Characterization of Intra-Quarry Variability in Pentelic Marble Using Stable Isotopes: A Case Study of the Parthenon

Abstract

This study presents a systematic analysis of stable isotopes (δ¹³C and δ¹⁸O) in Pentelic marble from the ancient quarries of Mount Pentelikon, Greece. A total of 610 samples were collected from 83 quarry pits, including all identified ancient extraction sites, to build a comprehensive reference database. Of those, stable isotope ratios of carbon and oxygen were measured for 384 samples. The results reveal significant variability in stable isotope values across different quarry regions, challenging the assumption of isotopic homogeneity in Pentelic marble. Notably, distinct δ¹³C and δ¹⁸O patterns allow differentiation between quarry areas and specific extraction zones. These findings provide a more refined isotopic framework for provenance studies of ancient artifacts. Application of the new database suggests that marble for the Parthenon’s West Pediment was sourced from the northern upslope quarries in Aspra Marmara, whereas structural elements of the Parthenon were extracted from the lower Spilia Divail quarry, as well as other pits across the ancient quarry zone. The results demonstrate that multiple quarry pits from both the lower slope and upper slope quarries were exploited for the construction of the Parthenon. This research highlights the benefits of extensive sampling and stable isotope analysis in identifying quarry sources, emphasizing the value of undertaking detailed and thorough field surveys and sampling programs to gain new perspectives into ancient resource use.

1. Introduction

Variations in the range of δ¹³C and δ¹⁸O values in discrete marble formations in the eastern Mediterranean have been used with limited success to distinguish between many different ancient white marble quarries [1,2,3,4]. The ability to quantitatively distinguish between the different quarries utilized in antiquity provides a tool for archaeologists, art historians, and museum curators to assign provenance to ancient Greek and Roman marble artifacts. The correct assignment of provenance assists scholars in their investigations of dating an execution of a sculpture, tracing ancient avenues of commerce, giving insight into evolving aesthetic values, and in determining modern forgeries, ancient copies and dissociated fragments [5]. There is a rich literature of archaeometric studies focused on determining the provenance of marble and other antique stones that use multiple physical and chemical techniques [4,6,7,8,9,10,11,12,13].

The quarries on Mount Pentelikon, Attica, Greece, have been a primary source of white marble since the beginning of the fifth century BCE [14]. The quarries were first exploited on a large scale for the 5th century BCE Periclean building projects, including the construction of the Parthenon and other monuments on the Athenian acropolis. During the Roman era, the immensely popular Pentelic marble was widely exported throughout the Mediterranean for both architectural and sculptural uses [15,16,17]. The Romans were so fond of the material that Pausanias (i.19) writes that Herodes Atticus exhausted the quarries for his various building programs in the 2nd century AD, though Pentelic marble was still being exported to North Africa in the 3rd century CE [18].

Despite its importance in Greek and Roman archaeology, the ability to accurately, physically, and geochemically characterize Pentelic resources has been met with challenges. A prevailing assumption of marble quarry characterization studies is that each quarry region is relatively isotopically homogenous. While there may be isotope variation within the region, spatially the spread of data is evenly distributed. This approach is problematic for Pentelic marble, as the distribution of δ¹³C and δ¹⁸O is quite large, especially when compared to the ranges of other known ancient sources [19,20] (Figure 1). Research published in the 1990s suggested that the reference sample collection used to develop the stable isotope signature field for the Pentelic quarries inadequately represented the entire quarry region. Matthews et al. [21] showed that marble samples of sculptural and architectural elements from the Parthenon and other monuments currently housed at the British Museum, which are historically documented as being Pentelic [22], have a higher oxygen isotope ratio than those in the established databases. Matthews et al. further suggested that there may be some spatial patterning or trends of stable isotope variation within the Pentelic quarry field.

Kane et al. [24] provided further support for the proposed idea of isotopic spatial patterning in the Pentelic quarries through their analysis of Pentelic sculptures from the Nymphaeum of Herodes Atticus at Olympia, Greece. Ancient financial records confirm the Pentelic origin of the marble used in this monument [25]. Stable isotope analyses of the marbles from various statues, their disassociated bases, and construction phases indicated that the materials may not all derive from the same quarry or individual block but instead from isotopically distinct blocks or possibly different quarries within the Pentelic region. However, the existing databases lack sufficient spatial information of the exact sampling locations, limiting their capacity to map isotope variations accurately across the quarry area.

To address these gaps, a comprehensive study was conducted to systematically characterize the isotope profile of Mount Pentelikon’s ancient quarry region. The project involved detailed survey, sampling, and analysis phases. The pedestrian survey was necessary to create a detailed geologic and topographic map that also delineated quarry features. The map was used as a guide to for an extensive sampling program throughout the quarry field that collected 610 samples [26]. Of those, 384 unaltered samples were analyzed using stable isotope analysis. This report focuses on the results from the spatially-rectified δ¹³C and δ¹⁸O isotope database of the ancient white marble quarries of Mt. Pentelikon.

2. Materials and Methods

To successfully model the isotopic profile of the ancient quarries on Mount Pentelikon, it was critical that a representative marble reference collection be obtained. Only this way can the reference sample collection be assured that every quarry is represented and that, if present, spatial isotopic trends can be mapped. To this end, an extensive topographic and geologic field survey was conducted. A detailed description of the survey and sampling program, the methodology employed, and findings are reported in Pike [26]. A summary of the results of the survey and sampling are reported here.

