Mons Claudianus and the problem of the 'granito del foro': a geological and geochemical approach.

Mons Claudianus and the problem of the 'granito del foro': a geological and geochemical approach. Granito del foro is a distinctive igneous rock igneous rock:see rock. igneous rockAny of various crystalline or glassy, noncrystalline rocks formed by the cooling and solidification of molten earth material (magma). , in fact agranodiorite granodioriteMedium- to coarse-grained rock that is one of the most abundant intrusive rocks. It contains quartz and is distinguished from granite by having more plagioclase feldspar than orthoclase feldspar; its other mineral constituents include hornblende, biotite, and rather than a granite, long known and named for its use inbuildings of the Roman Forum. Exactly what is it? Where does it comefrom? Where else was it used and not used? What does the granito delforo say about ownership and empire?IntroductionMons Claudianus in the Eastern Desert of Egypt was an importantsource of Roman monumental stone. Since 1987 an international teamworking under the aegis of the Institut Francais d'ArcheologieOrientale has investigated the site with extensive excavations and alsoa detailed survey, which has included mapping the quarry field (Peacock1993). To place this new work in its wider context, it is important tostudy the distribution and use of the Mons Claudianus rock, previouslyattempted by Kraus & Roder (1962) and more recently by Dodge (1988),Lazzarini (1987; 1992) and Galetti et al. (1992).While the rock of the Mons Claudianus quarries has often beenerroneously referred to as a granite, it has been classified variouslyas granodiorite (Lazzarini 1987), quartz diorite Quartz diorite is an igneous, plutonic (intrusive) rock, of felsic composition, with phaneritic texture. Feldspar is present as plagioclase (typically oligoclase or andesine) with 10% or less potassium feldspar. Quartz is present at between 5 to 20% of the rock. (Troger in Kraus &Roder 1962: 119), granodiorite/quartz diorite dioriteMedium- to coarse-grained igneous rock that commonly is composed of about two-thirds plagioclase feldspar and one-third dark-coloured minerals, such as hornblende or biotite. (Klemm & Klemm 1981;1992), tonalite Tonalite is an igneous, plutonic (intrusive) rock, of felsic composition, with phaneritic texture. Feldspar is present as plagioclase (typically oligoclase or andesine) with 10% or less alkali feldspar. Quartz is present as more than 20% of the rock. (Harrell 1989) and granodiorite/tonalite (Sabet et al.1972; Greenberg 1981). New chemical analyses carried out for the presentpaper show the Mons Claudianus rock to be granodiorite following theclassification of MacKenzie et al. (1982: 77), and this term will beused here.To architectural historians, it has long been known as the granitodel foro, because of its frequent occurrence in the Roman Forum. Thegranito del foro has been regarded as one of the most distinctive andeasily recognized of the Roman ornamental stones; in theory it should bea simple matter to plot its distribution. However, thin-sectionexamination of some typical specimens revealed an interesting point.While all were closely similar in mineralogy, some had markedly zonedplagioclase plagioclaseAny member of the series of abundant feldspar minerals that usually occur as light- to medium-grey-coloured, transparent to translucent grains or crystals. Plagioclase ranges in composition from albite to anorthite. feldspars. Samples are available from all of the 130quarries comprising the Mons Claudianus complex, and in these zoning isabsent or weakly developed. Since zoning results from compositionalchanges in the parent magma when the rock was forming (and not, forexample, from recent chemical alteration or weathering), it seemsinescapable that two or more rock sources have been conflated under onename. The object of this paper is to explore ways in which grey plutonicrocks of similar appearance can be distinguished, and to examine theimplications for Mons Claudianus.Until recently, the scientific characterization of plutonic rocksused in antiquity has lagged far behind that of coloured or whitemarbles, but recent papers by Lazzarini (1987; 1992) and Galetti et al.(1992) have made significant advances. The latter, in particular, haveattempted to define the chemical and petrographic pe��trog��ra��phy?n.The description and classification of rocks.pe��trogra��pher n. traits of samples fromall the major quarry sites. Our approach to the problem is different andcomplementary. We have concentrated on grey plutonic rocks of similarappearance, from a wide range of localities, in an attempt to finddistinguishing criteria and to define sources. In contrast to Galetti etal. our approach included a wide range of promising outcrops, regardlessof evidence for ancient exploitation. Our study includes visualcomparison of rock samples with archaeological artefacts, complemented,where possible, with petrographic and chemical work. Fortunately thelatter confirm the former, enabling us to use visual data in assessingthe importance of the Mons Claudianus rock, although distinction betweenother sources is rarely possible without further analysis.The grey plutonic rocks of the MediterraneanFIGURE 1 shows the distribution of grey acid and intermediateplutonic rocks in the Mediterranean region; in other words Adv. 1. in other words - otherwise stated; "in other words, we are broke"put differently potentialsources of material that might be confused with that of Mons Claudianusand hence be classed as granito del foro. The outcrops are classed underthree headings:1 those examined and included in this study;2 those unlikely to be of relevance for reasons given below; and3 those which we have been unable to examine.With the exception of Mons Claudianus, all the sources considered arein coastal locations or with river access to the sea. We have omittedoutcrops away from the Mediterranean coast, such as those in centralGaul, in central and western Iberia, in Cappadocia or in the near East(e.g. Jordan, Saudi Arabia Saudi Arabia(sä`dē ərā`bēə, sou`–, sô–), officially Kingdom of Saudi Arabia, kingdom (2005 est. pop. etc.) as it was necessary$to limit thefield-work within reasonable bounds, and at present these more remotesources seem unlikely to be relevant.In Egypt, only one outcrop was considered, the granodiorite of MonsClaudianus. Diorites were exploited at Semna, Barud, Umm Shegilat andUmm Balad in the Eastern Desert as well as a black or grey diorite orgranodiorite near Aswan (Gnoli 1971; Lazzarini 1987). However, those allhave a distinctive colour or texture which clearly differentiates themfrom the granito del foro. The granite of Wadi Fawakhir is somewhatsimilar in appearance to the Mons Claudianus granodiorite, but thepinkish feldspars serve to discriminate it unambiguously.Plutonic rocks are common in Turkey, but only those in westernTurkey, near the sea, were considered. Grey plutonic rocks are to befound comprising Kozak Dag (the granito misio of Lazzarini 1992),Eybekdag near Edremit, and in the area around Evciler. Outcrops in theTroad near Kestambol and Kocali were extensively quarried in the Romanperiod (Cook 1973; Lazzarini 1987), and the presence of purple feldspar feldspar(fĕl`spär, fĕld`–)or felspar(fĕl`spär), an abundant group of rock-forming minerals which constitute 60% of