|USFDC Home | USF Electronic Theses and Dissertations||| RSS|
This item is only available as the following downloads:
xml version 1.0 encoding UTF-8 standalone no
record xmlns http:www.loc.govMARC21slim xmlns:xsi http:www.w3.org2001XMLSchema-instance xsi:schemaLocation http:www.loc.govstandardsmarcxmlschemaMARC21slim.xsd
leader nam Ka
controlfield tag 001 001469427
007 cr mnu|||uuuuu
008 040524s2004 flua sbm s000|0 eng d
datafield ind1 8 ind2 024
subfield code a E14-SFE0000328
Archambeault, Marie Jeanette.
Sourcing of marble used in mosaics at Antioch (Turkey)
h [electronic resource] /
by Marie Jeanette Archambeault.
[Tampa, Fla.] :
University of South Florida,
Thesis (M.A.)--University of South Florida, 2004.
Includes bibliographical references.
Text (Electronic thesis) in PDF format.
System requirements: World Wide Web browser and PDF reader.
Mode of access: World Wide Web.
Title from PDF of title page.
Document formatted into pages; contains 127 pages.
ABSTRACT: Artifacts made of durable materials, such as stone, can provide valuable clues to reconstruct the past. Marble sourcing, in particular,provides information about contact, trade, and other activities in the greater Mediterranean area. The Worcester Art Museum of Massachusetts (WAM) initiated a provenance study by requesting that an analysis of several marble artifacts occur at the University of South Florida's Archaeological Science Laboratory. The 55 marble samples used in this study are from the Worcester Art Museum's collection of Antioch mosaics. Positive results might reveal: 1) preferred sources of tesserae, 2) information about trade of specialized stone, 3) changes in preferred sources during different chronological periods, and 4) workshop preferences. The requested analysis had two objectives. First, once the provenance of the materials is determined, then the results could reveal meaning behind the images contained within the mosaic floor. Second, the results could reveal new trade routes in the Mediterranean. The first step in this analysis was X-ray diffraction (XRD),which differentiates dolomite and calcite marbles. The second step used stable isotope ratio analysis (SIRA), which measures carbon-13 and oxygen-18 isotopic ratios. These two steps have helped to identify Mediterranean marble sources in previous studies. Most of the ancient Mediterranean marble sources have been identified. They have different isotopic values and other characteristics that allow for differentiation. Only one source of dolomite marble exists, which is located in the eastern Mediterranean. It has been identified through XRD in previous studies. Many of the calcite marble sources have different carbon and oxygen isotopic values, which were provided from the SIRA. Those marble artifacts with overlapping carbon and oxygen values can be further analyzed using archaeological, historical, and other information and by using other scientific techniques including cathodoluminescence, electron paramagnetic resonance, and strontium isotope analysis. This thesis discusses the methods used to prepare the samples and analysis conduction; it also discusses the results of the analyses, and presents interpretations regarding the provenance and trade of the marble used for mosaics at Antioch. The results of the SIRA and XRD analysis showed that the materials used for mosaic tesserae come from a variety of sources. Although no definitive matches were found, the results provide the basis for the collection of a colored marble database of sources and artifacts.
Adviser: Tykot, Robert H.
X-ray Diffraction (XRD).
Stable Isotope Ratio Analysis (SIRA).
x Applied Anthropology
t USF Electronic Theses and Dissertations.
Sourcing of Marble Used in Mosaics at Antioch (Turkey) by Marie Jeanette Archambeault A thesis submitted in partial fulfillment of the requirements for the degree of Master of Arts Department of Anthropology College of Arts and Sciences University of South Florida Major Professor: Robert H. Tykot, Ph.D. E. Christian Wells, Ph.D. William M. Murray, Ph.D. Sheramy D. Bundrick, Ph.D. Date of Approval: April 9, 2004 Keywords: Stable Isotope Ratio Analysis (S IRA), X-ray Diffraction (XRD), Tesserae, Daphne, Turkey Copyright 2004, Marie J. Archambeault
ii Acknowledgments I would like to thank Dr. Robert H. T ykot, my advisor, for his countless efforts and his unwavering support in th is endeavor and my education. I also wish to thank my committee members for their insightful comments: Dr. E. Christian Wells, Dr. William M. Murray, and Dr. Sheramy D. Bundrick. I want to acknowledge the following or ganization and individuals for their support of this project: the Worcester Art Museum, Dr. Ch ristine Kondoleon, Dr. Paula Artal-Isbrand, and Dr. Lawrence Becker. In addition, I wish to thank the following organizations for the financial support of th is endeavor: the Worcester Art Museum, the Interdisciplinary Center for Hellenic Studies at USF, a nd the USF Graduate Student Organization. I also wish to specially thank my moth er, Dr. Betty J. Conaway, and my father, Dr. William G. Archambeault, for their de votion to and support of my education.
iii Table of Contents List of Tables v List of Figures vi Abstract x Chapter One Introduction 1 Chapter Two History and Methods of Ancient Marble Extraction 4 Characteristics of Marble 4 Marble Extraction 6 Development of Quarrying Technology 8 Importance of Marble in Roman Construction 14 Chapter Three Archaeology of the Area Studied 18 Geography and Geology 19 Site History 21 Excavation History 22 Mosaic Production and Function 24 Workshops 30 Mosaic Destruction 32 Individual Mosaics: Images Contained and Symbolism 33 Drinking Contest Mosaic 38 Aphrodite and Adonis Mosaic 39 Dionysos and Ariadne Mosaic 41 Funerary Symposium, Agor a, and Eukarpia Mosaics 42 Hermes and the Infant Dionysos Mosaic 45 Ktisis Mosaic 46 Worcester Hunt Mosaic 48 Peacock Mosaic 49 Chapter Four Scientific Analysis of Marble 52 Marble Identification Techniques 53 Visual Marble Identification 53 Thin-Section Petrology 54 Cathodoluminescence 55
iv X-ray Diffraction 56 Electron Paramagnetic Resonance 57 Instrumental Neutron Activation Analysis 58 Stable Isotope Ratio Analysis 59 Application of Marble Provenance Techniques 64 Combination Analysis 66 Mosaic Analysis 70 Things to Consider and Commentary on Techniques 72 Chapter Five Analysis and Results 76 Chapter Six Discussion 88 Possible Sources 93 Chapter Seven Conclusions 102 Findings for Research Goals and Objectives 102 Limitations of This Study 105 Future Research 106 References 108
v List of Tables Table 1. Location, Date, and Color of Mo saic Samples Included in Analysis 33 Table 2. Antioch Mosaic SIRA and XRD Results 80
vi List of Figures Figure 1. Map of Turkey (a fter Turkey.com 2004) 1 Figure 2. Aliki peninsula: marble hills ide completely extracted during Roman times (Photo by Author 2003) 7 Figure 3. Lyre Player: Cycladic (second millennium B.C.) white marble sculpture (Mannoni and Mannoni 1986: 157) 8 Figure 4. Evidence of isolation at A liki Quarries, Thasos, Greece (Photo by Author 2003) 10 Figure 5. Quarrying techniques: the left side shows the hand-cut vertical grooves that were used to split the block from the parent rock; and the right shows the different ki nds of groove marks left on the parent rock (after Manno ni and Mannoni 1986: 73) 13 Figure 6. Roman and Greek tools used for cutting stone (after Mannoni and Mannoni 1986: 73) 13 Figure 7. Map of colored marble s ources used during Roman period: 2giallo antico 3-Carrara, 11rosso antico 12-Thasos, 13-Proconnesos, 14portasanta 16-Paros, 17cipolinno rosso 18-Aphrodisias, 20pavonazzetto (after Anderson 1989: 10) 15 Figure 8. Map of white marble quarries (after Moens 1992: 112) 15 Figure 9. Antioch in the Medite rranean (Kondoleon 2000: xiv) 20 Figure 10. Antioch map (Kondoleon 2000: xiv) 20 Figure 11. Corinth, Centaur Bath, detail of centaur, end of the fifth century B.C. (Dunbabin 1999: 6) 26 Figure 12. Detail of Lion Hunt pebble mosaic, Pella, Greece (Dunbabin 1999: Plate I) 26
vii Figure 13. Hybrid pebble and tesserae mo saic from the third century B.C., Lebena, Asklepieion (Dunbabin 1999: 19) 27 Figure 14. Cave Canem Beware of Dog, from Pompeii house doorway (Dunbabin 1999: 60) 29 Figure 15. Asarotos Oikos or Unswept Room, from Rome (Dunbabin 1999: 27) 29 Figure 16. Antioch city limits map (Kondoleon 2000: x) 35 Figure 17. Atrium House triclinium pavement (Kondoleon 2000: 63) 36 Figure 18. Drinking Contest mosaic (Kondoleon 2000: 171) 39 Figure 19. Aphrodite and Adonis mosaic (Kondoleon 2000: 175) 40 Figure 20. Phaedra and Hippolytus (h ttp://www.loggia.com/myth/phaedra.html ) 41 Figure 21. Dionysos and Ariadne mosaic (Photo Cour tesy Worcester Art Museum) 42 Figure 22. Funerary Symposium mosaic (Kondoleon 2000: 121) 43 Figure 23. Eukarpia mosaic (Photo Courtesy Worcester Art Museum) 44 Figure 24. Agora mosaic (Photo Courtesy Worcester Art Museum) 44 Figure 25. Hermes and the Infant Dionysos mosaic (Photo Courtesy Worcester Art Museum) 46 Figure 26. Ktisis mosaic (Kondoleon 2000: 67) 47 Figure 27. Worcester Hunt mosaic (Kondoleon 2000: 66) 49 Figure 28. Peacock mosaic detail (Kondoleon 2000: 209) 50 Figure 29. Peacock mosaic floor completed thro ugh computer regeneration, colored sections still exist (Kondoleon 2000: 209) 50 Figure 30. Original 13C and 18O variations (Craig and Craig 1972: 401) 62 Figure 31. The updated graph with additiona l white marble samples (after Herz 1987) 63
viii Figure 32. Comparison of the isotopic co mpositions of ancient marble quarries in the eastern Mediterranean with limestones (Wenner et al 1988: 326) 74 Figure 33. Comparison of isotopic signatu res of Classical marble sources and limestone from Neapolis and Corinth (Wenner and Herz 1992: 202) 75 Figure 34. SIRA results with calcitic samples labeled with squares and dolomitic samples labeled with diamonds 82 Figure 35. SIRA of mosaics color-coded by house 84 Figure 36. SIRA of mosaics grouped by color 85 Figure 37. Boxplots showing the range of carbon and oxygen isotope values 86 Figure 38. Aphrodisias white marble data base compared to mosaic samples (after Gorgoni et al 2002) 89 Figure 39. Carrara white marble database compared to mosaic samples (after Gorgoni et al 2002) 89 Figure 40. Dokimeion white marble data base compared to mosaic samples (Gorgoni et al 2002) 90 Figure 41. Naxos white marble database compared to mosaic samples (Gorgoni et al 2002) 90 Figure 42. Paros white marble database compared to mosaic samples (Gorgoni et al 2002) 91 Figure 43. Penteli white marble database compared to mosaic samples (Gorgoni et al 2002) 91 Figure 44. Prokonnesos white marble databa se compared to mosaic samples (Gorgoni et al 2002) 92 Figure 45. Thasos white marble database compared to mosaic samples (Gorgoni et al 2002) 92 Figure 46. Isotopic signatures for various colored marble from opus sectile mosaics (Capedri et al 2001: 14-21) 96
ix Figure 47. Comparison of the mosaic isotopi c values with the isotopic values of limestone quarries in the eas tern Mediterranean (after Wenner et al. 1988: 326) 99
x Sourcing of Marble Used in Mosaics at Antioch (Turkey) Marie J. Archambeault ABSTRACT Artifacts made of durable materials, such as stone, can provid e valuable clues to reconstruct the past. Marble sourcing, in pa rticular, provides information about contact, trade, and other activities in the greater Mediterranean area. The Worcester Art Museum of Massachusetts (WAM) initiated a provenan ce study by requesting that an analysis of several marble artifacts occur at the University of South Fl oridas Archaeological Science Laboratory. The 55 marble samples used in this study are from the Worcester Art Museums collection of Antioch mosaics. Positive results might reveal: 1) preferred sources of tesserae, 2) information about trade of specialized stone, 3) changes in preferred sources during different chronological periods, and 4) workshop preferences of stone material. The requested analysis was had two objectives. First, once the provenance of the materials is determined, th en the results could reveal meaning behind the images contained within the mosaic floor Second, the results could reveal new trade routes in the Mediterranean. The first step in this analys is was X-ray diffraction (XRD), which differentiates dolomite and calcite marble s. The second step used stable isotope ratio analysis (SIRA), which measures car bon-13 and oxygen-18 isotopic ratios. These two steps have helped to iden tify Mediterranean marble s ources in previous studies.
xi Most of the ancient Mediterra nean marble sources have b een identified. They have different isotopic values and ot her characteristics th at allow for differentiation. Only one source of dolomite marble exists, which is lo cated in the eastern Me diterranean. It has been identified through XRD in previous stud ies. Many of the calcite marble sources have different carbon and oxygen isotopic valu es, which were provided from the SIRA. Those marble artifacts with overlappi ng carbon and oxygen values can be further analyzed using archaeological, historical, and other information and by using other scientific techniques includi ng cathodoluminescence, electron paramagnetic resonance, and strontium isotope analysis. This thesis discusses the methods used to prepare the samples and analysis conduction; it also discusses th e results of the analyses, a nd presents interpretations regarding the provenance and tr ade of the marble used for mosaics at Antioch. The results of the SIRA and XRD analysis showed that the materials used for mosaic tesserae come from a variety of sources. Although no definitive matches were found, the results provide the basis for the collection of a colore d marble database of sources and artifacts.
1 Chapter One: Introduction Archaeologists have examined the impor tance of interregional contact through trade routes within the Mediterranean S ea for many decades (Craig and Craig 1972; Renfrew 1972; Coleman and Walker 1979; Grimanis and Vassilaki-Grimani 1988; Anderson 1989; Herz 1990; Rapp 1998). Limite d by available sources, archaeologists typically have focused on historical documen ts, artifacts from foreign cultures, and ethnographic information. More recently, with the advent of elemental and isotopic analysis, archaeologists have begun cataloguing the different va lues for sources of clay, obsidian, marble, and several other durable artifacts. Marble sourcing has provided archaeologists with pieces of the larger Medi terranean trade route puzzle. The samples used in this study come from the Roman site of Antioch in south central Turkey near the border of Syria (Figure 1). The Worcester Art Museum (WAM), which currently houses Figure 1. Map of Turkey (after Turkey.com 2004) Antioch
2 the artifacts, initiated the minimally destruc tive analysis of X-ra y diffraction (XRD) and stable isotope ratio analysis (SIRA) of samples that they collected. Archaeologists have studied the proven ance of marbles for over one hundred years, basing their analyses on color, grain size, and even smell (Moltesen et al 1992). Provenance studies today incorporate multiple analytical methods, including visual analysis, SIRA, and historical and archaeologi cal resources. For this study, the author began by using XRD and SIRA of 13C and 18O to examine several marble tesserae, squared mosaic pieces, from several different Antioch mosaic floors dated to the Roman and the Early Byzantine occupational periods (300 B.C. A.D. 565). A mosaic is defined as a grouping of stone, marble, glass, or terracotta that is jo ined by a binder to form a unit (Bergamini and Fiori 1999). This study of marble mosaic tesserae focused on the following research questions: 1. What is the source of the materials used in mosaic tesserae from Antioch? 2. Which of the sample s have similar results? 3. Is there a temporal or spatial re lationship between the source and the importance of the image created? 4. Is there a correlation between the impor tance of the materials used with the distance that the materials traveled? Positive results might 1) reveal preferred sources for tesserae of specific characteristics (color, grain, luster, etc.), 2) reveal informa tion about trade in specia lized stone, 3) reveal change in preferred sources in different chr onological periods, and 4) reveal workshop preferences in stone selection. In addition, the results coul d reveal how mosaicists at Antioch selected stones for use as tesserae.
3 The results of the analysis, presented in subsequent chapters, highlight the importance of a multi-disciplinary approach to archaeological questions. Without additional historical and arch aeological information, the XRD and the SIRA results only increase our questions about ma rble sources, rather than answer the existing questions. The use of SIRA, in combination with XRD and visual identification, can help identify the sources that were used for marble mo saic floors at Antioch. The information obtained from this study will add to the growing body of knowledge concerning ancient Late Roman and Early Byzantine cultures.