A quarry location map identifies 172 discrete quarries (Figure 2). Tool marks and unfinished worked blocks indicate that no fewer than thirty quarries were actively worked in antiquity [27,28] (Figure 3). The ancient quarries are distinguished by red traces that outline surviving quarry walls. In contrast, quarries traced in blue lack any physical evidence of ancient extraction activity and are displayed as modern quarry pits. For the purposes of this study, “modern” refers to the period following Greece’s independence from the Ottoman Empire in the mid-19th century. However, it is important to note that many modern quarries expanded into abandoned ancient extraction sites, thereby altering or obliterating physical evidence of earlier activity. Consequently, it is plausible that several quarries currently classified as modern were, in fact, exploited in antiquity. This complicates efforts to differentiate between ancient and modern operations solely based on surface observations.

The survey also documented key observations related to the geological structure of the quarries. Geologists studying northern Attica and Mount Pentelikon describe the mountain as an autochthonous, low- to medium-grade metamorphic complex that has been folded and uplifted into a northeastward-plunging anticlinorium. The primary geological composition of the mountain includes two distinct marble formations—the Upper Marble and the Lower Marble—separated by an intermediate schist layer known as the Kaisariani Schist. This schist is believed to extend eastward to Mount Hymettus. Within the Kaisariani Schist, a significant marble lens, often referred to as the Middle or Intermediate Marble, has been identified [29,30,31,32].

The quarries are predominantly located on the eastern limb of the plunging anticlinorium (Figure 4). A prominent ridge, trending northeast–southwest through the quarry area on the southern slopes of Mount Pentelikon, aligns with the axis of a major anticlinal fold. This central fold is flanked by parallel axes of minor anticlines and synclines, creating a complex structural framework. The surviving ancient quarries, along with most modern quarry pits, are situated along the limbs of this central ridge, reflecting both the geological accessibility and marble quality in these zones. These structural features not only influenced ancient quarrying strategies but also provide critical context for understanding spatial patterns in the isotopic variation of Pentelic marble.

Figure 2. LiDAR-generated digital terrain model of the ancient quarry region on the south slope of Mount Pentelikon outlining the surviving boundaries of quarry pits [33]. Those highlighted in red are ancient quarries as indicated by surviving traces of ancient activity. Blue boundaries are modern quarries, and the orange quarries to the north were identified after the initial survey and have not been sampled for this study.

Figure 2. LiDAR-generated digital terrain model of the ancient quarry region on the south slope of Mount Pentelikon outlining the surviving boundaries of quarry pits [33]. Those highlighted in red are ancient quarries as indicated by surviving traces of ancient activity. Blue boundaries are modern quarries, and the orange quarries to the north were identified after the initial survey and have not been sampled for this study.

Figure 3. Tool marks and unfinished worked blocks abandoned in quarries that were used to identify antiquity of the quarry operations. (A) Garland-type tool marks associated with the long-handle pick (Quarry Π54); (B) horizontal tool marks associated with the Archaic and early Classical short-handled pick (Quarry Π44); and (C) unfinished column capital from the floor of Quarry Π23 [34].

Figure 3. Tool marks and unfinished worked blocks abandoned in quarries that were used to identify antiquity of the quarry operations. (A) Garland-type tool marks associated with the long-handle pick (Quarry Π54); (B) horizontal tool marks associated with the Archaic and early Classical short-handled pick (Quarry Π44); and (C) unfinished column capital from the floor of Quarry Π23 [34].

Figure 4. A geologic map showing the lateral distribution of marble and schist interbeds in the upper portion of the Lower Marble formation. The younger units are towards the NW. Note the geologic distinction between marble groups 1, 2 and 3.

Figure 4. A geologic map showing the lateral distribution of marble and schist interbeds in the upper portion of the Lower Marble formation. The younger units are towards the NW. Note the geologic distinction between marble groups 1, 2 and 3.

The geologic map (Figure 4) displays three NE–SW trending marble units exposed at the surface, representing the uppermost section of the Lower Marble, with progressively younger rocks observed toward the west. Markoulis et al. [35] caution that these units, while appearing distinct, may actually represent a single geological formation. Recumbent folding and preferential weathering processes likely caused the surface exposures to appear as separate units. Nevertheless, for the purposes of this localized study, the three units are treated individually and are labeled as Marble Unit 1, Marble Unit 2, and Marble Unit 3 on the geologic map. The majority of ancient quarries are concentrated within Marble Unit 3, with a smaller number located in Marble Unit 2. No evidence of ancient quarrying activity has been observed in Marble Unit 1; although, modern quarry pits are prominent at lower elevations within this unit. With the aid of the quarry location (Figure 2) and geologic (Figure 4) maps, a sample reference collection of 610 samples was collected. For a detailed description of the field survey and acquisition of the Pentelic marble reference collection refer to Pike [26].

Reference samples were collected from 83 of the 172 identified quarries, including all quarries with evidence of ancient activity. To avoid contamination, all samples were taken directly from exposed quarry walls, ensuring accurate representation of their specific locations. No material was collected from loose debris on quarry floors or from rubble piles. Each sample’s location and elevation were recorded and are marked on a detailed quarry map. Where possible, multiple samples were taken from each individual quarry pit to assess potential lateral or vertical variations in texture, geochemistry, and isotopic composition. Some quarries were excluded from the sampling due to inaccessibility or the absence of fresh marble surfaces. Standard geological tools, such as steel chisels, hammers, and picks, were used for sampling, with great care taken to avoid damaging or altering any surviving ancient worked surfaces.

