Minerva in Colours: First Results on a Polychrome Roman Sculpture from Carnuntum (Pannonia)

Round 1

Reviewer 1 Report

General comments

The paper addresses an original topic, not sufficiently covered up to now in literature, referring to the polychromy of ancient Roman sculptures from the western and northern European provinces of the Roman Empire, for which evidence of paint residues on excavated sculptures or architectural elements is rare, if any. Therefore, this paper is of high impact for a wide audience interested in Polychromy studies.

While the study presented in the paper focuses on the materials and techniques used in a particular object, it would indeed be interesting to determine whether the findings follow the main, to our knowledge, stream of ancient Roman polychromy techniques and materials or whether they represent regional characteristics, attributed to a provincial style. The authors should consider and refer to this question, in the discussion of the results. And, also, to what extent this study brings new knowledge to our current understanding of the polychromy of ancient sculptures, more generally?

Referring to the followed methodology, while the proposed methodological approach in the paper includes both noninvasive imaging techniques and microanalytical methods applied to microsamples, it appears that the results obtained mainly come from the microanalytical approach, from the observation and interpretation of the analysed samples and respective cross-sections. The data obtained and the potential of the noninvasive approach is less emphasized or exploited.

 

Specific comments

3.1. Results of non-invasive methods

-          MSI is presented as an important step for the mapping of colour remains and the screening of pigments. In the set of images presented in Fig. 10, in certain of the images, in particular the (b), (c) and (d) corresponding to ultraviolet reflected light (UVR); infrared reflected
light (IRR);  and visible induced visible luminescence (VIVL); there is an intense colour prevailing all over the object, related to the acquisition and WB correction that is likely acting as a mask limiting the readability of the images, and therefore the extraction of any relevant information on the original colour traces. How do the authors deal with this handicap, and how the images could be meaningfully interpreted?

-          In the experimental session, [51] is given as a reference for the MSI acquisition protocol. Is this exact protocol faithfully followed?

-          Apart from the VIL images, which are self-evident for mapping the presence of Egyptian blue in the blue and green areas, what other information was obtained from the rest of MSI image set, for the characterization of the pigments? This is not clear in the text. How the MSI images were explored and interpreted? Related to this last comment, the following remarks need to be explained. How are the spectral features extracted from the images?

-          In Lines 331– 332: “…The spectral features revealed by MSI are in agreement with calcium carbonate….”

-          in Lines 401-403: In the text “…However, the latter is characterised by widely distributed remnants exhibiting the expected spectral features of green earth pigments in VIS, IRR and UVR images…”.

-          And in Lines 420-421: “… Some spectral signals (especially in IRR and UVR) of red and black areas…”

Some other points that need to be clarified:

-          Lines 403-405: “…Also, on the upper side of the plinth, there are large areas of preserved green earth applications (cf. fig. 15). Allocated within these areas, there are distinct areas of high VIL signal of Egyptian blue and typical VIS / UVR spectral characteristics of yellow ochres….” The discussion is not clear here: have the authors found in the green areas both green earth and a mixture of EB and yellow ochre? This point was possibly discussed in lines 689-691 but needs additional clarity.

-          Lines 410-411: “… a trial measurement with a FORS spectrometer that was available for testing (fig. 13) … “ This sentence needs to be rephrased and explained. FORS was not part of the methodological approach?

-          Line 435: “refurbished” part??? Is this an appropriate term? Possibly it could be replaced.

3.2. Chemical analysis on cross-sections

 Lines 506-518: the interpretation of the stratigraphy of the sample presented in Fig. 18 is not well-documented. The text needs to be revised and discussed based on the data.

Unless the authors are based on data that are not presented in the paper, but, in this case, any related data have to be referred to.