4 Chapter Two: History and Methods of Ancient Marble Extraction Marble has been used for a variety of purposes including architectural elements, decorative inlays, and sculptures. The fi nal marble products are affected by the individual characteristics of the different ma rble types. The history of ancient marble extraction methods, the characteristics of marble, and the use of marble in Roman construction will be discussed in the following paragraphs. Characteristics of Marble Over time, marble has been defined in many contradictory ways. Today, modern geologists define marble as: A well-known metamorphic rock composed predominately of calcite or dolomite; its grain size ranges from fine to coarsely granular. Marble results from either contact or regional metamorphism of limestones or dolostones. Pure marble is snowy white or bluish, but varieties of a ll colors exist because of the presence of mineral impurities in the parent sedimentary rock. The softness of marble, its uniform texture, and its various colors has made it the favorite rock of builders and sculptors throughout history (Monroe and Wicander 1997: 177). The definition of marble has not always been so precise. The term marble comes from the Greek word, marmaros which means a snow white and spotless stone; the
5 adjective marmoreos means resplendent, and the verb marmairo means to shine (Mannoni and Mannoni 1986: 10). This Greek term for marble is vague, leaving room for the inclusion of non-marble and eliminati ng all colored marble from the definition. The lack of precision in the Greek definition continues to cloud m odern understanding of ancient texts referencing marble. Many schol ars have expressed a serious distrust of ancient literature that refers to marble, becau se the definition includes limestone that can take a high polish (Herz 1988). While li mestone that takes a high polish might aesthetically resemble marble, its physical st ructure has not been ge ologically altered. Scientifically, limestone cannot be included in marble analysis, because its sources may or may not have vastly different characteri stics from marble. Consequently, limestone has not received the intense anal ysis and source characterization that marble has received. To further complicate the issue, modern i ndustrial and commercial developments often classify all ornamental rocks, including limestone and dolomite, as marble. Marble is formed through a combination of heat, pressure, and fluid activity. Calcite and dolomite can become marble thr ough pressure of a few thousand atmospheres or at a temperature of about 400 C. Regardless of the formation process, all marble has similar structure, physical composition, a nd working behavior (Mannoni and Mannoni 1986). The main variations in marble come fr om impurities, which affect the color of the material. Aesthetic quality and variation in co lor greatly affect the ornamental and commercial value of marble. The impurities in marble affect not only color, but also the physical characteristics of the stone. A physi cal characteristic of special interest to mosaicists would have been resistance to wear and tear of foot traffic. Color variations
6 seem to coordinate with variabilities in durability to weatheri ng and consistency of coloration after contact with air (Mannoni and Mannoni 1986). Color variation derives from either minerals or pigm entation. Commonly found colors of marble minerals are: white (feldspar, calcite, and dolomite), blac k (biotite, hornblend, augite), green (chlorite, epidote, actinote, diallagio, diops ite, olivine, and serpentine derivative), and clear (quartz, muscovite, and mica) (Anderson 1989: 11; Mannoni and Mannoni 1986: 54, 58). Pigmentation colors include yellow to orange, red, and violet, which do not exist in pure minerals. Iron oxides (hematite) usually make marble red. Green iron oxides (bivalent iron) are rare, but form in an environmen t with no oxygen. Hydrous environments cause a brown to yellow coloration (limonite). Manganese oxides caus e purple. Residual organic matter causes the more common allochro mic colors (pale gray to black). All of these naturally occurring vari ations in marble made some stones more suitable for specific building projects and provided the motiv ation for long distance transportation of stones. Variations in color make some marble especially desirable for mosaic images. Marble Extraction Marble exportation incr eased exponentially from th e Greek to Roman periods; therefore, addressing Antioch marble sourci ng requires a broad examination of marble quarries throughout the central and eastern Mediterranean ar ea, including Italy, Greece, Turkey, Israel, Egypt, and the northeastern bord er of the African continent. Many studies have focused on Greek marble, because the Greeks began industrial scale production and trade of marble. The Romans generally co ntinued exploiting Greek resources until all usable marble was extracted and then opene d new quarries to meet demand. An example
7 of such Roman extraction procedures is evid enced in a photo taken on Thasos, Greece, of a section of the Aliki quarry (Figure 2). Marble carving has existed in Greece since c. 5000 to 4500 B.C., when the Neolithic societies began carving anthropomorphic marble figures. Although Neolithic Greece never acquired the techniques necessary to extract marble for architectural means, it developed the skills that pr oduced a long tradition of marb le figurines. The Cycladic societies continued this carving, which burgeone d into the well-known figures associated with the Bronze Age Cyclades (Figure 3) Since quarrying technology was not yet prevalent in Greece, most of the material us ed for figures was composed of pebbles and boulders partially worn by tidal movements (Waelkens et al 1990: 47). The use of Figure 2. Aliki peninsula: marble hillside completely extracted during Roman times (Photo by Author 2003)
8 Figure 3. Lyre Player: Cycladic (second mille nnium B.C.) white marble sculpture (Mannoni and Mannoni 1986: 157) collected materials, as opposed to extracted mate rials, limited the size of the final artifact. Prior to the invention of bronze tools, sculpt ors used various materi als for sculpting and smoothing figures, including emery, obsidian, sa nd, and pumice. With the invention of bronze tools, similar in form to crowbars, sc ulptors were able to break off larger chunks of stone already separating from the outcrop through erosion. Development of Quarrying Technology While the Neolithic and Bronze Age culture s of the eastern Mediterranean did not possess methods of marble extraction, such t echniques did exist in contemporary Egypt. Egypt had invented extracti on tools that enabled them to develop techniques for
9 extraction of large blocks. Beginning agai n with collected or broken material, the Egyptians undoubtedly invented real qua rrying, which appeared during the Early Dynastic period (c. 3100-2686 B.C.), with th e culmination of dressed stone used for architectural purposes (Waelkens et al 1990: 48). The First and Second Dynasties produced underground tombs and large stelae with progressively improving quality of dressed limestone and granite. During the Third Dynasty (c. 2686-2613 B.C.), these dressed stones were increasingly used in above ground architect ure (Robin 1997: 40). King Djosers step pyramid at Saqqara was the first Egyptian architectural projects to be made completely out of dressed stone (Waelkens et al 1990: 48). In addition, monumental stone sculptures be gan to appear during this tim e. The Fourth Dynasty (c. 2613-2494 B.C.) rulers began shipping large gran ite blocks from Aswan to Giza for the construction of the pyramid complexes (Waelkens et al 1990). As the demand for larger stone blocks increased, the technology for extraction changed as well. While much is still unknown about the earliest quarrying techniques, current theory suggests that Egyptians were the firs t to quarry by cutting narrow trenches around a block of stone in an effort to separate it from the parent rock (Figure 4) (Waelkens et al 1990: 48; Mannoni and Mannoni 19 86: 75). For soft stones, quarrymen initially used copper tools, which were replaced during the New Kingdom (c. 1500 B.C.) with bronze tools, and then replaced again during the Ptolemaic (323-330 B.C.) period by iron tools (Waelkens et al 1990). The quarrying process left groove marks, which changed through time, depending on the tools that were used. Today one can use these marks as a dating method for the period of extraction. Copper tool s left short irregular marks on the parent rock. Bronze tools initially le ft longer marks in a herringbone pattern. Over time bronze
10 Figure 4. Evidence of isolation at Aliki quarri es, Thasos, Greece (Photo by Author 2003) tools left longer and stronger marks, almost parallel and slightly interrupted (Mannoni and Mannoni 1986: 75). Iron tools left long and parallel marks on the parent rock. For harder stones, such as granite and marb le, archaeologists are still debating quarry techniques, but it is generally assumed that harder stones were cu t by pounding with hard hammers. Many believe that the parent rock wa s heated up and then splashed with water, systematically weakening sections of the stone (Waelkens et al 1990: 49). This practice has been connected to an inscription from the Wadi Hammamat. Another theory suggests that changes in wedge marks had little to do with chronological adva nces in technology, but rather adjustments to speci fic quarrying problems (Waelkens et al 1990). Whatever the different quarry marks mean, Egyptians certainly developed quarrying techniques, which were then adopted by neighboring countries.
11 Quarrying technology spread to the Ae gean via the Minoans, c. 1900 B.C., although the Minoans still only quar ried softer stones (Waelkens et al 1990: 51). Evidence of the technology does not exis t on mainland Greece until the Mycenaean civilization (c. 1600 1200 B.C.), seems to ha ve disappeared completely for several centuries with the collapse of this civili zation, c. 1200 B.C. Besides Egypt, quarrying technology continued only among the neo-Hittite civilizations of s outhern Turkey and northern Syria. The Hittit es quarried a variety of ma terials including limestone, conglomerate rock, and basalt. Waelkens et al (1990) suggests that the Greeks were influenced by the post-Hittite culture, seen at Bo azky. The Greek quarry instrument was the pick, not the punch instruments used by the Egyptians. In addition to carving techniques, orientalizing sculpt ure styles were reintroduced from the general area of Syria (Waelkens et al 1990: 54). The earliest surviving Gr eek sculptures, made of limestone, were found on Crete and date to the ninth centu ry B.C. The orientalizing styles and the carving techniques of the scul pture suggest a direct link to the Phoenicians. The seventh century B.C. witnessed the beginnings of a re-opening of Egypt to the Greeks with the establishment of Naukra tis, a Greek trade colony on the delta of the Nile (Hurwit 1985; Whitley 2001). The abr upt stylistic changes, which were unquestionably influenced by the Near East ranged from ceramic vessel shape and decoration techniques to arch itectural styles (Hurwit 1985: 184). Syro-Phoenician influence was also directly responsible for the Cretan, Daedalic style, which spread throughout Greece. This style led to the Greek development of marble sculptures, which were very thin and flat, in cluding the statue of Nikandr e and the Naxian colossus on Delos (Waelkens et al 1990: 54; Whitley 2001: 215). The seventh century B.C. also
12 witnessed the first large-scale stone temples in Greece, which were constructed at Corinth and composed of dressed limestone (Hurwit 1985: 181). With the advent of megalithic architecture on the mainland and in Ionia, Greeks began to improve their quarrying techniques to obtain larger blocks of stone and harder materials like marble. The Siphnian Treasury at Delphi, erected c. 525 B.C., was one of the first structures on mainland Greece built entirely of marble. Evidence also suggests that all of the ma jor quarries of the Greek and Anatolian world were fully active by the end of the sixt h century B.C. Absolute dating of quarries and quarry sections still remains a problem, with the continual modern extraction of marble from larger quarries; however, some an cient evidence still surv ives (Figures 4 and 5). Greek extraction suggests particular quarrying of specific dimensions and finishes, with no industrial co llecting similar to Roman hoarding. Greek quarry workers do not seem to have taken more than they needed. Some preserved quarry marks (which are very regular, almost horizontal, or only s lightly curved grooves, consisting of shallow ledges, the result of crushing) were most likely produced by a l ong-handled, light pick, possibly resembling the tykos of modern Greek quarry work ers (Figure 6). Possibly, the tool is the latomis of ancient Greek sources (Waelkens et al 1990). The tool did not penetrate very deep after each strike, and created a horizontal, crushed groove. The traces of this tool are found on quarry walls that date from the early sixth century B.C. through the Roman Imperial period (Waelkens et al 1990). Although the light pick was well suited for extracting marble blocks of specific dimensions, the method was time consuming. The teams had to be small to allow for continuous movement along a line, and operations must have been run by privat e individuals with experience and knowledge
13 Figure 5. Quarrying techniques: th e left side shows the handcut vertical grooves that were used to split the block from the parent rock; and the right shows the different kinds of groove marks left on the parent ro ck (after Mannoni and Mannoni 1986: 73) Figure 6. Roman and Greek tools used for cutting stone (after Mannoni and Mannoni 1986:73)
14 of stone cutting. The virtual elimination of the light pick in favor of the bulkier pickhammer occurred during the first century A.D. The pick-hammer produced deep, strongly curved grooves, in a garlandlike pattern or festoni suggesting continuous action from one position (Waelkens et al 1990: 59). The Roman quarrying techniques produced an irregular quarry face and led to a greater loss of material. The high demand and cost of marble extraction suggests free quarryworkers pa ssing the knowledge on through generations. It is thought that slave labor would have been minimal, mostly used for dumping wasted material (Waelkens et al 1990: 62). Modern quarrying in Italy and Turkey still operates on a familial basis. The Greeks were largely responsible fo r carrying the quarrying knowledge into the Roman world. In addition to iron tool s, wooden wedges have also been found in some Roman quarries. Although no artifacts have been found in the Greek quarries, wooden wedge holes have been found. Rega rdless of the tools used, the skills and knowledge of the Greek quarrywork er were extremely important in his endeavor. Greek slaves, or technit s and their skill and advanced tec hnologies were largely responsible for the flourishing of the marble industr y during the Roman Empire (Mannoni and Mannoni 1986: 78). Willingly or unwillingly, the Greeks passed on their knowledge and skills to the Romans, who are credited with quarrying on a scale that has never been repeated (Waelkens et al 1990). Importance of Marble in Roman Construction Throughout history the use of marble has held many meanings. The Roman Republic exploited marble resources across the Mediterranean (Figure 7 and 8). In the
15 Figure 7. Map of colored marble sour ces used during Roman period: 2giallo antico 3Carrara, 11rosso antico 12-Thasos, 13-Proconnesos, 14portasanta 16-Paros, 17cipolinno rosso 18-Aphrodisias, and 20pavonazzetto (after Anderson 1989: 10) Figure 8. Map of white marble qua rries (after Moens 1992: 112)
16 beginning of the second century B.C., white marbles were imported to Italy from numerous quarries, including Carrara, Pr okonnesos, Dokimeion, and Aphrodisias. Vitruvius (3.2.5) writes that the Temple of Jup iter Stator (146 B.C.) was the first structure in Rome made entirely of marble. Cons truction of marble m onuments and buildings increased, and soon colored marbles began to be used. In the second century B.C., giallo antico a yellow marble with red ve ining quarried in Tunisia, and pavonazzetto a yellowish-white marble with gray to purple ve ining quarried in Asia Minor, began to be used for statues of barbarians as a means of separation from elit e individuals (Anderson 1989). Demands for particular colors arose as artisans began to use certain colors for specific representations (Gregarek 2002). Giallo antico was preferred for representations of Dionysos himself, recalling the theater costume of the god or the color of saffron, which is often connected to him. Rosso antico was favored for satyrs, recalling the red color of th e wine and the color of th e tanned body (Gregarek 2002: 212). These changing marble demands affected the cost of some marble types. Strabo (9.5.16) writes that the increase in trade for colo red marble actually led to the decrease in prices for white marble. As the market demand continued to change, colored marble began to be used for multiple purposes (G uidobaldi and Salvatori 1988). Plutarch writes that the first colored marble victory monument wa s displayed on the Capitoline Hill in Rome c. 91 B.C. ( Moralia 32). When these extravagant stones made their way into the Roman Republican world, the public campaign against luxuries was at its height. The fascination with embellishment and decoration was viewed as a threat to the Republics worldview (Anderson 1989: 13). Roman trade in marble increased during the first century B.C. It was common for individuals to adorn public buildings with costly
17 materials from distant lands as a means of displaying ones political strength (Anderson 1989). Adornment of public bu ildings immediately affected the decoration of private residences with an increase in the use of marble in scul pture, mosaics, and inlay. Architectural and sculptural ma terials were constantly reused as a means of saving money and time (Giuliano 1989). During the Roman Empire, marble repr esented luxury, wealth, and power; and therefore, marble had a royal asso ciation (Fant 1988). Suetonius ( Augustus 28.3) writes, Augustus so embellished Rome, a city not ad orned in proportion to the greatness of its empire and prey to fires and floods that he was able to boast deservedly that he was leaving to posterity a city clad with marble where he had found one of brick. Augustus commissioned an enormous network of quarri es, which continued to flourish until the late first or early second century A D. (F ant 1988; 1999). For example, evidence of a quarry from this period can be seen at the Aliki peninsula in Figure 2. The demand for marble was so great that quarries like Aliki were exploited to the extreme, so much so that at Aliki the entire pe ninsula was removed. Although th e original in tent of the quarries was not commercial, during the late first or early second century A.D. the quarries acquired a more economic role. A system of business class marble entrepreneurs arose, not as a result of higher demand for ma rble, but rather as a change in attitude towards fiscal independence of the realm (Fant 1988: 148).
18 Chapter Three: Archaeology of the Area Studied One of many Roman-period Mediterranean cities, Antioch-on-the-Orontes was known for the grandiose lifestyle of its resident s. With large avenues, stylish buildings, healing spring-fed waters, and the availabil ity of goods, both exotic and luxurious, at local markets, Antioch operated at a le vel congruent with Rome, Alexandria, and Constantinople (Jones 1981; Kondoleon 2000). De spite its size and complexity, we still know relatively little about the Roman city of Antioch. Ten Antioch mosaic floors were included in this study. Titles for each panel reflect early interpretations of the images portrayed. For ease of descrip tion in this thesis, the titles have been retained. The Antioch marble mosaic floors sample d include the following named panels: Worcester Hunt (WAM 1936.30), Funerary Symposium (WAM 1936.26), Agora (WAM 1936.39), Eukarpia (WAM 1936.38), Drinking Contest (WAM 1933.36), Aphrodite and Adonis (upper section: Princeton University 40.156; lower section: Wellesley College Museum/WAM 1933.10), Hermes and the Infant Dionysos (WAM 1936.32), Ktisis (WAM 1936.90), Dionysos and Ariadne (WAM 1936.25), and Peacock (WAM 1936.23). A total of 55 samples were taken from the mosa ic floors. Typically subjects for analysis of art historians, mosaics recently have received more scientific analysis in an effort to aid conservation and restoration efforts (Ber gamini and Fiori 1999). This study attempts to ascertain the provenance of the materials used in the Antioch mosaics in an effort to
19 1) establish the preferred sources of tesser ae, 2) determine information about trade of specialize stone, 3) reveal chronological cha nges in source preference, and 4) establish workshop preferences of stone in the Mediterranean. Geography and Geology The site of Antioch, which is today calle d Antakya, is located in modern-day Turkey near the border of Syria. The anci ent city dominated settlements at Daphne and Seleucia (Figures 9 and 10), which acted like suburbs. The primary driving force behind the development of Antioch was the environm ental advantages of the site. Ideally situated, Antioch is on the eastern side of the navigable Orontes River (today called the Asi River) (Jones 1981). Located within a bout 25 km, or a days sail, from the Mediterranean port at Seleucia Pieria, Antio ch gained economic advantages. To the southeast, Mount Silpio s (with an elevation of 500 m) provided defensive advantages. The Amuk plain and the lower Orontes valley we re extremely fertile, and in combination with a temperate climate and Roman tec hnology, Antioch was fully supplied with necessaries and luxuries. Local agriculture supplied grain, produce, oil, and wine. The local springs were modified with aqueducts, t unnels, and dams to nurture crops and meet public and private demands. The local geology, a combination of calcareous rocks, including basalt and limestone provided building material s (Downey 1963: 19). During the fourth and third centuri es B.C., the area was used as a limestone quarry. The complete environment allowed for an autonomous city to thrive into a metropolitan area.
