pp. 755–785. Trajan had the end of the Quirinal cut away and terraced to provide room for his forum. On the use of retaining walls by the Romans, cf. Vitruvius 6.8.7. MacDonald (1976, p. 58) argued that “the Market hall is a rectilinear forerunner of the Pantheon.”

8 MacDonald 1976, pp. 54–55. The width of the Domus Aurea is given by Wilson Jones (2000, p. 218) as 49.49 feet or 14.65 meters measured from corner to corner.

9 On the bath buildings at Baiae, see Yegül 1992, pp. 107–110, and Amedeo Maiuri, Phlegraean Fields from Virgil’s Tomb to the Grotto of the Cumaean Sybil, Rome 1947, pp. 61–73. The dates assigned for these structures vary widely, being based on constructional comparisons thought to be more securely dated, but it is often uncertain how long methods of construction persisted. For example, Yegül dates the Temple of Mercury from the late Republican to the early Julio-Claudian period, and likewise the “Temple of Apollo” at Lake Avernus, which has a span of 36.6 meters. However, Maiuri dates the latter to the reign of Nero (Maiuri 1947, pp. 139–141), though it has also been dated to the Antonine period (M. E. Blake, Roman Construction in Italy from Nerva through the Antonines, Philadelphia 1973, p. 270).

10 The Domitianic hall at Albano has a dome with an oculus, but like those at Baiae it formed part of a larger structure. The Forum Baths at Pompeii has an earlier example of an oculus, but this lit a conical vault (Licht 1968, pp. 211–215). The oculus was also used to regulate the temperature by partially closing it (Vitruvius, 5.10.5). No dome of a bath building is known to have had coffers.

11 MacDonald and Pinto 1995, pp. 28–29 (site description). For detailed accounts of excavations, see S. Aurigemma, Villa Adriana, Rome 1961. For dates of construction, see A. C. G Smith, “The Date of the Grandi Terme of Hadrian’s Villa at Tivoli,” Papers of the British School at Rome 46, 1978, pp. 73–93.

12 Dimensions for these buildings are given by Wilson Jones (2000, pp. 217–219). Although Renaissance architects such as Palladio reconstructed such tholoi with domes, I have been unable to find any dome fragments for these three examples. The entablatures of the two tholoi in Rome have been destroyed, and at Tivoli nothing of the wall survives above the level of the entablature (Richard Delbrueck, Hellenistiche Bauten in Latium, Strassburg 1907, Figs. 18 and 28). Unlike the Pantheon, the tholos in the Largo Argentina has a podium and an altar in front of its steps and so was unquestionably a temple. Although it has been considered a possible source for the Pantheon, the date when its portico was added is uncertain, and it did not have a transitional block. MacDonald (1976, pp. 67–68; Figs. 77–78) judged this tholos to be a “remote precedent,” but he observed that adding a portico to a drum “had not been done before” [the Pantheon]. For Licht, this tholos was not “either a basis or even a model,” but was worthy of note (Licht1968, pp. 218–219). He knew of no definite evidence for an earlier rotunda that was freestanding (p. 207). The Hellenistic Arsinoeon at Samothrace was large (about 60 feet in diameter), cylindrical, and freestanding, but it was roofed with wood, and it was not nearby, not contemporary, and not well known (cf. MacDonald 1976, p. 49 and Figs. 42 and 51).

13 Godfrey and Hemsoll 1986, p. 204.

14 For persuasive evidence that the Pantheon was intended to have a taller portico with a gabled roof, see Paul Davies, David Hemsoll, and Mark Wilson Jones, “The Pantheon: Triumph of Rome or Triumph of Compromise?” Art History 10, 1987, pp.133–153, and Wilson Jones 2000. Chapter Seven here provides further evidence (including foundations that could support larger columns and the set-back in the profile of the transitional block that ties in with the upper pediment).