The only analytical evidence for the two layers of different colour is that are both Pb-based. But the exact composition and nature of the two Pb pigments cannot be obtained by SEM-EDX elemental analysis. Additionally, the very thin dark layer in the interface looks rather like a dirt/soot layer and not like a pigment layer that has been altered. The fact that some Pb particles are in this layer is normal as it is in the interface of the two lead-based layers. The attribution of the black particles to plattnerite or lead sulfide is arbitrary, for the latter, there is not any stated correlation between sulfur and lead in this thin layer?! In the SEM images, the interface layer is also dark, therefore supporting the assumption that it is a dirt layer.   

 Lines 546-549: the stated clear correlation between lead and phosphorus in the painted area, should be further investigated before drawing any conclusion or interpretation. As the amount of phosphorus looks homogeneous, closely correlated with the lead, and at a relatively high coverage, the argument for attributing its presence to the binding medium is rather weak and potentially misleading.

4. Discussion

The “schematic virtual reconstruction” Lines 644 onward, as it is presented, is very questionable. First of all, a methodological note is mandatory, to be added in section 2. Materials and methods describing the purpose and the principles of this proposal as well as how this reconstruction was realized. The reconstruction process should be documented, including a description of the methodology, sources of information/evidence, and assumptions made during the reconstruction. Then, although the proposed, simplified, appearance of the colour in the clothes could rely on the identified pigment mixtures, and could help the reader to have a complete hypothetical view of the pigments and colours used in the polychromy of the sculpture, for the colour of the skin, it is rather arbitrary to propose any colour tone based on the analytical results obtained, therefore in this case, the aim of the reconstruction is unclear.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

·        Line 252: “Employed was a specially customized ILCE-7R camera…”

It would be good to mention the manufacturer of the camera and lens.

·       Line 287: Replace “Analyses were carried on 287 a light microscope…”

with “Analyses were carried out on 287 a light microscope …”

·         Line 209: Replace “Another sample …” with “an additional sample …” or “one sample …”

·         Line 287: “Analyses were carried out on 287 a light microscope (LM)” How were the LM images taken? bright field, crossed polars, dark field?

·         Line 292: Analyses were --- How about “Two analyses on this sample were performed …”?

·         Line 318, Figure 10:

UVR, IRR, and VIL should be grayscale and not color images. See Dyer, Verri, Cupitt [51]

·         Line 324: ”Surficial” seems to me a geological term. How about “the distribution of pigments on the surface”? Doble check with a native speaker.

·         Line 324/325: The applied non-invasive methods provided information on the surficial distribution of pigments and their identification.  

It is not correct that MSI can provide the chemical identification of pigments.

·         Line 331: “The spectral features revealed by MSI are in agreement with calcium carbonate.” Please elaborate. Is the MSI also in agreement with calcium sulfate, barium sulfate, other white pigments, or not?  

·         Line 336: “Also consisting mainly of calcium carbonate, though chemically very impure and admixed with a number of other phases …” how do you know?

·         Line 388: Figure 12 f-o should be grayscale.

·         Line 456, line 490: “green earth” identified by what? If identified by SEM-EDS, what are the elements detected?

·         Line 459: “With the exception of red lead (see 459 chapter 3.3), all pigments were applied in lime as binding medium.” Can you elaborate on this conclusion? Often in ancient polychromy the organic binding medium is not detected due to the age of the paint. Could that be the case here?

·         Line 471, Figure caption 15d and e: these are backscatter images.

·         Line 577: “The organic binders are complex materials enabling pigment grains to adhere to each other and form a coherent paint film to the surface of the support.”

Maybe better to say: pigment grains are dispersed in the binder …

·         Line 602: Does “egg” mean “whole egg” or “egg yolk”? Which reference are the standards taken from?, same for Table 2

·         Line 708: The “rounded feature at the Genius’ chest” is not visible in Figure 27 a.

·         Line 709: “stroking light” should be replaced by “raking light”

·         Line 719: Figure 27a should be grayscale.

·         Line 741: “purple pigment” Are you talking about Tyrian purple?

 

It seems as if only one sample (sample N) was analyzed for fatty acids and amino acids with GCMS. The authors conclude that only this paint sample has an organic binder, and in contrast the other paints were bound by lime. Only one analysis is not enough for this conclusion, or it has to be explained better.

