GENE WADDELL

CREATING THE PANTHEON Design, Materials, and Construction

«L'ERMA» di BRETSCHNEIDΕR

Interior of the Pantheon by Giovanni Paolo Ρannini with its original attic design, c. 1750 (FU 12526F)

GENE WADDELL Creating the Pantheon

© Copyright 2008 «L'ERMA» di BRETSCRNEIDER Via Cassiodoro, 19 - 00193 Roma

http://www.lerma.it

Progetto grafico: «L'ERMA» di BRETSCHNEIDER

Tutti i diritti riservati. È vietata la riproduzione di testi e illustrazioni senza il permesso scritto dell'Editore.

Special funding for the research and publication of this volume was provided by:

College of Charleston

Evans & Schmidt, Architects

A conscientious attempt has been made to determine the rightful owner of all copyrights. Copyright claimants are asked to provide proof of ownership and an invoice for appropriate compensation.

ISBN 978-88-8265-493-1

Contents

Abbreviations 9 1. Concrete 39 Preface 11 2. Proportion 39

3. Space 40 4. Enframements 40

PART I. INTRODUCTION 5. Scale 40 6. Precedents 41

1. Preliminary Considerations 15 7. Embellishment 41 A. Brief Description of the Pantheon ... . 15 8. Light 41 B. Date of Construction 17 9. Variety 42 C. Confirming the Identity of the Three 10. Permanence 42

Pantheons 21 D. Purpose of the Existing Pantheon 22 4. Concrete Construction 43 E. Main Problems to Be Solved A. Advantages 43

by the Architect 23 B. History 44 F. The Client and the Architect 23 C. Materials 44

1. Pozzolana 44 2. Major Advances in Knowledge 27 2. Aggregate 44

A. Serlio 27 D. Early Types of Facings 45 B. Palladio 27 E. The Use of Bricks 45 C. Desgodetz 28 1. Brick-faced Concrete 45 D. Francesco Piranesi 28 2. Through-courses 46 E. Leclère 29 3. Concrete Voussoirs 47 F. Lanciani 29 4. Tile Linings 48 G. Chedanne 29 F. Construction Methods 49 H. Beltrami and Armanini 30 G. Curing 50 I. Terenzio 31 J. MacDonald 31 5. General Sources of Design 51 K. Licht 32 A. Related Building Types 51 L. Wilson Jones 32 Baths 51

Tombs . 52 3. Basilicas 53

PART II. ROMAN DESIGN AND 4. Round Temples 53 CONSTRUCTION 5. Triumphal Arches 54

6. Theatres 54 3. Standard Design Procedures 35 B. Related Design Features 54

A. Vitruvius' Basilica at Fano 36 1. Domes 54 B. What the Pantheon's Architect Took 2. Internal Buttressing 55

for Granted 38 3. Attached Porticoes 56

6

6. Specific Sources of Design and 4. The Need for Models 84 Construction 57 5. Interlocking Spirals 85 A. Sources for Design Elements 57 D. Forecourt and Basilica of Neptune ... 86

1. Dome with Coffers and an Oculus 57 2. Free-standing Building with an

Attached Portico and Forecourt .. . 58 PART IV. CONCRETE CONSTRUCTION 3. Intermediate Block with Stairs 58 4. Cylindrical Drum with Solid Walls 58 9. Lower Drum and Block 91 5. Various Separate Elements in A. Site Preparatio 91

Combination 59 B. Materials and Methods 6. Proportions 60 of Construction 91

B. Sources for Structural Elements 61 1. Foundations 91 1. Dome with Steprings and 2. Walls 91

Through-courses 61 3. Vaulting 92 2. Wall Chambers Open to the Outside 4. Load-bearing Columns 94

on Multiple Levels 61 C. Surveying the Site 95 3. Concrete with Horizontal Layers

of Graded Aggregate and Preparing Foundations Construction Methods Used for the

95

Through-courses 62 Lower Two-thirds of the Drum and 4. Concrete Voussoirs 62 Block 95 5. Concrete Vaults with Tile Linings . 62 1. Lower Section 95

C. Unparalleled Features of the Dome ... 62 2. Middle Section 97

10. Upper Drum and Block 99 PART III. PRELIMINARY DESIGN PHASE A. Upper Third of the Drum 99

1. Exterior 99 7. The Site 67 2. Spaces Within the Drum 100

A. Site Selection 67 3. Buttress 100 1. Campus Martius 67 B. Upper Block 100 2. Vicinity 67 1. Exterior 100

B. Orientation 68 2. Rooms 101 C. Relative Altitudes 68 3. Stairs 102 D. Existing Conditions 71

1. Agrippa's Foundations 71 11. Dome 103 2. Other Existing Conditions 73 A. Lower Dome 103

1. How the Dome Could Have Been 8. Structural Design 75 Constructed 103

A. Plan 75 2. Steprings 104 1. Proportions 75 3. Materials 105 2. Circular Plan 77 4. Constructing the Coffers 106 3. Exedras 78 B. Upper Dome 106 4. Interior Attic 79 C. Oculus 106 5. Transitional Block 79 D. Steps 107

B. Elevations 81 E. Bronze Roofing 107 1. Proportions 81 F. Cracks in the Dome and Walls 107 2. Exedras 81 G. Conclusion 109 3. Exterior 82

C. Coffers 83 Alignment 83 PART V. EMBELLISHMENT Precedent 83 The Need for Computation 84 12. Comparisons of the Orders 113

7

Α. Proportions 113 Columns 113 Intercolumniations 116 Plans 117 Entablature 118

Β. Materials 120 Differences Between the Interior and Portico 121 Construction 121

13. The Porticoes 123 Α. Architectural Evidence 123

1. Description of the Present Portico 123 2. Structural Relation of the Portico

to the Block 124 Β. Archaeological Evidence 126 C. Further Evidence of Change 128

1 Inscription 129 2. Entrance 129 3. Upper Block 130 4. Pediments 130 5. Reuse of Material 132 6. Solecisms 132 7. Later Buildings Influenced by the

Pantheon 133 D. Proposed Explanations for Changes 134

Some Early Explanations 134 Sixty-foot Columns 135

E. Conclusion 135

14. Finishing 139 A. Main Story 139

Exedras 139 Aedicules 140 Wall Facing 140

B. Attic Story 141 C. Floor 141 D. Color Combinations 143 E. Dome 143

Interior 143 Exterior 143

F. Furnishings 144 G. Drum Exterior 144 H. Block Exterior 145 I. Entrance 145 J. Portico Ceiling 146 K. Sculpture 146 L. Forecourt 146

Conclusions 149 Notes 151

Bibliography 165 Index 177

ILLUSTRATIONS

Frontispiece: Pannini's View of the Interior of the Pantheon

1. Photographs of the Pantheon General views, pls. 1-7 Details, pls. 8-29 Construction, pls. 30-48, 49Β Comparable Hadrianic Sculpture, pls. 49Α, 50-55

2. Drawings and Prints of the Pantheon Peruzzi to Dosio, c. 1490-c. 1569 Peruzzi and others, c. 1490-c. 1535; pls. 56-58, 60B, 70 Heemskerck, 1535-1536; pls. 59-60 Α Serlio, 1540; pl. 61 Dosio, c. 1550-1569; pl. 62 Palladio, c. 1550-1570; pls. 63-69 Desgodetz, 1682; pls. 71-73 Piranesi, c. 1750-1790 G. B. Piranesi, c. 1750; pls. 74-75 Francesco Piranesi, 1790; pls. 76-104 Fea to Lanciani, 1806-1892 Fea, 1806; pl. 105E Leclère, Choisy, and others, 1813-1873; 105Α, 106-111, 115 Β Lanciani, 1882-1901; pls. 112-113 Middleton, 1892; pls. 114-115 Α Chedanne, Beltrami, and Armanini, 1891-1898; pls. 116-125 Gismondi to Terenzio Gismondi and others, 1927-c. 1930; pls. 126-127, 132Β-133 Beltrami, 1929; pls. 128-129 Terenzio, c. 1932; pls. 130-132 Α Licht to Wilson Jones; 1966-2000 Licht, 1966; pls. 134-137 Pelletti, 1989, and others; pls. 138-139, 144, 145Β Wilson Jones; 2000, pls. 140-143, 145Α

