Greek Marbles: Determination of Provenance by Isotopic Analysis Author(S): Harmon Craig and Valerie Craig Source: Science, New Series, Vol

Greek Marbles: Determination of Provenance by Isotopic Analysis Author(s): Harmon Craig and Valerie Craig Source: Science, New Series, Vol. 176, No. 4033 (Apr. 28, 1972), pp. 401-403 Published by: American Association for the Advancement of Science Stable URL: http://www.jstor.org/stable/1734394 . Accessed: 08/04/2014 11:45

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at . http://www.jstor.org/page/info/about/policies/terms.jsp

. JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide range of content in a trusted digital archive. We use information technology and tools to increase productivity and facilitate new forms of scholarship. For more information about JSTOR, please contact [email protected].

American Association for the Advancement of Science is collaborating with JSTOR to digitize, preserve and extend access to Science.

http://www.jstor.org

This content downloaded from 147.52.249.113 on Tue, 8 Apr 2014 11:45:54 AM All use subject to JSTOR Terms and Conditions the other hand, petrofabric studies are difficult and time consuming, require relatively large amounts of material, and depend on statistical treatments of what are, in these marbles, ratherpoor- ly developed fabrics. Our approach to this problem was Reports the idea that the 13C/12C and 18O/16O variations in Greek marbles probably provided the best chance for unique characterization by locality. The isotopes 13C and 180 vary by at least 10 and 20 per mil (1 to 2 percent), re- Greek Marbles:Determination of spectively, in marbles from different areas (9), so that with two isotopes Provenanceby Isotopic Analysis there is a reasonable probabilityof obtaining some distinctive differences. Abstract. A study has been made of carbon-13 and oxygen-18 variations in Moreover, we expected that the very Greek marbles from the ancient quarry localities of Naxos, Paros, Mount Hymet- process which originally formed the tus, and Mount Pentelikon. Parian, Hymettian, and Pentelic marbles can be clearly marblesmight well smooth out the vari- distinguished by the isotopic relationships; Naxian marbles fall into two groups ations within a single quarryingarea by characterized by different oxygen-18/oxygen-16 ratios. Ten archeological sam- local isotopic equilibration with fluid ples were also analyzed; the isotopic data indicate that the "Theseion" is made of phases. In general both these expecta- Pentelic marble and a block in the Treasury of Siphnos at Delphi is probably tions were fulfilled. Parian marble. We made detailed collections in the four major quarrying areas which pro- The problem of identifying the local- sen (7) were able to characterizeMar- vided marble for statuary and building ity from which a given specimen of mara and Parian marble as low in Mn, purposes in ancient Greece (10). In Greek marble was quarried has been none of these methods has been gen- each area we concentrated our sam- studied for almost 100 years (1). In erally very successful (8). Trace ele- pling at the sites of ancient quarries, as addition to the obvious applications ment studies suffer inherently from the indicated by old chisel marks, worked such as detection of copies and for- high variability in concentration (fac- faces, grooves, and other evidence. geries (2, p. 189) there are more gen- tors of over 100 for Na and Mn) Samples were collected over the maxi- eral problems of great importance due to highly localized interactionwith mum extent of each locality and from which could be solved by a method inclusions and surrounding rocks. On representativephases, in order to de- capable of uniquely assigning a piece of marble to its of The place origin. 7 major quarryinglocalities in use from preclassical to Hellenistic and Roman periods shifted successively from the Cycladic islands of Naxos, and later Paros, to Mount Pentelikon on the At- tic mainland, and finally during Roman times, to nearby Mount Hymettus (1; 2, p. 137; 3). Thus in many instances the specification of the provenance of a piece of statuary or a building can be used to date the object from the known periods of quarryingoperations in each area (4). Finally, there are geological problems at hand such as the question of whether the Pentelic and Hymettian marbles represent two formations or a single formation repeated by folding or faulting (5). Methods applied to Greek marbles for provenance determination in the past include petrographic techniques (1, 3), petrofabric studies (3, 6), and Na and Mn analysis by neutron acti- -9 -8 -7 -6 -5 -4 -3 -2 -1 vation (7). Although Herz and Pritch- 8018 (%o) ett (5) were able to solve a problem of Fig. 1. Carbon-13and oxygen-18variations in marble samples from ancient Greek a mismatch of inscription fragments by quarriesrelative to the PDB isotopic standard.Triangles denote some of the ar- a foliation study, and Ryback and Nis- cheologicalsamples listed in Table 1. 28 APRIL 1972 401