the earth's crust. phenocrysts makes this granito violetto (in its most typical form) easyto recognize. However, in the Troad there are outcrops of anon-porphyritic variety, used in modern building. There is no evidenceof ancient exploitation of the non-porphyritic type, but we haveincluded some samples from modern building debris in the Troad, togetherwith the more porphyritic por��phy��rit��ic? also por��phy��rit��i��caladj.1. Of or containing porphyry.2. Containing relatively large isolated crystals in a mass of fine texture.Adj. 1. variety. Roman columns of a very pale violettotype were seen in Cyprus in 1992, and work is continuing on these todetermine whether or not they also originate within the Troad.There are also outcrops of grey plutonics around the Sea of Marmara.Samples were obtained from Kapidag and Uludag, both of which support amodern stone industry, and from near Armutlu. The outcrops to the eastof Lapseki and near Karabiga comprise mainly crumbling, very alteredmaterial; it is unlikely that they would have been exploitable. Thesource near Balcik can be eliminated for similar reasons, as well as itspinkish colour and very coarse texture. A further outcrop, in EuropeanTurkey near Demirkoy, is again much degraded but does contain a few hardexploitable outcrops.Plutonic rocks outcrop in Macedonia and Thrace. Those in Macedoniaare again very altered, but workable material is to be found to the eastof Xanthi, although the rock has a pink tinge which usually serves todistinguish it from the granito del foro. A sample of purple-feldspargranodiorite was collected near Sandanski in Bulgaria.Grey 'granitic' rocks are also found in limited outcrop inCalabria in southern Italy. They are extensively altered and rarelyusable, but the existence of an ancient quarry at Nicotera, apparentlywith a local distribution, shows that a workable facies facies/fa��ci��es/ (fa��she-ez) pl. fa��cies ? [L.]1. the face.2. surface; the outer aspect of a body part or organ.3. expression (1). does exist(Solano 1985). Further north, grey biotite biotite(bī`ətīt'), iron-rich variety of phlogopite, most abdunant of the mica minerals. biotiteor black micaSilicate mineral in the common mica group. granites comprise the islandsof Giglio and Monte Cristo Monte Cristo(mŏn'tē krĭs`tō, Ital. môn`tā krē`stō), unpopulated, rocky island, 6 sq mi (15.5 sq km), belonging to Italy, in the Tyrrhenian Sea between Corsica and the Italian coast. as well as the massif mas��sif?n.1. A large mountain mass or compact group of connected mountains forming an independent portion of a range.2. of Monte Capanne Monte Capanne is the highest mountain of the italian island of Elba. onElba. Giglio has long been considered as a source of Roman decorativestone Geographical list of stone used for decorative purposes in construction and sculpture; currently or historically produced in various countries. *For generic petrological rock types, see: List of rock types BelgiumLimestone , and evidence for quarrying has been claimed along the easterncoast (Lazzarini 1992). Unfinished columns are to be seen in the areaaround Cavoli and Vallebuia on the southern slopes of Monte Capanne, butit is uncertain whether they are Roman or Pisan. The unfinished fountainknown as 'La Nave' is certainly post-Roman (Waldeck 1987:141-5).Roman exploitation of granites which outcrop extensively in Sardiniaand Corsica has recently been discussed by Wilson (1988) and byMassimetti (1991), but we have little information on the distribution ofthe grey facies. The column quarry at Capo Testa on the north shores ofSardinia is well known, and its products easily distinguished from MonsClaudianus and other grey granitic rocks by the distinctive pinkfeldspars. However, Wilson (1988) suggests that ahornblende-biotite-granodiorite was exploited on the island of SanBainzo off Corsica in the Straits of Bonifacio, and he has kindly madeavailable samples for this study.The plutonics of the French Massif Central Massif Central(mäsēf` säNträl`)[Fr.,=central highlands], great mountainous plateau, c.33,000 sq mi (85,470 sq km), S central France, covering almost a sixth of the surface of the country. , unlikely to be relevantbecause of their inland location, have not been included, but graniticrocks also occur in the Maures region of southern France. The rock isgrey, but strongly porphyritic with large (over 1 cm) potassium feldspar potassium feldsparA type of alkali feldspar that contains the molecule KAlSi3O8. Microcline and orthoclase are types of potassium feldspar. crystals. Since similarly porphyritic granites are found in Elba (Poli1992), a sample of Maures rock was included.Granites abound in Iberia, mostly in western Spain and Portugal.Further plutons are found in Catalonia, but the majority are soft andunworkable. Hard granitoid rocks occur in the Pyrenees, but had theybeen exploited to any significant extent we would expect to findgranitic material on sites such as Ampurias, which is not the case.Canto can��to?n. pl. can��tosOne of the principal divisions of a long poem.[Italian, from Latin cantus, song; see canticle. (1978), surveying the sources of decorative stone in Spain, drewattention to granite exploitation in the Basque country Basque Country(băsk, bäsk), Basque Euzkadi, Span. País Vasco, comprising the provinces of Álava, Guipúzcoa, and Vizcaya (1990 pop. , Galicia andAndalucia, all of which seem to be of local importance.In this paper, for the reasons given above, we consider material fromthe following outcrops: Mons Claudianus, Kozak Dag, Eybekdag (nearEdremit), Evciler, the Troad area, Kapidag, Uludag, Armutlu, Karabiga,Demirkoy, Xanthi, Giglio, Elba, San Bainzo, Bulgaria and Maures. Most ofthese rocks fall chemically within the monzonite-granodiorite range, butthe rocks of Elba, Giglio, Uludag and Armutlu, while remarkably similarin appearance, are granites. Multiple samples (2-7 samples) werecollected from most sources, particularly those of large extent such asthe Kozak Dag, chosen to include all variations of mineralogy/textureseen in the field. One hundred and thirty samples were collected fromMons Claudianus, one from each quarry site. Single samples only wereavailable from Maures, near Sandanski (Bulgaria) and Xanthi.Visual analysisSmall pieces of typical samples from the above outcrops, about 10 x20 mm, were polished and mounted on a wooden stick for easytransportation to archaeological sites, so that direct comparison couldbe made with artefacts. This simple but effective approach provides avaluable first step in studying these materials, which are spread overwide distances and through many present-day countries, with differentattitudes to sampling for scientific analysis. In the past, visualidentifications have been attempted by both archaeologists andgeologists claiming to 'know' a particular rock or rocks, andcarrying information about colour and texture 'in the head'.This is both difficult and fraught with error, because both theseparameters are difficult to judge in isolation.Comparison of samples from different outcrops was in itselfinteresting. The granodiorite of Mons Claudianus proved to be visuallyremarkably homogeneous throughout the 130 quarries, but appeared quitedifferent from any of the other outcrops in this study, except that ofSan Bainzo off Corsica (a source which may not have been exploited to agreat extent; chemical