3. Comparable Buildings A. Republican Period, 500-29 BC

c. 100 BC, Round Temple by the Tiber, pl. 146E c. 80 BC, Temple of Fortuna,

Praeneste (Palestrina), pl. 147Α c. 80 BC, Round Temple at Tivoli, pls. 147B-148Β Augustan Period (29 BC-14 AD), c. 30 BC, Basilica at Fano, pl. 148Β 28 BC, Mausoleum of Augustus (first used), pl. 149 c. 25 BC, Tomb of Caecilia Metella, pl. 150 16 BC, Maison Carrée, pl. 151 Α c. 11 BC, Theatre of Marcellus (completed; begun c. 44 BC), pl. 151B 2 BC, Forum of Augustus completed

(including the Temple of Mars Ultor), pls. 152-156ΑΑD 6, Temple of Castor and Pollux (rebuilt), pl. 156B 10, Temple of Concord , pl. 157Α 10-27, Triumphal Arch at Orange, pl. 158Α 38-c. 50, Aqueducts and Temple of Mercury, Baia, pl. 158-159 Nero and the Flavians, AD 54-96 c. 64, Nero's Aqueduct, pl. 161 Α c. 64, Nero's Golden House (Domus Aurea), pls. 161 B-162 70-80, Colosseum (interior and upper story largely rebuilt from c. 217-240;

often repaired), pls. 163-165 Α c. 80, Flavian Amphitheatre, Pozzuoli, pl. 165E C. 90, Domitianic Baths at Albano, pls. 166-167Α c. 90, Domitíanic Great Hall,

Roman Forum, pl. 167E Trajanic Period, AD 98-117

104-109, Trajan's Baths, pls. 168-174

107-112, Trajan's Forum (with the Basilica Ulpia and Trajan's Column, pls. 175-177 107-110, Trajan's Markets, pls. 178-184 C. 110, Trajan's Arch at Beneventum, pl. 185 Hadrianic Period, AD 117-138, pls. 186-213 c. 123-130, Hadrian's Villa, pls. 186-201

Serapaeum, pls. 190-191 Roccabruna, pls. 192-194 Α

c. 126, Piazza D'Oro Vestibule, pls. 1 94B- 195 c. 126, Small Baths, pls. 197 C. 129, Large Baths, pls. 197 Α, 198-199 c.123-c. 138, Misc. Other Buildings, pls.

189, 196, 200-201, 209E c. 130, Temple of Venus and Rome (largely completed in. c. 135; platform

survives; rebuilt c. 307), pl. 202-203 C. 130, Hadrian's Library, Athens, pl. 204-205Α c. 130, Triumphal Arch at Antinoopolis, pl. 205Β C. 130-139, Hadrian's Tomb and Bridge (Pins Aelius), pls. 206-209Α C. 130, Temple of Trajan and Plotina, pl. 210 c. 130, Hadrianic Buildings at Ostia, pl. 211 c. 130, Other Hadrianic Buildings at Rome and Baia, pls. 212-214 Antonins Period (AD 138-193) c. 138, Temple of the Deified Hadrian

(Hadrianeum), pl. 215Α 140, Temple of Faustina (dedicated also to Antoninus

Pius in c. 160), pl. 215B Severan Period (AD 193-235), pls. 216-222 c. 203, Arch of Septimius Severus, pl. 2 1 6Α C. 203, Portico of Octavia (reconstructed by Septimius Severus and Caracalla), pls. 216E-217 211-216, Caracalla's Baths, pls. 218-221

c. 230, additions to Agrippa's Baths by Alexander Severus, pl. 222 c. 230-240, Round temple at Ostia, pl. 223, 224E Diocletian and Maxentius, AD 284-312

C. 300, Temple of Minerva Medica (Licinii garden pavilion), pls. 226-227 c. 300, Tor de' Schiavi, pls. 228-229

305-306, Diocietian 's Baths dedicated (eventually renovated as S. Maria degli

Angeli), pls. 230-232, 233 Β 306, Basilica Nova (largely built by Maxentius; reoriented by Constantine),

pls. 233Α, 234-235 c. 307, Temple of Venus and Rome

(rebuilt by Maxentius), pls. 236-237 Α Miscellaneous, pls. 237 Α-240

8

Abbreviations

AAR—American Academy in Rome

AJA—American Journal of Archaeology

BComm—Bullettino della Commissione Archeologica Comunale di Roma

CIL— Corpus Inscriptionum Latinarium (Dresse11891)

EUR—Museo della Civiltà Romana, Esposizione Universale di Roma

FU—Fototeca Unione (American Academy in Rome)

GW—illustrations owned by the author

ICCD—Istituto Centrale Per il Catalogo e la Documentazione (ICCD; Fototeca Nazionale, Ministero per I Beni e le Attività Culturali; including the Aerofoteca or Laboratorio per la Fotointerpretazione e la Aerofotogrammetria)

JRA Journal of Roman Archaeology

JRS—Journal of Roman Studies

JSAH—Journal of the Society of Architectural Historians

MAAR—Memoirs of the American Academy in Rome

PBSR—Papers of the British School at Rome

QuadIStA—Quaderni dell'Istituto di Storia dell Architettura

RM—Mitteilungen des Deutschen Archäologischen Instituts: Römische Abteilung

S. A. R. A.—Società Aerofotografie e Rilevamenti Aerofotogrammetrici

SBAAL—Soprintendenza ai Beni Ambientali e Archittonici di Lazio (which in 1990 divided its architectural photographs into separate groups for Lazio and Rome and transferred the Rome photographs to SBAPCR. In the 30's all of Lazio including Rome was under the Soprintendenza Monumenti del Lazio.)

SBAPCR—Soprintendenza per i Beni Architettonici ed il Paesaggio per il Comune di Roma (including architectural photographs created for the Soprintendente aí Monumenti del Lazio; cf. SBAAL)

Preface

This book is about everything that was required to create the Pantheon. It covers the design and construction of the building, and it includes information on sources of design, methods of de-sign, materials, and methods of construction.

In addition to being one of the most influential buildings ever created, the Pantheon is one of the most controversial. Every as-pect of its design and construction has been a subject for scholarly debate, and many continue to be. Various scholars have dated it from c. 25 BC to AD 180. Its portico has been called definitely original and unquestionably added. Its purpose has been consid-ered sacred and secular. Its dome has been said to rely primarily for support on brick ribs and to be monolithic. Its interior col-umns are thought to be structural or decorative. The facing of the attic story was removed for having been supposedly added yet was an integral part of the overall design.

A great deal is still unknown about the Pantheon. There con-tinues to be disagreement about whether or not some parts such as the portico are original. Little archaeological work has been done, and what has been done continues to be reinterpreted. Al! of the building has never been thoroughly and systematically ex-amined to determine how parts of it were constructed. Most of the upper surface of the dome has not been uncovered in centu-ries. Cores have never been taken to determine if brick arches ex-tend entirely through the walls. Despite many uncertainties, more than enough is known about the building and about closely related buildings for conclusions to be reached that are likely to be con-firmed. What is concealed beneath the surface of the largely in-tact building can be inferred with a high degree of probability on the basis of a careful examination of the parts that are accessible and on the basis of numerous comparisons with buildings that are in many respects similar.

For any building as poorly documented as the Pantheon, it is necessary to make a thorough examination of the building it-self. For three decades, I have examined most parts of the build-ing carefully and repeatedly. At various times, I have been able to visit every type of space in the drum, Transitional Block, and adjacent grottoni. I have been on the roof and within the walls. I believe I have been able to determine what parts are o riginal, which were added at various times during a period of nearly two thousand years, and how all parts of the building were construct-

ed. On the basis of this knowledge, I have attempted to recon-struct how it was designed and how it was intended to look.

I have considered the stages of design and construction in the or-der in which decisions needed to be made. Every decision has been considered from the point of view of the architect in an attempt to determine what he could have learned from earlier buildings, what he was attempting to accomplish, what decisions he needed to make, what options he had, why some options were preferable, what prob-lems arose during construction, what vestiges of redesign are present, and why redesign was sometimes necessary. Although some design sources may not have survived, some buildings constructed shortly before have so many of the same features that they are likely to have been directly influential, and they help to reconstruct how the Panthe-on was designed and constructed.