This content downloaded from 147.52.249.113 on Tue, 8 Apr 2014 11:45:54 AM All use subject to JSTOR Terms and Conditions Table 1. Isotopic analyses of archeological marble samples. where R is the ratio 13C/12C or 18O/ 160 and R+ is the ratio in the PDB Sam- /ac80 Probable ple Description (per mil) (per mil) prove- standard. The mass spectrometric No. nance methods and calculations have been de- A-1 Athens: "Theseion" (Hephaesteion) 2.63 -7.90 Pentelic scribed (14). The analytical precision is in the Agora, column on north + 0.05 per mil. side On the 13C-180 diagram, the marbles A-2 Athens: basal slab on south side of 2.63 -8.03 Pentelic fall into well-defined groups with the "Theseion" possible exception of certain Naxian D-1 Delphi: carved block lying inside 3.89 -3.10 Parian and Pentelic varieties. The Parian mar- Treasury of Siphnos bles are unique in having the highest D-2 Delphi: drum of column of the 1.96 -4.15 ? 813C values of the Greek specimens Naxian Sphinx and all of the marbles studied by Baer- E-1 Epidaurus: carved block (lintel?), 2.01 -1.12 ? 10 m from Tholos tschi (9). These marbles resemble algal carbonates and travertines in their high E-2 Epidaurus: Corinthian capital next 2.16 -0.78 ? to the "triple circle" of Tholos 13C values, and this resemblance to- gether with the high purity leads us to E-3 Epidaurus: cut slab of white marble, 2.58 -3.96 ? 13 m from Tholos suspect that they originated as a chemi- cal N-43 Naxos: Apollon Gate, Palatia, 1.67 -5.51 Naxian precipitate (15). Naxos Harbour; flake from inside Pentelic and Naxian marbles are of gate much lower in 180 than Parian and N-46 Naxos: Apollon Gate, large block at 1.64 -5.35 Naxian Hymettian marbles, a feature which is foot of hill toward Naxos generally due to interaction with me- C-1 Caesarea, Israel: Corinthian capital 3.51 -2.64 ? teoric waters at elevated temperatures of grey marble (9). The Pentelic marbles are also higher in 13C than almost all of the marbles studied by Baertschi, and the termine the variability in each area. A Hymettian marble, Mount Hymettus, Naxian and Hymettian marbles fall in total of 170 quarry samples and some southeast of Athens, on the western the upper part of his range. Only the 20 archeological samples were taken. slope (13). The marbles are grey, white, Hymettian marbles fall within the gen- The marbles studied, in order of age and blue, and are the "upper marble" of eral 13C-180 trend outlined by the mar- of the workings, were as follows. Lepsius (1). Seven ancient quarries ble samples in Baertschi's work. It is Naxian marble, on the island of were located and sampled. noteworthy that the isotopic data clearly Naxos in the Cyclades (6th and 7th General descriptions of these mar- confirm Lepsius's conclusion (1) that centuries). The ancient Naxian quarries bles are given in the references (1, 3, Hymettian and Pentelic marbles ("up- occur in two main localities, in two 5). The marbles are often (though not per and lower" marbles) are different formations (11). Near Kinidaros, in the always) distinguishable with high prob- geological formations. central interior, ancient quarries (prob- ability in outcrop or fresh hand speci- The Naxian marbles so far studied ably used only locally) occur in the men (Naxian marble is generally coarse fall into two groups with respect to Keramoti marble. On the north point grained, Hymettian marble is often grey 818O values. Each of these isotopic sub- of the island, at Apollon, the ancient to blue and banded), but cannot be dis- groups contains samples from both the quarries are in the Amomaxi marble, tinguished with certainty from each Naxian localities studied. Lepsius (1) which is separated from the older Kera- other, or from other marbles, in build- noted that there were two types of moti marble by Komiaki schist. Each ings and statuary. Both Pentelic and marble in the vicinity of the Apollon of these ancient sites is marked by the Hymettian marbles occur on each of Kouros, one of which closely resem- presence of a very large recumbent these mountains (1). bles the roof tiles of the "Old Temple" figure (Kouros), the earliest Greek stat- In this initial study a statistical sam- of the Acropolis and the Temple of uary. pling of about 35 percent of the sam- Zeus in Olympia. The isotopic data Parian marble, on the Cycladic island ples was selected for analysis. Further indicate that both these marbles prob- of Paros (5th and 6th centuries). The studies, including measurement of trace ably outcrop at both Kouros sites on Lychnites marble occurs at Aghias elements such as Sr and Mg are con- Naxos. Minas near Marathi, in the interior of tinuing. However, the isotopic results Similar uncertainties in the case of the island (12). The quarries are under- are so promising that we present them the Attic marbles require further study. ground and are reached by two steeply now, in the hope that interested work- Herz and Pritchett (5) placed most of dipping shafts that follow the bedding. ers will send us marbles from other the Hymettian quarries in the "lower" This is the finest statuary marble known. classical quarry areas and from archeo- or Pentelic formation (which provides Pentelic marble, Mount Pentelikon, logical materials whose place of origin the Pentelic marble on Pentelikon). 18 km northeast of Athens (5th century is of interest. However, the isotopic data show to Roman era). The ancient quarries Figure 1 shows the isotopic results clearly that the two sets of Attic mar- begin at the cave of Spilia on the south on the quarry specimens and some of bles represent quite different forma- we slope and extend up the mountain. Six the archeological samples. The data are tions, and that the quarries studied in ancient and two recent quarries were lo- given as 8 values relative to the PDB on Mount Hymettus are probably the The cated and sampled. The white Pentelic isotopic standard (14): "upper marble" of Lepsius. geo- are marble which was quarried is the logical relationships at Hymettus the - of entire "lower marble" of Lepsius (1). a (per mil) = [(R/R+) 1] X 1000 most complicated the region. SCIENCE, VOL. 176 402

