Carving processes

We documented carving marks on the sculptures’ surfaces using a digital camera and then associated these with the tools used to make them. The results are reported based on their micro-chronological relationships and functions so as to reconstruct the sequence of tool use with reference to the traditional patterns of work (Adam Reference Adam1966; Rockwell Reference Rockwell1993: 12; Palagia Reference Palagia and Palagia2008; Wootton et al. Reference Wootton, Russell and Rockwell2013). Limited evidence was found for the early stages of marble working because the incremental progression through the sculptural phases usually erases each preceding step. Point-chisel marks were observed, however, on the underside of blocks (see Adam Reference Adam1966: 14–15 for comparable marks on Figures WP N (Iris?) and EP E (Demeter?); Figure 1a–b). Here, the roughing-out with the use of a point chisel was subsequently smoothed over by a claw chisel (Palagia Reference Palagia1993: 21; Figure 1c).

Figure 1. a–b) Evidence for use of the point chisel on the underside of Figure EP G, and front of E and F, respectively; c) evidence of use of the claw chisel on the back of Figures EP L and M; d) extensive use of the drill in the drapery folds of Figure M (figure by Katarzyna Weglowska) ©Trustees of the British Museum.

These two chisels were used in association with a third tool, the drill. This was used for removing stone during rough-shaping, then drawing out the depth necessary to create the three-dimensional forms (see Palagia (Reference Palagia1993: 41) for the drill on the WP; Figures 1d, 2a–d). The carvers made use of at least three different-sized drills: small, medium and large, from 3–10mm. Adam (Reference Adam1966: 51) identified drills of 4mm and 8mm on AcrM 1363, compared with 2mm and 4mm on the West Pediment at Olympia. During rough shaping, a series of large holes were drilled both perpendicular to the surface and on the diagonal to create a honeycomb effect; the bridges between these spaces were then knocked out with a point chisel (Adam Reference Adam1966: 50–57). The desired depth was attained by repeating this process. Finer drills were used for more detailed areas. Traces survive in the undercut folds of the drapery, and within the folds themselves, where it was presumably deemed unnecessary to remove these working marks completely (Figure 2a).

Figure 2. a) Drill marks surviving in the folds of Figure EP F; b–c) the drill and the claw used in tandem for the drapery of Figures EP L and M; d) evidence of the drill followed by fine-tool finishing and use of the rasp on Figure EP M (figure by Katarzyna Weglowska) ©Trustees of the British Museum.

The claw chisel was used for both rough- and fine-shaping. Marks are apparent on the drapery, smoothing the rippling forms created by the deep cuts of the drill, and on flatter areas such as the backs of the figures (compare with Adams’ (Reference Adam1966: 20) observations on EP A). The size of the claw used depended on the degree of detail needed, with the number of teeth ranging from five to seven (Paga Reference Paga2015: 171; Figures 2b–c).

The carvers worked in a non-linear manner during the fine-shaping of the sculptures. Rather than bringing the whole block gradually towards its final form, they worked by area, taking each towards a high level of finish before moving onto the next section. The tools and working processes were adjusted according to the intended form (Figure 2d). This is most obvious when comparing the heavily draped mantles and the more subtly curving tunics. On Figure EP L (Dione?), the finishing of the mantle demonstrates fewer fine-tool marks, such as those made by the rasp and small channelling tools; however, these are more apparent on the tunic, where the more delicate tooling accentuates a lighter, flowing linen material. In contrast, the mantle was carved to give a greater sense of volume; it is robust in its execution producing a form and finish more akin to woven wool.

Evidence for surface polishing on Figures EP L and M (Aphrodite?) was more difficult to discern due to weathering, with the exception of the neckline of M, which is very smooth, indicating that the flesh areas were further differentiated from the drapery. This smoothness can also be seen on the neck of Figure EP A (Helios), which has a polished surface (Haynes Reference Haynes1975: 131; Jenkins Reference Jenkins2001: 23). There is evidence for the point chisel being used late in the production sequence, with the backs of EP L and M apparently modified during the original installation in the pediment. In addition, there are drill holes for fixing attachments to the sculptures, such as those for a bracelet and a necklace on Figure EP M.

The figures therefore demonstrate that the carvers adjusted their approach according to the specific subject matter, for example, considering the different garments and their materials, in particular their texture, weight and hang. The focus of the carvers was on truth to form. The final stage was the addition of colour. Carving and polychromy had a symbiotic relationship (Heilmeyer et al. Reference Heilmeyer, Massmann, Germann, Koch-Brinkmann and Brinkmann2014; Blume Reference Blume2015); there was a correlation between a surface's finish and the application of colour, but a special preparation such as ‘keying’ was not required to facilitate the adhesion of pigments (Adam Reference Adam1966: 74–5).