For poorly documented buildings, a sequence of numbers must be reconstructed that is capable of generating all major di-mensions. Since architects design using numbers, architectural historians have to understand how and why specific numbers were used to be able to reconstruct a design process correctly. To de-termine this, it is necessary to consider dimensions in Roman feet (0.296 meters), and unless otherwise indicated, all dimensions that are given in feet are in Roman feet.'

To reconstruct the design of the Pantheon, I have tried to ex-amine every surviving building that might have influenced it. I used an approach similar to the one that was necessary to deter-mine how the Parthenon was designed. 2 What the architects of the Parthenon were attempting to accomplish could only be de-termined by comparing all 43 of the best preserved Greek Doric temples that have peripteral plans. This was the only way to re-construct a design process capable of generating the dimensions needed at each stage of construction for the great majority of these temples. Similarly, how the Pantheon was designed required tak-ing into consideration all evidence about the building itself and related buildings and accounting for all of it.

Hundreds of books and articles have been written about the Pantheon, and a large part of them attempt to prove mutually ex-clusive theories. My conclusions rely to a large extent on the in-sights of earlier writers, but by using a more comprehensive and systematic approach, I have reached significantly different con-clusions that account more fully for the available evidence.

12

Acknowledgements

My greatest debt is to studies of the Pantheon by Kjeld de Fine Licht and by William L. MacDonald. Much of what is known about the building was assembled by each of these schol-ars independently and at about the same time, publishing their principal works in 1965 and 1966. Together they provided an in-valuable basis for the study of all aspects of the Pantheon.

Licht's magnificent monograph has great and permanent value. No other study is as comprehensive or as well ilustrated. I have cited his book far more frequently than any other, and my debt to him is greatest of all. However, he used a different ap-proach and came to many conclusions that I have been unable to accept. 3 Rather than starting with the foundations, he started with the portico and worked his way from north to south and to the basilica behind the Pantheon. He might well have reached different conclusions if he had considered the problems that had to be solved in the order in which they needed to be solved. He presents a large number of the most important sources of design, but only after he discussed the building.

MacDonald's section on the Pantheon in his seminal study of the development of Roman architecture gave further consider-ation to specific sources of design and to how the Pantheon fits within the context of Roman architecture as a whole. After eval-uating the same and some other sources for the building's de-sign and construction, I reached somewhat different conclusions about how and why they were important. What he wrote about the building's construction has been largely superceded, but his writings have had great influence and continue to be essential reading for the Pantheon and for Roman architecture. His later book on the Pantheon has long provided the best overview of its place in the history of world architecture. 4

A forthcoming collection edited by Mark Wilson Jones and Tod Marder for Cambridge University Press will add further to what is known about the impact of the Pantheon and will also add to what has been discovered about the building through re-cent studies. My chapter on specific design sources was adapted from an article prepared for their volume.

I have often had the benefit of discussing the building on site with the architect Randolph Martz and the architect and ar-chitectural historian Mark Wilson Jones. Mark has also provided a great deal of constructive criticism for the chapters on design sources and on the portico.

Part of the research in Rome was made possible by sabbati-cals provided by the Getty Center for the History of Art and the Humanities in 1991 and by the College of Charleston in 2007-2008. Four travel grants from the College of Charleston Library and a Provost's Grant enabled me to spend nearly a year and a half in Rome to examine all relevant buildings thoroughly and to do research in ancient design and construction methods. David Cohen and Marie Ferrara have supported my research by provid-ing flexible work and vacation schedules and by approving re-search leave and a sabbatical.

I am grateful to Will Evans and Joe Schmidt of Evans & Schmidt, Architects, who made a generous contribution to the Friends of the College Library that helped to fund this publica-tion. John M. Rivers, Jr., helped greatly to make this book pos-sible.

John Poindexter read and commented on the entire text and improved it in innumerable ways. Mel Smith has provided good advice for many years. Katina Strauch urged me repeatedly to apply for a sabbatical to complete this study. Their friendship and encouragement have been indispensible.

Many items in the bibliography were made available through the Interlibrary Loan Department of the College of Charleston, and I owe a great debt to Mike Phillips and his high-ly eftícient staff. The rarest titles were available at the American Academy in Rome, the Deutsches Archäologisches Institut, the Biblioteca di Archeologia e Storia dell'Arte, and the Biblioteca Hertziana.

Historic photographs have been provided primarily by the Soprintendenza per i Beni Archite ttonici ed il Paesaggio per il Comune di Roma (including photographs of Roman buildings created for the Sop rintendente ai Monumenti del Lazio); by the Istituto Centrale Per il Catalogo e la Documentazione (ICCD; Fototeca Nazionale, Ministero per I Beni e le Attività Culturali; including the Aerofototeca or Laborato rio per la Fotointerpre-tazione e la Aerofotogrammetria); and by the Fototeca Unione (American Academy in Rome). Nearly all photographs taken by Alinari, Brogi, and Anderson are from original albumens that I have been able to acquire through many years of collecting.

For access to areas of the Pantheon that are ordinarily inac-cessible, I am grateful to Armando Schiavo, Architect and Presi-dent of the Pontificio Accademia dei Virtuosi al Pantheon; Gi-usseppe Simone tta, Architect; Marco Bernardi; and Giovanni Belardi, Architect and Director of the Pantheon.

Chapter 1: Preliminary Considerations

(lA) Brief Description of the Pantheon

The Pantheon consists of a cylindrical drum, an integral Tran-sitional Block, and an attached portico (pls. 1 & 4). The Transi-tional Block is located in between the drum and the portico, and it will usually be referred to simply as the Block (pls. 3 Α & 77-80). The drum and Block share the same foundation and were initially connected to one another, and they were thus largely constructed at the same time. The portico was constructed soon afterwards, and it rests partly on foundations of the earlier building that had been constructed on the site by Agrippa. There are many indica-tions that the portico was redesigned both before and after the rest of the building was completed.

Α basilica on the south side of the Pantheon was built at the same time as the drum and was connected to the south side of the drum by a buttress. This "bridge" stabilized the clearstory of the basilica, which was initially an entirely separate building (p1.48). Α series of narrow rooms (the grottoni) were later constructed in between the drum and the basilica (pl. 131).

What is evident from the materials and methods of construc-tion has been largely confirmed by brickstamps. The rotunda, Block, and basilica were built at the same time, and they were constructed during the reign of Hadrian, but partly of mate rial that had been manufactured earlier. The grottoni were added later, but are also Hadrianic. The foundations for the present portico were almost certainly added during the Hadrianic Period, but since they are level, they were inserted after the stone foundations of Agrippa's Pantheon had been reused for the Block and had settled unevenly.

The interior of the rotunda is 150 Roman feet in diameter and in height (frontispiece). The Block is square and overall is 108 feet in height and width.

The interior walls are half of the total height of the interior; the dome rests on a cornice that is 75 feet above the floor, and the dome itself is also 75 feet high. The interior drum and dome thus relate to one another as 1:1, and each relates to the overall height as 1:2. The interior drum is subdivided into a main story that is 45 feet high with an attic story that is 30 feet high, and the attic relates to the main story as 2:3.

The principal ratios used to design the interior and exterior of the Pantheon are the simple ones of 1:1, 1:2, and 2:3. The princi-pal dimensions divide evenly by 5, 10, 25, and/or 50 feet. Simple

proportions and measurements divisible by five or ten were char-acteristic of imperial Roman architecture.

The interior wall is divided into two levels, and above its foundation the exterior wall is divided into three levels (pl. 130). Two of the three levels of the exterior correspond to the interior. The exterior has an additional story at a higher level. The founda-tion has a brick facing for about 7 of its 15 feet that was evidently intended to be exposed to the weather, but has been largely con-cealed by silt. The taller attic of the exterior starts at the level of the springing of the dome, and it resembles the surcharge of many domes, but like the lower stories of the drum, it contains numer-ous chambers that lightened rather maximized its weight.