15 Blake 1973, pp. 12–15, n. 65; p. 17. Spartianus, Hadrian 19.9–10. Rodolfo Lanciani, Forma Urbis Romae, Rome 1893–1901, repr. Rome 1988.

16 Jürgen J. Rasch, Das Mausoleum bei Tor de’ Schiavi in Rom, Mainz 1993. C. Briggs, “The ‘Pantheon’ of Ostia (and Its Immediate Surroundings),” Memoirs of the American Academy in Rome 8, 1930, pp. 161–169; frontispiece and Plates 51–57. R. Meiggs,Roman Ostia, 2nd ed., Oxford 1973. On the possible exception, the tholos in the Largo Argentina, see n. 12 herein. Wilson Jones 2000 (p. 182 and Fig. 9.6) notes similar plans of earlier circular buildings at Stymphalos and Athens, but neither is known to have been domed.

17 Lynne Lancaster, “Concrete Vaulted Construction: Developments in Rome from Nero to Trajan,” Ph. D. diss., Oxford 1996, pp. 103–104, 247. The plan of Trajan’s Baths by Gismondi (Fig. 5.2 herein) designates the three best-preserved exedras differently from the plan in Licht’s 1974 monograph; Gismondi’s B, F, and C are Licht’s H, L, and D, respectively.

18 Square coffers had been used centuries earlier to articulate flat ceilings in Greek temples, and were adapted and used nonstructurally by the Romans for concrete structures. Roughly, square coffers had been used as early as about 100 BC on a double-curving barrel vault at Praeneste (Palestrina). In the Pantheon, coffers are a major unifying element of the interior. As MacDonald (1976, pp. 72–74) pointed out, their corners align in such a way as to create the impression of interlocking spirals. The effect achieved is similar to that of the diamond-shaped coffers in the apses of the Temple of Venus and Rome as rebuilt around AD 300.

19 Charles Cameron, The Baths of the Romans Explained and Illustrated: with the Restorations of Palladio Corrected and Improved, London 1772, Plate 7. Yegül (1992, p. 146) gives the somewhat larger dimension of 27 meters.

20 Licht 1974, Supplementum, Fig. 8 “S.”

21 The Tor de’ Schiavi and the Rotunda at Ostia have similar arrangements.

22 Wilson Jones (2000, pp. 194–196) lists other means used to emphasize axes inside the Pantheon, some of which are so subtle as to be easily overlooked.

23 M. E. Blake, Roman Construction in Italy from Tiberius through the Flavians, Washington, DC 1959, pp. 89–90. Rose granite was used for the Temple of Peace and gray for the Basilica Ulpia.

24 Packer 1997, p. 434; Wilson Jones 2000, p. 223.

25 Some city gates had similar arrangements with openings aligning rather than orders of different sizes (cf. Wilson Jones 2000, p. 116 and Fig. 6.12). On the exterior of the Pantheon, the largest cornice is at the top of the wall (as for the Colosseum), but on the interior, the largest cornice was used with the larger order, and the smaller order was treated more like the attic story of a triumphal arch (cf. Licht 1968, Figs. 99 and 100).

26 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) discussed convincingly how the attic story of the Pantheon’s interior reflects the structure of the building.

27 Licht 1974, Plates 1 and 3. On size and proportion in Roman architecture, see Wilson Jones 2000.

28 The width between the centers of end columns on the front of the portico is 32.03 meters and the height of the transitional block is 32.23 meters (Wilson Jones 2000, p. 220).

29 Packer 1997, folio 24. MacDonald and Pinto 1995, p. 82. In his book on the Pantheon, MacDonald makes a detailed comparison of the Forum of Augustus and the Pantheon, particularly in regard to the possible meanings that both buildings had for the Romans (MacDonald 1976, pp. 84–91). Most notably, the exedras of the Forum have a radius of 75 feet, the same radius as the Pantheon, but were not vaulted and were not fully hemispherical. Their influence would have been primarily in terms of scale.