analysis, below, assigned no artefacts to thissource). The granodiorite of Mons Claudianus is characteristicallymedium- to coarse-grained, with marked clustering of leucocratic andmafic minerals and a distinct foliation foliationPlanar arrangement of structural or textural features in any rock type, but particularly that resulting from the alignment of constituent mineral grains of a metamorphic rock along straight or wavy planes. . Given these features, it may bedifficult to see how it could be misidentified, but in cross-sectionscut parallel to the foliation of the rock, the sole distinguishingfeature would be the grain size and distribution, difficult to identifyin the field without comparative outcrop samples to hand.There is considerable visual ambiguity between the other sourcematerials Noun 1. source materials - publications from which information is obtainedsource - a document (or organization) from which information is obtained; "the reporter had two sources for the story" studied. The granite from Giglio is closely similar to thatfrom Uludag; both are similarly siliceous siliceousrelating to or made of silica or a silicate. biotite-bearing granites. Moresurprising is the close visual similarity between the granodiorite ofKozak Dag and some non-porphyritic granites of Elba; small differencesof grain size which might serve to differentiate the rocks could easilybe missed in field examination.In sum, visual analysis is of significant value in distinguishing thegranodiorite of Mons Claudianus from its most common analogues, but isof more limited value in differentiating the products of the otheroutcrops. The results of visual comparison of artefacts are thereforestated in terms of a simple 'no' or 'yes'identification with Mons Claudianus. In the comments column we notecorrections to previous identifications. References to Troad rock implytypical 'granito violetto' with characteristic purple feldsparphenocrysts, and the designation 'grey', unless otherwisespecified, includes artefacts which are suspected, but not yet proven,as being from the Kozak Dag (microprobe microprobe/mi��cro��probe/ (mi��kro-prob?) a minute probe, as one used in microsurgery. microprobea minute probe, such as one used in microsurgery. analysis, below, has confirmed asmall number of source assignments to Kozak). An attempt is made to givean impression of the relative importance of the rock types in terms ofnumbers of columns/artefacts, but quantitative data are patchy. Wherenumbers are quoted, they should be used with caution because countingcolumn fragments is unsatisfactory. A better way might be to sum thelengths of fragments, but it was not attempted in this preliminarysurvey. It is important to note the confirmation of the selected visualassignments to Mons Claudianus by the chemical work described below.Sites which were included in that work are indicated on TABLE 1 by anasterisk. The table does not include examples of espolio in Rome.Petrography pe��trog��ra��phy?n.The description and classification of rocks.pe��trogra��pher n. Representative samples from outcrops were selected for petrographicstudy in thin section. Results are summarized in TABLE 2 where sourcesamples are listed next to archaeological samples which were assigned tothat source by chemical analysis.Petrographic features can aid in source characterization. Samplesfrom Mons Claudianus have little zoning within the plagioclasefeldspars. Thin-section examination of material from other sourcesconfirms that many have indeed much more prevalent zoning than MonsClaudianus rock. The bluish-green colour of the amphiboles is a furthernoticeable feature of Mons Claudianus, but it is shared by rocks fromKapidag, Eybekdag (near Edremit), and, interestingly, the visuallysimilar San Bainzo source. Common accessory minerals including apatite apatite(ăp`ətīt), mineral, a phosphate of calcium containing chlorine or fluorine, or both, that is transparent to opaque in shades of green, brown, yellow, white, red, and purple. ,zircon zirconSilicate mineral, zirconium silicate, ZrSiO4, the principal source of zirconium. Zircon is widespread as an accessory mineral in acid igneous rocks; it also occurs in metamorphic rocks and, fairly often, in detrital deposits. and sphene sphene?n.A titanium accessory mineral in some granite and metamorphic rocks, CaTiSiO5, occurring in usually small brown or yellow crystals and sometimes used as a gemstone. Also called titanite. do not provide clear source discrimination. Lazzarini(1992) noted the prevalence of epidote epidoteAny of a group of colourless to green or yellow-green silicate minerals with the general chemical formula A2B3(SiO4)(Si2O7)O(OH), in which A is usually calcium (Ca) and B in Mons Claudianus rocks (thoughit is present in other sources), and commented on the distinctivereddish colour of the biotites in Italian (Elba and Giglio) granite, afeature noted also in our samples.Mineralogy observed in thin section may therefore give some clues toartefact See artifact. source but is unlikely to be completely diagnostic. Thinsections of a medium-coarse grained granite/granodiorite require asample ideally of 3 x 2 cm, in order to obtain representativemineralogy, amounts which may be difficult to obtain from finishedartefacts. It is noteworthy that archaeological samples listed in TABLE2, while generally showing excellent matches with proposed sources, dodisplay some differences. For example, C3 lacks the epidote generallycharacteristic TABULAR DATA OMITTED TABULAR DATA OMITTED of MonsClaudianus samples. This is almost certainly due to the small size ofthe sample of C3 examined.Chemical analysisSince many of the granitoid sources studied are, as noted above,visually and mineralogically similar, chemical analysis was used forsource characterization and provenancing of archaeological samples fromcolumns. This complemented the visual work on Mons Claudianus columns,and also provided confirmation of several assignments to that source.Two types of analysis were used: X-ray fluorescence and electronmicroprobe The electron microprobe is an analytical tool used to non-destructively determine the chemical composition of small volumes of solid materials. It uses a high-energy focused beam of electrons to generate X-rays characteristic of the elements present within a sample volumes 1 to 3 . Only one archaeological sample (from Ephesus) was largeenough for X-ray fluorescence analysis, so the greater part of theprovenancing was done using microprobe analysis.X-ray fluorescenceWhole rock chemical analysis by X-ray fluorescence (XRF XRF X-Ray FluorescenceXRF X-Ray FlashXRF Cross ReferenceXRF Extended Recovery Facility (IBM)XRF Extended Reliability FeatureXRF Cross Reference FileXRF External Reference ) was carriedout on 34 geological samples in order to confirm rock types andestablish compositional variation within the sources studied. It wasalso considered useful to make a preliminary assessment on the basis ofthese data of the potential of XRF for provenancing purposes.MethodThirty-four geological samples from 16 source areas, and onearchaeological sample from Ephesus, were analysed at the Open Universityusing energy-dispersive XRF (for trace elements Trace elementsA group of elements that are present in the human body in very small amounts but are nonetheless important to good health. They include chromium, copper, cobalt, iodine, iron, selenium, and