Although the drum and Block began to be constructed to-gether, they began to be constructed separately at the level of the barrel vaults over the entrance. The entrance consists of two bar-rel vaults with different heights. The barrel vault inside springs from an entablature supported by columns 36 feet tall and outside above columns that are 48 feet tall. The relative difference in the height of these columns is 3:4 (both dimensions being divisible by 12). The barrel vault on the inside is part of the drum alone, and it extends upward and cuts through the attic story (pl. 85). The bar-rel vault on the outside is entirely a part of the Block, and the rest of the Block was constructed separately from the drum.

Opposite the entrance is an apse that is four feet wider and that has a semi-dome. The entrance arch and the apse establish a strong axis from north to south, and this is the principal axis of the building. This axis was further emphasized by columns that flank the apse and by other design features including the orthogonal pat-tern of the floor and the coffer pattern of the dome.

On the cross axis (to the east and west) are large, semi-circu-lar exedras with the form of apses, but their semi-domes are con-cealed by the attic. These two exedras and four others are fronted by columns in antis, and the diagonal axes (southeast, southwest, northeast, and northwest) have similar exedras, making a total of six exedras and eight openings in the main story. The entrance is two-feet narrower than the exedras and the apse two-feet wider. The four exedras on the diagonal axes have straight sidewalls and barrel vaults (rather than half-domes), and these barrel vaults are also concealed by the attic.

There are eight broad structural piers in between the eight major openings of the interior, and attached to the front of these piers are aedicules that resemble small porticoes. The pairs of

16

aedicules flanking the two main axes on the cardinal points have

triangular pediments. The pairs of aedicules flanking the diagonal

axes have pediments with segmental arcs. The materials used to

construct these two sets of aedicules also differ and further distin-guishes one set from the other.

The attic story of the interior has 14 windows. There is a

window above every exedra and above every aedicule, but none

above the entrance and apse. The six windows above the exedras

function as a clearstory that allows light from the oculus to enter

the tops of the exedras (frontispiece). The eight windows above

the structural piers are false windows that were added to maintain

a regular pattern, but behind these false windows are wall cham-bers, and they originally opened into these chambers at least until

the concrete had cured; all other wall chambers have openings to

the outside. Centered on all 16 openings in the attic are flank-ing pairs of pilasters (the bases of which were cut through by the

arches of the entrance and apse).

The dome has a single oculus 30 feet in diameter at its apex, and this opening is one-fifth of the overall diameter. It provides the principal light for the interior, but light also enters through the entrance, which is equivalent in size, and the doors were undoubt-edly kept open at least when trials were being conducted; Roman trials were open to the general public. The surface of the dome has five rows of 28 nearly trapezoidal coffers for a total of 140 coffers. Each coffer is subdivided into a series of increasingly smaller and deeper coffers, and the corners of the coffers align and create the impression of interlocking spirals.

The exterior of the dome has a recessed vertical wall at its base that resembles a podium. Above this wall are six concentric steps or steprings, and above the set of steprings is a saucer dome (p1. 4Β).

The Transitional Block is an essentially hollow structure with roughly triangular stairwells at each end. On its front is a pair of apses flanking the entrance (pls. 130 & 75). In the center is the larger arch of the front entrance. The foundations of the Block and of the drum within the portico were made of concrete without a facing and without the use of formwork; the concrete was sim-ply laid up in thin layers within a trench that had been dug through alluvial clay. The walls of the drum have no facing on the interior, but they have a brick facing on the exterior. Small amounts of opus retículatum (a tufa facing) have been found at the base of the south side, but nowhere else. The absence of a facing in front of the Block indicates that some kind of portico was planned to protect the north part of the foundation.

The rotunda was initially covered with bronze roof tiles. Af-ter being stolen, the tiles were eventually replaced with lead, large quantities of which had been used by the ancient Romans for wa-ter pipes.

The present portico has a tall pediment supported by eight columns in front and four columns or piers on the sides with two additional internal columns and a pair of piers to either side of the

entrance (pl. 131). There are a total of 16 columns and 4 piers, and the piers are adjacent to the Block. The outer columns m i-tially all had shafts made of gray granite; the inner columns have shafts of red granite; and the piers have shafts of white marble. These elements are arranged to create three distinct spaces: a broader space opposite the entrance (three interaxials wide) and two aisles that align with the apses on the front of the Block.

The columns across the front of the portico rest on travertine foundations that survived from a previous Pantheon (pls. 118-121). These stone foundations are very nearly 150 feet across (the width of the interior of the rotunda), and since the portico is 108 feet across, the sidewalls of the stone foundation extend roughly 20 feet beyond each side of the portico. These earliest surviving foundations are located well below the present ground level and are not visible; these rectangular foundations were discovered dur-ing the only major excavations of the Pantheon that took place in the late 19th Century.

The columns of the front of the portico rest on the north side of these pre-existing stone foundations, and the front of the

Block rests on the south side. Within the rectangle of stone were

inserted four foundation walls of shuttered concrete to support the

portico's four rows of columns and piers. These inner founda-tions were constructed within wooden formwork that left board

impressions. Since the foundations of shuttered concrete were not

pulled down with the stone foundations when the Block and drum

settled, they were added after the construction of at least much of

the body of the building (p1. 121). The present portico was constructed with columns that have

monolithic granite shafts, and the piers were made of marble

drums (segmented). The columns and piers have marble bases

and capitals, and marble was used for the entablatures. Brick

arches on stone pedestals were used to support the roof structure

(p1. 14Β). The existing marble trim of the portico and Block is

mitered in ways that indicate that the existing marble trim was

created at the same time. The roof structure originally consisted

of bronze roof trusses, which were replaced with wood.

The walls of the drum are about 20 feet thick, and like the Block, they are largely hollow (pls. 77-79). In addition to the eight

major openings in the main story of the interior, there are eight ap-sidal chambers within the walls with a chamber behind each aedi-cule. The attic story of the drum is similarly hollowed out by the

concealed barrel vaults above the six exedras in addition to a visible barrel vault above the entrance and a semi-dome above the apse. Above the east and west exedras are half-domes. Above the four exedras on diagonal axes are curving barrel vaults, each of which is subdivided by two smaller arches running from front to back. In between the arches of the exedras is another set of apsidal chambers at the attic level that align with the lower set of chambers (p1. 78). On the exterior, the uppermost part of the drum that conceals the base of the dome contains sixteen chambers that correspond in posi-tion to lower chambers or vaulted openings in the walls (p1. 79).

17

The interior of the nearby basilica to the south of the Panthe-on is 150 feet long (like the interior of the Rotunda and the width of the stone foundations of the earlier Pantheon; pl. 76). Although this basilica was begun as a separate structure, it was later con-nected to the top of the Pantheon's drum. An integral projection that survives on the south side indicates that the Pantheon was intended to buttress the Basilica of Neptune (pls. 48 & 82-83). Clearstories were often buttressed, and a smaller clearstory cannot have been expected to have buttressed a larger dome.

After at least the walls of the Pantheon and the Basilica of Neptune had been completed, eight parallel walls was constructed on entirely separate foundations in between the two buildings. The foundations of these walls are shuttered concrete like the foundations for the inner columns of the portico, and the walls that rest on them are concrete faced with brick that have stamps dating from the Hadrianic and Severan periods. The walls were constructed during the Hadrianic Period, and arches in the walls were reinforced during the Severan Period.

Concrete, brick, stone, and bronze were used to create a fire-proof building on the site of two earlier Pantheons that had both burned. The Rotunda was centered axially in relation to the pre-existing stone foundations, and the front of the portico aligns with these foundations (even though they are 5 degrees west of North rather than correctly oriented). Considering that the portico aligns with the front of these stone foundations, that the interior of the Rotunda has the same width and orientation as these stone foun-dations, and that the name of the building and of its initial builder were reused, the existing remains on the site can be considered the starting point for the design.

(1B) Date of Construction

To be able to distinguish a potential source (an earlier build-ing) from a parallel example of influence (a later building), it is necessary to indicate with as much precision as possible when the Pantheon was constructed. There has been much disagreement about when the Pantheon was constructed, and its dates continue to be disputed.

For 17 centuries, from around AD 200 until 1882, the inscrip-tion on the frieze of the existing building stating that Agrippa built it was taken at face value. A study of brickstamps by Chedanne indicated, though, that all parts of the building were constructed using bricks that were made during the reign of the Emperor Hadrian or shortly before. Chedanne's research was confirmed by the excavations of Beltrami and Αrmanini.'