Stable isotope analysis was conducted at the Stable Isotope Laboratory, Department of Geology, University of Georgia, using a Finnigan MAT-delta E gas-source mass spectrometer. Ten–twenty mg aliquots of white marble powder were reacted with anhydrous phosphoric acid in a vacuumed reaction tube at 50 °C, following the protocol outlined by Swart, Burns, and Leder [36]. The resultant CO₂ gas was then purified in a glass extraction line and analyzed with the mass spectrometer. The obtained isotope ratios ¹³C/¹²C and ¹⁸O/¹⁶O were then compared to the PDB standard of the University of Chicago (Belemnitella americana, PeeDee formation, Cretaceous, South Carolina) and transformed to the standard delta notation by using the following equations [37]:

$$\mathit{\partial }{}^{13}\mathrm{C}=\left[{\displaystyle \frac{{\left(\frac{{}^{13}\mathrm{C}}{{}^{12}\mathrm{C}}\right)}_{\mathrm{sample}}-{\left(\frac{{}^{13}\mathrm{C}}{{}^{12}\mathrm{C}}\right)}_{\mathrm{standard}}}{{\left(\frac{{}^{13}\mathrm{C}}{{}^{12}\mathrm{C}}\right)}_{\mathrm{standard}}}}\right]\times 1000‰$$

$$\mathit{\partial }{}^{18}\mathrm{O}=\left[{\displaystyle \frac{{\left(\frac{{}^{18}\mathrm{O}}{{}^{16}\mathrm{O}}\right)}_{\mathrm{sample}}-{\left(\frac{{}^{18}\mathrm{O}}{{}^{16}\mathrm{O}}\right)}_{\mathrm{standard}}}{{\left(\frac{{}^{18}\mathrm{O}}{{}^{16}\mathrm{O}}\right)}_{\mathrm{standard}}}}\right]\times 1000‰$$

During the analyses, the mass spectrometer was programmed to measure the mass of the sample gas eight times. The multiple readings allow for the assessment of the homogeneity of the CO² gas and provide an important check on the precision of the analysis.

Two methods were employed to prepare sample aliquots prior to isotopic analysis. The first method utilized a variable-speed dental drill fitted with a carbide drill bit. To minimize the potential for isotopic fractionation caused by high frictional temperatures interacting with ambient room moisture [38], the drill operated at the lowest possible speed. Additionally, the outer 2 mm of each sample was removed and discarded to avoid contamination from weathered or isotopically altered marble surfaces. Drilling is preferred for marble stable isotope analysis as it mirrors the conditions under which archaeological samples are typically collected and allows for highly precise sampling of specific artifact surfaces while minimizing damage. However, drilling is not suitable for all physio-chemical characterization techniques. For instance, electron spin resonance (ESR) analysis produces a prominent “drilling peak”, attributed to electrons partially filling defects generated during drilling, regardless of speed [39].

The second method involved crushing samples using an agate mortar and pestle. Small fragments were first separated from the hand samples with a two-pound crack hammer, ensuring that only unweathered portions from the core of the samples were used. The agate mortar and pestle were designated exclusively for carbonate materials to prevent contamination from other substances. This method provided an alternative approach to prepare samples while maintaining the integrity and purity of the material for analysis. All samples, whether crushed or drilled, were reacted with anhydrous phosphoric acid in a vacuumed reaction tube at 50 °C, following established protocols [36]. The resultant CO₂ gas was then purified in a glass extraction line and analyzed by a Finnigan Delta mass spectrometer.

Of the 610 samples in the reference collection, stable isotope analyses were conducted on 384 samples. The remaining samples were excluded from analysis due to evidence of diagenetic alteration. Signs of alteration included chloritic mica veins permeating the samples and deep pink or orange discoloration plumes. Such features suggest potential isotopic fractionation between metamorphic fluids and calcite grains at the calcite–hydrous mineral interface [40]. Including these altered samples in the Pentelic marble database could compromise its integrity and accuracy. To ensure analytical precision, several samples were reanalyzed in separate runs. In all cases, the resulting values were consistent, falling within a single standard deviation, thus confirming the reliability of the analytical procedures.

3. Results

Table A1. Stable isotope data of δ¹³C and δ¹⁸O for all analyzed samples from the Pentelic marble sample reference collection.