This content downloaded from 147.52.249.113 on Tue, 8 Apr 2014 11:45:54 AM All use subject to JSTOR Terms and Conditions We were unable to obtain References and Notes cussions; Dr. E. Vanderpool of the American permission School of Classical Studies who led us to the from U.S. authorities to use the most 1. R. Lepsius, Abh. Akad. Wiss. Berlin (1890); Hymettian quarries; and H. Alt, F. Tsanos, detailed U.S. Service H. S. Washington, Amer. J. Archaeol. 11, 71 and the officers of the Societe Anonyme des Army Map maps; (1898). Carrieres du Marbre Dionyssos-Pendelis, for thus, we could not resolve this ques- 2. G. M. A. Richter, The Sculpture and Sculp- most helpful assistance on Mount Pentelikon. tion in the time available. tors of the Greeks (Yale Univ. Press, New 11. A geological map of Naxos is given by I. Haven, Conn., 1950). Papastamatiou, The Emery of Naxos (Geo- In Table 1 we list the archeological 3. N. Herz, Trans. N.Y. Acad. Sci. Ser. 2 17, logical and Geophysical Surveys, Athens, 499 (1955); Amer. J. Sci. 253, 299 (1955). 1951). samples collected for comparison with 4. Naxian marble (the earliest quarried) could 12. J. Anastopoulos, "Geological Survey of Lych- the quarry results. Figure 1 shows that have been used in later times, but Pentelic nites, Paros," stencil report (Institute of Geo- marble is of better quality and was much logical Subsurface Research, Athens, 1962). some success was obtained on these more accessible to Athens. Parian marble was The name "Lychnites" refers to the lamps first "unknowns." The "Theseion" of always in demand because of its superior used underground by the ancient quarry luster and grain, but our studies of the under- workers. the Athenian Agora (about 450 B.C.) ground mine from which the Lychnites marble 13. The ancient quarries are on the north bank Paros was obtained showed that Pentelic at of mining of Kakorhevma gorge, and are difficult to is almost certainly marble, stopped because the unfractured Lychnites locate. The gorge runs east to west at the site least on the north and south sides. The was totally removed. About 50,000 tons of of the quarries, which are indicated on U.S. this beautiful marble was ramped up two Army Map Service Sheet NJ 34-12 (Athinai) carved blocks inside the Treasury of narrow shafts dipping 35? before the bed was at coordinate reference GH 4505. The abandoned. Siphnos at Delphi are quite clearly of Acropolis bears N67?W from the easternmost 5. N. Herz and W. K. Pritchett, Amer. J. ancient quarry; the radar dome on the oppo- Parian marble. (The Treasury itself is Archaeol. 57, 71 (1953). The geology of the site bank is S85?E. Attic marbles is reviewed in this to be of marble.) The paper. 14. H. Craig, Geochim. Cosmochim. Acta 12, 133 reputed Siphnian 6. L. E. Weiss, Amer. J. Sci. 252, 641 (1954). (1957). The PDB is a Cretaceous belemnite. isotopic compositions of the rest of 7. L. Rybach and H. U. Nissen, Radiochem. 15. The 13C is enriched in carbonate precipitating Methods Anal. Proc. Symp. 1964 1, 105 from solution by escape of CO2 which is de- these samples do not correspond to any (1965). pleted in 13C. established for the 8. N. Herz has pointed out to us that most of 16. A advantage of the isotopic method is of the ;ranges so far these studies confirm that the Greeks great simply that only 10 to 20 mg of CaCO3 are required four areas studied. It is worth noting used the purest and most isotopic marbles for both analyses. We prefer to receive larger that two different marbles are that could be obtained. samples in order to ensure obtaining a clean quite 9. P. Baertschi, Schweiz. Mineral. Petrogr. Mitt. unaltered portion for analysis. 73 H. Geochim. Cosmo- lying in close proximity to the Tholos -37, (1957); Craig, 17. We thank K. Podvin and J. Brown for per- chim. Acta 3, 53 (1953). the and Professor of Fieldwork was done the fall of forming isotopic analyses (or labyrinth) Epidaurus. 