Brickstamps had been recorded from the Pantheon as early as the 18th Century, but their full significance was not realized (pls. 103-104). All of the evidence available to 1936 was reviewed in detail by Guey, who found that more brickstamps from the Pan-them occur at Hadrian's Villa than in any other major building

or group of buildings? A. C. G. Smith later confirmed that the Pantheon is contemporary with Hadrian's Villa.'

Bloch's summary of Guey's evidence indicated that of 120 dat-able stamps found in situ in the Pantheon and adjacent grott οni, 115 were probably made during Hadrian's reign (AD 117-138) 4 Bloch concluded that the Pantheon was constructed during the first half of Hadrian's reign beginning around 118 or 119, when Hadrian was first in Rome after becoming emperor, and is likely to have been dedicated from 125 to 128, when he was again in Rome.

Some aspects of the evidence of brickstamps are controver-sial, and insufficient consideration has been given to the strati-graphic significance of the specific locations of finds. There is no question, though, that most of the building including its dome is Hadrianic rather than Augustan, Trajanic, or Severan. There is good reason to believe that the interpretations of the evidence of brickstamps by Guey, Bloch, and Smith are correct and that the building was wholly constructed during the Hadrianic Period. This evidence will be discussed in detail later in this section.

The evidence relating to the date of the portico is still more difficult to determine than for the drum and Transitional Block since its columns rest on shuttered concrete rather than on con-crete made using bricks that can be dated, but the evidence as a whole favors a Hadrianic date: Identical foundations used for the grottoni support brick-faced walls with Hadrianic brickstamps. Large monolithic columns of Egyptian granite began to be used widely around the beginning of the 2nd Century AD rather than during the Augustan Period. Only the stone foundations of the portico are probably Augustan. If Domitian had built new foun-dations, they would probably have been made of concrete, which he used for his nearby stadium and for his palace on the Palatine

The number of years required to construct the Pantheon is uncertain, but a more crucial factor was how long concrete needed to cure rather than how many workmen could be employed at one time. Some parts of the walls are about 20-feet thick, and since thick layers of concrete had to be given more than sufficient time to set up in order to support additional layers, the construction of the building probably took at least several years.

This estimate of the time required is comparable to the num-ber of years known to have been needed to construct other large buildings primarily of concrete. For example, the Colosseum was constructed from AD 70-80 (p1. 164), Trajan's Baths from AD 104-109 (pl. 168), and the Baths of Caracalla from AD 211-216 (p1.218). 5 Since the Pantheon was a smaller project, it may have taken less than five years to build, but considering that its walls are so much thicker than usual even for large buildings, the walls needed an unusually long time to cure before they could securely support the immensely heavy dome. The length of time that was considered necessary for concrete to develop sufficient strength to support vaulting could well have been several years.

Like other building materials, bricks seem to have been given time to prove their soundness. Vitruvius wrote:

18

with regard to burnt brick, nobody can tell offhand whether it is of the best or unfit to use in a wall, because its strength can be tested only after it has been used on a roof and exposed to bad weather and time—then, íf it is good it is accepted. If not made of good clay or if not baked sufficiently, it shows itself defec-tive there when exposed to frosts and rime. Bricks that will not stand exposure on roofs can never be strong enough to carry its load in a wall. Hence the strongest burnt brick walls are those which are constructed out of old roofing tiles.'

Vitruvius stated that even stone needed to be stored for two years before using,' and Pliny stated that lime should not be used for three years. 8 Wood was also allowed to dry for long periods be-fore being used. Considering the lack of control in the quality of production that is reflected in Vitruvius' distrust of bricks, they too must have been set aside until it could be determined if they were likely to weather well. Although bricks are now used as soon as they have been fired, they are made of clay that has been tested chemically, that has been mixed uniformly, and that has been fired at a consistent temperature, and random samples are tested for compressive strength. Having none of these methods to ensure uniformity, the Romans had to rely on how well each batch of bricks weathered.

Bricks were sometimes used long after having been manu-factured, and the reasons include lack of demand and stockpiling to wait for better prices. Some bricks were also salvaged from buildings that had burned and were reused. It is necessary to con-sider as many bricks as possible and to evaluate every brickstamp from any building that is being dated. As Lanciani cautioned:

When we examine, for instance, the grain warehouses at Os-tia, or Hadrian's villa at Tivoli and find that their walls have never undergone repairs, that their masonry is characteristic of the first quarter of the second century, that their bricks bear the dates of Hadrian's age and no others, we may rest assured that the stamps speak the truth. Their evidence is, in such a case, conclusive. But if the bricks are variously dated, or bear the names of various kilns, and not of one or two only, then their value as an evidence of the date of a building is diminished, if not lost altogether. 9

In the case of the Pantheon, "the bricks are variously dated" and "bear the names of various kilns," and the evidence as a whole has to be evaluated rather using any portion of it that might seem to be more reliable. Even bricks that have the names of consuls and that can be dated to the year they were manufactured cannot be taken at face value unless they occur in sufficient numbers by themselves, and even then the bricks are unlikely to have been used for at least several years after being manufactured.

Archaeological evidence provides confirmation that bricks were ordinarily used several years after being made. Bloch states that the Temple of Serapis, Ostia, was dedicated 24 Jan. 127. 10 Its walls "...seem to be faced almost entirely with bricks of Julius Lupin made in 123...."" Thus, the temple was "almost entirely"

constructed during the three years before it was dedicated. Later, he added that "within the Serapaeum and its precinct, to be sure, brick-stamps of 123 are in the overwhelming majority, but a [sin-gle] stamp of 124 (S. 98) appears in the cella wall in a place which cannot possibly be a latter addition. " 12 One other stamp dated 125 was found in the street entrance, and two stamps dated 126 were found in what is "unquestionably an addition to the somewhat higher structure behind" the temple. Not a single stamp later than 124 and only one that was dated 124 was found in the temple ded-icated at the beginning of 127. Although Bloch himself insisted that the street entrance was surely constructed before the temple was dedicated, he indicated that this was a conjecture. Regardless of whether or not this was the case, he concluded that "We shall therefore be allowed in the future to be more confident in assign-ing buildings in which great masses of brick-stamps of 123, 124, and 125 occur to the period immediately following these years, that is to the years from 126 to 128." He did not think that bricks were used in the year they were manufactured, and in fact, he spe-cifically states that there was an average of a three-year delay in their use. Lancaster found definite evidence that some bricks in the same level at Trajan's Market were used about a decade or more after being manufactured, and she also found that five dif-ferent brickstamps were used for one set of rooms, a circumstance that "would suggest they had been stockpiled together. "' 4

Bloch gave special consideration to why 13 dated brick-stamps found in the Pantheon were manufactured by Lupus dur-ing the Trajanic Period. 15 "It was the high official of the Emperor Trajan, Marcus Rutilius Lupus, who in 110 A.D. first dated his stamps. From 114 to 117, while he was viceroy of Egypt, he resumed this practice, obviously in order to control the brick production of his estates more easily during his absence."" He added that "the example which Lupus gave for the organization of brick production was soon imitated. In the year 123, almost all brick producers of Rome marked their bricks with dated stamps, probably also for fiscal reasons...."" In a later publication Bloch noted, "my interpretation that he [Lupus] must have disposed of considerable reserves from all four years [114-117] upon his re-turn from Egypt still stands and has not been upset by later ar-chaeological discoveries. " 18 Thus, the small number of bricks found in the Pantheon that were manufactured during the Trajanic Period can be accounted for by the usual delay in their use to en-sure that they had been tested sufficiently against the effects of weathering. In addition, he pointed out the likelihood that they were stockpiled during the Parthian War (AD 113-116) in order to get a better price after the war ended.