zinc. Trace elements are also called micronutrients. , 25 samples), andwavelength-dispersive XRF (for major elements, 29 samples). Samples werecleaned to remove any weathered surfaces and crushed in agate forpreparation of powder pellets (c. 8 g of sample needed) and fused glass Fused glass is a term used to describe glass that has been fired (heat-processed) in a kiln at a range of high temperatures from 593o C (1100oF) to 816o C (1500oF). There 3 main distinctions for temperature application and the resulting effect on the glass. discs (c. 1 g of sample needed). The analysis followed the methods andprinciples described in Potts et al. (1984) and Potts & Webb (1992),using Link Systems Ltd Meca 10-44 and ARL ARL - ASSET Reuse Library 8420 spectrometers andassociated software. For trace elements, precision at 2[Sigma] istypically 2-8% relative, rising to 15-17% for Cu and Ga, and varyingfrom 1-30% for Th depending on abundance. Ni precision, not measured byPotts et al. (1984), is estimated at up to or better than 10% relative1[Sigma]. Accuracy is close to or within precision. For major elementsprecision and accuracy are both typically batter than 3% relative1[Sigma].Results and source discriminationRepresentative XRF analyses are given in TABLE 3 (full data areavailable from one of the authors (OWT OWT Omega World Travel (travel agency)OWT Overweight (finance)OWT One World Trust (London, UK)OWT Old Wives TaleOWT Operation Warrior Trainer ) on request). FIGURE 3, of Yplotted against Zr for the geological and one archaeological samples,demonstrates clear discrimination between most sources. Variationbetween samples from the same locality exceeds precision for someelements including Zr, implying some variation within source. Furtherdistinction between Elba, Kozak and Edremit, close together on FIGURE 3,is provided by Sr data (212, 221 in Elba; 646, 653 in Kozak; 856,877 inEdremit; all in parts per million parts per millionmg/kg or ml/l; see ppm. (ppm); cf. TABLE 3). Thearchaeological sample from Ephesus lies close to the Kozak source(within 2 and 4 standard deviations) on FIGURE 3, and most otherelements analysed in this sample are within or close to precision ofKozak. Differences in Pb (Ephesus 85 ppm, Kozak 31-33 ppm) may be due tocontamination of the column sample from metal fittings.The XRF data, a first test of the method, show good potential forprovenancing. It is expected that further analyses of geologicalmaterial would increase analytical spread. However, the wide range ofrock types and composition represented within the sources studiedimplies feasibility for artefact provenancing. The archaeological samplefrom Ephesus, whose source at Kozak is confirmed below by micro-probeanalysis, was clearly weathered and weighed only a few tens of grams(extremely small in terms of whole rock characterization of granitoids).Considering these factors, the agreement with Kozak data is extremelygood.The main difficulty with XRF analysis of artefacts is the relativelylarge amounts of material required. While XRF can be done on samples ofunder 10 g, such amounts are unlikely to be representative samples ofcoarse granitoid rocks. This major impediment to whole rock analysisprompted the further analysis by electron microprobe, using very smallsamples, described in the next section.It is noteworthy that the Kozak sample analysed by Galetti et al.(1992: table 4, marmo misio, a sample from a 'small ancient quarrynear Kozak') is not of the same rock as our own Kozak samples, butrepresents a different rock type with higher Si[O.sub.2]. This Misiosource sample has not yet been shown to correlate with archaeologicalartefacts.Electron microprobe analysisThe electron microprobe permits analysis of sample areas only a few[[micro]meter] in diameter, allowing study of small parts of individualcrystals. In this work a sample area of about 15 [[micro]meter] wasselected, to avoid decomposition of the sample which might affectpotassium and sodium analyses. In a fresh unaltered rock it is possiblesuccessfully to analyse samples composed of, for example, only two orthree biotite mica and/or amphibole amphibole(ăm`fəbōl'), any of a group of widely distributed rock-forming minerals, magnesium-iron silicates, often with traces of calcium, aluminum, sodium, titanium, and other elements. crystals. In practice a rock orartefact size of about 5 x 3 x 1 mm is generally needed to ensure thatenough crystals are present to give representative analysis. As we showlater, plagioclase crystals are not good discriminators of source, andsamples composed primarily of this mineral will not be suitable.Microprobe analysis has the further advantage that, for granites andrelated rocks, the composition and characteristics of individualminerals reflect closely the composition of the granite magma and thepressure--temperature conditions under which the intrusion formed, thusgiving the best chance of relating composition to locality, and definingunique discriminating features of each source.MethodAnalyses of minerals described here were carried out at the OpenUniversity using a Cambridge Instruments Microscan 9 electronmicroprobe, following the method described in Potts et al. (1983).Precision is 1-3% relative (1[Sigma]) for elements present in mineralsat an abundance of 1% or greater, and typically 5-15% for elements below1% (figures based on 54, 12, and 11 repeat analyses of, respectively,glass, plagioclase and pyroxene pyroxene(pī`rŏksēn), name given to members of a group of widely distributed rock minerals called metasilicates in which magnesium, iron, and calcium, often with aluminum, sodium, lithium, manganese, or zinc occur as X in the chemical reference materials; Potts et al. 1983:1240). Accuracy, determined by comparison of repeat analyses ofreference materials with published data, is generally within 2[Sigma]precision limits for elements present at abundances of 1% or greater,except for [Al.sub.2][O.sub.3] and [Na.sub.2]O (1-4[Sigma], and4-6[Sigma] respectively) (Potts et al. 1983: 1239 & table 2).Analysis of geological source materialTwenty-four samples from 18 localities were selected asrepresentative. One hundred and fifty microprobe analyses were carriedout, comprising 46 analyses of plagioclase feldspar plagioclase feldspar(plā`jēəklās'): see feldspar. , 48 of amphibole(hornblende hornblende:see amphibole. hornblendeAny of a subgroup of amphibole minerals that are calcium-iron-magnesium-rich and monoclinic in crystal structure. ), 54 of biotite and 2 of tourmaline tourmaline(tr`məlĭn, –lēn), complex borosilicate mineral with varying amounts of aluminum, iron, magnesium, sodium, lithium, potassium, and sometimes other elements, . Tourmaline was found ininsufficient samples to be useful for source discrimination. Theanalysed samples, chosen to TABULAR DATA OMITTED be visually andpetrographically representative of the sources, included multiplesamples where available and for sources which we suspected would be ofmajor importance, for example those with known Roman quarries.The plagioclase analyses were performed on 21 different crystals in 9samples from 7 source localities in Turkey and Egypt. The analysedsamples are characterized by varied combinations of normal/oscillatoryzoning which are reflected in the microprobe analyses. Chemical analysesof the plagioclases show overall zonation zo��na��tion?n.1. Arrangement or