In this month’s ‘Science Notes’, we will look at the way three-dimensional (3D) imaging can be used to study the accuracy of plaster casts created by 19th-century archaeologists to record and preserve ancient monuments. A recent study published in Antiquity (https://doi.org/10.15184/aqy.2019.179) compared casts taken of parts of the Parthenon in Athens in the early 1800s and the 1870s with the original sculptures in their current state, in order to determine the reliability of the casts and to help monitor the sculptures’ deterioration over time.

Deviation maps for the Elgin casts of figures 5 and 6 of the West Frieze
On the left, deviation maps show Figure 5 (above) and Figure 6 (below) of West Frieze III at 5mm maximum deviation: greyscale >5mm deviation; green <1.5mm deviation. On the right, the images show the same figures at 1mm maximum deviation. Data applied to Elgin Cast. [Image: E Payne. 3D imaging conducted courtesy of the Trustees of the British Museum and the Acropolis Museum. Antiquity Publications Ltd.]

The 3D scanning and digitisation of plaster casts has become more common in recent years, as their value has become more widely recognised. However, these projects rely on the assumption that the casts are accurate substitutes for the original monuments. The study conducted by Emma Payne, King’s College London, was designed to test this assumption, as well as assessing whether the casts preserve information that has since been lost from the original sculptures.

The accuracy of the casts was assessed using 3D imaging because it allows quantitative comparisons of surface morphology to be made without being affected by other factors that hinder photographic comparisons, such as lighting. It also meant that it would be easier to compare the original sculptures directly with the casts. The focus of the study is the West Frieze of the Parthenon. Lord Elgin commissioned moulds and plaster casts of this section in 1801-1802 (at the same time as he removed many of the originals). By 1872, however, the earlier moulds had become worn and a new set of casts was created by Charles Merlin. Five sections of the frieze were identified that showed differences between the casts and originals, and 3D imaging of these sections was carried out on both sets of casts and on the original sculptures, which remained in situ on the temple until 1993, when they were moved to the Acropolis Museum.

The 3D imaging was accomplished using a Breuckmann SmartSCAN with XY (horizontal) resolution of up to 140 microns. This device uses structured white-light scanning to create extremely accurate point clouds of data from the surface of an object, digitally capturing its exact shape and size. This data was processed using OPTOCAT software, Breuckmann’s 3D-image-processing software, to create stereolithography (STL) files, which describe the surface geometry of an object and were used to produce a 3D representation.

The three sets of 3D representations were compared to ascertain how much the casts and originals differed. In addition, the surface texture of the statues was examined using Gaussian curvature, a calculation of curvature that characterises surface-roughness (where 0 is perfectly smooth, positive numbers indicate concave features, and negatives indicate convex features), and mean curvature, which calculates the average curvature so as to illustrate larger features more effectively.

3D model of figure 23 with Gaussian curvature indicated
3D models of Figure 23 with Gaussian curvature indicated: at the top, Elgin Cast; in the middle, Merlin Cast; and at the bottom, the original sculpture. In the casts, the textured surface of the clothing and hair is clearly revealed; in the original, surface noise caused by the reflective quality of the marble prevents effective characterisation of texture. [Image: E Payne. 3D imaging conducted courtesy of the Trustees of the British Museum and the Acropolis Museum. Antiquity Publications Ltd.]

The analysis of the images suggested subtle differences between the casts and the original sculptures as a result of moulding practices, as well as a small amount of weathering of the originals after the moulds were taken and more significant losses through vandalism of the original sculptures. However, both sets of casts were found to be generally very accurate copies of the original friezes at the time they were moulded. The Merlin Casts were found to be the most accurate, although the Elgin Casts are also within a 1mm level of deviation from the originals when areas of greater deviation caused by vandalism or weathering rather than poor moulding practice are excluded.

It was believed that the deterioration of the original statues occurred more rapidly in the period between the creation of the Elgin and the Merlin Casts, as the earlier casts are crisper and sharper, with pieces present that are missing from both the Merlin Casts and the present-day originals. However, the study of the surface texture of the Elgin casts has revealed areas with a clay-like texture in a few statues, suggesting that additions may have been made to the moulds in order to improve their appearance and replace losses that occurred before the casts were created. Despite this, most of the Elgin Casts do appear to be accurate replicas of the originals as they were in the early 19th century.

This study shows that 3D imaging can be used effectively to measure and visualise differences between original statues and casts, as well as supporting the use of 19th-century casts as an archaeological resource.


This article appears in issue 360 of Current Archaeology. To find out more about subscribing to CA magazine, click here.

Leave a Reply