Heilmeyer interpreted the Pantheon's brickstamps as indicat-ing that the building was begun during the reign of Trajan, but he consistently preferred the earliest possible date that a brick could have been manufactured rather than, like Block, taking an aver-age within a range of possible dates as more likely than either extreme. Heilmeyer noted that only two brickstamps definitely

19

post-date Hadrian's reign, and renovations by Antoninus Pius and

Septimius Severus are dοcumented. 19 Bloch asked the question, "Why were bricks dated? He

did not know why they suddenly began to be dated, stopped be-ing dated, and resumed being dated. He wrote, "Axel Βο thius in a kind and generous review of my original work on Roman

brick-stamps and at the same time in an article in Eranos sug-gested that the dates were used to assure the buyer that the bricks

had seasoned enough. This hypothesis was inspired by Vitruvius'

recommendation for testing baked bricks by using them for some time as roof tiles and thus exposing them to the weather." 20

Bricks were not made during the Augustan period specificai-ly to face concrete. Tufa was largely used as a facing, and when bricks were used, they were roof tiles that had been salvaged as when buildings were pulled down to be replaced or had been burned in one of the numerous fires that destroyed large parts of Rome. Roof tiles were cut down, and their edges were ground flat for reuse. When brick were found to weather better than tufa, bricks began to be manufactured specifically to be used as a fac-ing for concrete. 21

In his summary of the brick industry, Bloch notes that dated stamps were used "frequently between 110 and 164 AD," but rare-ly afterwards:

It is difficult to see why producers would have stopped using stamps if they had been useful to them, but in any case there

was little consumption between 164 and about 298, and the

brick industry came to be so entirely owned by the emperor that controls could be applied more effectively without relying on the honesty of a private brick maker.... The inconsistent use

of dating cannot be readily explained by any single reason, but

the best reason to date bricks was to ensure that they had been

exposed to weathering for several years.

The largest bricks, bipedales, were often dated during the reign of Hadrian (pls. 35 Β and 49Β). Bloch noted that brickstamps from the numerous Hadrianic structures at Ostia had brickstamps with consular dates for 11 years of his reign: 120, 123-130, 134, and 138. 22 New stamps were thus being regularly made and used throughout Hadrian's reign rather than outdated stamps continuing to be used.

Although stamps dating from 123 are encountered far more frequently than for any other year, this cannot be taken as evi- dence that the tools made for that year continued to be used indef- initely. Considering that Hadrian was an architect, it is likely that an unusually large quantity of brick had been manufactured on his orders in anticipation of his return from his initial period of travel.

Still greater caution must be used when considering evidence such as the thickness of bricks and mortar, color, and aggregates. All of this type of evidence can be useful as, for example, dis- tinguishing repairs when bricks cannot be removed to check for brickstamps. Although there are many exceptions, Lanciani noted

that "the fundamental rule for the chronology of brick structures is this: the thinner the bed of cement between the layers of bricks, the older the structure." 23 In general, Hadrianic brickwork (AD 117-138) is characterized by bricks that are even in their thickness and by mortar joints about a third as thick as the bricks. Severan brickwork (AD 193-235) can usually be distinguished from Hadri-anic, and it is characterized by uneven bricks with mortar joints about half as thick as the bricks. Brickwork dating from as late as the reign of Diocletian (AD 284-305) can be more readily dis-tinguished from much earlier brickwork by its use of mortar joints that are about half as thick as the bricks, and Christian repairs have joints about as thick as the bricks. Finer distinctions than these are difficult to make.

Hadrian became emperor on 11 August 117, but he was in Antioch when Trajan died, and he did not arrive in Rome until nearly a year later in July 118. 24 He was in Rome from 118-121, and it was during this period that he probably arranged for the manufacture of a large number of bricks in AD 123 to provide far greater quantities of brick for the numerous buildings he wanted to create. More bricks have been found dating from this year than from any other.

Only two bricks dated AD 123 have been found in the Pan-theon, and considering that nearly 200 brickstamps have been found there, so small a number dating from this year is undoubt-edly significant. The Pantheon must have been largely completed by AD 126 or it would have more bricks dating from AD 123. If a fragmentary stamp identified as dating from AD 123 actually came from about one meter above the floor of the Pantheon, the rules of stratigraphy would require that everything above one me-ter to have been created afterwards. 25 As Wilson Jones argues in a forthcoming article, this single stamp is likely to reflect the in-stallation of trim during the final stages of construction. Further investigation may result in more bricks dating from AD 123 being found in the dome, where relatively few bipedales have been re-moved to be checked for brickstamps.

Smith's article on Hadrian's Villa provides highly valu-able evidence for dating the Pantheon in relation to the vi11a. 26

He notes that 11 brickstamps occur both in the Pantheon and in buildings at the Villa, the buildings of which were created almost entirely during Hadrian's reign. One of the stamps widely used at the villa (number 315) was found in locations throughout the Pantheon. His summary of Block's evidence makes untenable Heilmeyer's assertion that most of the Pantheon was constructed earlier, and his conclusion is well supported: "I feel therefore that the evidence of the brickstamps support the conventional dat-ing of the Pantheon in the early years of Hadrian's reign rather than an earlier dating." 27 He noted that five stamps occur in both the grottoni of the Pantheon and at the Villa; thus, there can be no question that the grottoni are primarily Hadrianic (with the exception of the west wall, which is Severan and which he does not mention). Hadrianic stamps were found in both stories of the

20

grottoni and in the arch in between the Pantheon and the Basilica of Neptune.

The building at Hadrian's Villa with features that most close-ly resemble those of the Pantheon is the Serapaeum, which is part

of the Canopus (pls. 190-191; Chapter 6B2). Domenico Faccenna

identified the 279 brickstamps found in the excavation of the Ca-nopus from 1950-1955 as having the following consular dates:

189 from AD 123; 11 from AD 124; 17 from AD 125; 61 from

AD 126; and 1 from AD 127. 28 This key complex of buildings was

thus constructed from about AD 126-129. It cannot have been

constructed before 123 or long after 130. Mostly likely, it was

begun as the Pantheon was being completed, but its dome could

predate the Pantheon's dome. The building methods required to

construct the Pantheon's dome must have been tested somewhere on a smaller scale before being used on a building with an unprec-edented span. For this to be confirmed, all brickstamps from the

dome of the Pantheon and from the dome of the Canopus would

need to be compared, and sufficient information is not now avail-able. Some additional samples may also need to be taken. It will

eventually be possible to use brickstamps to settle this question,

but the level where every brick was found needs to be known to

determine which dome was begun and completed first. During the Hadrianic period, stamps were more consistent-

ly used than earlier or later and were used on most of the larg-est bricks (bipedales, which have stamps from 50 to 100 percent

of the time depending on the manufacturer). Only about 1 in 10 smaller bricks were ordinarily stamped, and many of these do not

have the names of consuls and can only be dated approximately

on the basis of their occurrence in buildings of known date. In

the case of the Pantheon, they do not provide better evidence than

numerous stamps with the names of the consuls, but they must

be taken into consideration. Although dated stamps cannot have

been used earlier than the date on them, they can have been used

much later than when they were manufactured and so must also be

interpreted with caution. They cannot, for example, be assumed

to have been used immediately after manufacture for the construc-tion of any building.

Bloch usually summarized his findings numerically by brick-stamp for major Roman buildings, and his summary facilitates

comparing the 43 types of stamps found in the Pantheon with

finds for the same stamps in earlier and contemporary buildings:

Trajan's Baths: 3 identical stamps, one of which is H αdri-anic. 29

Trajan's Markets: 2 identical stamps, one of which also oc-curs in Trajan's Baths and the other of which may occur in

another part of Trajan's Forum. 30 Trajan's Forum (Basilica Ulpia and Bibliotheca only): 2 iden-tical stamps, one of which occurs in Trajan's Baths and Mar-kets. 31 Combined total for Trajan's Baths and Forum (including its Markets): 3 identical stamps (2 of which are variants of the

same stamp with the possibility of an additional variant also

being represented). 32

Combined total for Hadrian's Villa: 11 identical stamps. 33-36

This simple comparison indicates that Hadrian's Villa pro-duced about three times as many stamps used in the Pantheon as were used in Trajan's Baths and Forum (including the Markets). It provides an example of how brickstamps can be compared more objectively. Although the size of the samples is small, the results are consistent, and variant stamps were represented in proportion-ate quantities at each site. All that can be concluded on the basis of so small a sample is that the Pantheon has significantly more in common with buildings of the Hadrianic Period than of the Traja-nic Period. Brickstamps alone do not suffice for any definite con-clusions about when the Pantheon was begun and completed, but help greatly to narrow the range of possible dates. The evidence of brickstamps will become increasingly valuable when all data on where every type has been found is more readily available and when more objective comparisons can be more readily made us-ing a searchable database. There is no doubt that brickstamps can be even more valuable than they are as evidence, but not until the methods for their interpretation involve fewer assumptions.