Quarry	Sample ID	δ13C	δ18O
4	4-1	4.2	−7.2
4	4-2	3.8	−6.5
11	11-1	4.0	−7.8
12	12-1	4.3	−7.1
12	12-2	2.7	−8.2
12	12-3	3.8	−9.1
13	13-1	3.9	−7.6
13	13-2	3.6	−8.6
14	14-1	4.1	−7.6
15	15-1	4.6	−7.0
15	15-2	4.5	−7.4
16	16-1	4.7	−7.6
16	16-2	4.8	−7.1
17	17-1	4.8	−7.6
20	20-1	5.0	−7.1
20	20-2	4.1	−9.9
21	21-1	4.6	−6.9
22	22-1	5.0	−6.6
22	22-3	3.4	−8.4
23	23-1	2.8	−8.4
25	25-1	3.3	−8.2
25	25-2	3.2	−8.2
26	26-1	2.7	−8.3
27	27-1	4.0	−7.4
27	27-2	4.3	−6.7
27	27-3	3.3	−7.6
27	27-4	3.2	−7.2
35	35-1	4.7	−6.2
36	36-1	4.2	−8.3
36	36-2	4.0	−6.4
38	38-1	4.7	−5.7
38	38-2	4.4	−6.5
39	39-1	2.9	−8.5
40	40-1	2.8	−7.8
43	43-1	2.6	−8.0
46	46-1	2.7	−8.8
46	46-2	2.7	−8.6
46	46-3	3.0	−7.3
46	46-5	2.8	−5.4
46	46-6	2.7	−5.8
46	46-7	2.8	−6.8
48	48-1	2.6	−8.2
48	48-2	3.1	−8.5
49	49-1	3.5	−7.2
53	53-2	2.5	−7.1
54	54-1	2.9	−7.7
54	54-2	2.9	−8.0
54	54-3	2.9	−7.9
54	54-4	4.1	−7.7
54	54-5	4.2	−7.0
54	54-6	2.7	−8.9
54	54-7	4.7	−8.0
55	55-1	2.9	−8.4
55	55-2	3.0	−8.4
55	55-3	3.1	−9.1
55	55-4	2.8	−8.7
55	55-5	2.7	−8.7
55	55-6	2.9	−4.7
56	56-2	3.0	−7.0
56	56-3	3.2	−7.2
56	56-4	3.1	−7.1
64	64-1	2.8	−7.8
64	64-1	2.8	−7.8
64	64-2	3.0	−8.1
64	64-2	3.0	−8.1
64	64-3	3.0	−7.9
64	64-3	3.0	−7.9
64	64-4	2.7	−6.0
67	67-1	2.8	−6.9
67	67-2	2.7	−5.5
67	67-3	2.8	−4.8
67	67-4	3.0	−5.3
67	67-5	3.3	−5.5
67	67-5	3.4	−5.6
67	67-6	2.9	−6.2
67	67-7	3.0	−8.0
67	67-8	2.8	−9.0
67	67-9	2.8	−9.1
67	67-10	3.0	−7.9
67	67-11	3.0	−8.3
67	67-12	2.8	−7.8
69	69-1	3.1	−5.6
69	69-2	3.3	−4.9
69	69-3	3.1	−5.1
69	69-4	2.8	−5.3
69	69-5	3.1	−5.3
69	69-6	3.4	−4.7
69	69-7	3.0	−6.8
69	69-8	3.0	−8.7
69	69-10	2.9	−7.0
69	69-11	2.7	−6.4
69	69-12	2.9	−7.1
69	69-13	3.2	−5.9
69	69-14	3.1	−7.2
70	70-1	2.7	−5.2
70	70-2	3.0	−8.3
70	70-3	2.7	−8.3
70	70-4	2.7	−6.7
70	70-5	2.7	−7.4
70	70-6	2.9	−4.1
71	71-1	2.5	−7.1
72	72-1	2.5	−7.5
79	79-1	2.7	−7.5
79	79-2	2.8	−7.5
81	81-1	2.7	−9.1
81	81-2	2.9	−8.8
82	82-1	2.8	−8.4
82	82-2	3.6	−8.2
83	83-1	2.2	−9.7
83	83-2	3.3	−8.6
83	83-3	2.5	−8.9
83	83-4	2.7	−9.0
84	84-1	2.6	−7.1
84	84-2	2.8	−9.1
84	84-3	3.0	−7.0
84	84-4	2.8	−7.1
84	84-5	3.2	−7.6
85	85-1	2.9	−8.3
85	85-2	2.9	−5.6
85	85-3	2.7	−6.9
85	85-4	2.8	−5.6
85	85-5	2.7	−6.2
85	85-6	2.8	−5.5
85	85-7	3.1	−8.5
85	85-8	3.0	−4.3
85	85-9	2.7	−8.3
85	85-10	4.2	−7.5
87	87-1	2.9	−6.2
87	87-2	2.8	−6.6
87	87-3	2.5	−8.2
87	87-4	2.6	−7.2
87	87-5	2.5	−7.6
87	87-6	3.0	−6.4
88	88-1	2.7	−4.3
88	88-2	2.7	−4.9
88	88-3	2.7	−4.0
88	88-4	2.8	−4.1
88	88-5	3.0	−4.5
88	88-6	3.0	−4.3
88	88-7	2.8	−4.3
88	88-8	2.9	−4.3
88	88-9	2.7	−4.4
88	88-11	2.7	−6.2
88	88-12	2.6	−6.4
88	88-14	2.5	−6.8
88	88-15	2.6	−5.9
89	89-1	2.6	−7.7
89	89-2	2.7	−5.7
89	89-3	2.6	−5.6
89	89-4	2.5	−5.8
89	89-5	2.4	−5.5
89	89-6	2.7	−4.6
89	89-7	2.9	−4.7
89	89-8	3.0	−5.6
89	89-9	2.2	−5.5
89	89-10	2.7	−5.0
89	89-11	3.0	−5.1
89	89-12	2.7	−4.6
89	89-13	2.6	−4.4
89	89-14	2.6	−4.9
90	90-1	3.0	−5.7
90	90-2	3.1	−5.4
90	90-3	3.2	−4.6
90	90-4	2.7	−4.4
90	90-5	3.0	−5.3
90	90-6	2.8	−6.1
90	90-7	2.9	−5.9
90	90-8	3.0	−4.4
90	90-9	3.1	−4.7
90	90-10	2.5	−4.8
90	90-11	2.7	−4.6
90	90-12	2.6	−5.6
90	90-13	2.7	−5.7
90	90-14	2.7	−5.7
90	90-15	2.8	−4.7
90	90-16	3.0	−5.6
90	90-17	2.7	−5.0
90	90-18	2.8	−6.0
90	90-19	2.7	−5.5
91	91-1	2.7	−5.4
91	91-10	2.8	−5.8
91	91-11	3.0	−8.8
91	91-12	2.8	−9.0
91	91-13	2.7	−8.3
91	91-14	2.8	−9.4
91	91-15	2.8	−7.2
91	91-16	3.0	−8.4
91	91-11a	1.7	−7.1
91	91-13a	3.0	−8.5
91	91-15a	2.9	−6.8
91	91-17	3.1	−8.5
91	91-18	2.9	−8.6
91	91-19	2.7	−8.4
91	91-2	2.5	−5.6
91	91-20	2.6	−7.8
91	91-22	2.3	−8.5
91	91-23	2.9	−6.4
91	91-24	2.9	−6.6