Polychromy and its uses

We detected Egyptian blue, in combination with two white pigments, gypsum and probably bone white, on 11 pedimental sculptures and one figure from the frieze. A selection of the most significant occurrences is presented here in detail with images and a complete list of instances in which pigments were identified is reported in Table S1. A purple pigment, at this stage tentatively identified as an anthraquinone-containing colourant, was also discovered on Figure EP M.

Water: Egyptian blue was detected on the area of water from which Figure EP A is shown rising (Figure 3). Blue for waves is commonly found on terracotta productions; for example, the wavelets on the Aphrodite in a Shell at the National Archaeological Museum in Athens (inv. no. EAM 2060).

Figure 3. Figure EP A: a–b) visible and c) VIL images of the back of the sculpture (figure b–c by Giovanni Verri) ©Trustees of the British Museum.

Snake skin: a few grains of Egyptian blue were also observed on the serpent tail of Figure WP B (Kekrops?). While the interpretation of extremely small quantities of pigment can be problematic, the use in antiquity of blue for snakeskin and sea creatures is well established as noted, for example, on a ritual basin from Ascoli Satriano, Italy (Wallert Reference Wallert1995: 178).

Background/empty space: a small amount of Egyptian blue was also detected in the background space between the rear legs of the stools/boxes on which Figures EP E and F (Demeter? and Persephone?) are shown sitting. This area probably represents ‘empty space/air’ between the legs of the seat. A considerable quantity of Egyptian blue was found beneath the peplos (dress) of Figure EP G (Hebe?; Figure 4), but no pigment on the dress itself. Despite the interpretative complications caused by the partial cleaning of this figure in the 1930s (Jenkins Reference Jenkins2001: 24–5), it is possible that the pigment was not used for the peplos, but only to represent the empty space below it. This use of blue is well attested in antiquity. Solid-blue backgrounds behind figural representations have been found, for example, in the space between the legs of the Ascoli Satriano basin mentioned above, and on the so-called ‘Throne of Eurydice’ from Vergina, Greece, to indicate the spaces between its legs. It is also applied in a more ‘naturalistic’ manner as a backdrop to the image on the backrest of that throne, as well as to the landscape of the hunt frieze from the ‘Tomb of Philip II’ at the same site. The appearance of blue on the Parthenon's pedimental sculpture may be related to the original colour of the architectural background against which the sculptures were placed.

Figure 4. Figure EP G: a) visible and b) VIL images (figure (b) by Giovanni Verri) ©Trustees of the British Museum.

Textiles: extensive areas of Egyptian blue were detected on garments. In particular, the mantle of EP L shows clear traces of Egyptian blue on the front of the figure (Figure 5), as well as the side and rear (Figures S1–S4 in OSM). We interpret these traces as depicting a woven design, representing dyed wool yarns in tapestry weave, but other techniques cannot be excluded (Spantidaki Reference Spantidaki2016: 78–96). The reading of the woven composition is complicated by the loss of areas of the original surface to erosion, and by other areas covered with a black crust where the blue pigment cannot be detected with the instrumentation used in this study. Nonetheless, all surviving traces of blue are exclusively found at the same distance from the lower edge of the mantle, suggesting a wide border. While the exact characterisation of the woven design is unclear, a cautious interpretation of the results can be attempted.

Figure 5. Figure EP L: visible and VIL images of the front. Details and line drawings of the design are in Figures S1–S4 in the online supplementary material (figure by Giovanni Verri) ©Trustees of the British Museum.

The analysis of the outlines of the various shapes visible on the mantle suggests that Egyptian blue may have been used as a background to a decorative design. On the garment over Figure EP L's left knee, it possible to discern the outline of bent human legs, as if running to the right (see athlete on the vase at the Boston Museum of Fine Arts (inv. no. MFA 96.720) for a near identical position). Over the top and side of EP L's right leg, the garment is painted with parallel bands of Egyptian blue. By the figure's knee, what appears to be an outstretched hand is visible, and a foot wearing a shoe is represented on the garment covering EP L's calf (See Figures S1-S4 in the OSM).

The shapes defined by the Egyptian blue background suggest that the woven decoration depicted was figurative, rather than geometric or floral. The absence of blue along the edge of the garment also suggests that the decoration was divided in woven panels or friezes, such as that on a vase from Chiusi, Italy (inv. no. Chiusi 1831). The iconography includes Penelope at her loom with a woven motif of a running figure with bent legs, similar to that visible on EP L's knee. The dimensions of the legs of the winged figure represented on Penelope's textile relative to Penelope's own knee roughly match those proportions of EP L and the leg shown on the mantle. The most convincing comparative example is found on an Attic vase at the British Museum (inv. no. BM 1873,0820.375; Figure 6). It shows Demeter wearing a mantle decorated with horizontal bands representing, among other decorative elements, winged human figures running with bent legs. There, the painter has depicted the figures on Demeter's mantle as they would have behaved on a physical garment, with at least some figures disappearing into the folds.