In summary, since manpower and funding available to a Ro-man emperor were almost unlimited, since the need for materials had been anticipated, and since Hadrian was highly knowledge-able about architecture, the period of construction is more likely to have been shorter rather than longer than usual. It is unlikely that the Pantheon took longer to construct than much larger build-ings such as imperial baths, but it cannot have taken a great deal less time. Crucial consideration were the thickness of walls and the weight of vaulting. Wall thickness determined how long con-struction would take, and thicker walls took longer to cure be-cause all heat from the chemical reaction that was involved had to escape. The length of time required for one layer of concrete to reach surncient strength for another layer to be added was crucial (Chapter 4G). Since this had long been known through trial and error, the only delays in construction are likely to have been caused by design problems which arose as a building with an unprecedented span was being constructed. Since there is also evidence of changes in design during construction, some further delays are likely to have added to the total period of construction, but the Pantheon cannot have taken longer to construct than far larger buildings.

It is most likely that Bloch was correct that the Pantheon was constructed from AD c. 118-126. It was in AD 118 that Hadrian first reached Rome after becoming emperor, and it was around 126 that the large number of brickstamps manufactured in 123 first began to be used at Ostia. His interpretation of the dates pro-vided by brickstamps is consistent with the historical and archi-tectural evidence. Although even dated brickstamps do not enable buildings to be dated with precision, they are particularly useful for establishing relative dates, and this will also be useful when

21

considering whether a particular building was designed earlier or later than the Pantheon and consequently if it could have been a source for the Pantheon's design.

(IC) Confirming the Identity of the Three Pantheons

The best evidence that the present building was referred to in antiquity as the Pantheon are two specific references with descrip-tions:

Dio Cassius wrote in c. 229 that the "Pantheion" has a "vault-ed roof... [that] resembles the heavens." 31 Ammianus Marcellinus wrote in c. 380: "the Pantheon [Pan-theum] like a rounded city-district, vaulted over in lo#y beau- ty "38

These two references indicate that the name Pantheon was being used from at least c. 200-c. 380 AD to refer to the existing build-ing, which by 200 was about 75 years old.

The Severan inscription on the Pantheon's portico can be dat-ed to AD 202 or within a couple of decades of when Dio was writ-ing about the Pantheon. 39 The name Pantheum has been repeatedly said to be part of this inscription, and it probably was when the inscription was transcribed in the 15th and 16th centuries. Now, only the letters "PANT..." are visible (with the rest of the word having been lost through spalling; figs. 94-95), but the amount of space left where the letters have eroded suffices for the full name. Multiple copies of this inscription differ so much that they were undoubtedly made independently of one another, and several made by reliable scholars include "PANTHEUM." The meaning of these different transcriptions is essentially the same, differing only in minor wording.

The most widely accepted version in the Corpus Inscriptio-num Latinarium was based primarily on a copy by Poggio Brac-ciolini made in the 15th Century. Letters that are not now visible are in italics:

IMP CAES L SEPTIMIVS SEVERVS PIVS PERTINAX ARABIC VS ADIABENICV S PARTHICVS PONTIF MAX TRIB POTEST XIMP XI COS III P P PROCOS ET/ IMP CAES M AVRELIVS ANTONINVS PLUS FELIX AVG TRIG POTESTAT V COS PROCOS PANTHEVM VETVSTATE CORRVPTVM CVM OMNΙ CVLTV RΕSΤΙΤVERVΝΤ.40

Leaving out most titles, the inscription translates, "The Emperor L. Septimius Severus... and the Emperor M. Aurelius Antoninus [Caracalla]... restored with every care the Pantheon, repairing the ravage of time." 41

Andrea Palladio's version of this inscription is essentially the same. He wrote out "PANTHEVM" fully, but omitted "COR-RVΤVΜ. "42 It differs sufficiently in minor points that it was definitely not copied from Poggi ο 's43 The inscription was often recorded later in very nearly the same way as in Desgodetz's tran-

scription, which was not copied either from Poggio or Palladio. 44 These three readings must have been taken from the inscription itself, and increasing deterioration in the 15th, 16th, and 17th centuries suffices to account for minor differences. 45 Although the inscription is now badly eroded, parts are well preserved, and judging by the earlier transcriptions, by the initial letters, by the spaces without letters, and by the sense of the inscription, there is good reason to believe that the word "PANTHEVM" was part of the inscription. Although the transcriptions do not provide better evidence to identify the building, better evidence is not needed.

Palladio did not hesitate to abbreviate or to expand abbrevia-tions. For the well preserved inscription on the frieze, he wrote, "Μ. AGRIPPA L. F. COS. III. FECIT," and he abbreviated TER-TIUM as "III" (pl. 11). This translates as, "Marcus Agrippa, son of Lucius, consul for the third time, built this." 46 The implication is that the building was constructed in or somewhat later than 27 BC, but it was probably begun in 25 BC. 47 The inscription on the frieze had bronze letters added in the 19th Century, but the bronze accurately follows the outlines of letters that show distinctly in earlier photographs.

The existing building was believed to have been built by Agrippa until Chedanne's discovery of Hadrianic brickstamps in 1892. Dii's description of AD c. 229 undoubtedly refers to the present building, and even he was evidently misled by the inscrip-tion into thinking that this building was the Pantheon constructed by Agrippa even though he knew, but had forgotten, that it burned in AD 80, was rebuilt by Domitian, and burned again in AD 1 10 48 Moreover, since the present building is fireproof, it is definitely not the building constructed by Agrippa and rebuilt by Domitian.

There is no definite evidence to tie the stone foundations under the portico of the present Pantheon to either Agrippa or Domitian. Considering, though, that construction in solid stone was characteristic of the Augustan Period and rare in the Domiti-anic Period, it is more likely that the stone foundations were built by Agrippa. Considering that inscriptions on the portico refer to Agrippa and that they referred also to the Pantheon, the indirect evidence suffices to place all three buildings on the same site and to identify the pre-existing foundations as Augustan. The extent of Domitian's work on the site is wholly unknown, and although he may well have rebuilt Agrippa's building, he is unlikely to have rebuilt its foundations.

The reason why Hadrian put Agrippa's name on his Pan-theon or allowed it to remain on some reused parts of Agrip-pa's Pantheon is likely to have been in emulation of Augustus. The inscription by Augustus that was placed at the entrance of his mausoleum (his Res Gestae or list of accomplishments) in-cluded the statement that he "restored [the Theatre of Pompey] at great expense without the inscription of my name." 49 By contrast, Domitian was criticized among other reasons for put-ting his name on buildings that he merely restored. Hadrian made it a regular practice not to add his name to any building

22

he built or restored with the exception of the Temple of Trajan and Plο tina. 58

Gordon considered the inscription to Agrippa and concluded that there was "nothing in the lettering to interfere with a Hadri-anic date. However, he added that "no reliable criteria have yet been established for distinguishing [the lettering] of one century from another" earlier than the 4th Century ΑD. 51 There seems to be no significant difference between the forms of the letters of the inscription in honor of Agrippa and the inscription by Severs or in the inscription by Hadrian on the bridge to his tomb.

(1D) Purpose of the Existing Pantheon

Dio Cassius wrote that Hadrian "held court with the assis-tance of prominent men now in the palace [Palatine] or again in the Form, the Pantheon, and in many other places, always on a platform, so that what was done was open to public inspection." 52

This indicates that the Pantheon was used as a courtroom and that it was open to the public so that anyone who wished to could see that justice was being administered impartially. This is the sole purpose the building is known to have served during the reign of Hadrian.

It was not unusual for emperors to preside over major trials. Trajan is known, for example, to have presided over the court of the centumviri. Pliny the Younger, a lawyer as well as an author, described a meeting of this court in the Basilica Julia in which 80 judges sat together to hear a case of exceptional importance. Or-dinarily, this court was divided into sections with 20 jurors hear-ing cases of lesser importance. Pliny wrote that on the occasion of the larger trial, the Basilica Julia was barely sufficient for all the people who wanted to attend.S 3 Such large attendance in some cases indicated the need for still larger basilicas, and successively the Basilica Ulpia, the Pantheon, and the Basilica Nova fulfilled this need.