formation in zones; zonate structure.2. Ecology The distribution of organisms in biogeographic zones. (including core-rim analysesof individual crystals) ranging from An 18%-An 58%. The Mons Claudianussamples are characterized by less complex oscillatory oscillatorycharacterized by oscillation.oscillatory nystagmussee pendular nystagmus. zoning and weakercompositional zoning than many of the Turkish sources analysed (cf.petrography, above). As variation within plagioclase crystals was oftengreater than variation between sources, reducing its effectiveness as adiscriminant dis��crim��i��nant?n.An expression used to distinguish or separate other expressions in a quantity or equation. , remaining sources were not analysed for this mineral.The biotite analyses were done on 53 different crystals in 24 samplesfrom 18 source localities in Turkey, Egypt, Italy, Corsica, France andGreece. Compositional zoning in biotites was not significant, andanalysis of core and rim of one biotite in sample EDR (Enhanced Data Rate) An acronym occasionally used to reflect an advancement in transmission or transfer speed. For example, Bluetooth 2.0 + EDR provides a dramatic increase in speed over the previous version. See Bluetooth. 2 showeddifferences within precision, of 0.2-2.4% relative except for[Na.sub.2]O for which the difference was 29% relative at an abundance of0.07%. Repeatability of analysis of two different crystals in the samesample was good, with differences typically of 5% relative (within ornear precision) for most elements. Ti[O.sub.2] was exceptional inshowing higher differences in several samples, of 18-36% relative; othersamples showed much better Ti[O.sub.2] repeatability, of between 5 and10% relative. Homogeneity within, and differences between, sources isdiscussed below in relation to the discrimination graphs.The amphibole analyses were done on 44 different crystals in 18samples from 12 source localities in Turkey, Egypt, Corsica and Greece.No amphiboles were present in the samples from Italy (Elba, Giglio),France (Maures Massift and from Uludag (Turkey). Analysis of cores andrims of three amphibole crystals showed slightly higher differences thanfor biotite, of 0.45-6.7% relative (for sample KOZ3). Repeat-ability ofanalysis of two different crystals in the same sample was generallywithin or near precision, between about 5-10% relative. Exceptional tothese figures are amphibole crystals which have formed around relict RELICT. A widow; as A B, relict of C D. pyroxene cores. Such crystals, not always identifiable by microscopicexamination of the probe slides, may be unintentionally analysed inaddition to or instead of 'true' (primary) amphiboles.Analytically, this is apparent because such crystals are significantly(and variably) depleted in Ti[O.sub.2], [Al.sub.2][O.sub.3], FeO,[Na.sub.2]O and [K.sub.2]O, and to a lesser extent in CaO, and enrichedin MnO, MgO and to a lesser extent in Si[O.sub.2]. Out of 44 amphibolecrystals analysed in geological samples, it is apparent that 6 areamphiboles mantling relict pyroxene cores. Since such crystals couldeasily form part or all of an analysis of a very small archaeologicalsample in which choice of crystals to analyse may be very limited, it isimportant that these amphiboles are recognized and defined as part ofthe source where relevant. They may occur in all sources considered; sofar, they are apparent in samples from the Troad, Edremit, Karabiga andKozak Dag, and also in thin sections of samples from Xanthi, Kapidag,and Evciler. Analysis NTD NTDNeural tube defect, see there , within a Kozak Dag sample, provides anexample of a microprobe analysis strongly affected by a relict pyroxene.Representative examples of biotite and amphibole analyses in sourcesare given in TABLE 4; full data are available on request from one of theauthors (OWT).Discrimination of geological sourcesAnalyses of amphibole and biotite crystals were used in sourcediscrimination. Since some archaeological and geological samples wereanalysed for amphibole or biotite only, discrimination by both mineralswas important. Within the amphiboles [K.sub.2]O and MnO and MgO and FeOwere selected as the most effective discriminants. Almost all sourcesstudied are separated by a plot of [K.sub.2]O against MnO; onlyDemirkoy, north of the Sea of Marmara is overlapping in composition withthe field of non-porphyritic Troad rocks. Samples from Asagicuma areincluded with the Kozak group because the village of Asagicuma lieswithin the Kozak Dag. The Mons Claudianus granodiorite is well separatedfrom the Turkish sources, and from Xanthi in Greece and San Bainzo inCorsica. Italian granites are not included in this diagram as amphiboleswere not found in these samples. Variation within source is [+ or-]1-2[Sigma]. Point NTD, within a sample from the Kozak Dag, shows thelower [K.sub.2]O and higher MnO typical of an amphibole formed around arelict pyroxene core (cf. above). FIGURE 5, of MgO plotted against FeO,enhances the discrimination between the rocks of Kozak and thenon-porphyritic material from the Troad, though other sources overlap onthis diagram. NTD has higher MgO and lower FeO than other Kozak samples,as expected. Variation within source is again within [+ or -]1-2[Sigma].Ti[O.sub.2] in amphiboles may also be used to discriminate between theTroad and Kozak, falling between 0.75-1.08 % for non-porphyritic Troadrocks, and between 1-15% and 1-45% for Kozak in primary amphiboles(those not formed on relict pyroxene cores). MOC2 KOZ2 KOZ3(*) ASA4 EDR2 EVC2Si[O.sub.2]% 43.76 46.27 50.27 46.35 45.33 46.96Ti[O.sub.2] 1.32 1.30 0.55 1.28 1.25 1.13[Al.sub.2][O.sub.3] 9.12 7.59 4.33 7.52 8.14 6.84FeO 17.69 14.26 12.55 14.33 16.45 15.03MnO 0.45 0.47 0.54 0.42 0.56 0.65MgO 10.37 12.93 15.00 13.02 11.65 12.87CaO 11.16 11.64 11.50 11.70 10.58 11.88[Na.sub.2]O 1.27 1.42 0.88 1.50 1.26 0.98[K.sub.2]O 1.03 0.85 0.36 0.82 0.85 0.69total 96.17 96.73 95.98 96.94 96.07 97.03 TROAD DEM1 BAIN1a XAN1 KAP2 non-porph.Si[O.sub.2]% 46.72 48.51 43.75 47.78 44.81Ti[O.sub.2] 1.08 1.12 1.35 0.87 1.03[Al.sub.2][O.sub.3] 6.54 5.94 8.85 5.37 8.35FeO 13.85 12.98 20.56 18.54 19.88MnO 0.48 0.50 0.85 1.31 0.94MgO 13.63 14.34 8.38 11.26 9.26CaO 12.01 11.74 11.00 9.88 10.82[Na.sub.2]O 1.43 1.34 1.48 1.45 1.54[K.sub.2]O 0.72 0.64 0.91 0.43 0.67total 96.46 97.11 97.13 96.89 97.30Sample localities are: MOC Mons Claudianus, KOZ Kozak Dag, ASA Asagicuma, EDREybekdag, Edremit, EVC Evciler, DEM Demirkoy, BAIN San Bainzo, XAN Xanthi, KAPKapidag.Troad non-porph. is non-porphyritic Troad material.* KOZ3 is point NTD (amphibole with composition affected by a relict pyroxenecore).Totals do not add up to 100% because of the presence of water in amphibole.