91	91-25	2.9	−9.1
91	91-26	2.6	−7.9
91	91-27	2.7	−4.8
91	91-28	2.7	−5.1
91	91-29	3.0	−5.9
91	91-3	2.5	−6.1
91	91-30	3.2	−5.9
91	91-4	2.4	−5.7
91	91-5	2.9	−5.8
91	91-6	2.8	−6.7
91	91-7	2.6	−5.7
91	91-8	2.6	−5.1
91	91-9	2.9	−5.6
92	92-2	2.3	−5.4
92	92-3	2.7	−5.6
93	93-1	2.8	−8.4
93	93-10	3.4	−7.0
93	93-11	3.1	−8.7
93	93-12	2.7	−6.4
93	93-13	2.7	−8.4
93	93-3	2.7	−7.5
93	93-4	3.3	−6.8
93	93-5	2.8	−8.1
93	93-6	3.0	−8.0
93	93-7	2.8	−7.6
93	93-8	3.0	−8.2
95	95-1	3.2	−7.7
95	95-10	2.8	−8.3
95	95-11	2.6	−9.4
95	95-12	2.4	−11.0
95	95-13	2.7	−9.7
95	95-14	2.6	−7.9
95	95-2	3.2	−8.1
95	95-3	4.3	−8.3
95	95-4	3.0	−6.8
95	95-5	3.9	−7.6
95	95-6	4.1	−8.4
95	95-7	3.2	−7.4
95	95-8	4.1	−7.8
95	95-9	3.6	−7.5
96	96-1	2.5	−8.7
96	96-2	3.1	−8.0
96	96-3	2.9	−8.1
96	96-4	3.5	−6.7
96	96-5	3.4	−6.7
96	96-6	2.6	−6.4
96	96-7	2.9	−8.2
96	96-8	2.6	−8.5
97	97-10	3.5	−7.6
97	97-11	3.4	−7.7
97	97-12	4.3	−6.5
97	97-13	4.3	−7.9
97	97-2	3.5	−8.1
97	97-3	3.7	−7.6
97	97-4	4.1	−7.6
97	97-5	3.4	−7.9
97	97-6	3.4	−7.7
97	97-7	3.9	−8.2
97	97-8	4.0	−7.2
97	97-9	3.3	−7.2
98	98-1	2.7	−6.9
98	98-2	2.5	−7.3
98	98-3	2.6	−7.5
98	98-4	2.7	−8.7
98	98-5	2.5	−7.8
98	98-6	2.3	−9.4
103	103-1	2.9	−8.7
103	103-2	2.8	−7.6
103	103-3	2.7	−7.8
103	103-4	2.7	−9.1
103	103-5	2.6	−8.6
104	104-2	4.5	−8.1
104	104-3	3.7	−6.4
104	104-4	3.8	−6.8
105	105-1	2.9	−8.0
105	105-2	2.8	−7.3
105	105-3	2.8	−8.2
105	105-4	3.1	−8.3
105	105-5	2.7	−9.2
105	105-6	2.6	−9.3
105	105-7	3.0	−7.5
106	106-12	2.9	−9.9
106	106-14	4.0	−7.9
107	107-1	3.9	−8.7
107	107-10	2.8	−8.4
107	107-11	3.1	−8.5
107	107-12	2.7	−8.2
107	107-13	2.9	−8.9
107	107-14	2.7	−8.4
107	107-2	3.2	−6.7
107	107-3	2.9	−8.4
107	107-4	2.9	−8.2
107	107-5	3.2	−8.1
107	107-6	2.9	−8.2
107	107-7	2.9	−8.4
107	107-8	2.8	−7.7
107	107-9	2.2	−8.4
109	109-1	4.3	−8.1
109	109-2	3.6	−6.9
109	109-3	3.4	−6.7
109	109-4	3.5	−6.7
109	109-5	4.4	−6.7
109	109-6	3.5	−6.5
109	109-7	4.2	−6.5
122	122-1	3.1	−7.9
122	122-2	2.9	−8.0
124	124-1	2.8	−7.3
125	125-1	3.3	−6.6
125	125-2	2.7	−7.2
127	127-1	3.1	−7.5
129	129-1	4.6	−8.3
129	129-2	4.4	−7.8
130	130-1	2.9	−9.2
130	130-2	2.8	−6.1
130	130-3	3.2	−8.1
130	130-4	3.1	−9.2
130	130-5	2.6	−11.7
130	130-6	4.0	−9.0
130	130-6	3.3	−8.4
131	131-1	2.8	−7.2
131	131-2	4.0	−7.1
131	131-3	3.7	−7.1
132	132-1	2.7	−7.9
132	132-10	4.3	−5.8
132	132-2	2.9	−7.7
132	132-4	3.2	−7.7
132	132-5	4.1	−6.9
132	132-6	4.0	−6.7
132	132-8	3.7	−6.2
132	132-9	3.5	−6.0
133	133-1	3.6	−6.2
133	133-2	2.8	−6.8
134	134-1	2.9	−5.9
134	134-2	2.9	−6.5
135	135-1	3.8	−5.8
135	135-2	3.1	−8.8
135	135-3	3.6	−6.4
135	135-4	2.7	−6.0
136	136-1	4.3	−5.0
136	136-10	4.1	−6.4
136	136-11	4.5	−5.5
136	136-13	3.6	−6.6
136	136-14	3.5	−6.6
136	136-15	4.2	−5.7
136	136-16	3.4	−6.0
136	136-18	3.4	−5.6
136	136-2	3.4	−5.1
136	136-3	4.2	−5.8
136	136-4	3.5	−6.3
136	136-5	4.1	−5.5
136	136-6	3.9	−7.5
136	136-8	3.4	−6.3
136	136-9	4.2	−5.7
137	137-1	4.2	−11.4
137	137-2	4.0	−8.1
142	142-1	3.7	−9.6
142	142-4	3.9	−9.6
154	154-1	3.2	−6.5
154	154-2	3.6	−7.1
155	155-2	3.0	−8.3
156	156-1	3.1	−6.2
156	156-2	3.2	−7.0
157	157-1	3.4	−6.8
158	158-2	2.9	−7.7
158	158-3	3.4	−7.4
158	158-4	2.9	−6.0
159	159-1	4.6	−8.3
162	162-1	3.4	−6.7
162	162-2	3.4	−8.0
164	164-1	3.4	−9.3
164	164-2	2.8	−5.4
164	164-3	2.6	−6.4
164	164-4	2.7	−6.1
164	164-6	2.8	−8.9
164	164-7	2.7	−5.8
165	165-1	2.6	−5.0
168	168-1	3.3	−7.8
169	169-1	4.5	−7.7
169	169-2	4.7	−7.3
169	169-3	4.4	−5.7
169	169-4	4.4	−7.2
169	169-5	4.9	−7.0
169	169-6	5.1	−6.5
171	171-1	4.1	−7.3
171	171-2	3.9	−8.4
172	172-1	3.5	−8.0