20 Figure 9. Antioch in the Mediterranean (Kondoleon 2000: xiv) Figure 10. Antioch map (Kondoleon 2000: xiv)
21 Site History Antioch served as the governmental center of Syria and the capital of the eastern region of the Roman Empire. Seleukos I officially founded Antioch in 300 B.C.; although, it had already existed as a Greek polis for many years (Jones 1981). So rich in Hellenic culture, Antioch even had a school of rhetoric led by Liba nios in the fourth century A.D. Antioch proved to be a consum er city, interacting with the ports of the Mediterranean to the west, the Euphrates to th e east, Ephesos to the north, and Jerusalem to the south. Antioch was the city where the east met the west. Influenced from the east by Persia, and from the west by Rome, and ev ery place in between, the city of Antioch existed as a melting pot for economic a nd cultural trends (Dunbabin 1999; Kondoleon 2000). The Christian orator, John Chrysostom, captured in written hi story what life was like in Antioch during the fourth century A. D., and revealed that a small percent of Antioch society was poor, suggesting the ex istence of a large middle class (Kondoleon 2000: 3-4). Excavation reports by William Campbell (1936) described th e history of the northeast section of the ancient city, which re presents most of the buildings dating from the Early Roman Empire. Construction of bu ildings began in the second century B.C., which continued to be reused and rebuilt until the second century A.D. In addition to minor repairs during the Imperial period, ear thquakes in A.D. 115 and 526 caused major destruction. The earthquake of A.D. 115 near ly killed the emperor Trajan during an extended visit to the city. This earthquake nega tively effected the growth of the city until the third century. The fifth century A.D. saw architectural changes, while the sixth century saw great disasters that ultimatel y caused the demise of Antioch and Daphne
22 (Campbell 1936). Brickwork and masonry rebuil ding was characteristic of the reign of Justinian I (A.D. 527 565). Shortly af ter, Antioch was abandoned, possibly in connection with the invasion of Chosros (C ampbell 1936). Before the invasion, Antioch achieved greatness as the capital of Eastern Rome. Within the Aurelian walls, it was actually larger than Rome (Campbell 1934: 201). Sections of the city were used in the Middle Ages and an apse was used as a potte ry kiln for glazed wares before the site eventually became a cultivated field. Although geographical location led to Antio chs greatness, it appears that the accessibility of the site, natural disasters, a nd active tectonics also led to the citys demise. In addition to a series of earthqua kes, which reduced the strength of the city, Antiochs proximity to the Mediterranean made it a continual target of the Persians. Flash floods were also a consta nt threat. A series of disasters within a short period of time, a fire in 525, an earthquake in 526 and then again in 528, the Persian invasion of 540, and the bubonic plague in 560, led to the ultimate collapse of Byzantine Antioch during the seventh century. Despite the multiple disasters, a relatively large number of Roman mosaics survived the tumultuous sixt h century A.D. and are preserved today. Excavation History The first excavation of Antioch-on-the-Or ontes began in 1932 and continued until 1939. Under the leadership of Professor Ch arles Rufus Morey of Princeton University, the Committee for the Excavation of Antioch and its Vicinity was formed to organize the numerous sponsors, committees, museums, and universities willing to help the
23 excavation efforts of the enormous site. These institutions included the Worcester Art Museum (WAM), the Muse du Louvre, the Baltimore Museum of Art, Princeton University, and Wellesley College. In 1939, th e Fogg Art Museum at Harvard University and its affiliate Dumbarton Oaks join ed the committee (Campbell 1934; Jones 1981; Kondoleon 2000), which included nine members from seven different institutions. The committee members were responsible for obt aining proper clearance from the various governmental institutions. At the end of Wo rld War I, the Ottoman territories of Hatay (including Antakya) and Cilicia, just to the north, were placed under French mandate. The French High Commissioner gr anted permission for excavati on, with the approval of then director of antiquities for the Syrian G overnment, M. Henri Seyrig. Work at the site was postponed in 1939 due to World War II and the region was annexed to Turkey after a League of Nations vote on June 23, 1939 (Downey 1963; Jones 1981; Kondoleon 2000). Several individuals were ac tive in the preliminary survey and early excavation process. In addition to dire cting field crews, Campbell was also largely responsible for the early publications of the site excavations The extensive staff changed from year to year in response to altering research goals and demands of the site (Campbell 1934; 1936). The excavations explored a large ar ea of the region including Antioch proper; Daphne, about 8 km south of Antioch; the port city of Seleucia Pieria; Yakto; and a few isolated sites in the area. Th e initial excavation research goa l was to locate a series of large, elaborate structures and monument s including the palace, the hippodrome, the Forum of Valens, the octagonal Golden Church of Constantine, and the round Church of the Virgin of Justinian (Kondoleon 2000). None of these structures or monuments was found, and this caused conflicts between the major parties concerned with the project,
24 including committee and crew members. To whatever extent the original committee was disappointed, excavation of the site did yi eld over 80 small buildings and nearly 300 mosaic floors in this region during the excavations between 1932 and 1939. Because none of the public and private structures we re discovered and because the majority of major finds were mosaics, the research goals were adjusted to focus on salvation and conservation of the mosaics. The majority of the buildings were used as private residences; therefore, the majority of inform ation derived from site excavation focused on Antiochs private elite (Kondole on 2000: 63). Mosaic floors were most common in elite homes. The plethora of mosaics also suggest ed the enormity of the elite population in Antioch. All of the mosaics analyzed in this study were from residences in the Antioch and Daphne area. Mosaic Production and Function Mosaic materials have received little an alytical attention, because mosaics were considered an unimportant art form for ma ny years; however, mosaics provide a glimpse into the world of Roman art forms that no longer exist, like wall paintings (Dunbabin 1999). The emphasis of the art historical analysis has generally focused on the iconography of the images rather than the materials used. Most wall paintings have collapsed or have been dest royed, particularly in the Ea stern Roman Empire. Mosaic floors were more durable and captured some of these same images. Mosaic production had evolve d greatly over time. The first mosaics were formed out of pebbles rounded by running water (Dunba bin 1999: 5). Early examples of black and white, patterned mosaic floors are locate d at Mira in Mesopotamia (2000 B.C.), at
25 Gordion in Phrygia, Anatolia (800 B.C.), and the Assyrian palaces of Arslan-Tash and Til Barsip in Northern Syria (800 B.C.) (Berga mini and Fiori 1999: 199). In Greece, the earliest surviving decorative mosaics date from th e late fifth century B.C. In contrast to the plain pebble floors, which were found in temples, the mo saics of the late Classical period (early fourth century to c. 340 B.C) were found almost exclusively in private houses (Dunbabin 1999: 6). The pebble mosaic became a true art form by the end of the fifth century B.C. in Greece. Examples of the pebble mosaic art form can be found at Corinth, as well as Olynthos and Pella in northern Greece (Figures 11 and 12). Many changes occurred in mosaic design and production during the late fourth and early third centuries B.C. To create continuous lines that pebbles could not mosaicists began employing thin pieces of lead to outline figures (Bergamini and Fiori 1999). Eventually, emblemata or selfcontained panels, were created at workshops and then brought to their final destination (Dunbabin 1999: 29). Color ranges increase d, adding grays, reds, and yellows, achieving the artistic effect of a pain ting. In addition to artistic changes, mosaics spread geographically to as far east at the palace of Ai Khanoum in Af ghanistan, during the Hellenistic period. The Hellenistic infl uence continued through the Roman period. During the third century B.C., mosaicists began using hybrid techniques, such as opus tessellatum (tesserae work), with pebbles for th e border and background and cut marble for the central figure (Figure 13). Mosaicists continued to refine their work with square tesserae to create the opus vermiculatum technique (wormlike work). Opus vermiculatum refers to the mosaic technique that uses fine gradations of color creating outlines and shadows in much the same way as the art medi um of paint (Bergamini and Fiori 1999).
26 Figure 11. Corinth, Centaur Bath, detail of centa ur, end of the fifth century B.C. (Dunbabin 1999: 6) Figure 12. Detail of Lion Hunt pebble mosaic, Pella, Greece (Dunbabin 1999: Plate I)
27 Figure 13. Hybrid pebble and tessera mosaic fr om the third century B.C., Lebena, Asklepieion (Dunbabin 1999: 19) Most of the Antioch mo saics are composed of opus tessellatum Several common mosaic styles include black and white geometric designs, color geometric designs, twodimensional black and white images, and twoand threedimensional color images. Eventually glass and ceramics were used to increase the color range further, adding most notably Egyptian faience blue. Stones used for mosaic tesserae had to fulfill specific
28 requirements including low hardness, homoge neous color, and compact fine-grained texture. The material had to be hard enough not to break during use as a floor, but also soft and fine-grained enough to allow clea r cutting of the right size for the image (Bergamini and Fiori 1999). Roman mosaics served many purposes. The elaborate displays in doorways, dining areas, and gardens expressed the ow ners personality. Many displays are associated with religious affiliation, while othe rs suggest a warning to strangers, such as the Pompeian Cave Canem or Beware of Dog, mosaic (Figure 14) (Dunbabin 1999). Besides self-expression, mosaics served a pr imary, yet simple utilitarian function. Lavagne (1988) described mosaics as a functional art, which extended to wall decorations, thus attaining aesthetic qualities. They also served as camouflage for dirt and food debris on floors (Dunbabin 1999: 7). Es pecially in the dining room, the mosaic floor provided a distraction for visitors critic al eyes. The best example of this visual distraction is the Asarotos Oikos or Unswept Room, mosaic that ironically shows everything a good host would not want to see on their floors (F igure 15). Visitors to this dining room would see shells, bones, foodstuffs, and even a mouse. A variety of styles were used in Roman mosaic floors. Many ar ticles describing the Antioch mosaics refer to their style as copies of Hellenistic paintings (Hanfm ann 1939; Jones 1981; Kondoleon 2000). Campbell (1934) suggested that by view ing the mosaics in the House of Ge one experiences four centuries of style in ancient pain ting. This, of course, only refers to colored mosaics with detailed images. Other st yles continued to be used, such as a black and white combination to depict geometric de signs or simple two-dimensional images. Many of the framed images included in mosaic pavements at Antioch were created using
29 Figure 14. Cave Canem, Beware of Dog, from Pompeii house doorway (Dunbabin 1999: 60) Figure 15. Asarotos Oikos or Unswept Room, from Rome (Dunbabin 1999: 27)
30 small emblemata surrounded by two borders: one bor der made of common motifs such as fish or birds and a second exterior borde r of a geometric design (Dunbabin 1999). In a few cases, the division between emblema and border was several millimeters wide, suggesting the emblema was created at a workshop and then set into a floor (Campbell 1938). Styles varied across the vast Roman Empire In Rome, the preferred style did not include emblemata or painting-like designs, but was closer to a carpe t or tapestry with an overall decorative design (Dunba bin 1999). Reflective of styles in Sicily and the rest of Italy, mosaics in Punic Carthage employed a signina technique, with mortar and aggregates of crushed pottery or tile form ing a red-toned pavement. In addition, black and white patterned designs were common in Sicily, Italy, and Punic Carthage (Dunbabin 1999). The Palestine and Transjordan regions refl ect the Hellenistic styl es that are visible at Antioch. In fact, at Sepphoris, Israel, one mosaic contains an image of the drinking contest between Dionysos and Herakles (D unbabin 1999: 188). This same contest is represented at Antioch in the Drinking Contest mosaic of the Atrium House. Mosaics of Asia Minor, Cyprus, and Constantinople were definitely influenced by Hellenistic styles; but around the first century B.C. Italian signina and black and white patterned designs began to replace Hellenistic styles (Dunbabin 1999). Workshops One aspect of understanding mosaic imag es and materials used is the workshop. Sourcing the materials used fo r tesserae may reveal new wo rkshops, as well as solidify
31 information about known workshops. Depe nding on the size and location of the workshop, different types of material were avai lable. Workshops are difficult to identify due to a lack of survival both of records and signatures associ ated with mosaics. Several techniques exist to associate mosaics with workshops: 1) color c onnections, 2) similar geometric or ornamental designs, 3) images connections, and 4) similarities in technique. The primary factor of associating mosaic s with the workshop of origin is often geographical proximity. Sheila Campbell (1979) suggested that similarities of color or geometric motifs often lead to mistakes in links between workshops and mosaics; however, repeated themes or s ubject matters that are not sim ilar in appearance, but cover multiple rooms might suggest a connection. Besides geography and color or image connections, a third workshop identificati on method exists. Often artists used a combination of patterns or a va riation of standard motifs as a signature for their work. Other factors include transmission of id eas through pattern books and itinerant workmen. One could argue that the similarities in mosaic images of the drinking contest between Dionysos and Herakles, represente d at both Sepphoris, Israel, and Antioch, Turkey, suggest the two sites had at least one workshop in common. The similarities do not, however, inform archaeologists as to where the workshop might be. Provenance studies can aid in this identif ication. Campbell (1979: 288) sugge sted three stylistic traits that can be used to identify workshops: (1) variations on standard geometric forms; (2) repeated combinations of geometric forms; and (3) repeated themes or iconography.
32 The first two could have been transferred th rough pattern books of itinerant artists or workmen. We must remember that many mo saics were lost through destruction over multiple time periods; therefore, the survival of excavated, intentionally or otherwise, mosaics is essential for a more complete anal ysis of material remains. All of these methods are imprecise; however, because su rviving pavements repr esent only a fraction of the whole corpus. An additional met hod may prove to be more precise. The identification of exotic or lo cal types of stone may also ai d in workshop identification. Mosaic Destruction The Antioch expedition of 1936 realized its obligation to preserve mosaic pavements that were discovered accidentally by locals (Campbell 1936). In addition to saving mosaics, the expediti on recorded evidence of ear lier destruction, including fragments of broken pavements in terrace wa lls, excavated pavements with most of the scenes chipped out, and the testimony of locals who had either broken up pavements themselves or had witnessed their dest ruction (Campbell 1938: 208). William Campbell (1938) recorded every possible destruction mechanism from planting trees to road construction during the expeditions many years at the site. Most of the later mosaics close to the surface were destroyed through modern cultivation and planting. Another way mosaics were destroyed was through m odern road construction. During the construction of the road from Antioch to Daphne the road builders broke through the mosaic floor of a long colonnaded hall with a continuous central panel representing a series of five pairs of animals grouped he raldically (Campbell 1936: 8). Although many
33 mosaic floors were destroyed by modern activit ies, the excavation managed to collect and salvage almost 300 floors. Individual Mosaics: Images Contained and Symbolism The Antioch marble mosaic floors samp led in this study include the following named scenes: the Worcester Hunt the Funerary Symposium Eukarpia Agora the Drinking Contest Aphrodite and Adonis Hermes and the Infant Dionysos Ktisis Dionysos and Ariadne and the Peacock (Table 1). These floors, which range in date from the second century to the sixth century A.D., were found in houses and baths at Antioch and its suburb Daphne (Figure 16). Basic descriptions of the mosaic panels contain vital information about the context of the samples incl uded in this analysis. An evaluation of the colors and types of tesser ae included in each of the mosaics provides important clues for understanding the source of the tesserae. Table 1. Location, Date, and Color of Mosaic Samples Included in Analysis USF#Museum #Mosaic NameHouseCityCenturyColor 61151936.30Worcester Hunt MosaicHouse of Worcester HuntDaphneSixthwhite 61161936.30Worcester Hunt MosaicHouse of Worcester HuntDaphneSixthred 61171936.31Worcester Hunt Mosaic East BorderHouse of Worcester HuntDaphneSixthwhite 61181936.31Worcester Hunt Mosaic East BorderHouse of Worcester HuntDaphneSixthred 61191936.29Agora, BorderNecropolisAntiochFourthwhite 61201936.29Agora, EmblemaNecropolisAntiochFourthwhite 61211936.29Agora, EmblemaNecropolisAntiochFourthred 61221936.29Agora, BorderNecropolisAntiochFourthred 61231936.28Eukarpia, BorderNecropolisAntiochFourthwhite 61241936.28Eukarpia, EmblemaNecropolisAntiochFourthwhite 61251936.28Eukarpia, EmblemaNecropolisAntiochFourthred 61261936.28Eukarpia, BorderNecropolisAntiochFourthred Continued on next page
34 Table 1 (continued) USF#Museum #Mosaic NameHouseCityCenturyColor 61271936.26Funerary Symposium, BorderNecropolisAntiochFourthwhite 61281936.26Funerary Symposium, EmblemaNecropolisAntiochFourthwhite 61291936.26Funerary Symposium, EmblemaNecropolisAntiochFourthred 61301933.36Drinking Contest, AAtrium HouseAntiochSecondwhite 61311933.36Drinking Contest, BAtrium HouseAntiochSecondwhite 61321933.36Drinking Contest, CAtrium HouseAntiochSecondwhite 61331933.36Drinking Contest, DAtrium HouseAntiochSecondwhite 61341933.36Drinking Contest, EAtrium HouseAntiochSecondwhite 61351933.36Drinking Contest, FAtrium HouseAntiochSecondwhite 61361933.36Drinking Contest, GAtrium HouseAntiochSecondred 61371933.36Drinking Contest, HAtrium HouseAntiochSecondred 61381933.36Drinking Contest, IAtrium HouseAntiochSecondred 61391933.36Drinking Contest, JAtrium HouseAntiochSecondred 61401933.36Drinking Contest, KAtrium HouseAntiochSecondred 61411933.36Drinking Contest, LAtrium HouseAntiochSecondred 61421933.36Drinking Contest, MAtrium HouseAntiochSecondred 66091933.36Drinking Contest, NAtrium HouseAntiochSecondBrown 66101933.36Drinking Contest, OAtrium HouseAntiochSecondBlack 66121933.10Aphrodite and Adonis, AAtrium HouseAntiochSecondRed 66131933.10Aphrodite and Adonis, BAtrium HouseAntiochSecondRed 66141933.10Aphrodite and Adonis, CAtrium HouseAntiochSecondRed 66151933.10Aphrodite and Adonis, DAtrium HouseAntiochSecondRed 66161933.10Aphrodite and Adonis, EAtrium HouseAntiochSecondWhite 66171933.10Aphrodite and Adonis, FAtrium HouseAntiochSecondWhite 66181933.10Aphrodite and Adonis, GAtrium HouseAntiochSecondWhite 66191933.10Aphrodite and Adonis, HAtrium HouseAntiochSecondWhite 66201933.10Aphrodite and Adonis, IAtrium HouseAntiochSecondBlack 66211933.10Aphrodite and Adonis, JAtrium HouseAntiochSecondBrown 66221936.32Hermes and the Infant Dionysos, ABath DAntiochFourthWhite 66231936.32Hermes and the Infant Dionysos, BBath DAntiochFourthWhite 66241939.90Ktisis, AHouse of GeDaphneFifthWhite 66251939.90Ktisis, BHouse of GeDaphneFifthWhite 66261939.90Ktisis, CHouse of GeDaphneFifthRed 66271939.90Ktisis, EHouse of GeDaphneFifthBlack 66281936.25Dionysos and Ariadne, AHouse of the Sun-DialDaphneThirdWhite 66291936.25Dionysos and Ariadne, BHouse of the Sun-DialDaphneThirdWhite 66301936.25Dionysos and Ariadne, CHouse of the Sun-DialDaphneThirdWhite 66311936.25Dionysos and Ariadne, DHouse of the Sun-DialDaphneThirdRed 66321936.25Dionysos and Ariadne, EHouse of the Sun-DialDaphneThirdRed 66331936.25Dionysos and Ariadne, FHouse of the Sun-DialDaphneThirdRed 66341936.25Dionysos and Ariadne, GHouse of the Sun-DialDaphneThirdBlack 66351936.23Peacock Mosaic, AHouse of the Bird RinceauDaphneSixthWhite