[H.sub.2]O cannot be determined by the electron microprobe technique.TABLE 4(a). Microprobe analyses of geological samples -- amphiboles.FIGURE 6, showing Ti[O.sub.2] plotted against [Al.sub.2][O.sub.3] inbiotites, includes the Italian sources of Elba and Giglio and the MauresMassif in France, clearly separated from each other and from theTurkish, Egyptian, Greek and Corsican localities. The Turkish source ofUludag, in which amphiboles were not seen, is also separated on thisgraph. Separation of these sources is confirmed on FIGURE 7, showingTi[O.sub.2] plotted against MnO in biotites. The spread of analyseswithin one source is greater in relation to precision than foramphiboles, at between [+ or -]1-3[Sigma] for [Al.sub.2][O.sub.3] and upto about [+ or -]2[Sigma] for MnO. Variation of Ti[O.sub.2] withinsource is up to c. 12 times precision at 10 ([+ or -]6[Sigma]),reflecting both the high level of precision for this element in biotite,and its more variable behaviour. As Ti[O.sub.2] can be subject toalteration leading to formation of rutile rutile,mineral, one of three forms of titanium dioxide (TiO2; see titanium). It occurs in crystals, often in twins or rosettes, and is typically brownish red, although there are black varieties. and heterogeneous distributionof Ti, data require careful monitoring. Despite this, Ti[O.sub.2]remains one of the most effective discriminating elements for biotites. MOC99 KOZ2 ASA2 ULU2 KARAB1Si[O.sub.2]% 36.32 36.88 37.25 35.50 36.23Ti[O.sub.2] 2.89 4.20 4.20 3.27 4.79[Al.sub.2][O.sub.3] 15.27 13.77 13.34 17.25 13.84FeO 18.70 17.75 16.98 20.61 18.05MnO 0.28 0.37 0.32 0.49 0.43MgO 11.36 13.04 13.35 8.37 11.98CaO 0.01 0.01 0.01 0.01 n.d.[Na.sub.2]O 0.11 0.18 0.14 0.05 0.09[K.sub.2]O 9.51 8.80 9.15 9.70 9.52total 94.45 95.00 94.74 95.25 94.93 ARM1 TROAD CAV1 VALL1 GIG2 porph.Si[O.sub.2]% 36.97 36.95 36.38 36.37 35.35Ti[O.sub.2] 4.21 3.74 3.16 2.90 3.66[Al.sub.2]O.sub.3] 13.89 14.33 16.23 16.17 18.26FeO 17.86 17.66 19.18 17.51 21.15MnO 0.58 0.65 0.37 0.31 0.29MgO 12.47 12.45 10.03 11.19 7.90CaO 0.01 0.01 0.01 0.01 0.01[Na.sub.2]O 0.05 0.11 0.09 0.11 0.15[K.sub.2]O 9.48 9.28 9.28 9.03 9.41total 95.52 95.18 94.73 93.60 96.18Sample localities are: MOC Mons Claudianus, KOZ Kozak Dag, ASA Asagicuma, ULUUludag, KAR Karabiga, ARM Armutlu, CAV Cavoli (Elba), VALL Vallebuia (Elba),GIG Giglio Castello.Troad porph. is porphyritic Troad material.n.d. is not detected.TABLE 4(b). Microprobe analyses of geological samples -- biotites.Porphyritic and non-porphyritic samples from the Troad may bedistinguished from each other, though they have some importantsimilarities in whole rock analysis, as determined by XRF; both haveunusually high U, Th and Sr, reflecting their origin in the same generalarea.Provenancing of archaeological samples to sourcesThirty four analyses were carried out on 11 archaeological samplesfrom 8 sites, including 17 analyses of amphiboles and 17 of biotites;representative analyses are given in TABLE 5. Analyses were comparedmanually with all source analyses and are plotted on FIGURES 4-7. Amatch for unaltered elements with a source group of within 1-2[Sigma]was taken to indicate a likely provenance at that source. The samplesare from columns at 8 sites in Italy, north Africa, France and Turkey,and are listed with geological provenances in TABLE 6. For samples fromItaly, we re-examined thin sections obtained from monuments by ProfessorE. Troger in 1962 (in Kraus & Roder 1962: 117-20).FIGURE 4 shows that samples from Hadrian's Villa Hadrian's VillaHadrian's country residence, built (c. AD 125–34) at Tivoli near Rome. A sumptuous imperial complex with parks and gardens on a grand scale, it included baths, libraries, sculpture gardens, theaters, alfresco dining areas, pavilions, and private , the Pantheon,the Baths of Caracalla The Baths of Caracalla were Roman public baths, or thermae, built in Rome between 212 and 216 AD, during the reign of the Emperor Caracalla. The extensive ruins of the baths have become a popular tourist attraction. and from a Roman column re-used in the PalazzoBorghese Palazzo Borghese is the main seat of the Borghese family in Rome; it was nicknamed il Cembalo ("the harpsichord") due to its unusual trazezoidal groundplan; its short "keyboard" front (illustration) faces the Tiber. all have compositions consistent with an origin at the MonsClaudianus quarry, but also partly within error limits of Kozak. Threesamples, from a further column at Hadrian's Villa, from Carthageand from Ephesus lie close to or within the Kozak group, with oneanalysis of an amphibole in the Carthage sample falling within theEdremit source, but still within 1[Sigma] of Kozak. On FIGURE 5 theidentity of these three samples with Kozak is clear, and Edremit lies 3standard deviations away from the limits of Kozak. FIGURE 5 alsounderlines the identity of the Mons Claudianus artefacts with thatsource, since they are now clearly separated from Kozak.FIGURE 6 includes further samples for which amphiboles were not foundand biotites only were used in provenancing. Samples SG1 and A1 from StGilles and Arles respectively lie within the field of Elba rocks closeto those from Cavoli, and two standard deviations or more away fromother sources. PB1 from the Palazzo Borghese lies at the edge of theElba group and within c. 1[Sigma] of it, but also within 2[Sigma] ofUludag. FIGURE 7 confirms the identity of all three samples with Elba,again closest to Cavoli, and shows distinction from Uludag. Samplesassigned to Mons Claudianus and Kozak on FIGURES 4 & 5 remain withinthese groups on the biotite graphs. Some of the samples assigned to MonsClaudianus are absent from biotite graphs because biotites were notanalysed in these samples; this mineral is not necessarily absent fromthe rock (and indeed should not be absent because it is present in theassigned source of Mons Claudianus) but was not seen in the very smallsections available for probing.In summary, five samples have compositions consistent with an originat the Mons Claudianus quarries, three with the Kozak Dag area ofTurkey, and three with Elba, particularly the Cavoli locality. Sourcingcan thus be done by microprobe analysis alone.Discussion and conclusionsOur attempt to characterize the grey plutonic rocks of theMediterranean region has proved very successful. Visual assessment ofthe petrography enables us to distinguish the most distinctive outcrops,while chemical analysis confirms this, and allows discrimination betweenvisually similar sources. It is particularly gratifying that theelectron microprobe analysis has been successful, for it is now possibleto determine the origins of a column by simply taking a few biotite oramphibole crystals, which is, to all intents, non-destructive.Our work has demonstrated that visual determinations withoutreference materials to hand are hazardous, even when undertaken byworkers with wide experience of the rocks concerned. In the case of greyplutonic rocks, where the colour is neutral and mineral contents aresimilar, such identifications have of necessity to be based solely ongrain size and distribution, parameters which are notoriously difficultto hold in mind. On the other hand, the use of a portable reference setof source rocks improves matters dramatically; our results show completeagreement between our visual assignments to Mons Claudianus and chemicalassignments to that source.The use of comparative reference standards is by no means a universalpanacea, but is only of value when the source being sought is visuallydistinctive. In the case of Mons Claudianus we recognize the smallpossibility of confusion with San Bainzo material (cf. above) which mayhave been locally distributed and on which we are doing further studies.However, the San Bainzo source does not match any artefacts analysedchemically, and it seems reasonable to argue the distribution of MonsClaudianus material on the basis of all evidence presented here --visual, petrographic and chemical.A further major result is the suggestion of an important alternativesource of grey plutonic plu��ton��ic?adj.Of deep igneous or magmatic origin: plutonic rocks.