in the Appendix A presents the results of the stable isotope analyses, with averaged data from multiple runs marked by asterisks. These results are displayed graphically in the scatter-plot diagram in Figure 1. As illustrated, the sample distribution spans the entire 90% probability ellipsoid defined by Matthews et al. [21], while also extending the δ¹³C range further into positive values. A notable pattern emerges when the data are grouped according to their respective Marble Units (Figure 5). The scatter-plot reveals that quarries within Marble Units 1 and 2 exhibit higher δ¹³C values compared to those in Marble Unit 3. This observation is statistically supported by an independent sample t-test, which confirms—with over 95% confidence—that the mean δ¹³C value of the combined Marble Units 1 and 2 differs significantly from that of Marble Unit 3 (Figure 6).

While there is some overlap in δ¹³C values (2‰–3.5‰) among the three units, the elevated δ¹³C values of Marble Units 1 and 2 distinctly separate from the narrower range seen in Marble Unit 3. Since Marble Unit 1 lacks archaeological quarries, any archaeological Pentelic sample with a high δ¹³C value must have originated from Marble Unit 2. Furthermore, only Marble Unit 3 samples exhibit relatively high δ¹⁸O values (>−5.5‰) combined with relatively low δ¹³C values (<3.5‰), distinguishing this unit from the others.

The scatter-plot diagrams in Figure 7, Figure 8 and Figure 9 display the sample distributions of all the quarries identified as ancient during the initial field survey, as well as Quarry Π69. Quarry Π69 is included in this subgroup because of its very large size, its location in the apparent center of ancient quarrying activities, and because it most likely is an ancient quarry with no physical evidence preserved.

The scatter-plot diagram in Figure 7 represents all the upslope quarries within Marble Unit 3. Quarries Π88, Π89, and Π90, located in the northern upslope section known as Aspra Marmara [41], exhibit the highest δ¹⁸O values in the database and are denoted by open symbols. The Aspra Marmara marbles are qualitatively among the highest quality, whitest, and least impure marbles within the entire quarry field. Furthermore, the elevated δ¹⁸O values distinguish these quarries from all others in the region. Quarry Π85 has one sample with a high δ¹⁸O value (−4.3‰), but this is a statistical outlier, as it falls outside the 95% confidence interval for that quarry. Some samples from Quarry Π88 and one from Quarry Π89 trend toward the lower δ¹⁸O values more typical of other Pentelic quarries. Notably, the lower δ¹⁸O values from Quarry Π88 originate from its southern end, where the marble quality is poorer and exhibits significant weathering, alteration, and discoloration.