35 Figure 16. Antioch city limits map (Kondoleon 2000: x) The earliest room tested in this thesis is found in Antio chs Atrium House, which dates to the second century A.D. The Drinking Contest and the Aphrodite and Adonis mosaics were discovered in a triclinium or dining room, in the Atrium House. The triclinium was t-shaped and had five panels, which was very common for Antiochene dining rooms (Figure 17). The triclinium had evidence that the panels were created as emblemata prior to setting in the floor. The Aphrodite and Adonis mosaic, located the farthest away from the entrance to the room, was mostly destroyed. The Drinking
36 Figure 17. Atrium House t riclinium pavement (Kondoleon 2000: 63)
37 Contest mosaic was located closest to the entrance of the room. The other panels depict a dancing boy, a dancing girl, and the Judgment of Paris Chronologically, the next room included in this study is the House of the SunDial, which dates to the third century A.D. The Dionysos and Ariadne mosaic, discovered in the House of the Sun-Dial, is located on the outskirts of Daphne. The next rooms included in this study are located at the Necropolis and Bath D, which date to the fourth century A.D. The Funerary Symposium mosaic and its two side panels, Agora and Eukarpia were found in the Necropolis. The Hermes and the Infant Dionysos mosaic was found in Bath D at Antioch. The Ktisis mosaic was discovered in the fifth century A.D. House of Ge, which was located in Daphne. The Peacock mosaic and the Worcester Hunt mosaic date to the sixth century and were both discovered in Daphne. The Peacock mosaic was discovered in the House of the Bird Rinceau. The Worcester Hunt mosaic was discovered in the House of the Worcester Hunt. Several colors of tessera e were sampled in this study (see Table 1). White samples were taken from each of the mosaics. Seven red samples, six white samples, one brown sample, and two black samples were taken from the Drinking Contest mosaic. Four red samples, four white samples, one black sample, and one brown sample were taken from the Aphrodite and Adonis mosaic. Three white samples, three red samples, and one black sample were taken from the Dionysos and Ariadne mosaic. One red sample and two white samples were taken from the Funerary Symposium mosaic. Two red samples and two white samples were taken from the Agora panel. Two red samples and two white samples were taken from the Eukarpia panel. Two white samples were taken from the Hermes and the Infant Dionysos mosaic. Two white samples, one red
38 sample, and one black sample were taken from the Ktisis mosaic. Two red samples and two white samples were taken from the Worcester Hunt mosaic and its border. One white sample was taken from the Peacock mosaic. Each of the mosaics is described in further detail below. Drinking Contest Mosaic The mosaic of the Drinking Contest of Herakles and Dionys os was discovered in the dining room of the Atrium House in Antioch (Figure 18). Stylistically it dates to the early second century A.D. a nd measures 1.84 x 1.86 m. The Drinking Contest mosaic is composed of marble, limestone, and glass tesse rae. As mentioned above, this mosaic panel was part of a five-image triclinium that measured 7.20 x 4.80 m (Elderkin 1934). The triclinium was composed of five individual emblemata (Levi 1947: 15). The Drinking Contest mosaic would have been the first image seen upon entrance into the dining room. This mythological scene reveal s the problems associated with challenging a god. Dionysos has obviously won the drinking contest, having finished his cup, while Herakles desperately tries to empty his cup, clenching on to the drapery for support as he haphazardly leans backwards. Kondoleon (200 0: 170) describes the scene thus, The composition captures the essence of the str uggle between mortal and immortal, the elegant repose of the god and th e unbalanced human. The panels symmetry is complete with a female flute player behind Herakles; an Eros-type figure poi nting out the obvious winner; and a Silenos with white hair, also ce lebrating Dionysos victory. The mosaicist accomplished an array of graded colors through the use of light and dark tesserae. The mosaic style of the five figures arranged from foreground to background, and the
39 Figure 18. Drinking Contest mosaic (Kondoleon 2000: 171) shadows created by the drinking vessels and even musculature, suggest this mosaic is a copy of an earlier, lost painting (Kondole on 2000: 170). The use of multiple borders surrounding the panel also conveys the effect of a framed painting. Aphrodite and Adonis Mosaic The mosaic of Aphrodite and Adonis was the third major panel in the Atrium Houses triclinium (Figure 19). Unlike the Drinking Contest mosaic, the Aphrodite and Adonis panel faced diners arrayed on their couches. Stylistically it also dates to the early second century A.D. The Aphrodite and Adonis mosaic is composed of marble,
40 limestone, and glass tesserae; it measures 1.60 x 1.90 m. The construc tion of a later wall destroyed the upper section of the mosaic. Initially, due to poor preservation, Campbell (1934) believed that the imag e of Aphrodite and Adonis actua lly represented Phaedra and Hippolytus, characters from another love stor y who often are posed in this manner. The panel depicts a female figure seated on a th rone, with a nude male figure seated on her right. His spear and dog suggest the male figur e is a heroic hunter type (Levi 1947: 25). Although Phaedra and Hippolytus often are depict ed with a dog, they almost always are portrayed with the other major characters i nvolved in their love triangle (Figure 20). Even though the image is badly destroyed, its size suggests that no ot her characters were included. On the other hand, Aphrodite and Ad onis often are represented alone. Given Figure 19. Aphrodite and Adonis mosaic (Kondoleon 2000: 175)
41 Figure 20. Phaedra and Hippolytus (http:// www.loggia.com/myth/phaedra.html) the individuals shown in the other panels in this hous e, the Aphrodite and Adonis combination fits the grouping better. The mo st intriguing aspect of the Atrium House mosaics is the central panel: where a mortal faces the trial of fate and the deities. Although the so-called Judgment of Paris (in the middle panel) is not included in this study, it depicts the mythological trial of Paris who was forced to judge a beauty contest between Hera, Athena, and Aphrodite. The pain terly effects of the panels extend to the borders of the Judgment of Paris and the Aphrodite and Adonis panels. Dionysos and Ariadne Mosaic The mosaic panel of Dionysos and Ariadne was discovered on the outskirts of the suburb of Daphne in 1935 (Figure 21) and serves as an example of a mosaic that was
42 Figure 21. Dionysos and Ariadne mosaic (Photo Courtesy Wo rcester Art Museum) salvaged during the 1930s project. The Dionysos and Ariadne mosaic dates to the third century A.D. (Jones 1981). Found in the Hous e of the Sun-Dial, the mosaics in this house were mostly destroyed except for th e surrounding panels (Stillwell 1938). The image of Dionysos and Ariadne existed in di fferent mosaics around the area of Antioch and Daphne. This panel depicts the bust of a male and female. Both figures are crowned with wreaths of leaves. The male wears a white tunic with gray shading and wears a necklace. The female carries a spear and w ears a dark brown tunic with dark red, gray, and white highlights. The panel is surrounde d by geometric panels of triangles on the right and stars on the left (Stillwell 1938: 202). The ma in image that this panel surrounded was destroyed completely. Funerary Symposium, Agora, and Eukarpia Mosaics The Mnemosyne mosaic, or the Funerary Symposium and its side panels Agora and Eukarpia were discovered on the edge of Antioch s city limits in the Necropolis, or cemetery (Figures 22 through 24). Each of the panel images is an emblema which was
43 formed at a workshop and then set in place (Levi 1947: 295). The entire group dates to the fourth century A.D., measures 1.77 x 2.69 m, and is composed of limestone, marble, and glass tesserae. The central mo saic reveals a womens funerary AI XIA, or banquet, most probably honoring a woma n whose name, Mnemosyne, appears above a large cloth or textile pinned to the wa ll in the background (Kondoleon 2000: 121-122). In total, six women are attending Mnemosynes banquet: one sits on a low stool while holding a scroll, two recline on a curved couch, two are entering the room with wineskins as a probable offering, and another (a servant) has entered from the right with a jug and basin. The mosaic floor was discovered in a sma ll chamber surrounded by tombs. Benches similar to the one depicted in the mosaic we re uncovered in the ex cavation of this room, suggesting a connection between the scene a nd actual events. Alternatively, Kondoleon Figure 22. Funerary Symposium mosaic (Kondoleon 2000: 121)
44 Figure 23. Eukarpia mosaic (Photo Courtesy Worcester Art Museum) Figure 24. Agora mosaic (Photo Courtesy Worcester Art Museum)
45 (2000: 122) suggests that the room may have served as a meeting place for women in a funerary collegium and the inscription could also be tr anslated as memory. To the left and right of the Funerary Symposium mosaic are two female personifications, one of Agora (the Marketplace) and the other of Eukarpia (Abundance) (Levi 1947: 296; Campbell 1988). Although this was the only pa vement recovered in the cemeteries of Antioch, the funerary banquet was a common decoration for Roman tombs. The Roman funerary banquet was also an im portant ritual surrounding death. Hermes and the Infant Dionysos Mosaic The mosaic panel of Hermes and the Infant Dionysos was discovered on the eastern side of Room 3 in Bath D at Antioch (Figure 25). Stylistica lly this mosaic dates to the early fourth century A.D. (Campbell 1988). The surviving mosaic measures 2.25 x 3.25 m and is composed of marble, limestone, and glass tesserae. The original mosaic was more than 15 meters in length. A wi de band of ornamental designs surrounds the rectangular panel depicting Hermes carry ing the infant Dionysos to the nymphs (Campbell 1934). Hermes looks to his right, but moves to his left suggesting he is running from something (Levi 1947: 286; Ca mpbell 1988: 17). He wears only a cloak and has two wings projecting from both a nkles. Dionysos has a Christ-like pose, balanced on Hermes right hand with a nimbus behind his head and a wreath in his hair. Dionysos is identified through an inscription above his head: ION[Y O ]. The rest of the surviving mosaic (n ot shown) was separated by a large gap. On the other side of the gap, another inscription, refers to the nymphs (i.e., [NY]M[ AI]) to whom Hermes
46 Figure 25. Hermes and the Infant Dionysos mosaic (Photo Courtesy Worcester Art Museum) carries the child (Campbell 1988). The fragme ntary image shows a broken pillar with a leafless branch behind it, next to a wreathed female head. Ktisis Mosaic The mosaic image of Ktisis was discovered in Room 4 of the House of Ge in the suburb of Daphne in 1936 (Figure 26). Dating to the fifth century A.D., the House of Ge contained a collection of female images repres enting abstract ideas su ch as life, earth,
47 Figure 26 Ktisis mosaic (Kondoleon 2000: 67) spring, and winter. The mosaicists of Antioch often created female personifications of concepts such as KTICIC (Foundation) or H (Earth) or BIOC (Life) (Kondoleon 2000: 65). Fifth century floors frequen tly used medallions with a bust image, like that of Ktisis, surrounded by an octagon or a st ar-pattern (Morey 1938). The inscription divided in two parts by the female bust is KTICIC. Ktisis has a crown of large round, red and green jewels separated by a vertical series of two pearls (Levi 1947: 347). The womans hair is pulled back into a loose mass at the nape of the neck. The figure has earrings with a triangular shape hanging from thick gold hoops. Ktisis wears a violet tunic with a red mantel thrown over her shoulders (Levi 1947: 347). The bust is enclosed in a golden
48 octagon. The square panel that surrounds th e bust of Ktisis cont ains a multi-colored continuous pattern of diamonds tangent on the co rners and enclosing fo ur-pointed stars. A border with large birds and flowers surr ounds the entire panel (Levi 1947: 347). Worcester Hunt Mosaic The Worcester Hunt mosaic was discovered in Daphne at the House of the Worcester Hunt (Figure 27). Stylistically the Hunt mosaic dates to the sixth century A.D. and measures 6.26 x 7.11 m. The Hunt mosa ic is composed of both marble and limestone tesserae. One of the largest floors from Antioch, the Worcester Hunt mosaic portrays various hunting scenes. The co mplex scene shows hunters on foot and horseback using sword, spear, or bow and arro w to hunt lions, tigers, deer, antelope, rabbits, a wolf, a panther, and a bear with great success, except for one hunter who is saved by the spear of a horseman after being attacked by a lion (Morey 1938; Levi 1947). A company of animals in various poses flanks the central hunter. This central figure calls the viewers attention be cause he is larger than the other hunters depicted in the scene. Morey (1938: 41) compared this image to the hunter image depi cted on third century A.D. sarcophagi of western Asia Minor and suggested a Persian influence on the pavement makers. Morey (1938) noted the attention to detail given to the animals, as well as the lack of detail given to the hunt ers, and suggested th at Persian taste was responsible for the design. Although all of the figures rev eal action, the animals are the only figures with musculature. Morey argues that a more traditi onal Greek design would have depicted the opposite: de tailed human figures and va gue animal figures; however,
49 Figure 27. Worcester Hunt mosaic (Kondoleon 2000: 66) the detailed depicted are complex. Furt hermore, the design reflects Hellenistic emblemata styles of carpet or tapestry (Kondoleon 2000: 158. Peacock Mosaic The Peacock mosaic was discovered in the Hous e of the Bird Rinceau at Daphne (Figure 28). Stylistically it date s to the sixth century A.D. The Peacock mosaic measures 1.17 x 3.81 m, but is part of a much larger floor, measuring 65 m2. A computer
50 regeneration of the complete Peacock mosaic floor, which was divided among sponsoring institutions, is shown in Fi gure 29 (Kondoleon 2000). The mosaic is composed of marble and limestone tesserae. The image shows a gr ape vine scroll, entwin ed with birds and Figure 28. Peacock mosaic detail (Kondoleon 2000: 209) Figure 29. Peacock mosaic floor completed through computer regeneration, colored sections still exist (Kondoleon 2000: 209)
51 animals, growing from an urn in each corner. The vines give the impression of a fluttering ribbon (Levi 1947; Kondoleon 2000). The fragment included in this study shows two peacocks surrounding a basket of grapes. These two birds are the only peacocks in the entire floor, suggesting an intentional importance. Although the Romans viewed peacocks as linked to immortality and eternal life, the motif ca n also be viewed as Christian, or Early Byzantine. Paired peacock s, inhabited vines, grapes, or wine vessels were popular in early Christian art showi ng the beauty of Gods creation (Kondoleon 2000: 209). A review of the physical and artistic c ontext of the mosaic tesserae included in this analysis provide s important background informati on that may aid in the final conclusions about the source of the tesserae material. The mosa ics included in this study exemplify a range of images from multiple time periods from Roman Antioch.
52 Chapter Four: Scientific Analysis of Marble Scientific analysis of ar chaeological materials began as early as the Italian Renaissance and escalated into the peri od of scientific di scovery known as the Enlightenment. In 1798, for example, Mart in Klaproth analyzed the chemical composition of Roman glass and bronze mi rrors. Michael Faraday (1791-1867) and Humphrey Davy (1778-1829) were involved in early analytical work on chemical analyses of Egyptian blue (i.e. faiance) a nd an opaque red vitreous material (Henderson 2000: 8). The scientific analysis of archaeologi cal stone has mostly focused on obsidian, chert, flint, and marble. Most often stone material has been analyzed through mineralogy, microscopic struct ure, texture, and inclus ions like fossils. These characteristics, while aiding in analysis of the environment of the rock structure, have helped deduce the provenance of the materials (Henderson 2000: 297). Archaeologists needed a more exact met hod for describing the various types of ancient marble (Coleman and Walker 1979: 107) For the purpose of the thesis, only scientific techniques re levant to marble analysis are di scussed here. Most scientific analysis of archaeological stone from th e Roman period has occurred on marble, which was widely used and traded. Historicall y, the three main goals of Greek and Roman
53 marble analysis since the Rena issance have been to ascertai n: 1) provenance, 2) correct association of separated fragments, and 3) authenticity (Herz 1990: 101). Marble Identification Techniques A plethora of physical, chemical, isotopic, and trace-element analysis techniques has arisen and proven successful in the past two decades including: thin-section petrology (Bergamini and Fiori 1999; Polikreti a nd Maniatis 2002), cathodoluminescence (CL) (Moens 1992; Blanc 1995), X-ray diffr action (XRD) (Lloyd 1988; Herrmann 1990), electron paramagnetic resonance (Attanasio a nd Platania 2000; Polik reti and Maniatis 2002), instrumental neutron ac tivation analysis (Grimanis and Vassilaki-Grimani 1988; Rapp 1998), and stable isotope ratio analysis (SIRA) (Craig and Cr aig 1972; Herz 1990; Gorgoni et al 2002). Although each technique has a dvantages and disadvantages, the ultimate analytical program may involve a combination of techniques. Several techniques have proven more successful when paired with an additional test, such as combining spectrometry and neutron activati on analysis, because di fferent techniques measure different elements. Archaeologists have used many techniques for marble source determination. Visual Marble Identification Visual identification of stone can provid e basic descriptive an alysis of an object including geological typology. More detailed an alysis of marble artifacts through visual identification of color has been us ed for nearly a century (Moltesen et al 1992); however, this method creates problems for per ception and description. Although some
54 archaeologists rely solely on visual identification, the prob lems of communicating the perception of color force most archaeologist s to seek out other forms of analysis. Although some marble has distinct colors many marble quarries produce pure white marble, making visual source id entification difficult. In a ddition, homogeneity of color is not guaranteed. Color charts often are used as a guide, but color is only one attribute associated with marble. Many archaeologi sts acknowledge the difficulty of visual analysis and have re-evaluated previously grouped materials. In the past, subjective aesthetic conclusions about a stones source were drawn and objects were given place names as adjectives. Obviously, many contr oversies, which remain unresolved, arose, and the literature is plagued with contradictory descriptio ns of the same pieces (Herz 1990: 101). Thin-section Petrology Another technique used to source marble is thin-section petrology, or petrofabrics. Using a mounted section of the study material under an optical light microscope, thinsection petrology examines a representative se ction of the material for arrangement of inclusions, along with their size, shape, frequency, and composition (Henderson 2000: 12). The sections are c. 30 m in thickness, allowing polarized light to pass through the materials and highlight irregular ities and variation in color, which can then be used to identify the source of the material. An auxiliary lens and various comparative thinsections (quartz, gypsum, mica, etc.) help iden tify crystal minerals and their orientations. This technique is one of the less expensive ways of examining marble; however, many inexperienced researchers mistake normal vari ations for something more extraordinary.