30 On the use of whole number or “round” dimensions, see Mark Wilson Jones, “Principles of Design in Roman Architecture: The Setting out of Centralised Buildings,” Papers of the British School at Rome 57, 1989, pp. 106–151; Wilson Jones 2000, pp. 71–84.

31 The Pantheon’s diameter from wall facing to wall facing is 43.57 meters (147.20 Roman feet) and from column center to column center is 44.52 meters (150.41 Roman feet); Wilson Jones 2000, p. 220.

32 The order of the attic including its running pedestal is 30 feet tall, and the order of the main story is 45 feet (Wilson Jones 2000, p. 185).

33 Wilson Jones 2000, pp. 135–156.

34 A ratio of about 1:8 was considered sufficient for most domes and about 1:10 for most barrel vaults (Wilson Jones 2000, pp. 82 and 233 n. 38, citing Janet DeLaine, The Baths of Caracalla in Rome: A Study in the Design, Construction, and Economics of Large-Scale Building Projects in Imperial Rome (Journal of Roman Archaeology, Supplement 25), Portsmouth, R.I., 1997; Jürgen Rasch, “Zur Konstruktion spätantiker Kuppeln vom 3 bis 6 Jahrhundert,” Jahrbuch des deutsches Archäologischen Instituts 106, 1991, pp. 311–383).

35 Through-courses had earlier been used intermittently in the Colosseum and in Trajan’s Baths (Lancaster 1998, p. 285). In the Mausoleum of Augustus, horizontal sections of concrete were separated by layers of limestone chips (personal observation; cf. H. Von Hesberg and S. Panciera, Das Mausoleum des Augustus. Der Bau und seine Inschriften, Munich 1994). In Domitian’s Nymphaeum at Albanum, the sections were separated by layers of pozzolana (Blake 1959, p. 138).

36 Lancaster 1996, vol. 1, p. 209, and Fig. 110B.

37 D. Moore, The Roman Pantheon: The Triumph of Concrete, Wyoming 1995. Francesco Piranesi, Seconda parte de’templij antichi che contiene il celebre Pantheon, Rome 1790, Plates 10 and 29, illustrated two different structures for the dome of the Pantheon. In the earlier version, he showed brick ribs, but in the later one, he showed that the step-rings were separated by through-courses. Horizontal layers of bricks (tegole) had been shown previously on the intrados by Antonio da Sangallo il Giovane (Waddell 2008, Fig. 58B).

38 The span of the Serapeum is about 54 feet or approximately one-third as much as the Pantheon (R. Vighi, Villa Hadriana, trans. J. B. Ward Perkins, Rome 1958). Another similarity shared with the Pantheon is that of nonalignment; the seven segments of the vault of the Serapaeum do not align with the nine openings in its walls (MacDonald and Pinto 1995, p. 112).

39 Being so much smaller, the Serapeum could conceivably have been completed early enough to have been a source for the Pantheon’s dome. For the excavation of the Serapaeum and the surrounding area of the Canopus, see S. Aurigemma, “Lavori nel Canopo di Villa Adriana,” Bollettino d’Arte 39, 1954, pp. 237–341. Most of the principal buildings at Hadrian’s Villa, including the Serapaeum and Roccabruna, have bricks stamped with the names of consuls for the year 123, and many have stamps from 124 (Herbert Bloch, I bolli laterizi e la storia edilizia romana. Contributi all’archeologia e alla storia romana (1936–1938), Rome 1947, pp. 137–141). The Piazza d’Oro and the domes of the Large and Small Baths have brick manufactured a few years later. A total of 279 brickstamps with the names of consuls were found while excavating the Canopus from 1950 to 1955, and they were analyzed by Domenico Faccenna: 189 were manufactured in AD 123, 11 in AD 124, 17 in AD 125, 61 in AD 126, and 1 in AD 127 (Aurigemma1961, p. 127).