The δ¹⁸O signatures of the other Marble Unit 3 upslope quarries—Π84, Π85, Π87, and Π91—do not display the same high values or patterns as Quarries Π88, Π89, and Π90. Represented by solid symbols in Figure 7, these quarries exhibit δ18O values ranging from −9.4‰ to −5.5‰, with Quarry Π85 containing a single statistical outlier (−4.3‰). The maximum δ¹⁸O value for this group is still higher than most other reference samples in the database. Quarry Π91, an extensive quarry, shares a δ¹⁸O range similar to the grouping of Quarries Π84, Π85, and Π87.

The δ¹³C values for Quarries Π84, Π85, and Π87 show less variance, ranging from 2.7‰ to 3.2‰, with one outlier measuring 4.2‰. For Quarry Π91, δ¹³C values range from 2.4‰ to 3.2‰, with one anomalous sample at 1.7‰, which falls well outside the 95% confidence interval for the quarry. No distinguishable physical features were identified in the quarry space to account for either of the δ¹³C outliers.

The scatter-plot diagram in Figure 8 represents the remaining ancient quarries in Marble Unit 3. The closed symbols indicate the quarries in the mid-slope region, the crossed symbols represent the three ancient quarries on the lower slope of the quarry field, and the open symbols reflect Quarry Π55, the renowned Spilia Divali quarry. Quarry Π55 is isolated, as it has been identified by M. Korres as having been the quarry to supply essentially all the marble for the construction of the Parthenon [34]. Comparing the isotope profiles of the large, mid-slope quarries reveals that they share a similar profile with one another. The δ¹³C ratios representing Quarry Π70 fill a narrow range between 2.7‰ and 3.1‰ and are roughly 0.2‰ below the average δ¹³C value for Quarry Π69. The δ¹⁸O values range between –8.7‰ and –4.1‰. Further downslope ancient quarries, Π64 and Π67, have a very similar distribution of δ¹³C and δ¹⁸O ratios.

Quarry Π55 (Figure 8, open symbols) exhibits a narrow range of five measurements with one anomalous sample showing a relatively higher δ¹⁸O value. The box plot in Figure 6 highlights the extent to which this value deviates as an outlier. The main cluster of samples lies within a δ¹³C range of 2.7‰ to 3.1‰ and a δ¹⁸O range of −9.1‰ to −8.4‰. Two independent analyses of the outlier produced an averaged stable isotope signature of 2.9‰ for δ¹³C and −4.8‰ for δ¹⁸O. Upon reconciling sample locations with the isotope data, it is evident that the outlier originated from the quarry’s northeast wall. This section, lacking evidence of ancient extraction, was likely exposed in the modern area, possibly to supply marble for the construction of the royal palace in Athens in 1846 [42]. The remaining grouped samples were collected from the quarry’s west wall, which is marked extensively by ancient tool grooves.

Further examination of Figure 8 reveals that the small ancient Quarries Π39, Π48, and Π49, located on the lower flanks of Mount Pentelikon’s southern slope and within Marble Unit 3, exhibit more negative δ¹⁸O values combined with δ¹³C values near 3.0‰. In contrast, Quarries Π35 and Π38—the lowest sampled ancient quarries in the study area—deviate from this trend. These quarries display higher δ¹³C values, ranging between 4.4‰ and 4.7‰, and δ¹⁸O values between −6.6‰ and −5.7‰, distinguishing them from the broader Marble Unit 3 dataset.

Figure 9 presents the δ¹³C–δ¹⁸O scatter-plot diagram for all the ancient quarries in Marble Unit 2. Quarries Π95 and Π96, located at the top of Mount Pentelikon’s southern slope and represented by open symbols, display a significantly higher δ¹³C range compared to the upslope quarries of Marble Unit 3 (Figure 5). Additionally, these two quarries exhibit lower maximum δ¹⁸O values of −6.8‰ and −6.4‰, respectively, when compared to neighboring Marble Unit 3 quarries. To the west and north of Quarry Π70, across the thick mica schist layer that separates Marble Unit 2 from Marble Unit 3, lie Quarries Π103 and Π105. These quarries show a narrow δ¹³C range and a relatively small δ¹⁸O range, with values at the lower end of the Pentelic marble spectrum: δ¹³C ranges from 2.6‰ to 3.1‰, and δ¹⁸O ranges from −9.9‰ to −7.3‰. Further downslope, along the schist interlayer, Quarry Π107 shows a similar δ¹⁸O range (−8.9‰ to −7.7‰) but a wider δ¹³C range (2.2‰ to 3.9‰). The highest value (3.9‰) falls outside the 95% confidence interval for this quarry.

The seven samples from Quarry Π54 exhibit consistently low δ¹⁸O values (−8.9‰ to −7.0‰) but show two distinct δ¹³C groups. One group has δ¹³C values below 3.0‰, while the other has values above 4.0‰. A reconciliation of sample locations with isotopic data reveals that the lower δ¹³C values correspond to samples from the quarry’s eastern wall, adjacent to the schist interlayer. In contrast, the higher δ¹³C values were collected from the western wall, indicating a significant δ¹³C increase of over 1.0‰ across the narrow 15–20 m width of the quarry. A similar pattern is observed in Quarry Π107, where the sample Π107-1, collected from the west wall furthest from the schist interlayer, displays the highest δ¹³C value (3.9‰).