55 While thin-section petrology is a quantitative analytical t echnique, this technique is destructive, requiring a large sa mple of the material, and is not very informative by itself for marble sourcing (Polikreti and Maniatis 2002: 1). One way to eliminate subjective conclusions is to base analys is entirely on more scientif ic, objective data. To avoid subjectivity as a result of inexperience, th in-section petrology was not used in the examination of the Antioch tesserae samples. Cathodoluminescence (CL) Cathodoluminescence microscopy (CL) wa s first used to source marble successfully in 1987 by Danielle Decrouez and Vincent Barbin (Moens 1992). An electron beam bombards the mineral mount ed on a thin-section, here calcite (CaCO3) and dolomite (CaMg(CO3)2) marble types, to reveal different colors (Barbin et al 1992; 1999). Impurities and lattice defects affect th e luminescence image. The visible colors, variations of blue and orange, are associat ed with a white marbles source. Each cathodomicrofacies generally char acterizes a given area (Barbin et al 1992; 1999). Barbin et al (1992) were able to discriminate di fferences between marble from quarries at Penteli, Dokimeion, Naxos, Thasos, Paros, Pteleos, Candoglia, Lasa, Crevola, Villete, and Doliana. In 1995, Blanc published an experimental use of CL attached to a spectrometer, which used a compressed powde r sample, approximately 3 mg, mixed with graphite and coated with carbon or gold-palla dium. A spectrum of wh ite marble displays two bands of energy, a variation attributable to manganese. Blanc (1995) suggested that the use of CL as an accessory to stable isotope s, which is discussed below, is especially valuable for provenance of white marble. Th is method is not widely used alone. Again,
56 CL can be subjective because of variances in color identification; therefore, CL was not used in the analysis of An tioch mosaic tesserae. X-ray Diffraction X-ray diffraction (XRD) spectrometry is a non-destructive spectrometric analytical technique (Lloyd 1988). This tech nique can be used to identify crystalline materials, such as calcite or dolomite, or to determine the degree of crystallinity (Henderson 2000: 10). The determination of marbles crystallinity provides valuable information for the sourcing process. For exam ple, if a white marble sample is composed of dolomitic marble, then the objects source is most probably Thasos, a quarry with a high dolomitic content. X-ray diffraction invol ves the emission of radiation wavelengths at the crystalline material (Herrmann 1990). The wavelengths bounce off the crystalline structure in spectra unique to the sample ma terial. The spectra have independent peak intensities represented graphi cally by height. Although XRD requires that the sample be in powder form, thus destroying the original st ructure, the size of the sample required is relatively small, and the chemical composition of the sample is not altered in the analysis. The sample can be re-used for another method or for a second analysis by XRD. Although little analysis has b een done on exactly how small a sample can be, the author experimented with the sample size of a known calcitic and a known dolomitic marble sample to determine the reliability of small sa mple sizes. It was determined that a sample size as small as 2.6 mg would produce reliable results. X-ray diffraction is a valuable technique for the analysis of mosaic tesserae, since it only requires a small sample size
57 and provides valuable source information. Mo st of the samples were large enough to be analyzed by XRD. Electron Paramagnetic Resonance Electron paramagnetic resonance spectrosc opy (EPR), or electron spin resonance (ESR), has been used for analyzing unpaired electrons in a molecule Unpaired electrons can aid in the determination of the age or the provenance of an object (L ambert 1997: 264). For the past 15 years, EPR spectra have been collected for different types of white marble from around the Easter n Mediterranean (Polikreti an d Maniatis 2002). Several different uses of this method of analysis have occurred. Polikreti and Maniatis (2002), in addition to several other archaeologists, used 10 parameters in order to discriminate between quarries, such as those at Penteli, Naxos, Hymettus, and Prokonnesos, with some degree of overlap between Paros and Pr okonnesos and Paros and Hymettus. Other quarries that have been identified through EP R include Seravezza and three quarries from Carrara. Researchers have compared different pairs of combinations of the parameters in order to ascertain a provenance. In addition, Polikreti and Maniatis used maximum grain size, measured with a stereomicroscope. Electron paramagnetic resonance characterizes marble by its impurities, manganese (Mn2+) or Iron (Fe3+). Manganese is diluted into the lattices of calcite or dolomite, which are th e main constituents of marble (Attanasio and Platania 2000). This constant irregularity allows for meas uring of a selected spectral feature of the impurity from both arch aeological samples and from known quarry samples. Electron paramagnetic resonance wa s not used in this analysis, because a
58 relatively large-sized sample is required to pe rform this analysis and little is known about the proportions of various impur ities of colored marble. Instrumental Neutron Activation Analysis Instrumental neutron activation analysis (INAA) is one of the most sensitive and accurate techniques available for the determination of a large nu mber of trace elements in different materials (Grimanis a nd Vassilaki-Grimani 1988; Mello et al 1988). Many chemical elements can be distinguished at th e low parts-per-million level and some at the parts-per-billion range (Rapp 1998: 148). Th e powdered sample (50 mg for metals, 200 mg for silicates) is placed in a capsule, which is irradiated in an atomic pile for a defined period of time (Rapp 1998). Decay of the el ements begins counting first the short-lived elements, followed by the long-lived elements The results, a spectrum of wavelength against peak intensity, are displayed graphically (Hende rson 2000). Luedtke (1978) discovered that if chert types were formed close in time and space they shared similar proportions of trace elements, as determined by INAA. Eventually, she was able to differentiate between three sources of chert in the North American midwest. Instrumental neutron activation analysis has now been used for multielemental analysis of marble specimens for provenance studi es; however, trace element composition of marbles is highly variable. Further analysis for rare earth elements (REE) is necessary. Rare earth elements are distributed more evenly in marbles than many other trace elements (Grimanis and Vassilaki-Grimani 1988: 275); however, for marble, INAA has limited capabilities in quarry discrimination a nd is rarely used al one (Polikreti and Maniatis 2002: 1). Mello et al (1988) showed that INAA could identify the marble
59 quarries of Paros, Aphrodisias, Marmara, Naxos, Penteli, Carrara, and Denizli; however, the quarries could only be determined through paired analysis. In strumental neutron activation analysis requires a large sample size and has not been used on colored marble; therefore, it was not incl uded in this analysis. Stable Isotope Ratio Analysis Stable isotope ratio analysis (SIRA) is another technique that is useful for the analysis of small samples. Isotopic analys es of marble have enabled archaeologists to reconstruct correctly marble artifacts and to determine the provenance of marble artifacts. Stable isotope ratio analysis uses mass sp ectrometry to determine ratios of certain elements. Harmon Craig and Valerie Craig de veloped a method in the early 1970s to test the isotopic composition of marble pieces (C raig and Craig 1972) Craig and Craig examined carbon and oxygen values, which are major components of calcite or calcium carbonate (CaCO3) that largely makes up marble, in order to compare the results. Elements differ by the number of protons in their nuclei. Many elements also have multiple isotopes that occur in nature. Oxygen and carbon ex ist in nature in several isotopic forms, all of which are stable ex cept for carbon-14. With six protons and six neutrons, carbon is a stable isotope, with an atomic number of 12. Carbon also exists with extra neutrons. Carbon-13 has one extr a neutron and is stab le. Carbon-14 has two extra neutrons and is not stable. Oxygens main form occurs in nature as oxygen-16. Other forms of oxygen are oxygen-17 and oxyge n-18. While Carbon-14 is unstable and therefore radioactive, none of the oxygen isotopes are unstable.
60 Proportions of various isotopes can differ fr om place to place due to variations in formation of the material. Marble is derive d from limestone, which is a sedimentary rock created from cementation of shells and othe r sea life. The differences in limestone formation are the foundation for variation among marbles. Vari ation in the three formation factors heat, pressu re, and fluid activity also le ads to variation in isotopic values. The processes involve d in isotopic composition can be described as follows: 1) mode of origin, 2) isotopic composition of water, 3) temperature of the metamorphism that converted the limestone into marble, a nd 4) later weathering history (Herz 1990: 105; Herz and Dean 1986; Gorgoni et al 2002). Atmospheric, geological, or biological processes (wind, water, and metabolism) can move substances cont aining lighter isotopes faster than those with heavie r isotopes. Water is the source of oxygen and carbon dioxide is the source of carbon in the raw material s that comprise rock s and stones. Local conditions can affect the isotopic values. When geological formation is complete, the isotopic proportions are sealed in place like a fingerprint of the source (Lambert 1997: 5). Craig and Craig (1972: 401) be lieved that the ratio of 13C/12C and 18O/16O in Greek marbles provided the best chance for unique characterization by locality. Their 1972 study collected samples from four major quarries Paros, Penteli, Hymettos, and Naxos which were used by the ancient Gree ks. Craig and Craig (1972) also suggested that the trace elements, strontium (Sr) and magnesium (Mg), could be used to analyze further a marble source. The isotope results are given as deviation ( ) values in relation to the isotopic standard reference material of a natural limestone, Pee Dee Belemnitella (PDB), in parts per mil (): () = [(R/R+) 1] x 1000
61 where R is the ratio 13C/12C and 18O/16O and R+ is the ratio in the PDB standard (Craig and Craig 1972; Herz 1987). This technique provides a precision of 0.05 per mil (Craig and Craig 1972: 402). The resu lts were then plotted on a 18O 13C diagram. The results showed that Pentelic and Na xian marbles were much lower in 18O than Parian and Hymettian marbles, probably due to interact ions with meteoric water at elevated temperatures. The Pentelic marbles were also higher in 13C than most of the marbles studied (Craig and Craig 1972: 402). Craig and Craig (1972) s uggested that the isotopic method for provenancing Greek marbles woul d be the most useful, especially if combined with other techniques. Norman Herz believes that, at present, SIRA is the most powerful analytical technique available for sourcing marble, stati ng that one of the main advantages is the small size of the sample required, only about 10 mg or less (Herz 1990: 103). This small sample size is beneficial wh en analyzing art work and archaeological artifacts. The technique developed by Craig a nd Craig has enabled archaeol ogical works of art to be attributed to specific quarries; however, sci entific contributions to archaeology often follow an uneven path of evolution (Lambert 1997: 7). The limitations of this technique were immediately obvious as the database of comparative samples began to grow. Craig and Craig examined only the four most promin ent Greek sources, and their separation in a plot was simple and easily distinguis hed the sources. He rz (1992), Matthews et al (1992), and Moens et al (1992) published databases created by sampling additional quarries. More sources were sampled and a dded to Craig and Crai gs original diagram (Figure 30), creating a much more complex picture (Figure 31) that has multiple problems. Up to six sources are possible for certain isotopic signatures. Several of the
62 tested quarries have isotopic signatures that overlap with multiple other sources. In addition, isotopic variation within a single quarry can be large. Many archaeologists argue that since carbon and oxygen SIRA alone fails to identify the provenance in many cases, then it should be replaced with another technique (Polikreti and Maniatis 2002: 2). Alt hough SIRA does not always determine a single source for every marble artifact sampled, to date it is still the most accurate and widely available method. The addition of historical information and non-scientific data can help differentiate between sources. In addition, other scientific techniques can be used. Two other measurable isotopes exist in marble. The isotopes of strontium and electron spin resonance of a particular form of manganese have been used, independently, to separate successfully some of the overlapping results. In addition, the proporti ons of a variety of elements, including chromium and antimony, have been used in a similar manner to Figure 30. Original 13C and 18O variations (Craig and Craig 1972: 401)
63 Fi g ure 31. The u p dated g ra p h with additional white marble sam p les ( after Herz 1987 )
64 separate overlapping results (Lambert 1997: 8). As for isotopic variation within a quarry, not all quarries exhibit a large disparity. Carrara, a white marb le source in Italy, shows a small variation of less than 0.5 for 13C and less than 2 for 18O within an outcrop (Herz 1987). Herz (1992) has collected data bases for isotopic values from multiple quarries from periods ranging from the Earl y Bronze Age to the Classical era. The application of the databases to further arch aeological questions ha s led to improvements of the analysis process (Gorgoni et al 2002: 116). Isotopic values of colored marble have been collected from several quarries ; however, colored marble quarries have not received the same analytical attention that white marble has. Although several studies have focused on rosso antico a red marble (Lazzarini 1990; Gorgoni et al 2002), no substantial colored marble database exists. Further description of analytical techniques and their application are discu ssed in the next section. Applications of Marble Provenance Techniques G. Richard Lepsius (1890) published the first systematic description of major marble quarries exploited during Classical ti mes. Limited to physical characteristics, Lepsius descriptions became and remain archaeological gospel (Herz 1990: 101). Some of his descriptions ar e as follows: Pentelic was a medium-grained, weakly foliated, sometimes micaceous marble; Hymettian was fine-grained and bluish; Parian mediumto coarse-grain ed, pure white, and translucent; and Naxian or merely island was a coarse-grained, white marble (Lepsius 1890:13-22; 77-85). Today, multiple analytical techniques are used to derive in formation from artifacts and sources. In the analysis of marble, thin-secti on analysis involves tedious microscopic study and produces
65 very detailed results; however, a database of comparative analyses of known sources does not exist, and a large sample is needed for this technique. While elemental analysis also provides detailed information, trace elemen ts can vary by factors of over a hundred within the same quarry due to localized in teractions with inclusions and surrounding rocks (Craig and Craig 1972; Herz 1990). Recently, the use of multivariate statistical treatment of elemental data ha s provided a means for partly overcoming the variability in the composition of the material. Discriminant analysis, scatter plots, and ellipses of the results of NAA and SIRA have been used to improve source determination (Matthews et al 1995). A provenance study using SIRA was pe rformed on a marble bust housed at Harvards Fogg Art Museum. The bust was said to be a representation of Antonia Minor (accession number 1972.306), the daughter of Mark Antony and the mother of Germanicus and Claudius (Erhart 1978). The bus t is composed of five separate parts: the head, the lower portion of hair, and three bus t pieces. Several scholars question whether or not these multiple pieces ac tually belonged to the same bust (Lambert 1997: 4). While the busts history can be traced back to the se venteenth century as a part of the collection of Wilton House in England, its prior hist ory is unknown (Erhar t 1978: 195). Herz (1990) analyzed all of the pieces of th e Antonia bust by examining the carbon-13 and oxygen-18 values of the marble. In both Greek and Roman antiquity all marble portraits of important persons were made of Pari an marble (Herz 1990: 105). The final conclusion was that only the head proved to be authentic, made of Parian marble. The lower hair-piece and one of the pieces of th e bust were of Carrara marble, a source in Italy. The other two fragment s were also made from Parian marble, but their isotopic
66 signatures were clearly different from the h ead. The results sugge st that throughout its history, the Antonia bust was reconstructed mu ltiple times with pieces of marble from different sources (see also Lambert 1997: 4-5). Combination Analysis Most archaeologists agree that a combination appro ach to the provenance of artifactual marble produces the best results (Moens et al 1988; Herz 1990; van der Merwe et al 1995; Attanasio and Pl atania 2000; Gorgoni et al 2002). Henderson (2000: 12) suggests that thin-sec tion petrology and XRD spect rometry provides the best analytical combination, producing crystal identification and distribution through the material. Herz (1990: 108) often uses a comb ination of SIRA and XRD, as he did for the allegedly ancient Greek kouros from the J. Pa ul Getty Museum in Malibu. Art historians are highly skeptical about the authenticity of th is piece, because this kouros is stylistically unique from most, and only 12 complete kouroi are known worldwide. If authentic, the kouros stylistically would date to c. 530 B.C. Stable is otope ratio analysis results showed: 18O = 2.37 ; 13C = +2.88 (Herz 1990: 108). These results, when compared to the database, suggest the follo wing quarries as possibl e sources: Denizli, Doliana, Marmara, Mylasa, and ThasosAkr opolis. X-ray diffraction results show a composition of 88 percent dolomite and 12 percent calcite. Through the process of elimination, only Denizli, Marmara, and Th asos-Akropolis contain dolomitic marble. Comparison of trace-element data eliminated Denizli. Dolomitic content determined by Cordishchi through EPR eliminates Marmar a (Herz 1990: 108). Since Thasos is composed of almost 100 percent dolomite a nd Marmara is composed of only 57 percent
67 dolomite, discriminant function analysis sugge sts that the kouros, with 0.9 probability, was composed of Thasian marble. Historical evidence concurs with the results, because Thasos-Akropolis has the oldest quarries of Thasos, which also produced kouroi during the seventh and sixth centuries B.C. Thus the results of th is analysis, even though they do not conclusively prove the Getty kouros to be authentic, revealed that a combination of multiple scientific and arch aeological techniques is required if we are to succeed in determination of marble sources (Herz 1990: 108). Several archaeologists have developed wh at they view as the correct methodology for analysis of archaeological marble. Many heated debates have arisen during ancient marble conferences, like the Association fo r the Study of Marble and Other Stones used In Antiquity (ASMOSIA), about the benefits and failings of the various techniques used to analyze marble. An example of a different technique used to acquire the same goal, provenance determination, came from a pers onal communication with Yannis Maniatis on October 9, 2003. He believes the best met hod includes the following series of steps: 1) a chip must be removed from the object in question, 2) the mi crostructures of the sample must be examined, 3) the grain size ra nge must be determined, 4) the texture must be described, 5) the sample should th en be ground for EP R spectroscopy of Mn2+ and Fe3+ impurities in the marble, and 6) SIRA of carbon and oxygen should be performed. The results of the EPR are quantitative, and do not destroy the material Maniatis (2003) believes that carbon and oxygen isotopic analysis is a blind test, which might produce results unrepresentative of the provenan ce area. While the results might be unrepresentative, this is true for any techni que and, furthermore, the chance of obtaining an unrepresentative sample is proportionally relative to the size of an object.
68 Attanasio and Platania (2000) recently have realized the importance of combination analysis. While they recogni zed the combination of INAA and carbon and oxygen SIRA to determine the provenance of marble, they used EPR spectroscopy, primarily focusing on the sp ectral features of the Mn2+ impurity. With either analytical technique, they emphasized the importance of petrographic and art hist orical information. For identification of joining fragments, A ttanasio and Platania suggest that while 13C and 18O isotopes have correctly identified joining fragments, SIRA results are extremely uncertain. They suggest that quarry variabil ity is substantial enough to make incorrect associations (Attanasio a nd Platania 2000: 322). Matthews (1988) recognized that variability must be relatively small over distances up to about one meter, but stated that fragment association is still possible, with some caution, and easily done with isotopic analysis. Matthews (1988) tested the variability of large sculptures of different marble types by taking samples from more than one place. The variability in one mausoleum frieze was shown to have up to 1.1 range of variation in 18O (Matthews 1988: 344). This variabil ity was assumed to be a result of weathering. Matthews (1988) suggests that discarding a greater amount of surface drillings produces a more reliable result. He concludes that multiple samplings of large objects should be taken in order to asse ss adequately an objects isotopic values (Matthews 1988: 345). For attribution of marble sculptures h oused in the Boston Museum of Fine Arts and the Sackler Museum, van der Merwe et al (1995) compared isotop ic results collected from the sculptures to that of the quarry database produ ced by Herz (1992), Matthews et al (1992), and Moens et al (1992). The results show that only two of the 83 sculptures
69 analyzed could be attribut ed unequivocally to a singl e quarry (van der Merwe et al 1995: 188). But many of the overlapping results we re resolved when additional information such as grain size, mineralogy, color, and hi storical data were ta ken into account. Moens et al (1988; 1992) favored an approach that combines thin-section petrology, carbon and oxygen SIRA, and traceelement analysis, using INAA, AAS, and CL. In a study published in 1992, Moens et al reported the results of analysis of 129 white marble artifacts. A core sample (diameter = 15 mm; length 50 mm) was extracted from half of the artifacts, chip s were taken from about 10 percent of the artifacts, and powdered samples were taken from the remaining ones (Moens et al 1992: 248). The core samples, which were taken fr om less visible areas of the sculpture during restoration, allowed for all thr ee methods of analysis to occu r. Isotopic and petrographic analysis (including CL) only occurred on the samples where chips were removed due to the size required for petrogra phic analysis. The powdered sa mples only received isotopic analysis. This combination of techniques proved highly successful. Of the 129 samples tested, 118 received attribution to a single quarry source (Moens et al 1992: 249). For the artifacts that could not be attributed, half were powdered samples, limiting analysis to SIRA. The different analytical methods yielded contradictory information for the rest of the unattributable artifacts, suggesting a quarry not in the database, which contains 14 different quarries (Moens et al 1992: 249). Until this combination approach, Archaic Naxian sculptures only received stylistic analysis (Kokkorou-Alevras et al 1995: 95). Using a combination of EPR and INAA, the authors attempted to determine th e provenance of the sculptures. Maximum grain size was determined from small frag ments removed from the sculpture with a
70 chisel. Drilling was not used b ecause, according to Kokkorou-Alevras et al (1995), it alters the EPR spectrum. Using the follo wing characteristics in EPR (Kokkorou-Alevras et al 1995: 96): the Mn2+ ions and the peaks with g-values equal to 14.25, 4.70, 4.32, 2.0044, 2.0037, 2.0056, 2.0020, 2.0000 and the following elements in INAA (Kokkorou-Alevras et al .1995: 96): Na, K, Sc, Fe, Co, Zn, As, Br, Sr, La, Ce Nd, Sm, Eu, Tb, Yb, Hf, Th, U, and the ratio Eu/Ce the authors were able to associate or unasso ciate the sculptures with specific marble quarries. Sixteen of the 27 sculptures that were analyzed were actually from Paros; only five of the sculptures were from Naxos. Th is analysis even allowed for differentiation between the two marble sour ces on the island of Naxos. Mosaic Analysis Two factors must be considered when studying mosaic supply sources: 1) the availability of materials near the site and 2) the possibility of the use of recovery materials, or secondary waste materials from larger works. The mosaicist often used architectural marble debris (Bergamini and Fiori 1999). At Antioch, limestone and basalt were available as building materials. The ma jority of mosaics at Antioch are composed of a combination of limestone, marble, and glass tesserae. While limestone was most likely obtained from the local quarry and gla ss was most likely produced on site at the mosaic workshops, marble was not av ailable from a nearby source.