40 Piranesi was probably also misled by the use of these ladder-like ribs during the Severan period for renovating a part of Agrippa’s Baths known as the Arco della Ciambella, particularly since it lies so close to the Pantheon and was also believed to have been created by Agrippa. See Licht 1968, p. 297.

41 MacDonald and Pinto 1995, p. 113. Martines also emphasizes the importance of these wall cells for the construction of the Serapeum and the Pantheon (Giangiacomo Martines, “La struttura del Pantheon velut regionem fornicatam,” Quaderni dell’Istituto di Storia dell’Architettura 41, 2004, pp. 3–16, p. 7 n. 29, and Chapter Four in this volume).

42 Von Hesberg and Panciera 1994, Figs. 1 and 2 and Plate 3c.

43 Lynne Lancaster, Concrete Vaulted Construction in Imperial Rome: Innovation in Context, Cambridge 2005, p. 167.

44 Alberto Terenzio, “La Restauration du Panthéon de Rome,” Museion 20, 1932, Plates 10–11.

45 J. DeLaine, “Structural Experimentation: The Lintel Arch, Corbel, and Tie in Western Roman Architecture,” World Archaeology 21, 1990, pp. 416–417.

46 Lancaster noted that the earliest “relieving arches ... of bricks or tiles” were used in the Augustan period and that bricks were used later to create “voussoirs” of concrete (Lancaster 2005, pp. 88 and 94). However, she usually interpreted vaults faced with bipedales as having a solid arch of bipedales rather than as being made of concrete voussoirs. She reconstructed the great arches of the Pantheon as consisting wholly of bipedales (2005, p. 97; Fig. 80).

47 The broad arch that serves as the main entrance to the main hall of Trajan’s Markets shows on the intrados that most of its bipedales extend much less than two feet into the concrete. A prerestoration set of photographs made of Trajan’s Markets shows that the arches of its hemicycle were also constructed using concrete voussoirs (Instituto Centrale per il Catalogo e la Documentazione, Ministero Beni e le Attivita Culturali). One photograph from this series is reproduced by Giovanni Teresio Rivoira (Roman Architecture and Its Principles of Construction Under the Empire with an Appendix on the Evolution of the Dome Up to the 17th Century, New York 1972, a translation of Rivoira, Architettura Romana, Milan 1925), Fig. 126. Arches with flat lintels were made of concrete as early as the end of the second century BC (DeLaine 1990, pp. 416–417).

48 This aqueduct was constructed of concrete by Nero and was reinforced using concrete in ca. AD 201 by Septimius Severus and Caracalla (Esther Boise Van Deman, The Building of Roman Aqueducts, Washington 1934, pp. 266–270).

49 The consensus of opinion is that the Pantheon’s great arches are made wholly of bipedales and mortar rather than concrete with a brick facing: Auguste Choisy, L’art de bâtir chez les Romains, Paris 1873; Rivoira 1925; Terenzio 1932; Licht 1968; MacDonald1982; Moore 1995; Lancaster 2005; Gerd Heene, Baustelle Pantheon: Planung, Konstruktion, Logistik, Düsseldorf 2004; Giovanni Belardi, Il Pantheon: storia, tecnica, e restauro, Viterbo 2006.

50 The function of a buttress is primarily to resist lateral thrust, while the function of an arch is to support and transfer weight to either side of an opening. For a diagram with the buttresses of the main hall in Trajan’s Markets see MacDonald 1982, Plate 93. How the buttress on the south side of the Pantheon terminated is uncertain. Licht (1968, p. 159) reconstructs a half arch extending upward against a wall of the Basilica of Neptune. Heene (2004, p. 63) reconstructs a half arch extending downward above the apse of the basilica.

51 J. H. Middleton, The Remains of Ancient Rome, 2 vols., London 1892, pp. 1, 60.

52 For the extensive use of brick linings for the concrete vaults in Trajan’s Markets, see Lancaster 1998, p. 283; Lancaster 2000; Lancaster 2005, pp. 31, 32, 176, and 207.

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