10. during 1968, N. Herz for many helpful discussions. Dr. this connection it should be supported by a research grant from the H. U. Nissen in Zurich was also most (In of San We help- University California at Diego. ful. The laboratory work was supported by pointed out that the Athenians scatter gratefully acknowledge the many kindnesses the National Science Foundation. of Pentelic marble around of the following: Dr. A. Panagos of the fragments University of Athens for many helpful dis- 20 August 1971; revised 17 November 1971 * the Parthenon each winter, in order to provide material for the insatiable pil- lage by tourists. This marble is from modern quarries and is isotopically dis- Electrical Conductivity and the Red Shift of tinct from that of the classical quarries. Pressure All of the samples in Table 1 are defi- Absorption in Olivine and Spinel at High nitely ancient artifacts.) A single sample of grey marble was Abstract. Above 100 kilobars the apparent absorption edges (approximately 3 collected at Caesarea, Israel, in the be- electron volts) of single-crystal and polycrystalline samples of the metastable lief that it might be Hymettian marble olivine and stable spinel forms of Fe2SiO, shift rapidly with pressure from the imported by the Romans. However, near-ultraviolet into the lower-energy infrared region. Simultaneously, an expo- this appears not to be the case (Table nential increase in electrical conductivity occurs. These effects are reversible as 1). pressure is reduced or reapplied and are not accompanied by a first-order phase It appears that the isotopic method change in olivine or spinel. These observations relate to fundamental concepts of for determining the origin of Greek electrical conductivity and photon absorption in complex transition-metal silicates marbles will probably be the most use- in that they cannot be readily interpreted in terms of an intrinsic band-gap model. ful of the tests so far devised, especially The intensity and energy changes are too great and the effect occurs at too low a if used in conjunction with other tech- pressure to be explained by processes such as spin-pairing and other crystal-field niques. Some of the archeological sam- effects. The results suggest that a new mechanism of conduction, perhaps symbiotic ples in Table 1 fall outside of the iso- and employing an efficient charge-transfer process, is induced at high pressure. topic provenances so far delineated, and it is clear that other quarrying areas Measurements of the optical and elec- 1 atm (4, 5) dealt mostly with crystals were used. Thus, coincidence of a sam- trical properties of fayalite (olivine, or- having magnesium-rich compositions in ple with one of the isotopic clusters in thorhombic) and spinel (cubic) forms the olivine series. High-resolution polar- Fig. 1 is not, in itself, a unique indica- of Fe2SiO4 have been made at pressures ized spectra for iron-rich olivines at 1 tion of provenance. We hope to estab- in the range from 45 to 300 kb. The atm have been reported by Burns (6) lish a variety of geochemical character- experimental apparatus is a modifica- and Mao and Bell (3). Several investi- istics of these and other marbles. We tion of the design employed in an ear- gators (7) have measured olivine spec- will welcome samples from areas in lier study (1, 2) in which the effects de- tra at high pressure, but curiously, as which other marbles used by the an- scribed herein were first discovered. in the investigations of electrical con- cient Greeks were quarried, in order Optical absorption, electrical conduc- ductivity mentioned below, they did to extend the isotopic method to as tivity, and x-ray diffraction are mea- not explore pressures high enough in many localities as possible (16). sured simultaneously with a diamond- combination with sufficiently iron-rich HARMON CRAIG windowed cell in single crystals and olivine to observe the major effects. VALERIE CRAIG polycrystalline powders under pressure Properties of both natural (Inyo Scripps Institution of Oceanography, (3). County, California) and pure synthetic University of California at San Diego, Earlier definitive studies of the opti- fayalite (Fe2SiO4) were measured. T,he La Jolla 92037 cal absorption properties of olivine at synthetic crystals of olivine were pre- 28 APRIL 1972 403

This content downloaded from 147.52.249.113 on Tue, 8 Apr 2014 11:45:54 AM All use subject to JSTOR Terms and Conditions