71 Capedri et al (2001) analyzed mosaics from the rooms of the Domus dei Coiedii at Suasa (Ancona, Italy) that dated archaeol ogically and stylistically between the end of the first century B.C. to the beginning of the first century A.D. and the second century A.D. to the first half of the third century A.D. Most of the floors were covered in mosaics. Some of the mosaics were destr oyed when Suasa was sacked and others were damaged by drainage. The mosaics are com posed mostly of stone, with a few glass tesserae. The older tesserae were less than 1 cm2, while the tesserae from the second period were larger than 1 cm2. Capedri et al (2001: 10) used a combination of petrofabric analysis and SIRA analysis to examine 81 tesserae, and determined that the opus tessellatum sections were made of mostly local stones, which belonged to the Umbro-Marchigiana Sedimentary Sequence. The white to pinkish and reddish tesserae were mostly limestone from the Scaglia Ro sata Formation. The dark to black tesserae were composed of non-fossiliferous marls a nd marly clay, which pr obably derive from the local Marne a Fucoidi Formation. The stones from the opus sectile sections, mosaics formed of geometric designs, were composed of sedimentary stone, magmatic stone, and marble. The sedimentary stones we re limestone, which belong to the Rosso Ammonitico, and occurred in the Umbro-Marchigiana Sedimentary Succession, black marls and marly clays similar to the stones in the opus tessellatum sections, and onyx marble. The magmatic stones were prophyrit es and gabbros. The marble stones were composed of several white marbles from Marmara and Carrara and colored marbles including: marmo cipollino (green), rosso antico (wine red), pavonazzetto (white with purple veining), portasanta (pink), giallo antico (white or pink and yellow), bigio antico
72 (dark gray with white and gray veins), lapis taenarius (dark gray), and brecce coralline (white set in reddish cement) (Capedri et al 2001: 7). Things to Consider and Commentary on the Techniques Attempts to source marble require proper sampling of both geological and artifactual materials. Weathering, ground wa ter, and carbon dioxide can cause variations in physical composition, including isotopic, cr ystalline, and elemental characteristics. Additional variation can occur w ithin a single quarry as a result of formation processes. All of these factors combined suggest the need for continued sampling and statistical treatment of the existing results. Some archaeologists have ar gued that when provenance que stions arise, the focus is on two or three quarries (Polikreti and Maniatis 2002); however, this focus adds an assumption into the analysis, which may prove false. It is important to include all quarries that might be involved rather than pick and c hoose a few to study, because trade routes have not been es tablished definitively. In addition to including all possible marble sources, we must consider the sources of other stone as well. Stable isotope rati o analysis results do not differentiate between marble and limestone. One problem is that is otopic compositions of marble often overlap with those of limestone, marbles protolith (Gorgoni et al 2002: 121). Wenner et al (1988: 325) suggest that the ge ological factors, such as metamorphism, that control variability of isotopic values in marble, also control variab ility in limestone. This, of course, means that any analyti cal technique designed to study the characteristic signatures of marble (trace element, spectra, or isotopic values) will not help differentiate between
73 marble and limestone. Wenner and Herz (1992: 199) also poin t out that while no database for the provenance of Classical limest ones exists, this is probably a result of the use of local limestone. Local stone resources, if available, would have been cheaper to use; however, commerce in certain limestone s existed during Greek and Roman times. Both Pliny and Theophrastus praised the poros limestone of Greece as being lighter than, but as attractive as, the famous lychni tes marble of Paros (Wenner and Herz 1992: 199). Corinth and Neapolis were both know n for limestone exportation by ship during Classical and Byzantine times. Wenner a nd Herz (1992) began the assembly of a database of isotopic signatures of limestone sources in Classical Greece by focusing on Corinth and Neapolis. Figure 32 shows the range in isotopic values of limestones, while Figure 33 compares the results of Wenner and Herzs study of the Neapolis and Corinth limestone quarries to a few of the near-by ma rble quarries. Wenne r and Herz (1992: 199) suggest that further isotopic, megascopic, a nd microscopic analysis should occur in order to understand better the extent of limestone exportation. Weathering is another explanation fo r contradictory information. Tykot et al (1999) show that weathering has a significant impact on isotopic signatures by using the combination of SIRA, XRD, gr ain-size determination, and arch aeological data to provide minimally destructive provenance informa tion. X-ray diffraction requires only a few milligrams of marble powder, and since XRD do es not alter the sample, SIRA can re-use the same sample. Because of the size of the sample, a highly weathered sample can produce unusual results. Weathered marble surfaces are likely to have an altered crystalline structure and alte red isotopic values (Tykot et al 1999). Understanding the formation of isotopic values and crystallin e structures can expl ain this phenomenon.
74 Figure 32. Comparison of the isotopic compositions of ancient marble quarries in the eastern Mediterranean w ith limestones (Wenner et al 1988: 326) Remembering that water affects the oxygen isot ope values and that carbon dioxide affects the carbon values, ground and meteoric wate r and atmospheric carbon dioxide exposure can lower isotopic signatures a nd recrystallize dolomite and calcite on marble surfaces; thus, a weathered marble sample requi res cleaning prior to analysis (Tykot et al 1999). The results of this study revealed that even small cracks or fissures can expose marble surfaces to weathering, and to be safe, even clean marble should be treated prior to
75 Figure 33. Comparison of isotopic signatures of Cl assical marble sources and limestone from Neapolis and Corinth (Wenner and Herz 1992: 202) sample collection. The resu lts of the work by Tykot et al (1999) reveal that a minimum of 1-2 mm of marble must be discarded prio r to collection of an XRD and SIRA sample from a weathered marble object. Stable isot ope ratio analysis and XRD were selected for this analysis because both techniques are minimally destructive, require small samples sizes, and have proven a successful combina tion in previous provenance determination studies.
76 Chapter Five: Analysis and Results The primary research objective of this study is source determination. The secondary objective is to evaluate the sour ce of the mosaic tesser ae in relation to the distance the material traveled and the image cr eated by individual stones. Positive results might reveal 1) preferred sources for tesserae, 2) trade routes of specialized stone, 3) chronological changes in pref erence and 4) workshop pref erence of stone material. Provenance determination can provide clues to help understand the meaning of the mosaic images and the possible trade routes that existed during the Roman period. X-ray diffraction (XRD) and stable isotope ratio analysis (SIRA) were selected for the determination of the provenance for 55 marb le mosaic tesserae from Antioch-on-theOrontes. Both techniques are minimally destruct ive and have been proven to be part of a successful combination method for determini ng the provenance of ma rble. The surface samples were collected by the Worcester Art Museum conservator at the time, Lawrence Becker, and sent for analysis in plastic c ontainers. The 55 samples include 22 red, 26 white, 4 black, and 2 brown tesserae and come from 10 mosaic floor images from 7 different rooms (see Table 1 for details). This analysis attempted to source mosaic tesserae with minimally destructive techniques of XRD and SIRA. In this study of tesserae fragments from Antioch mosaic floors, a combination analysis was applied. The samples were co llected by chipping
77 away at individual tesserae for samples of an appropriate size for SIRA and XRD analysis. First, the samples were described by color, weighed, and powdered for analysis. X-ray diffraction was used to determine th e crystalline compos ition of the sample, focusing on dolomite or calcite. Then, carbon and oxygen SIRA was performed on the same samples used in XRD. Comparative analysis of the samples 13C and 18O isotopic values to the isotopic values for qua rries that have been previously published occurred. The results were not definitive, but were informative. The isotopic databases that have been published, or made available, are limited to mostly white marble quarries. No technique applied alone can resolve all of the provenance ques tions, especially if there is no archaeological or art historical in formation available to restrict the number of quarries and confine the problem (Polikre ti and Maniatis 2002: 1). Discussion, comparative analysis, and further explanations of results are provide d in Chapter Six. X-ray diffraction uses x-rays to determine the primary mineral in an object. For marble objects, XRD can differentiate between dolomitic and calcitic marbles. Since several sources can be eliminat ed if we know the marble is dolomite, this is an easy way to determine what analytical technique s hould follow diffraction analysis. Because the ultimate question of XRD is to determine th e primary component of marble, dolomite or calcite, a shortened wavelength spectrum from 27 to 32 was selected to reduce analysis time. A calcitic sample would reflect the x-rays off the sample at around 29.4 and a dolomitic sample would reflect the x-rays at around 30.8. Some sa mples did not reflect the x-rays between these ranges suggesting one of to things. One po ssible explanation is that the sample is not primarily composed of either dolomite or calcite. Another possible explanation for no reflection in this range is that the sa mple is not large enough.
78 Stable isotope ratio analysis uses the i ndividual isotopic prop ortions of carbon and oxygen that are sealed in duri ng geological formation to fingerprint the source. The powdered sample, 100 g, was separated for the SIRA on a Finnigan MAT Delta Plus XL mass spectrometer equipped with a Kiel I II individual acid bath carbonate system. Isotopic values can also help eliminate se veral sources when trying to provenance an artifact. A combination of multiple pr ovenance techniques helps archaeologists reconstruct the variables that Romans cons idered when they collected marble for different uses. These variables include color, grain size, or size of marble that can be extracted. Based on the resulting databa se collected during analysis by several archaeologists including Norman Herz, Carlo Gorgoni, and Lorenzo Lazzarini, possible sources for several mosaic samples were determined. Some marble sources have multiple quarries, which have very different isotopic values. Paros, a small Cycladic Island, has three different marble quarries which were exploited (van der Merwe et al 1995; Gorgoni et al .2002). Sometimes this offers additional information about an artifact. Carrara, a marble source in Nort hern Italy, has one larg e quarry, but different sections were exploited duri ng different cultural periods (Herz 1987). The Classical period exploited one section of Carrara, while the Renaissanc e period exploited a different section (Herz 1987: 39). We know our samples are authentic Classical artifacts, and not replications, because they compare to the same isotopic values as the Classical section of the quarry in the database. This kind of information allows us to know how far marble traveled via artisans or tradesmen. The mosaics, which currently are housed at the Worcester Art Museum (with the exception of the lower half of the Aphrodite and Adonis mosaic image, which is housed
79 at the Wellesley College Museum), represent a broad time range, from the first century A.D. to sixth century A.D. Even with th e specific question about image meaning, the results of XRD and SIRA may reveal even more important information about the relationships that exist between the samples and the possible sources. Provenance of this material might reveal source preferences, pr eferences over time, information about trade of specialized stone, and workshop preference of stone material. The results of the XRD and SIRA are show n in Table 2. The XRD results reveal that all of the dolomitic samples are red. The rest of the tesserae are composed of mostly calcitic materials (Figure 34). Fifty-one of the 55 samples were large enough to run XRD analysis. The four samples too small for XR D analysis to occur were two red samples, one white sample, and one brown sample from the Aphrodite and Adonis mosaic panel. Of the 51 samples run, only one sample return ed an indeterminate result and one black sample was determined not to be marble. Of the 49 other samples, 11 were dolomitic and 38 were calcitic. All of the dolomitic results were from red tesserae, including six samples from the Drinking Contest mosaic, one sample from the Ktisis mosaic, and three samples from Dionysos and Ariadne mosaic. Nine of the 22 red samples were calcitic, 24 of the white samples were calcitic, both of the brown samples were calcitic, and four of the five black samples were calcitic. Although most of the mosaic samples appear to cluster in value with the other samples from the same mosaics, their sources cannot necessarily be connected to values existing in the white marble database. The Drinking Contest mosaic has the most variability in isotopic values, suggesting multiple sources were used. Although the results are not definitive, othe r tests can be run to eliminate some of the remaining
80 Table 2. Antioch Mosaic SIRA and XRD Results USF#Mosaic NameColorXRD Resultd13C PDBd18O PDB 6115Worcester Hunt Mosaicwhitecalcite-5.4-5.5 6116Worcester Hunt Mosaicredcalcite-7.0-6.8 6117Worcester Hunt Mosaic East Borderwhitecalcite-3.8-5.2 6118Worcester Hunt Mosaic East Borderredcalcite-3.0-4.6 6119Agora, Borderwhitecalcite-3.1-5.2 6120Agora, Emblemawhitecalcite-4.0-5.4 6121Agora, Emblemaredcalcite-5.0-5.2 6122Agora, Borderredcalcite-5.2-6.1 6123Eukarpia, Borderwhitecalcite-2.7-5.3 6124Eukarpia, Emblemawhitecalcite-4.5-5.6 6125Eukarpia, Emblemaredcalcite-5.7-5.9 6126Eukarpia, Borderredcalcite-3.8-5.5 6127Funerary Symposium, Borderwhitecalcite-3.2-5.5 6128Funerary Symposium, Emblemawhitecalcite-5.6-6.3 6129Funerary Symposium, Emblemaredcalcite-5.0-5.8 6130Drinking Contest, Awhitecalcite-2.7-6.4 6131Drinking Contest, Bwhitecalcite-2.9-7.2 6132Drinking Contest, Cwhitecalcite-2.9-6.5 6133Drinking Contest, Dwhiteindeterminant-4.4-4.2 6134Drinking Contest, Ewhitecalcite-2.6-6.6 6135Drinking Contest, Fwhitecalcite-2.7-6.9 6136Drinking Contest, Greddolomite-1.2-2.7 6137Drinking Contest, Hreddolomite-4.5-1.8 6138Drinking Contest, Ireddolomite0.60.4 6139Drinking Contest, Jreddolomite-0.52.7 6140Drinking Contest, Kreddolomite-0.52.6 6141Drinking Contest, Lreddolomite-0.62.7 6142Drinking Contest, Mreddolomite-6.2-2.9 6609Drinking Contest, NBrowncalcite1.3-3.0 6610Drinking Contest, OBlackcalcite-1.0-4.3 6612Aphrodite and Adonis, ARedcalcite-1.4-5.2 6613Aphrodite and Adonis, BRedn/a-3.3-6.2 6614Aphrodite and Adonis, CRedcalcite-3.6-6.1 6615Aphrodite and Adonis, DRedn/a-2.41.1 6616Aphrodite and Adonis, EWhitecalcite-3.2-5.6 Continued on next page
81 Table 2 (continued) USF#Mosaic NameColorXRD Resultd13C PDBd18O PDB 6617Aphrodite and Adonis, FWhitecalcite-3.5-4.4 6618Aphrodite and Adonis, GWhiten/a-3.4-6.4 6619Aphrodite and Adonis, HWhitecalcite0.1-3.2 6620Aphrodite and Adonis, IBlackcalcite-1.1-3.3 6621Aphrodite and Adonis, JBrownn/a0.6-2.5 6622Hermes and the Infant Dionysos, AWhitecalcite-3.8-5.2 6623Hermes and the Infant Dionysos, BWhitecalcite-4.0-5.2 6624Ktisis, AWhitecalcite-2.0-6.3 6625Ktisis, BWhitecalcite-2.4-6.5 6626Ktisis, CReddolomite-5.8-1.7 6627Ktisis, EBlackunknown-18.3-25.4 6628Dionysos and Ariadne, AWhitecalcite-4.9-5.7 6629Dionysos and Ariadne, BWhitecalcite-6.1-6.2 6630Dionysos and Ariadne, CWhitecalcite-6.2-6.2 6631Dionysos and Ariadne, DReddolomite-0.52.5 6632Dionysos and Ariadne, EReddolomite0.02.7 6633Dionysos and Ariadne, FReddolomite0.22.8 6634Dionysos and Ariadne, GBlackcalcite-0.5-4.6 6635Peacock Mosaic, AWhitecalcite-4.8-4.6 possibilities. Those marble artifacts with ove rlapping carbon and oxygen values can be further analyzed through quantitative analys es or by other techni ques including scanning electron microscope (SEM), CL, EPR, and stro ntium isotope analysis. These analytical techniques used in combination with SIRA can provide an attributi on to a more specific source. Figure 34 highlights the differences in isotopic values of the calcitic and dolomitic samples. The dolomitic samples have higher 13O values. Two clusters of dolomitic samples are apparent. One cluster of six samples, 6139, 6140, 6141, 6631, 6632, and 6638, suggests the samples match a single source. Three of those samples come from the Drinking Contest mosaic and three come from the Dionysos and Ariadne mosaic, which are part of the triclinium pavement in the Atrium House. Another cluster
82 Figure 34. SIRA results with calcitic sample s labeled with squares and dolomitic samples labeled with diamonds of dolomitic samples, whose values have a greater spread, suggest a single source for the following samples: two from the Drinking Contest mosaic (6137, 6142) and one from the Ktisis mosaic (6626). Two major clusters of calcitic samples are evident. One is composed of one red (6612), one white (6619), three black (6610, 6620, 6634), and two brown (6609, 6621) samples which come from the Aphrodite and Adonis Drinking Contest and Dionysos and Ariadne mosaics. This group has higher 13C values than the second cluster. The second cluster is co mposed of nine red samples (6116, 6118, 6121, 6122, 6125, 6126, 6129, 6613, 6614) and twenty-three white samples (6115, 6117, 6119, 6120, 6123, 6124, 6127, 6128, 6130, 6131, 6132, 6134, 6135, 6616, 6617, 6622, 6623, 6624, 6625, 6628, 6629, 6630, 6635), which include samples from each of the mosaics
83 tested in this study. Within th e second cluster there are few s econdary clusters that reveal pertinent information. With slightly higher 13C, a small cluster of white calcitic samples (6130, 6131, 6132, 6134, 6135, 6625, and 6624) shows very similar isotopic values. These seven samples come from two mosaics, the Drinking Contest and Ktisis The SIRA results are presented in Figur es 34 through 36 and reveal a range of carbon and oxygen values: for 13C = -7.1 to 1.3 and for 18O = -6.9 to 2.8 . Figure 35 shows the isotopic values of each of the samples included in this study with each of the houses represented by differing colors. In Figure 35, the isotopic values are labeled according to the USF laboratory num ber as shown on Table 1 and Table 2. Figure 36 shows the distribution of the isotop ic values of each the different colored samples. SPSS exploratory data analysis was used to crea te boxplots revealing the range of carbon and oxygen isotopic values (Figure 37). The boxplots reveals that the greatest range of isotopic values exists within the Drinking Contest (range of 13C = 7.5 ; range of 18O = 9.9 ) and Dionysos and Ariadne (range of 13C = 6.4 ; range of 18O = 9.0 ). The next largest range of values exists in the Aphrodite and Adonis (range of 13C = 4.2 ; range of 18O = 7.5 ) and the Ktisis (range of 13C = 3.9 ; range of 18O = 4.9 ). Further investigation of the exploratory data analys is reveal that both the oxygen and carbon values for the Dionysos and Ariadne mosaic are bimodal (modes for 13C = 5.3 and -.3 ; modes for 18O = -5.3 and 2.8 ) in distribution suggesting two sources of material used. One black sample now believed to be limestone, produced the unexpected results of 13C: -18.3 and 18O: -25.4 . This extremely low value suggests that the sample is not marble and was not included in the exploratory data
84 Figure 35. SIRA of mosaics color-coded by house
85 Fi g ure 36. SIRA of mosaics g rou p ed b y color
86 Figure 37. Boxplots showing the range of carbon and oxygen isotope values
87 analysis. Although the results fail to match up with the Classical white marble database, the results form patterns that suggest similar sources and accurate results. The results will be further discussed in groups by color since previous studies have focused on color as an identification technique.