[From Latin Pl decorative stone in western Turkey, withevidence pointing to the Kozak Dag area (the source region of thegranito misio of Lazzarini (1992)). This is certainly not the onlyalternative source, as some grey columns in southern France and in Italyoriginated in Elba, but it appears to have been important in the easternand central Mediterranean. Unfortunately, the Kozak Dag has yet to besubject to detailed archaeological study, but there are alreadyindications of exploitation during the Roman period. At Asagicuma in thewestern part of the intrusion, there is an area of ancient and modernquarrying with two broken column roughouts lying in the debris. A farmerin the village has used two column fragments, one clearly unfinished, asgateposts. By the road from Kozak to the sea is an unbroken unfinishedcolumn relating to relating torelate prep → concernantrelating torelate prep → bez��glich +gen, mit Bezug auf +accno known Roman building (Peacock 1993). This slightevidence is sufficient to suggest that the scientific predictions arealmost certainly correct. P2 H3 PB2 C1Si[O.sub.2]% 44.62 43.94 43.58 44.64Ti[O.sub.2] 1.27 1.31 1.16 1.35[Al.sub.2][O.sub.3] 9.11 8.84 10.31 8.88FeO 17.92 17.33 18.50 17.08MnO 0.42 0.40 0.42 0.40MgO 10.35 10.55 9.82 10.90CaO 11.74 12.03 11.57 11.86[Na.sub.2]O 1.40 1.20 1.40 1.24[K.sub.2]O 1.15 1.08 1.22 1.00total 97.98 96.68 97.98 97.35 C3 H1 B1 EPSi[O.sub.2]% 43.47 46.57 46.58 46.68Ti[O.sub.2] 1.40 1.37 1.37 1.43[Al.sub.2][O.sub.3] 9.39 7.59 7.68 7.70FeO 17.56 14.04 14.18 14.60MnO 0.39 0.47 0.47 0.50MgO 10.34 12.99 13.17 12.84CaO 11.56 11.71 11.71 11.58[Na.sub.2]O 1.31 1.50 1.57 1.47[K.sub.2]O 1.17 0.79 0.74 0.81total 96.59 97.03 97.47 97.61Sample localities are: P Pantheon, H Hadrian's Villa, PB Palazzo Borghese, CBaths of Caracalla, B Carthage, EP Ephesus.Gilles.TABLE 5(a). Microprobe analyses of archaeological samples -- amphiboles. H3 PB2 H1 B1Si[O.sub.2]% 36.53 36.93 37.03 36.91Ti[O.sub.2] 2.39 2.77 4.26 4.19[Al.sub.2][O.sub.3] 15.18 15.35 13.68 13.78FeO 18.24 18.84 17.41 17.33MnO 0.28 0.31 0.35 0.32MgO 11.94 11.66 13.04 13.27CaO 0.04 0.01 0.01 0.06[Na.sub.2]O 0.12 0.16 0.14 0.16[K.sub.2]O 9.56 9.57 9.44 9.12total 94.28 95.65 95.36 95.14 EP PB1 A1 SG1Si[O.sub.2]% 36.92 37.43 36.90 36.80Ti[O.sub.2] 4.27 3.48 3.26 3.32[Al.sub.2][O.sub.3] 13.66 16.29 15.92 15.76FeO 17.17 18.56 19.18 19.25MnO 0.34 0.34 0.35 0.35MgO 12.86 10.47 10.33 10.40CaO 0.01 0.01 0.01 0.01[Na.sub.2]O 0.16 0.14 0.09 0.13[K.sub.2]O 9.42 9.37 9.24 9.28total 94.81 96.09 95.28 95.30Sample localities are: H Hadrian's Villa, PB Palazzo Borghese, B Carthage, EPEphesus, A Arles, SG St Gilles.TABLE 5(b). Microprobe analyses of archaeological samples -- biotites.sample ref. site sourceC1 Baths of Caracalla, Rome Mons ClaudianusC3 Baths of Caracalla, Rome Mons ClaudianusP2 Pantheon, Rome Mons ClaudianusH3 Hadrian's Villa, Tivoli, nr Rome Mons ClaudianusPB2(*) Palazzo Borghese, Rome Mons ClaudianusH1 Hadrian's Villa, Tivoli, nr Rome Kozak DagB1 Carthage, Tunisia Kozak DagEP Ephesus, Turkey Kozak DagPB1(*) Palazzo Borghese, Rome ElbaA1 Arles, S France ElbaSG1 St Gilles, S France Elba* espolio.TABLE 6. Archaeological samples analysed and provenances assigned.The distribution of Mons Claudianus granodiorite is curious. Far fromthe pan-Mediterranean diffusion previously suggested, it appears to havea very tight and restricted usage. As our site examination is not yetcomplete, we have no data at all for certain parts of the Mediterraneansuch as the northern Levant Levant(ləvănt`)[Ital.,=east], collective name for the countries of the eastern shore of the Mediterranean from Egypt to, and including, Turkey. (though preliminary work in Israel hasrevealed no examples of Mons Claudianus material). However, a largenumber of Roman sites have been examined, and a pattern begins toemerge. Mons Claudianus material is found in Wadi Fatiri and Abu Zawalin the Eastern Desert of Egypt, and possibly in Hagia Sophia Hagia Sophia(hä`jə sōfē`ə, hā`jēə,)[Gr.,=Holy Wisdom] or Santa Sophia,Turkish Aya Sofia, inIstanbul. It is certainly present in Diocletian's Mausoleum mausoleum(môsəlē`əm), a sepulchral structure or tomb, especially one of some size and architectural pretension, so called from the sepulcher of that name at Halicarnassus, Asia Minor, erected (c.352 B.C. inSplit. All the remaining positive identifications are in Rome and inHadrian's Villa just outside Rome. What is more, if we exclude thesecondary usage in the churches and palazzi of Rome, Mons Claudianusrock appears to feature only in the more imposing public buildings,Trajan's Basilica Ulpia The Basilica Ulpia was an ancient Roman civic building located in the Forum of Trajan. It was named after Roman emperor Trajan whose full name was Marcus Ulpius Traianus.[1] being a prodigious example. Thedistribution and restricted use contrasts markedly with that of othergreat decorative stones such as the granite from Aswan, or marble fromthe Cipollino quarries in Euboea, all of which suggests that MonsClaudianus may have been a rather special stone, perhaps restricted tothe emperor himself. The relationship seems to be reflected in thewritten evidence. A number of ostraca os��tra��ca?n.Plural of ostracon. from Mons Claudianus, as yetunpublished, suggest that finished artefacts were to be reported to be spoken of; to be mentioned, whether favorably or unfavorably.See also: Report to theprocurator PROCURATOR, civil law. A proctor; a person who acts for another by virtue of a procuration. Procurator est, qui aliena negotia mandata Domini administrat. Dig 3, 3, 1. Vide Attorney; Authority. caesaris (information from A. Bulow-Jacobsen). If this is thecase, the rock seems to have been the subject of an imperial monopoly,and quite distinct from other stones involved in the Imperial ratiomarmorum or 'marble bureau' in Fant's (1992) apttranslation.It remains to note the date of this activity. There is a goodcorrelation between finds in Rome and dating evidence from theexcavation at Mons Claudianus itself, where the earliest dated piece isa Neronian ostracon os��tra��con?n. pl. os��tra��caAn inscribed potsherd.[Greek ostrakon, shell; see ost- in Indo-European roots.] dating to AD 68 (information from H. Cuvigny). Thelatest is a monumental ostracon which may have hung on the gate duringits last phase of use: a dedication by Cohors II Ituraeorum to'Zeus sun great Sarapis and the gods who share the temple on behalfof the emperor Severus Alexander'. It can thus be dated to AD222-235, soon after which the fort may have fallen into disuse dis��use?n.The state of not being used or of being no longer in use.disuseNounthe state of being neglected or no longer used; neglectNoun 1. ; the verylatest finds are single coins of Aurelian (AD 269/270-275) and Probus(AD 276-282).In Rome, the earliest known find of Mons Claudianus rock is in theNeronian (AD 54-68) Domus Transitoria; it occurs in Domitianic contexts,peaking under Trajan, and continuing into the reign of Hadrian in theearlier part of the 2nd century AD. Thereafter the picture becomes hazy;the finds in the Baths of Caracalla fit in well with the mention of thestone in P. Oxy. 3242 dated to An 214-215, and the evidence from thequarry site. Occurrences in Diocletianic contexts (AD 284-305) are moredifficult to explain, as is mention of the rock in his price edict A decree or law of major import