Figure 6. Red-figure skyphos showing figurative textile (inv. no. BM 1873,0820.375 c. 500–480 BC) (©Trustees of the British Museum).

The folds of the mantle on EP L's lap show a complex distribution of Egyptian blue, possibly corresponding to appearing and disappearing figures, only partially represented. Similar, but later, representations of figures with bent legs are found on textiles from archaeological contexts, such as the fragment preserved at the Boston Museum of Fine Arts (inv. no. MFA 53.18). The shape defined by Egyptian blue on EP L's calf is tentatively interpreted as a sandalled foot running parallel to, but not touching, the band where no Egyptian blue was detected. While straps are visible, the footwear does not seem to be a standard sandal, but rather a closed shoe with solid uppers and leather straps. A comparable example can be found on Figure EP N (the Seer) of the Temple of Zeus at Olympia (Dohan-Morrow Reference Dohan Morrow1985: 51). The network of straps, as well as the half-moon shape at the front of EP L's shoe, is executed using Egyptian blue, representing what is probably leather. The use of blue pigment in this way and mixed with other pigments to represent different hues (Verri et al. Reference Verri, Opper, Lazzarini, Liverani and Santamaria2014), is also found on the brown leather sandals of the philosopher pointing at a globe in the ‘Tomb of the Philosophers’ in Pella, Greece (Brecoulaki et al. Reference Brecoulaki, Verri, Kalaitzi, Maniatis and Lilimpaki-Akamati2023: fig. 19). Bright blue boots/shoes are represented using Egyptian blue on a ceramic lekythos in Athens (inv. no. EAM 2059). The half-moon feature at the front of the shoe could be an embellishment, as observed on Treu's drawing of leather straps decorated with a palmette (Dohan-Morrow Reference Dohan Morrow1985: 51) and the larger palmette at the front of the shoe of the Seer from Olympia (Brinkmann Reference Brinkmann2003: fig. 319.1).

EP L's garment appears to have been highly decorated, and therefore does not correspond to what is commonly observed in other representations of Athenian fashion contemporaneous to the construction of the Parthenon (Brinkmann Reference Brinkmann and Liverani2004: 154–6). Figure EP L represents a goddess but deities’ garments do not necessarily follow fashion rules observed for mortals. In this case, the use of elaborate decorative designs on the dress may hark back to earlier historical periods (e.g. Figure 6, which is dated to c. 490 BC), when decoration on textiles was common even for humans, or it might simply evoke the grandeur of a numinous presence. Even if not necessarily worn by mortals, elaborate textiles were nonetheless present during the middle of the fifth century BC; the tradition of annually weaving a new peplos with figurative elements, to be dedicated to Athena Polias, attests to a continued production of richly decorated textiles (Stamatopoulou Reference Stamatopoulou2012: note 1); epigraphic evidence from the Sanctuary of Artemis at Brauron confirms the existence of figured textiles in the Classical period (Llewellyn-Jones Reference Llewellyn-Jones2003: 140; Cleland Reference Cleland2005: 114–15). A desire to distinguish deliberately between the representation on a public monument of the garments worn by deities and the clothes worn by ordinary people on the streets of Athens may explain the decorative designs found on EP L's garment. The painted ornamentation may have even aided the reading of the pediments’ iconography.

Figures EP L and M are shown reclining on a rock softened by a draped cloth. Egyptian blue was detected on the reverse of this group. It forms a line equidistant from the edge of the cloth, suggesting a border running along its length (Figure 7a). A small sample taken from this line shows that the blue pigment was mixed with and layered over a phosphate-rich white matrix. Raman spectrometry reveals calcium phosphate under a thick, calcium oxalate layer, now orange because of iron impurities. Discussions on the nature of this orange layer, which is likely of natural origin and therefore the product of environmental deterioration, are beyond the scope of this study (Palagia Reference Palagia, Pike, Pensabene and Gasparini2015). It is worth noticing, however, that the oxalate layer formed when the Egyptian blue layer was already highly fragmentary (Figure S5), strongly supporting the natural origin of the orange layer. SEM investigations of this sample show a uniform distribution of phosphorus suggesting an intentional use of a phosphate-rich material, such as bone white pigment. The use of a nearly identical stratigraphy on the Mausoleum of Halicarnassus (Gratziu et al. Reference Gratziu, Jenkins and Middleton1989: 324) may not be coincidental and could be related to the pigments’ resistance to the elements for outdoor sculpture.

Figure 7. Figure EP L and M: VIL image of a) the back and b) visible and c) VIL image of a detail of the front by L's feet. The white arrow in (a) shows the sample location for Raman and SEM-EDX analysis (figure by Giovanni Verri) ©Trustees of the British Museum.