Although court cases were not invariably held in basilicas, providing a place for cases to be held was one of the primary purposes of basilicas as a building type, and it was not a primary purpose of any other building type. Basilicas were usually rectan-gular such as the basilica Vitrvius designed for Fano (p1. 148 Α; Chapter 3 Α), but for the Basilica Ulpia, Trajan added to each end of a rectangular space a pair of semicircular exedras with the diameter of the Pantheon, and together they had the same floor space as the Pantheon, but they were roofed with wood (pls. 175-176). Smaller law cases could have been held in exedras of the Pantheon or in the adjacent Basilica of Neptune, and Dio states that Hadrian heard cases elsewhere besides the Pantheon.

Vitrvius noted that his arrangements for his basilica at Fano ensured that court cases would not interfered with business, and he implies that the commercial function was the primary one and the legal function secondary and more for the convenience of the

public than the court. Law cases could be heard on the outside of temples, but even this was exceptional. Augustus sometimes heard cases in the peristyle (that is, outside) the Temple of Her-cules at Tivoli. 54 Law cases were not heard inside temples," and the Pantheon does not have many of the basic features common to temples. It was undoubtedly used for other purposes besides trials, but it cannot initially have been a temple.

As a building type, a basilica has been defined as "a large public building whose principal function was to house law courts," 56 and it invariably seems to have served this function, but others as well. It was not exclusively a courthouse. Whether or not a case was being heard, basilicas were open to the public to discuss commercial transactions, politics, or any other subject.

There is no evidence for commercial use in connection with the Pantheon, and it is unlikely that its few exedras were used as offkces or shops, but the main space was undoubtedly used as an exchange and as a place for individuals to have discussions. Similarly, the forecourt of the Pantheon was in between a form and a bath, and its colonnade is likely to have served as a form, but more as a place to meet or discuss business rather than as a place with shops. When voting ceased in the adjacent Saepta Ju-lia, it was converted into an immense form and used primarily for commerce, a function it continued to serve after Hadrian re-built at least the part of it that is adjacent to the Pantheon.S 7

Dio's separate account of Agrippa's Pantheon of c. 25 BC and his account of the Pantheon as it existed in c. AD 229 were so conflated that what he wrote needs to be sorted out. About the earlier building, he wrote that Agrippa had wanted to place a statue of Augustus inside, but Augustus had refused to permit this, and, instead, Agrippa placed statues of Augustus and of himself outside the building. If the building had been planned or dedi-cated as a temple, Agrippa would certainly not have suggested placing his own statue inside. Undoubtedly the interior contained statues of some, though by no means all, of the gods, but statues of gods could be set up anywhere including, for example, an ath-letic facility.

If Agrippa's or Hadrian's Pantheon had housed statues of all the gods or even of the 12 Olympic gods, there would have been no reason to doubt why it was called the Pantheon. It cannot have had either because Dio frankly admitted that he did not understand why the building was referred to as the Pantheon: "It has this name, per-haps because it received the images of many gods and among them the statues of Mars and Venus; but my own opinion is that the name is due to its round shape, like the sky." 58 He was writing about a century and a half after Agrippa's Pantheon had been destroyed, and he confused Hadrian's dome with Agrippa's building.

By around AD 229 when Dio wrote, the Pantheon then did house "many," but not all gods or even the principal ones. It probably had statues of gods that Agrippa had undoubtedly hon-ored: the deified Julius Caesar and his divine ancestress Venus. Hadrian's Pantheon may have added a statue of Augustus on the

23

interior, but neither building was a temple to all the gods or all deified emperors. After being deified, both Julius and Augustus were given separate temples for their worship, and Hadrian was given a separate temple when he was later deified. As Ziolkowski pointed out, the inscription with Agrippa's name would be unique if it were for a temple in that it provides no indication which god or gods the building was supposedly built to honor. 59

Pliny the Elder stated that Agrippa's Pantheon had a pedi-ment, and in this respect it resembled a temple, but Augustus had been allowed to place a pediment at the entrance of his house. Pediments had continually less religious significance, but pedi-mented porticoes continued to designate a temple. Similarly, the depiction of sacrificial instruments on a building did not invari-ably indicate that it had a sacred purpose.

Further evidence that the existing building was constructed as a basilica rather than as a temple is provided by the building itself. It has the largest interior space created during Antiquity, and temples were not gathering spaces. It had a forecourt similar to the one designed for the Basilica Ulpia rather than being placed within a court like the Temple of Venus and Rome or projecting into a court like the Temple of Mars Ultor. In its forecourt was a triumphal arch, and a triumphal arch had been placed at the en-trance to the Basilica Ulpia, but would have been inappropriate on axis with the front of a temple. Temples invariably had an altar in front of them placed near enough to be visible to a cult statue in-side, and no evidence of an altar has been found despite repeated excavations in front of the Pantheon. Roman temples character-istically were built on a podium whether they were rectangular or circular, and when circular, they had a peristyle, and these are two additional features the Pantheon does not have. With the ex-ception of a pedimented portico, the evidence of typology corre-sponds to the only documented use of the building, and evidence will be presented that the pediment was added to a previously un-pedimented portico (Chapter 13C4).

(1E) Main Problems to Be Solved by the Architect

The principal challenge for the architect was to design the largest domed space ever created. A major part of the dome's im-pressiveness was intended to be its size. Hadrian had personally designed the largest temple and built the largest tomb ever con-structed in Rome, and by doing so, he was following the example of earlier emperors who had successively created the largest cir-cus (Julius), the largest theatre (Pompey), the largest amphitheatre (Vespasian), and the largest forum (Trajan). These Romans hoped to surpass their predecessors and their successors, and all of them succeeded. In terms of scale, Frontinus pointed out that the aque-ducts of Rome surpassed the pyramids in the effort required and that they were far more useful. The Romans customarily built on the largest possible scale.

Size was not enough to create the impression of immensity. The surface of the dome had to be subdivided in some way that would enable its depth, shape, and scale could be readily per-ceived. This was accomplished better than it had been before or has been since through the use of a skillfully designed pattern of coffers.

Permanence was a consideration equally important with size. It had to be achieved structurally, and the building needed to be entirely fireproof. Concrete was chosen for both reasons—for its structural soundness even in earthquakes and for being fireproof—as well as for enabling an unprecedented span to be achieved.

Since the Pantheon had previously burned twice, one of the principal concerns was to design a completely fireproof building to replace it. Area fires were taken for granted. Since the adjacent Saepta Julia had a wooden roof, the walls of the Pantheon needed to be firewalls like the wall behind the Forum of Augustus. The need for fireproofing explains why the Pantheon had no windows in its exterior walls, why brick-faced concrete was chosen as its principal material, why it was given unprecedented roof trusses of bronze, and why the shafts of its columns were granite. Windows facilitate the spread of fire; wooden roof trusses can be where a fire originates if, for example, a building is struck by lightening; and marble columns turn to lime with intense heat. Concrete walls with a marble facing could have been seriously damaged by fires. Ideally, even the entablature and capitals should have been of granite, but they could easily have been replaced if necessary, and the use of marble for details did not threaten the structure of the building.

The greatest difficulty was to construct a dome of unprece-dented size as part of a free-standing building. Even most smaller domes had previously been constructed as integral parts of other buildings and used the walls of adjacent rooms to buttress them. There were some precedents for other building types that used buttresses constructed as parts of walls, but no precedents are def-initely known to have existed for the steprings that successively enable the space that needed to be spanned to be minimized.

(1F) The Client and the Architect

There is no record of who designed the Pantheon, but Hadri-an may well have been both client and architect. He was an archi-tect, and he designed the Temple of Venus and Rome, which was dedicated on 21 April 128 (pls. 202-203). 60 In addition to being a highly effective emperor, he had previously served as a general and provincial governor. He had first-hand experience in design, in construction, and in the organization of large-scale operations.

There is good evidence that before becoming emperor, Hadri-an had designed innovative domes. When he commented on a de-sign by Apollodorus, the architect of Trajan's Forum and Baths, he was told, `Be off and draw gourds. You don't understand any