88 Chapter Six: Discussion By far, white marble has received the most attention (in the literature) for source determination. Pure white marble was the preferred marble type for both Greek and Roman sculptures (Herz 1987: 35; Rapp 1998: 140). This is not necessarily true for mosaics. Using the stable isotope ratio analysis (SIRA) results from the mosaic samples, comparisons with the white marble database can aid in the provenance determination of the mosaic samples. Figures 38 through 45 compare the mosaic results with the published source fields for the white marble da tabase of stable isotopes for each of the major quarries as shown in Gorgoni et al (2002). Doted lines s how the SIRA results of artifacts tested in the analysis that Gorgoni et al (2002) performed. The dashed lines show the SIRA results of the quarries tested and the circles with white interiors show the previous analysis performed by Moens et al (1992). The gray areas show the differences between the Gorgoni et al (2002) data and the Moens et al (1992) data. The idea behind using these diagrams is to provide pictoria l correlations between the mosaic results and the known databases. The actual databases have not been provided to the author; therefore, statistical assessment of the samples cannot o ccur. Using a visual comparison, the majority of the results, with a few excep tions, fall outside of th e most commonly used white marble quarries. Those exceptions includ e samples from each of the color groups.
89 Figure 38. Aphrodisias white marble database co mpared to mosaic samples (after Gorgoni et al 2002) Figure 39. Carrara white marble database compared to mosaic samples (after Gorgoni et al 2002)
90 Figure 40. Dokimeion white marble database co mpared to mosaic samples (after Gorgoni et al 2002) Figure 41. Naxos white marble database compared to mosaic samples (after Gorgoni et al 2002)
91 Figure 42. Paros white marble database compar ed to mosaic samples (after Gorgoni et al 2002) Figure 43. Penteli white marble database compared to mosaic samples (after Gorgoni et al 2002)
92 Figure 44. Prokonnesos white marble database co mpared to mosaic samples (after Gorgoni et al 2002) Figure 45. Thasos white marble database compared to mosaic samples (after Gorgoni et al 2002)
93 Examining the eight most commonly used wh ite marble quarry values provides some clues as to the source of the mosaic samples. Possible Sources Two brown samples (6609, 6621), one white sample (6619), one red sample (6136), and one black sample (6620) overlap with the isotopic ellipses of the Aphrodisias quarry. Two additional black samples (6610, 66 34) match the results of the Aphrodisias quarry. These samples come from the Drinking Contest and the Aphrodite and Adonis mosaics, which were found in the same house, and from the Dionysos and Ariadne mosaic. These mosaics all date between the second and third century A.D. With the exception of sample 6612, these seven samples form one of the four clusters in the data. It is also important to note that the qu arry of Aphrodisias produced black and white colored marbles (Anderson 1989: 65). One brown samples (6609) falls just outside of the statistical ellipse of the Ca rrara quarry; however, the Ca rrara quarry does not compare with isotopic signatures of any other sample s in this study. One brown (6609), one black (6634), and one red sample (6612) match the is otopic ellipse of the Dokimeion quarry. A second black sample (6610) falls just outside the isotopic ellipse for the Dokimeion quarry. Dokimeion marble, also known as pavonazzetto exists in two forms: a finegrained all white marble and a yellow-white wi th gray to red to violet veining (Anderson 1989: 93). Three of the samples that match the Dokimeion quarry fa lls within the known pavonazzetto color variation: the red sample (6612) from the Aphrodite and Adonis mosaic and the black samples (6634) from the Dionysos and Ariadne mosaic and (6610) from the Drinking Contest mosaic. One brown sample (6609) also falls within the Naxos
94 quarry isotopic ellipse. Both brown sa mples (6609, 6621) and one red sample (6138) overlap the Paros quarry isotopic ellipse. Probably, the white sample (6619) also falls within to isotopic signatures of the Paros quarry. The brown sample (6609) also falls within the Prokonnesos quarry. None of the samples fall within the ellipses for the Penteli and Thasos. Although several of the mo saic results match the isotopic values of multiple white marble quarries, the comparisons are just a starting point for the ultimate source assignment for the mosaic samples. One must also keep in mind that colored marble might have a different isotopic valu e than white marble from the same source location (Gorgoni 1992). In addition, colored marble or colored limestone might match isotopic signatures from different locations. Although the isotopic va lues did not provide ultimate source determination for all samples, a few explanations for the results can be given. One explanation for the results not be ing definitive is that not enough isotopic values have been collected from colored marb le sources. Another explanation is that the samples might not be marble, but rather limest one. The analytical techniques used in this study do not differentiate between marble and limestone, which are both calcareous materials, suggesting furthe r testing is needed to draw conclusive results. Only 9 of the 55 samples correspond with values on the white marble databases shown. While further scientific techniques ar e available for determining the source of the materials used in mosaics at Antioch, they cannot be used here, because multiple samples cannot be extracted from the materials in quest ion. The size of the original artifacts, the individual tesserae, does not al low every type of analysis to occur. Bergamini and Fiori (1999: 200) also noted that the size of mosaic pieces do not enable multiple analytical techniques to be used. The limited number of samples that can be taken without
95 destroying the mosaic itself and the limited funding available to run further scientific techniques hinder further analysis. Previous studies of source determination for mosaics have shown that the majority of tested tesserae are not actually marble, but rather limestone. Petrographic analysis and chemical analysis of the opus tessellatum sections of the Domus dei Coiedii mosaic at Roman Suasa in Ancona, Italy, reve aled that most tesserae were made from colored limestones or marls and marly clays (Capedri et al 2001: 12). Another study using petrographic analysis al one reached the same conclusions. A survey of mosaic tesserae from different centuries and vari ous geographic localities revealed that 85 percent of the 100 tesserae sampled were ma de of calcareous sedimentary limestones, and only 15 percent of the tesserae were made of marble and magma tic and detritic rocks (Bergamini and Fiori 1999: 200). This might explain the inability to connect the mosaic tesserae results with any known white marble quarries. The study by Capedri et al (2001) on mosaics from the rooms of the Domus dei Coiedii at Suasa also included an examinati on of samples taken from sections of opus sectile fragments, or mosaics made of geomet ric patterns. The stones used in these sections were composed of several different lithologies including metamorphic, sedimentary, and magmatic. The metamorphic stones included white and colored marble. The white marbles were Prokonnesian ma rble and Carrara marble (Capedri et al 2001). Several colored marble mosaic te sserae were also tested, including cipollino verde pavonazzetto lapis taenarius giallo antico portasanta rosso antico and brecce coralline Figure 46 shows the isotopic signatures for each of the above-mentioned marbles. The graphic results do not suggest a connection with any of the mosaic samples
96 Figure 46. Isotopic signatures for various colored marble from opus sectile mosaics (Capedri et al 2001: 14-21) indicating that none of the marble types included in the analysis performed by Capedri et al (2001) exist in the tesserae sampled in this analysis. In addition to the metamorphic stones, Capedri et al. (2001) also examined the sedime ntary and magmatic stones. The sedimentary stones included reddish limest ones, dark gray to black marls, and onyx marble (alabaster). The magmatic stone s included green porphyrites and mediumgrained gabbros. No SIRA was performed on these particular samples in the study by Capedri et al (2001). The ultimate determination of provenance for the mosaic samples must come from the isotopic ratios, which are more typi cal of limestone than of marble. If the samples are marble, they do not match anyt hing in the known marble database. Herz
97 (1992: 188) noticed a similar trend in one of his many studies and determined that in all probability the samples he tested were local limestone. A reexamination of the limestone plot from Chapter Four provides some clue s as to possible answers for the results obtained from the mosaic samples (Figure 47) A dashed rectangular box is drawn over the area in which most marble isotopic signatures fall. The majority of the mosaic samples fall in the ranges of the average fr eshwater limestones and the common marine limestones. Although the comparative analysis fails to identify the sources of a large number of the mosaic samples, many observations can be made about the isotop ic results. When compared to the limestone quarries published in Wenner et al (1988), other sources can be determined (Figure 47). The isotopic valu es for the white and red tesserae tested from the Necropolis, including the Funerary Symposium (6127, 6128, 6129), Eukarpia (6123, 6124, 6125, 6126), and Agora (6119, 6120, 6121, 6122) mosaic pavements, reveal a clustering that suggests a similar source for a ll of the samples. Th e Necropolis tesserae are probably all freshwater or marine limestones, since they co mpare to the isotopic fields published by Wenner et al (1988). The samples taken from the Atrium House, including the Drinking Contest (6130, 6131, 6132, 6133, 6134, 6135, 6136, 6138, 6139, 6140, 6141, 6609, 6610, and 6611) and the Aphrodite and Adonis (6612, 6613, 6614, 6616, 6617, 6618, 6619, 6620, and 6621) mosaic pavements, al so appear to cluster together. These samples include red and white tesserae a nd appear to compare to isotopic ratios of the marine limestones. Another clustering comes from three red samples from the Dionysos and Ariadne mosaic pavements (6631, 6632, and 6633) and three of the red samples from the Drinking Contest mosaic (6139, 6140, and 6141). According to the
98 comparison figure these samples can be calle d deep limestones. The red and white Worcester Hunt mosaic samples 6115 and 6116 overlap the isotopic ellipses of freshwater limestones and the Worcester Hunt mosaic samples 6117 and 6118 overlap the isotopic ellipses of marine limest ones. The white samples from the Hermes and the Infant Dionysos mosaic (6622, 6623) match the isotopic ellipses of freshwater limestones. Two white samples from the Ktisis mosaic (6624, 6625) and one black sample from the Dionysos and Ariadne mosaic (6634) fall within the isotopic values of marine limestones. Most of the mosaics sa mpled in this study fall within one of the categories set up by Wenner et al (1988). One red sample from the Drinking Contest mosaic (6138) matches the values of ooze, a deep-sea sediment composed of shells of marine animals and plants (Monroe and Wi cander 1997: 616). Only six of the mosaic samples do not overlap with anything on the li mestone isotopic scatterplot, including two red Drinking Contest samples (6137, 6142), one red Ktisis sample (6626), one white Drinking Contest sample (6133), one red Aphrodite and Adonis sample (6615), and one black Ktisis sample (6627). Although many samples match the isotopic ellipses for limestone, it does not mean that all the sa mples are limestone or another sedimentary rock. The similar values simply highlight th at the possibility exists that the materials sampled are not marble. With this possibility, the limestone in the Antioch region should be tested and compared with the SIRA results provided in this thesis. If the samples are not limestone, then another explanation for unusual isotopic results may be weathering. Several researcher s have described unexpected isotopic ratios as possibly relating to weathering. The research studies by both van der Merwe et al (1995) and Margolis and Showers (1988) test ed surface and subsurface isotopic ratios
99 Figure 47. Comparison of the mosaic isotopic values with the isotopic values of limestone quarries in the eastern Mediterranean ( after Wenner et al 1988: 326.
100 and compared the results. van der Merwe et al. (1995: 194) revealed that the oxygen isotopic ratios in weathered surfaces of dolomitic marble can deplete 1-2 . Margolis and Showers (1988; 1990) also reported negative shifts in carbon isotopes of up to 13 and in oxygen isotopes of up to 3.2 in marb le sculptures. Both studies reveal that weathering can have an effect, usually a negative shift, on the stable isotopic ratios. Since the majority of SIRA results for the mosaic samples are more negative than what would be expected if the samples were marb le, the SIRA results may be indicative of weathering. Both carbon and oxygen isotopic va lues of groundwater tend to be more negative than those of marble, and could possi bly be the cause of th e isotopic values of the mosaic samples. It is important to note the exact nature of the samples accepted. Although the samples were solid, they were too tiny to remove a part of the surface; and therefore, eliminate the possibi lity of weathering. As with collections of sculptures, the mosaics sampled here had been cleaned thor oughly by museum curato rs, greatly reducing the probability of weathered surfaces. While it is possible that the isotopic values obtained in this study are the result of weat hered surfaces, it is unlikely given that no weathering was evident. Also if the isotopic values were th e result weathering, then the isotopic values would be scattered and w ould not cluster as they seem to do. Clearly, a single method of anal ysis is not particularly e ffective in determining the sources of colored mosaic tesserae. The combination of carbon and oxygen SIRA with the identification of dolomitic marble through XRD did not prove as successful in this study as it has in previous white marble st udies. Additional analysis with techniques such as SEM, CL, EPR, and trace element anal ysis could provide further answers. These methods are highly dependent on the extent of sampling allowed by museum curators.
101 Typically, chips and cores necessary for pe trographic and elemental analysis are not easily obtainable unless the samples are undergoi ng massive restorati on procedures (van der Merwe 1995: 195). In addition, the collect ion of a non-white marble database would significantly aid future research on colored mosaic tesserae. This discussion focused on identifying possible quarry sources for the Antioch mosaic tesserae. The values for some samples tested in this analysis were similar to the white marble database but other samples have very different values In addition, some samples overlap with isotopic values to se dimentary rocks, such as limestone.
102 Chapter Seven: Conclusions Since the isotopic values for the samp les presented in this study show considerable overlap, it is cl ear that visual, historical, and archaeological information regarding the mosaics and the samples are vi tally important. Visual analysis of each sample the presence of streaks and inclusi ons, the grain size, and color aids in the final conclusions about possibl e sources. Fifty-five mosaic tesserae from ten Antiochene mosaics were analyzed with XRD and SIRA. It was thought that SIRA combined with XRD would provide the information necessary to answer the research questions. The quantitative data obtained from carbon a nd oxygen SIRA in this study provide the archaeological community with the beginnings of an artifact database for both mosaics and colored marble. Findings for Research Goals and Objectives The research goals this study hopes to address focused on 1) using XRD and SIRA to determine the source of the materials used in mosaic tesserae from Antioch; 2) comparing results for similarity; 3) determ ining the temporal or spatial relationship between the source and the importance of the image created; and 4) determining if a correlation exists between the importance of th e materials used with the distance that the material traveled.
103 The research questions for this study fo cused on two main objectives. The results of provenance determination could reveal m eaning behind the mosaic images and could reveal new trade routes in the Mediterranean. These questions will remain unanswered, since the provenance of most of the mosaic tesserae was not determined. Although this study did not answer a ll of the proposed research questi ons, it does show the importance of integrating multiple techniques for provenance determination studies. The results obtained in this study revealed important in formation that adds to existing marble databases and provides a basis for further investigation of colored marble artifacts. The results of the SIRA and XRD analysis showed that the materials used for mosaic tesserae come from a variety of sources. Although no definitive sources were found, several possibilities exist. Fifty-one of the samples were large enough to run XRD. One returned an indeterminate resu lt, one sample was determined not to be marble, 11 samples were determined to be dolomite, and 38 were determined to be calcite. The results of the comparative SIRA showed a large degree of overlap. While Aphrodisias was determined to be a possi ble source for the following samples: 6609, 6610, 6612, 6619, 6136, 6620, and 6634, Carrara was also identified as a possible source for sample 6609, Dokimeion as a possi ble source for 6609, 6634, 6610 and 6612, Naxos as a possible source for sample 6609, the Paros quarry as a possible source for 6609, 6619, 6621, and 6138, and Prokonnesos as a possible source of 6609. The results of the SIRA present multiple sources for several of the tesserae (i.e. 6609), and failed to identify sources for many of the tesserae. Several explanations can be given for these results. First, it is important to note the possibility that the samples are not marble, but rather limestone. One way of addressing this possibi lity would be to perform SIRA on the local
104 limestones around the Antioch region. Second, th e SIRA values were more negative than expected if the results were marble suggesting the possibility of weat hering. Sample size makes this possible explanation more difficult to address. The second research goal examined whic h samples share sim ilar results. The results described in Chapter Five were comp ared and revealed several clusters that suggest similar sources. Four clusters were evident when the samples were graphed on a scatterplot: two dolomitic cluste rs and two calcitic clusters. The first dolomitic cluster includes three samples from the Drinking Contest mosaic and three from the Dionysos and Ariadne mosaic which have matching isotopic values. The second dolomitic cluster includes three samples: one from the Ktisis mosaic and two from the Drinking Contest mosaic. The first calcitic cluster includes six samples from the Aphrodite and Adonis the Drinking Contest and the Dionysos and Ariadne mosaics. The second calcitic cluster includes 33 samples from each of the mosaics included in this study. The third research goal focused on de termining the temporal or spatial relationship between the source and the importa nce of the image created. This question cannot be answered without fu rther testing, because source s were not determined for enough of the samples to derive any relati onship between the source and the importance of an image. Testing of additional samples from specific portions of images, testing of additional colored marble sources, and the incl usion of additional scie ntific and analytical techniques would help answer this research question. The fourth goal was to determine if a correlation existed between the importance of the materials used and the distance that the material traveled. Again, this question cannot yet be answered as the definitive so urce of each of the mosaic tesserae has not
105 been determined. One possible explanation for the results suggests that limestone was used rather than marble. This suggests that mosaicists were utilizing cheaper and more readily available resources for mosaics. It is important to note that the mosaic tesserae that have similar values to the known marb le sources are from the second and third centuries, the earliest centuries sampled here and that none of the mosaic tesserae from the fourth through sixth centuries had simila r isotopic values with the known marble database. This variance in isotope values is supported by the ra nge displays in the boxplots in Figure 37. Not all of the sample s from the second and third centuries are similar to the known marble sources, possi bly suggesting a transitional period where mosaicists used both limestone and marble resources. A complete picture of the correlation cannot be drawn between the importan ce of the materials and the distance that the material traveled wi thout further testing. Although some of the research goals were not attained in this study, answers are ultimately attainable through further analysis. Even though the source has not been determined for all of the samples tested, the spatial relationship between the sample values and other values on the isotopic data base helps in understanding the relationship between the mosaic images. Several cluste rs of values suggest common sources. Although the sources temporarily remain unknown, further anal ysis can build on the data provided here. Limitations of This Study Several limitations to this study exist. Fi rst, with samples that are smaller than 2.0 cm2, size limits the number of analytical t echniques that can be applied to any one
106 tessera. Characterization and classification of stone materials used in mosaics is a tedious and difficult task. For this reason, mosaic an alyses require scientif ic techniques with a high degree of accuracy for small sample sizes. These techniques often are more expensive and limit the amount of analysis that can occur. The second limitation to this analysis is money. Availability of finances proved to be a deciding factor in the amount of analysis that occurred in this project. The third limitation to this analysis is the complete absence of an existing database for colored marble. The lack of a colored marble database made it impossible to compare the results of this study with known colored marble quarries. Future Research Future work on mosaic tesserae analysis should begin with further analysis of colored marble sources. The existing SIRA databases focus primarily on white marble. The construction of a colored marble databa se would enable arch aeologists not only to source mosaic tesserae, but also colored sculptures, architectur al elements, and inlays. In addition, future analysis of mosaic tesserae should include more than just SIRA and XRD analysis. Strontium isotope analysis, thin section petr ography, SEM, or CL would enhance the analysis of mosaic tesserae a nd provide additional information when two techniques return conflicting re sults. These additional tests add to the cost of analysis; however, they also aid in the final source determination process. Archaeologists have studied interregional contact through trade routes within the Mediterranean Sea for many decades. This study provides further evidence of trade networks and suggests that trade was not just for essential materials, like metals, or food,
107 like bulk shipments of grain. The materials used for mosaic tesserae suggest that trade also occurred for decorative materials, such as mosaic floors. This thesis has addressed the history and methods of ancient marble extraction techniques, the archaeology of the Antioch region, the scientific methods of XRD and SIRA, and a discussion of the results of the analysis of the mosaic tesserae sa mpled. This study provides the archaeological community with the basis for a database of colored marble quarries and artifacts. The information provided here adds to the gr owing body of knowledge about the Late Roman and Early Byzantine world.