On the border of the same cloth at the front of the group, in the protected area under the knee of Figure EP L, we detected a palmette (Figure 7b–c). Lower down, several squares appear to run along the bottom edge of the textile. We observed a similar pattern on Figures WP T (Kreusa?) and EP K (Hestia?), where extensive traces of Egyptian blue were detected on these figures’ mantles. An example of a battlement pattern on a mantle, together with borders with palmettes, can be seen on a kylix representing Hera in the Berlin Antikensammlung (inv. no. Berlin 2685).

We also detected bright purple particles at the point where the cloth on which EP M sits meets the figure's mantle at the back of the block (Figure 8). It is impossible, however, to establish to which of the two textiles these purple traces belong. The particles show a translucency and gloss typical of organic colourants and are embedded in a white matrix. A variety of purple colourants—as well as combinations of red and blue—were available in antiquity and some are discussed in the Leyden and Stockholm papyri (a set of Greek texts of Roman imperial date from Egyptian Thebes which record various chemical and technical processes; Jensen Reference Jensen2008). The FORS spectrum of the purple colour, together with the absence of bromine in the XRF spectrum makes the use of true purple extracted from sea snails from the Muricidae family (also previously known as Murex)—one the most precious and coveted colours in antiquity—unlikely. When compared with other available purple dyes, the purple colour detected here shows a relatively good accordance with anthraquinones, but these may be only one of the components present. The investigations are ongoing. The desirability of the colour purple in the Classical period is well attested in textual sources (e.g. the records of Artemis Brauronia mentioned above), as well as in high-status Athenian coroplastic forms (e.g. inv. no. EAM 2059) and marble production (Brecoulaki et al. Reference Brecoulaki, Kavvadias, Verri, Østergaard and Nielsen2014).

Figure 8. Detail of the purple organic colourant on Figure EP L. The inset shows the location of the area. Absorption spectrum of the colourant in comparison to standards (for orcein/folium see Aceto et al. Reference Aceto, Angelo, Fenoglio, Idone, Gulmini, Picollo, Ricciardi and Delaney2014, for alkanet Angelini et al. Reference Angelini, Tozzi, Bracci, Quercioli, Radicati and Picollo2010) (figure by Giovanni Verri) ©Trustees of the British Museum.

We also detected Egyptian blue on the beltline of Figure WP N (Figure 9). This may correspond with the representation of a textile band or strap. Alternatively, the recessed carving of the beltline, in conjunction with a small hole at the back may indicate the fitting of a metal belt (for comparison, see the small hole intended to secure a metal belt on Metope North XXXII, Acropolis Museum, Athens, inv. no. AKR 20706). The hole may also be the remains of drill-work for a fold, rather than for an attachment. The presence of blue in an area that might have been covered by a metal attachment is difficult to understand. It could be a later repainting or a blue background for a perforated metal belt. Repainting seems less likely because the blue at the back is in an inaccessible area. Here, SEM and Raman analysis reveal the presence of gypsum, rather than bone white, as the white pigment with which the blue was mixed. Whether the presence of two different types of white pigments corresponds to different workshops, an intentional artistic choice or the result of a later intervention is currently unknown. Gypsum can form from calcium carbonate in the presence of sulphuric acid, as in the case of ‘acid rain’; however, while the paint layer appears well-bound and cohesive—unlike a typically powdery alteration product—the presence of calcium sulphate instead of calcium phosphate, as found in the other samples, presents the same questions and uncertainties highlighted here.

Figure 9. Figure WP N: VIL image. The white arrow shows the sample location for Raman and SEM-EDX analysis (figure by Giovanni Verri) ©Trustees of the British Museum.

Finally, we detected Egyptian blue in the recesses of the folds of Figure 32 on EF Block V, (Figure 10a–b). These may be remnants of a representation of a garment made with blue-dyed wool, as with WP N's belt above. The analyses of ancient painting practices, however, show the difficulty involved in interpreting scant traces of colour and this example could have a more complex interpretation. For example, on the Amathus stele in the British Museum (inv. no. BM 1894,1101.717), Egyptian blue was used to represent depth in the folds of both white and coloured textiles (Figure 10c–d). The blue found on Figure 32 could have been a solid application of colour or part of a shading technique; a sample indicates the presence of a phosphate-rich white pigment, with a stratigraphic distribution similar to that observed on the East Pediment.

Figure 10. Figure East Frieze, Block V, Figure 32. a) visible and b) VIL image. The white arrow in (b) shows the sample location for Raman and SEM-EDX analysis. Hellenistic stele from Amathus(?), Cyprus, showing two women (inv. no. BM 1894,1101.717, fourth–third century BC); c) visible and d) VIL images (figures b–d by Giovanni Verri) ©Trustees of the British Museum.