Minor editing of English language required. see above

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 3 Report

This is an interesting paper on a multi-analytical approach to investigating the polychromy of Roman sculpture in central Europe. Generally, the paper is well-structured and clear in its incentives.

In the introduction, the historical and geographical contexts of Roman art and sculpture in Dunavian regions are demonstrated. These also justify the importance of this research approach and the questions that needed to be answered. The analytical methodology, as the authors put it, mainly focuses on non-invasiveness employing imaging in the UV, visible, and infrared with the most possible sources. On the other hand, a decisive part of their investigation lies in invasive analysis on cross sections of detached samples, using SEM-EDS (elemental analysis aiming at identifying the pigments), as well as on GC-MS analysis aiming at investigating the organic binders.

Based on the fact that analysis of cross-sections with SEM-EDS typically needs additional techniques for a more accurate characterization of inorganic materials, the authors manage to combine elemental analysis results with their imaging results to reach their conclusions concerning pigments, which appear pretty convincing.

Considering GC-MS analysis, two different derivatization protocols were employed, one for the lipids and the other for proteins. Based on the combination of these separate analyses, the authors conclude that possibly egg was used as a binder. Regarding the reported fatty acid ratios (Table 1), some point towards egg and some towards casein, a fact that authors acknowledge; is there an explanation for this discrepancy?

At this point, a question arises regarding quantitation; the lipidic fraction of egg is expected to be similar or different than that in casein (especially when fat is removed during extraction of the latter)? As far as the methodology and the followed protocol show, there was no internal standard (IS) used in the analysis. So, can the results justify the relative amounts of the lipidic and the protein fraction? If not, and based on the fact that GC-MS analysis of silyl derivatives resulted in significant amounts of fatty acids (comparable to those of amino acids, as seen in Fig. 22b), can a lipid-protein quantitative relation be justified?

Moreover, additional analysis of reference standards could well contribute to this direction and is generally a typical asset for better establishing such conclusions. Interestingly, the authors mention ‘reference standards’ in Table 1, without further specifying if these are ones that were run by them, or based on the literature; please explain.

In all, the reviewer believes that this paper deserves to be published in HERITAGE but after revision.

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Round 2

Reviewer 1 Report

The authors have thoroughly considered all the comments and suggested corrections and the revised manuscript has been significantly improved as to the quality of the figures as well as to the soundness of the arguments in the discussion and interpretation of results.

There are still some points that can be further improved such as

-          Lines 373-376: The revised text lacks clarity, in particular in the sentence “provide first constrictions towards the identity of phases”

-          Referring to comments already given in the first review, the authors should correct the text, as MSI is not a spectral imaging technique, it does not allow to record “spectral features” but, only characteristic luminescence/reflectance features.

o   Such luminescence patterns, when very characteristic, as for Egyptian blue, can allow to map the pigment on the object’s surface and also provide strong evidence for its identification (Lines 400-411).  

o   However, in the case of pigments as calcium carbonate, Line 381, the response of the pigment in MSI reflectance/Luminescence cannot be considered as evidence for its identification. If the pigment is locally identified with an analytical technique, then MSI can assist in tracing and mapping its distribution on the surface.

-          Lines 570-614: (with reference to comments 12 and 13 of the First Review) The authors have revised the text, however, the presented arguments are not sufficiently supported by the data presented.

o   SEM-EDS provides evidence for the presence of a lead-based pigment, Red lead is a good candidate, but there is no analytical data for its identification. Further, in Figure 18b, how do the authors explain the colour difference between the two layers, in which both, the authors state that the pigment is red lead?

 

o   As to Comment 13 of the first review: there is evidence of a clear correlation between Pb and P that allows attributing P to the pigment or pigment mixture and not to the binding medium. A more careful discussion of these data is required. 

Author Response

Please see the attachment

Author Response File: Author Response.pdf

Reviewer 2 Report

All comments and suggestions have been addressed.

Author Response

There are no open questions, as far as we understand. Thank you for the positive comments.