108 References Anderson, Maxwell L. 1989 Colored Marble Sculpture in Roma n Antiquity. In (Maxwell L. Anderson and Leila Nista, Eds.), Radiance in Stone: Sculptures in Colored Marble from the Museo Nazionale Romano 11-22. Emory University Museum of Art and Archaeology, Atlanta. Attanasio, Donato and Rosario Platania 2000 ESR Spectroscopy as a Tool for Identifying Joining Fragments of Antique Marbles: The Example of a Pulpit by Donatello and Michelozzo. Journal of Magnetic Resonance 144 : 322-329. Barbin, Vincent, K. Ramseyer, D. Decrouez, S. J. Burns, J. Chamay, and J. L. Maier 1992 Cathodoluminescence of White Marble: An Overview. Archaeometry 34 (2): 175-183. Barbin, Vincent, Ellen Reeder and Terry Drayman-Weisser 1999 An Unidentified But Probably Identical Origin of Several White Marble Artifacts from the Walters Art Gallery (Baltimore, USA): Evidence from Petrographic, Cathodoluminescence and Stable Isotope Studies. In (M. Schvoerer, Ed.), Archomatriaux Marbres et autres roches, 39-43 ASMOSIA IV Proceedings, Bordeaux, 1995. Archetype Publications, London. Bergamini, Maria Letizia and Cesare Fiori 1999 Characterization of Limestones, Ma rbles and Other Stones Used in Ancient Mosaics. In (M. Schvoerer, Ed.), Archomatriaux Marbres et autres roches, 199207 ASMOSIA IV Proceedings, Bordeaux, 1995. Archetype Publications, London. Blanc, Philippe 1995 A Cathodoluminescence Spectrometer Built for an SEM JSM840A: First Results on Minerals and White Marble. In (Ya nnis Maniatis, Norman Herz, and Yannis Basiakos, Eds.) The Study of Marble and Othe r Stones Used in Antiquity 137-141. Transactions of the 3rd International Symposium of th e Association for the Study of Marble and Other Stones Used in Antiqui ty, Athens. Archetype Publications, London.
109 Campbell, Sheila 1979 Roman Mosaic Workshops in Turkey. American Journal of Archaeology 83 (3): 287-292. 1988 The Mosaics of Antioch. Corpus of mosaic pavements in Turkey : Subsidia mediaevalia 15. Pontifical Institut e of Mediaeval Studies, Toronto. Campbell, William Alexander 1934 Archaeological Notes: Excavati ons at the Antioch-on-the-Orontes. American Journal of Archaeology 38 : 201-206. 1936 Archaeological Notes: The Third Season of Excavation at Antioch-on-theOrontes. American Journal of Archaeology 40 : 1-9. 1938 Archaeological Notes: The Fourth and Fifth Seasons of Excavation at Antioch-on-the-Orontes: 1935-1936. American Journal of Archaeology 42 : 205-218. Capedri, Silvio, Giampiero Venturelli, Sandro De Maria, Maria Pia Mantovani Uguzzoni, and Gabriele Pancotti 2001 Characterisation and Provenance of Stones Used in the Mosaics of the Domus dei Coiedii at Roman Suasa (Ancona, Italy). Journal of Cultural Heritage 2 : 7-22. Coleman, M., and S. Walker 1979 Stable Isotope Identificati on of Greek and Turkish Marbles. Archaeometry 21 (1): 107-112. Craig, Harmon and Valerie Craig 1972 Greek Marbles: Determinati on of Provenance by Isotopic Analysis. Science 176 : 401-403. Downey, Glanville 1963 Ancient Antioch Princeton University Press, Princeton. Dunbabin, Katherine M. 1999 Mosaics of the Greek and Roman World Cambridge University Press, Cambridge. Elderkin, G.W. 1934 The Figure Mosaics. In (G.W. Elderkin, Ed.), Antioch-on-the-Orontes: The Excavation of 1932 42-48. Vol I. Princeton Univ ersity Press, Princeton. Erhart, K. Patricia 1978 A Portrait of Antonia Minor in th e Fogg Art Museum and Its Iconographical Tradition. American Journal of Archaeology 82 : 193-212.
110 Fant, J. Clayton 1988 The Roman Emperors in the Marb le Business: Capitalists, Middlemen, or Philanthropists? In (N. Herz and M. Waelkens, Eds.), Classical Marble: Geochemistry, Technology, Trade 147-158. Kluwer Academic Publishers, Dordrecht, The Netherlands. 1999 Augustus and the City of Marble. In (M. Schvoerer, Ed.), Archomatriaux Marbres et autres roches, 277-280 ASMOSIA IV Proceedings, Bordeaux, 1995. Archetype Publications, London. Giuliano, Antonio 1989 Colored Marble from Constantine to Napoleon. In (Maxwell L. Anderson and Leila Nista, Eds.), Radiance in Stone: Sculptures in Colored Marble from the Museo Nazionale Romano 23-34. Emory University Museum of Art and Archaeology, Atlanta. Gorgoni, Carlo, Lorenzo Lazzarini, Paolo Pallante, and Bruno Turi 2002 An Updated and Detailed Mineropetr ographic and C-O Stable Isotopic Reference Database for the Main Mediterranean Marbles Used in Antiquity. In (John J. Herrmann, Jr., Norman Herz, and Richard Newman, Eds.), Interdisciplinary Studies on Ancient Stone 115-131. Proceedings of the Fift h International Conference of the Association for the Study of Marble and Other Stones in Antiquity, Museum of Fine Arts, Boston, 1998. Archetype Publications, London. Gorgoni, C., I. Kokkinakis, L. Lazzarini, and M. Mariottini 1992 Geochemical and Petr ographic Characterization of Rosso Antico and Other White-Grey Marbles of Mani (Greece). In Marc Waelkens, Norman Herz, and Luc Moens (Eds.), Ancient Stones: Quarrying, Trade, and Provenance: In terdisciplinary Studies on Stones and Stone Technology in Europe and Near East from the Prehistoric to the Early Christian Period 155-165. Leuven University Press, Belgium. Gregarek, Heike 2002 Roman Imperial Sculpture of Colo red Marbles. In (John J. Herrmann, Jr., Norman Herz, and Richard Newman, Eds.), Interdisciplinary Studies on Ancient Stone 206-214. Proceedings of the Fifth Intern ational Conference of the Association for the Study of Marble and Other Stone s in Antiquity, Museum of Fine Arts, Boston, 1998. Archetype Publications, London. Grimanis, Apostolos P. and Maria Vassilaki-Grimani 1988 Provenance Studies of Greek Marb les by Instrumental Neutron Activation Analysis. In (N. Herz and M. Waelkens, Eds.), Classical Marble: Geochemistry, Technology, Trade 275-281. Kluwer Academic Publishers, Dordrecht, The Netherlands.
111 Guidobaldi, F. and A. Salvatori 1988 The Introduction of Polychrome Marbles in Late Republican Rome: The Evidence from Mosaic Pavements with Marb le Insertions. In (N. Herz and M. Waelkens, Eds.), Classical Marble: Geochemistry, Technology, Trade 171-175. Kluwer Academic Publishers, Dordrecht, The Netherlands. Hanfmann, George M. A. 1939 Notes on the Mosaics from Antioch. American Journal of Archaeology 43 : 229246. Henderson, Julian 2000 The Science and Archaeology of Materi als: An investigation of inorganic materials Routledge, New York. Herrmann, Jr., John J. 1990 Thasos and the Ancient Marble Trade: Evidence from American Museums. In (Marion True and Jerry Podany, Eds.), Marble: Art Historic al and Scientific Perspectives on Ancient Sculpture 73 100. Papers delivered at a Symposium organized by the Departments of Antiquitie s and Antiquities Cons ervation and Held at the J. Paul Getty Museum, April 28-30, 1988. The J. Paul Getty Museum, Malibu, CA. Herz, Norman 1987 Carbon and Oxygen Isotopic Ratios: A Database for Classical Greek and Roman Marble. Archaeometry 29 (1): 35-43. 1988 Geology of Greece and Turkey: Potential Marble Source Regions. In (N. Herz and M. Waelkens, Eds.), Classical Marble: Geochemistry, Technology, Trade 7-10. Kluwer Academic Publishers, Dordrecht, The Netherlands. 1990 Stable Isotope Analysis of Greek and Roman Marble: Provenance, Association, and Authenticity. In (Marion True and Jerry Podany, Eds.), Marble: Art Historical and Scientific Perspectives on Ancient Sculpture 101 110. Papers delivered at a Symposium organized by the Departments of Anti quities and Antiquities Conservation and Held at the J. Paul Getty Museum, April 28-30, 1988. The J. Paul Getty Museum, Malibu, CA. 1992 Provenance Determination of Neolith ic to Classical Medi terranean Marbles by Stable Isotopes. Archaeometry 34 (2): 185-194. Herz, Norman and Nancy E. Dean 1986 Stable Isotopes and Archaeological Ge ology: The Carrara Marble, Northern Italy. Applied Geochemistry 1 : 139-151. Herz, Norman and David B. Wenner 1981 Tracing the Origins of Marble. Archaeology 34 (5): 14-21.
112 Hurwit, Jeffrey M. 1985 The Art and Culture of Early Greece, 1100-480 B.C. Cornell University Press: London. Jones, Frances 1981 Antioch Mosaics in Princeton. Record of the Art Museum, Princeton University 40 (2): 2-27. Kokkorou-Alevras, G., V. Mandi, A.P. Grimanis, Y. Maniatis 1995 The Traditional Archaeological Char acterisation of Marble Sculpture and the Results of Modern Scientific Techniques. In (Yannis Maniatis, Norman Herz, and Yannis Basiakos, Eds.) The Study of Marble and Other Stones Used in Antiquity 95102. Transactions of the 3rd International Symposium of the Association for the Study of Marble and Other Stones Used in Antiquity, Athens. Archetype Publications, London. Kondoleon, Christine 2000 Antioch: The Lost Ancient City Princeton University Press and Worcester Art Museum, Princeton. Lambert, Joseph B. 1997 Traces of the Past: Unraveling the Se crets of Archaeology Through Chemistry Perseus Books, Reading, MA. Lavagne, H. 1988 Il Mosaico Attraverso i Secoli Longo Ravenna. Lazzarini, Lorenzo 1990 Rosso Antico and Other Red Marbles Used in Antiquity: A Characterization Study. In (Marion True and Jerry Podany, Eds.), Marble: Art Historical and Scientific Perspectives on Ancient Sculpture, 237-252. Papers delivered at a Symposium organized by the Departments of Anti quities and Antiquities Conservation and Held at the J. Paul Getty Museum, April 28-30, 1988. The J. Paul Getty Museum, Malibu, CA. Lepsius, G. Richard 1890 Griechische Marmorstudien Akademie der Wissenschaften zu Berlin vom Jahre 1890. Abhandlungen der Knigl. Preuss, Berlin. Levi, Doro 1947 Antioch Mosaic Pavements 2 vols. Princeton Univer sity Press, Princeton.
113 Lloyd, R. V., A. Tranh, S. Pearce, M. Cheeseman, and D. N. Lumsden 1988 ESR Spectroscopy and X-ray Powder Diffractometry for Marble Provenance Determination. In (N. Herz and M. Waelkens, Eds.), Classical Marble: Geochemistry, Technology, Trade 369-377. Kluwer Academic Publishers, Dordrecht, The Netherlands. Loggia.com 2003 Image of Phaedra and Hippolytus. Electronic image, http://www.logia.com/myth/phaedra.html, accessed on July 10, 2003. Luedtke, Barbara E. 1978 Chert Sources and Trace-Element Analysis. American Antiquity 43 (3): 413-423. Maniatis, Yannis 2003 Personal Communication at Demokritos on October 9, 2003 Mannoni, Luciana and Tiziano Mannoni 1986 Marble: The History of a Culture Facts on File Publications, New York. Margolis, Stanley V. and William Showers 1988 Weathering Characteristics, Age and Provenance and Determinations on Ancient Greek and Roman Marble Artifacts. In (N. Herz and M. Waelkens, Eds.), Classical Marble: Geochemistry, Technology, Trade 233-242. Kluwer Academic Publishers, Dordrecht, The Netherlands. 1990 Ancient Greek and Roman Marble Sculpture: Authentication, Weathering, and Provenance Determination. In (Mari on True and Jerry Podany, Eds.), Marble: Art Historical and Scientific Pe rspectives on Ancient Sculpture 283-299. Papers delivered at a Symposium organized by the Departments of Antiquities and Antiquities Conservation a nd Held at the J. Paul Getty Museum, April 28-30, 1988. The J. Paul Getty Museum, Malibu, CA. Matthews, Keith J. 1988 Variability in Stable Isotope Analysis : Implications for Joining Fragments. In (N. Herz and M. Waelkens, Eds.), Classical Marble: Geochemistry, Technology, Trade 339-346. Kluwer Academic Publishe rs, Dordrecht, The Netherlands. Matthews, K. J., M. N. Leese, M. J. Hughes, N. Herz, and S. G. E. Bowman 1995 Establishing the provena nce of marble using statisti cal combinations of stable isotope and neutron activati on analysis data. In (Ya nnis Maniatis, Norman Herz, and Yannis Basiakos, Eds.) The Study of Marble and Othe r Stones Used in Antiquity 171-180. Transactions of the 3rd International Symposium of the Association for the Study of Marble and Other Stones Used in Antiquity, Athens. Archetype Publications, London.
114 Matthews, Keith J., Luc Moens, Susan Walk er, Marc Waelkens, and Paul De Paepe 1992 The Re-evaluation of Stable Isot ope Data for Pentelic Marble. In (Marc Waelkens, Norman Herz, and Luc Moens, Eds.), Ancient Stones: Quarrying, Trade and Provenance: Interdisciplinary Studi es on Stones and Stone Technology in Europe and Near East from the Prehistoric to the Christian Period 203-212. Leuven University Press, Belgium. Mello, E., D. Monna, and M. Oddone 1988 Discriminating Sources of Medite rranean Marbles: A Pattern Recognition Approach. Archaeometry 30 (1): 102-108. Moens, Luc, Paul De Paepe, and Marc Waelkens 1992 Multidisciplinary Research and C ooperation: Keys to a Successful Provenance Determination of White Marble. In (Marc Waelkens, Norman Herz, and Luc Moens, Eds.), Ancient Stones: Quarrying, Trade and Pr ovenance: Interdisciplinary Studies on Stones and Stone Technology in Europe and Near East from the Prehistoric to the Christian Period 247-252. Leuven University Press, Belgium. Moens, L., P. Roos, J. De Rudder, P. De Paepe, J. Van Hende, and M. Waelkens 1988 A Multi-Method Approach to the Id entification of White Marbles Used in Antique Artifacts. In (N. Herz and M. Waelkens, Eds.), Classical Marble: Geochemistry, Technology, Trade 243-250. Kluwer Academic Publishers, Dordrecht, The Netherlands. Moltesen, M., Herz, N., and Moon, J. 1992 The Lepsius Marbles. In (Marc Waelkens, Norman Herz, and Luc Moens, Eds.), Ancient Stones: Quarrying, Trade and Pro venance: Interdisci plinary Studies on Stones and Stone Technology in Europe and Ne ar East from the Prehistoric to the Christian Period 277-281. Leuven University Press, Belgium. Monroe, J. S., and R. Wicander 1997 Physical Geology: Exploring the Earth Wadsworth Publishing Company, Belmont, CA. Morey, C. R. 1938 The Mosaics of Antioch Longmans, Green and Co., London. Polikreti, K. and Y. Maniatis 2002 A New Methodology for the Provenan ce of Marble Based on EPR Spectroscopy. Archaeometry 44 (1): 1-21. Rapp, Jr., George, and Christopher L. Hill 1998 Geoarchaeology: The Earth-Science Appr oach to Archaeologi cal Interpretation. Yale University Press, New Haven.
115 Renfrew, Colin 1972 The Emergence of Civilisation: The Cyclades and the Aegean in the Third Millennium B.C. Metheun, London. Robins, Gay 1997 The Art of Ancient Egypt Harvard University Press, Cambridge, MA. Stillwell, Richard (editor) 1938 Antioch-on-the-Orontes: The Excavations 1933-1936 Vol. II. Princeton University Press, Princeton. Turkey.com 2004 Turkish Map. Electronic image, http://www.turkey.com/travel/maps02.html, accessed February 4, 2004. Tykot, Robert H., Richard Newma n, and Nikolaas J. van der Merwe 1999 Weathering Surfaces on Classical Marble Sculptures: Isotopic and X-ray Diffraction Analysis. In (M. Schvoerer, Ed.), Archomatriaux Marbres et autres roches, 239-242 ASMOSIA IV Proceedings, Bordeaux, 1995. Archetype Publications, London. van der Merwe, Nikolaas J., John J. Herrma nn, Jr., Robert H. Tykot, Richard Newman, and Norman Herz 1995 Stable Carbon and Oxygen Isotope Sour ce Tracing of Marble Sculptures in the Museum of Fine Arts, Boston and the Sackler Musuem, Harvard. In (Yannis Maniatis, Norman Herz, and Yannis Basiakos, Eds.) The Study of Marble and Other Stones Used in Antiquity 187 197. Transactions of the 3rd International Symposium of the Association for the St udy of Marble and Other Stones Used in Antiquity, Athens. Archetype Publications, London. Waelkens, Marc, Paul De Paepe, and Luc Moens 1990 The Quarrying Technique of the Gr eek World. In (Marion True and Jerry Podany, Eds.), Marble: Art Historical and Scie ntific Perspectives on Ancient Sculpture 47-72. Papers delivered at a Sym posium organized by the Departments of Antiquities and Antiquities Conservation and Held at th e J. Paul Getty Museum, April 28-30, 1988. The J. Paul Getty Museum, Malibu, CA. Wenner, David B. and Norman Herz 1992 Provenance Signatures for Classica l Limestone. In (Marc Waelkens, Norman Herz, and Luc Moens, Eds.), Ancient Stones: Quarrying, Trade and Provenance: Interdisciplinary Studies on Stones and St one Technology in Europe and Near East from the Prehistoric to the Christian Period 199-202. Leuven University Press, Belgium.
116 Wenner, David B., Sherri Havert and Andrew Clark 1988 Variations in Stable Isotopic Compositio ns of Marble: An Assessment of Causes. In (N. Herz and M. Waelkens, Eds.), Classical Marble: Geochemistry, Technology, Trade 325-338. Kluwer Academic Publishe rs, Dordrecht, The Netherlands. Whitley, James 2001 The Archaeology of Ancient Greece Cambridge University Press: London.