promulgated by a king, queen, or other sovereign of a government.An edict can be distinguished from a public proclamation in that an edict puts a new statute into effect whereas a public proclamation is no more than a declaration of a law (Lauffer 1971), for there is no evidence of activity at the quarry siteas late as this. Perhaps minor operations were undertaken, administeredby Mons Porphyrites 50 km to the north, or perhaps by this late datethere was sufficient material in circulation to allow re-use of columnsin later edifices. In this respect, it should be noted that there isvery little Diocletianic material: two columns in his Baths at Rome(neither of which definitely belong to the original building of theBaths), and two small columns at Split. All these could have resultedfrom re-use of earlier material.Our work on the visual and chemical identification of Roman greyplutonic rocks has therefore led to a more precise definition of thedistribution of the Mens Claudianus rock within the Mediterranean area.Analysis has also indicated the use of other grey granodiorites andgranites from Kozak Dag in western Turkey as well as from Elba.Acknowledgements. This paper has arisen from work at Mons Claudianusand we are grateful for travel grants from the University ofSouthampton In the most recent RAE assessment (2001), it has the only engineering faculty in the country to receive the highest rating (5*) across all disciplines.[3] According to The Times Higher Education Supplement . Artefact samples were obtained from the Dame KathleenKenyon Dame Kathleen Mary Kenyon (5 January, 1906 – 24 August, 1978), important English archaeologist of Neolithic culture in the Fertile Crescent and excavator of a small area of Jericho in Israel from 1952 to 1958. collection (Ephesus) and from a collection by M. Rene Artois(France). For a sample from Carthage we thank INAA Tunis and M. A.Anabli. We are also grateful to Professor Wimenauer for providing accessto Professor Troger's samples.Chemical analysis was supported by a grant from the Fund for AppliedScience in Archaeology which we gratefully acknowledge. Part of thiswork was carried out under the tenure of grant no. F.269P;S903248 fromthe Leverhulme Trust to the Open University (for OWT). We thank Dr P.C.Webb and Mr J.S. Watson for XRF analyses, Mr A. Lloyd and Miss K.Knowles for drawing the figures, and anonymous referees for helpfulcomments on the text.ReferencesAUDIN, A. & P. QUONIAM. 1962. Victoires et colonnes del'autel federal des trois Gaules: donnees nouvelles, Gallia 1962:103-16.BULIC F. 1908. Materiale e provenienze delle pietre, delle colonne,nonche delle sfingi del Palazzo di Diocleziano a Spalato e delle colonneecc. delle basiliche cristiane a Salona, Bollettino di Archeologia eStoria Dalmato 31: 86-127.CANTO, A.M. 1978. Avances sabre la explotacion del marmol en laEspana Romana, Archive Espanol de Arquelogia 50-51: 165-88.COOK, J.M. 1973. The Troad: an archaeological and topographicalstudy. Oxford: Clarendon Press.DODGE, H. 1988. Decorative stones for architecture in the RomanEmpire, Oxford Journal of Archaeology 7(1): 65-80.FANT, J.C, 1992. The Roman Imperial marble yard at Portus, inWaelkens et al. (1992): 115-20.GALETTI, G., L. LAZZARINI & M. MAGGETTI. 1992. A firstcharacterization of the most important granites used in antiquity, inWaelkens et al. (1992): 167-78.GNOLI, R. 1971. Marmora Romana. Rome: Edizione Elefante.GREENBERG, J.K. 1981. Characteristics and origin of Egyptian YoungerGranites: summary, Geological Society of America The Geological Society of America (or GSA) is a nonprofit organization dedicated to the advancement of the geosciences. The society was founded in New York in 1888 by James Hall, James D. Bulletin 92: 224-32.HARRELL, J.A. 1989. An inventory of Ancient Egyptian quarries,Newsletter of American Research Centre in Egypt 146: 1-7.KLEMM, R. & D. KLEMM. 1981. Die Steine der Pharaonen. Munich:Staatliche Sammlung Agyptischer Kunst. 1992. Steine und Steinbruche inalten Agypten. Berlin: Springer-Verlag.KRAUS, T. & J. RODER. 1962. Mons Claudianus -- Bericht uber eineerste Erkundungsfahrt im Marz 1961, Mitteilungen des DeutschenArchaologischen Instituts, Abteilung Kairo 18: 80-120.LAUFFER, S. 1971. Diokletians Preisedict. Berlin: De Gruyter.LAZZARINI, L. 1987. I graniti dei monumenti Italiani e i loroproblemi di deterioramento, Bollettino d'Arte Supplemento : 157-72.1992. Des pierres pour l'eternite: los granits utilises dansl'Antiquite Classique, Les Dossiers d'Archeologie 173: 58-67.LEWIS, M.J.T. 1985. Roman methods of transporting and erectingobelisks, Transactions of the Newcomen Society 56: 87-110.MACKENZIE, W.S., C.H. DONALDSON & C. GUILFORD. 1982, Atlas ofigneous ig��ne��ous?adj.1. Of, relating to, or characteristic of fire.2. Geologya. Formed by solidification from a molten state. Used of rocks.b. Of or relating to rock so formed; pyrogenic. rocks and their textures. Harlow: Longman.MASSIMETTI, M.G.C. 1991. Lo sfruttamento del granito gallurese inepoca imperiale: risvolti economica e sociali, in L'Africa Romana,Atti dell'VIII convegno di studio Cagliari, 14-16 dicembre 1990:3-10. Sassari: Edizioni Gallizzi.MEREDITH, D. & L. TREGENZA. 1949. Notes on Roman roads andstations in the Eastern Desert I, Bulletin of the Faculty of Arts Historically the Faculty of Arts was one of the four traditional divisions of the teaching bodies of universities, the others being theology, law and medicine.[1] Nowadays it is a common name for the faculties teaching humanities. References1. FouadI University II: 1-30.PEACOCK, D.P.S. 1993. Rome in the Desert: a symbol of power.Southampton: University of Southampton.POLI, G. 1992. Geochemistry of Tuscan archipelago granitoids, centralItaly: the role of hybridization hybridization/hy��brid��iza��tion/ (hi?brid-i-za��shun)1. crossbreeding; the act or process of producing hybrids.2. molecular hybridization3. processes in their genesis, Journal ofGeology 100: 41-56.POTTS, P.J. 1983. On the precision of electron microprobe data: a newtest for the homogeneity of mineral standards, American Mineralogist 68:1237-42.POTTS, P.J. & P.C. WEBB. 1992. X-ray fluorescence spectrometry,Journal of Geochemical Exploration 44: 251-96.POTTS, P.J., P.C. WEBB & J.S. WATSON. 1984. Energy-dispersiveX-ray fluorescence analysis of silicate silicate,chemical compound containing silicon, oxygen, and one or more metals, e.g., aluminum, barium, beryllium, calcium, iron, magnesium, manganese, potassium, sodium, or zirconium. Silicates may be considered chemically as salts of the various silicic acids. rocks for major and traceelements, X-Ray Spectrometry 13: 2-15.SABET, A.H., S. EL-GABY & A.A. ZALATA. 1972. Geology of thebasement rocks in the northern parts of El-Shayib and Safaga sheets,Eastern Desert, Annals of the Geological Survey of Egypt 2: 111-27.SCAIFE, C.H.O. 1953. The origins of some Pantheon columns, Journal ofRoman Studies 43: 37.SOLANO, A. 1985. Su una cava romana di granito a Nicotera, StudiMiscellanei 26: 83-96.WAELKENS, M., N. HERZ & L. MOENS (ed.). 1992. Ancient stones:quarrying, trade and provenance. Leuven: Acta Archeologica Lovaniensa.Monograph 4.WALDECK, H. 1986. Die Inseln Elba. Berlin: Gebruder Borntraeger.WILSON, R.J.A. 1988. Ancient granite quarries on the Bocche diBonifacio, in N. Herz & M. Waelkens (ed.), Classical marble:geochemistry, technology and trade: 103-12. Kluwere Academic Publishers.