The stable isotope data from the smaller ancient quarries on the lower slopes of Marble Unit 2, represented by crossed symbols in Figure 9, provide further insight. Quarry Π20 shows a wide δ¹⁸O variation of nearly 2‰, yet both Quarries Π20 and Π21 maintain isotopic signatures characterized by relatively low δ¹⁸O and high δ¹³C values. The single sample from Quarry Π23 has a δ¹³C value comparable to those of the lower Marble Unit 3 quarries. The close proximity of these three quarries, combined with their wide isotopic range, suggests that the upper portion of Marble Unit 2 may not exhibit the same degree of isotopic distinctiveness as Marble Unit 3.

4. Discussion

There is ongoing debate regarding the specific quarries that supplied marble for the construction of the Parthenon. M. Korres, the former architect of the Parthenon reconstruction project, proposed that the marble used in the Parthenon was sourced from the renowned Spilia quarry, Quarry Π55. Korres suggests that the ancient Greeks systematically began quarrying on the lower slopes of Mount Pentelikon and progressively moved upslope as the lower quarries were exhausted [43]. In addition, Quarry Π55 is the lowest surviving large ancient quarry, notable for its widely spaced joints, which facilitated the extraction of large, dimensionally uniform marble blocks. Given that the Parthenon was the first monumental construction project to use Pentelic marble, Korres argues that it would have been logical to source its material from this lowest and most accessible quarry. This hypothesis contrasts with the findings of Matthews et al. [21], who propose that the marble used for the Parthenon, or at least parts of the sculptural program, was sourced from the upper quarries in Aspra Marmara.

To address this debate and to test the applicability of the Pentelic stable isotope database, stable isotope values from samples from the architectural blocks and sculptural members of the Parthenon were compared to those in the database presented here. Matthews et al. published the ∂¹³C and ∂¹⁸O values of seven south metope, seventeen frieze, and six west pediment samples from the Parthenon Marbles currently housed at the British Museum [21]. In addition to the published data, this report presents previously unpublished stable isotope data from recently collected and analyzed architectural blocks from the Parthenon, including material from the original colonnade, west wall, and northwest corner.

The scatter-plot in Figure 10 graphically shows the stable isotope values from the original Classical period colonnade, the west wall, and northwest architectural blocks in squares, each represented by orange, blue, and purple, respectively. The green triangles show the frieze values published by Matthews et al., as well as west pediment samples being reported here for the first time. Comparing the stable isotope ratios of the sculptural monuments to those of the database reveals that the high δ¹⁸O values obtained for the Parthenon sculptures correlate with the high values obtained for Quarries Π88, Π89 and Π90 (Figure 10), the Aspra Marmara quarries. The architectural blocks have more variety and span much of the Pentelic marble database, suggesting that the builders were using marble from various quarry pits. Yet, it is interesting that the northwest architectural blocks group relatively close to one another, with two outliers. Might this suggest, then, that marbles for the colonnade may have been preferentially selected from one quarry? Even more revealing is that the west pediment material groups nicely with the higher ∂¹⁸O values of the upper slope of Marble Unit 3. The Parthenon’s building accounts indicate that the temple’s superstructure was completed in time for the dedication of the cult statue Athena Parthenos in 438/437 BCE, yet the pediment sculptural program was not completed until at least 433/432 BCE [21,44]. The west pediment stable isotope data suggest, then, that the Aspra Marmara quarries were preferentially selected to supply marble for this final phase of the sculptural program, which occurred several years after the completion of the Parthenon’s superstructure. The dearth of architectural blocks from Aspra Marmara suggests that the high-quality marble from this area of Pendeli was preferentially limited to sculptural activity.

Importantly, the stable isotope data reveal that in the Greek Classical period, quarrying on Mt. Pentelikon did not take place systematically from the lower quarries to the upper quarries. Rather, as the isotope record of the Parthenon demonstrates, builders exploited at least two, if not more, regions in the greater Pentelic quarry area. Marble from Aspra Marmara, being purer and whiter, appears to have been favored for sculptural components of the monument, whereas other Pentelic quarries supplied marble for the superstructure. The Spilia Divali quarry, having relatively long distances between natural joints, appears to have been one of the quarries selected to supply the larger elements of the structure, such as column drums and lintels.

5. Conclusions

Prior to this extensive study of the Pentelic quarry region, the heterogeneous nature of the Pentelic stable isotope field was perceived to be a limiting factor on the resolution of the Pentelic quarry database. The wide ranges for δ¹³C and δ¹⁸O were seen as problematic, since the large signature field increased the likelihood that the Pentelic quarry field would overlap signature fields of other quarry districts. Overlapping signature fields make provenance assignments more difficult for those artifacts that fall within the overlap. However, this research has illustrated that the wide range of isotope values at Mt. Pentelikon can be an asset rather than a hindrance.

With detailed surveying and systematic and extensive sampling, the presented database improves the discriminatory ability of stable isotopic analysis for marble provenance studies. Rather than treating the entire Pentelic quarry region as a homogenous whole, the presented database distinguishes between three mappable and exploited marble units. By correlating individual samples within the Marble Units to their quarry sample spots it is possible to recognize stable isotopic sub-regions within the quarry area. In some special cases individual quarries or groups of quarries can be identified. The utility of the database is illustrated by determining the probable quarry or group of quarries from which the Parthenon sculptures were extracted. The extensive field survey and sampling of the quarry region provided the information necessary to investigate and identify intra-quarry isotopic variation. Future characterization projects of other ancient quarry districts should undertake similar extensive sampling protocols.