Obsidian Artifacts from Multiple Sources and Subsources at Valdesi in Western Sicily

Round 1

Reviewer 1 Report

The paper is very interesting, well organized and the results are very useful in the specific field of obsidian archaeology and archaeometry.

In my opinion the paper is worth to be published as it is.

I would just suggest to give some more information in the captions: particularly Table 1 where the samples could be better presented not only by abbreviations.

Also the units of Figg. 5, 6 and 7 should be introduced.

Author Response

Changes made - ratios for Figs. 5&6, ppm for Fig. 7

Reviewer 2 Report

The paper “Obsidian artifacts from Multiple sources and sub-sources at Valdesi in Western Sicily” reports an intriguing analysis of provenance of obsidian Prehistoric artifacts, highlighting the incredible economical network and circulation of these materials in the central Mediterranean area during Prehistory.

The topic is mastered by the authors as demonstrated by the long tradition of studies on provenance recognition of obsidian artifacts. The non-invasive technique for collecting geochemical data is one of the most appreciable aspects of the methodological approach. Calibrated pXRF investigations of already-analyzed obsidian sources by XRF and ICP-MS demonstrate the possibility of collecting valuable trace elements without damaging the tools. This is a great advantage as it allows to analyzed small finds stored in private collections or Museums, usually inaccessible and extremely safeguarded by owners and conservators.

The structure of the paper is good and the provenance results extremely interesting and intriguing in particular regarding the tool from Sardinia (mt. Arci), that testifies forms of open-water connections and commercial networks between Sardinia and Sicily since Prehistory.

The paper presents only one main issue, that is actually the visual representation of the provenancing of the Valdesi obsidian tools in the scatterplots.

In fact, even if some of the principal discriminant scatterplots adopted for provenance recognition of obsidians are reported in the draft, they just present the distribution of the geological markers, but I do not see the actual plot of the obsidian tools under investigation, that is the most important part of the research. I think that these graph were just edited from figures 3-4 of Tykot et al. 2017, Obsidian Studies in the Prehistoric Central Mediterranean: After 50 Years, What Have We Learned and What Still Needs to Be Done?, Open Archaeology 2017; 3: 264–278.

Therefore, I suggest to edit these images with new ones where the archaeological tools are displayed together with the obsidian source markers. Moreover, the graphical layout of the scatterplots should be improved, as they are not coherent: one report Sr over Nb vs Rb over Nb (that should be Sr/Nb vs Rb/Nb, in order to match with the one at figure 6). For the sake of clarity, the labelling of these figures should be explained in a legend clearly reporting the full description of, for example, BdT (Balata dei Turchi I suppose), LdV1 and LdV2. The same should be done for table 1.

Having considered this, I also think that the provenance determination needs to be cross-correlated with other types of discriminant graphs, such as Y vs Rb/Nb, adopted in Tykot, Foresta Martin 2020, Analysis by pXRF of Prehistoric Obsidian Artifacts From Several Sites on Ustica (Italy): Long-Distance Open-Water Distribution From Multiple Island Sources During the Neolithic and Bronze Ages, Open archaeology Open Archaeology 2020; 6: 348–392. I suggest to use this scatterplot as the authors stated that Y was acquired from archaeological tools and that this element was used for XY graphs (line 240), but actually it was not used to plot any of the figures. Other useful provenancing graphs could be based on non-factionary elements such as Nb vs Zr or Zr vs Sr, as reported in A.M. De Francesco, G.M. Crisci, M. Bocci 2008, Non-destructive analytic method using XRF for determination of provenance of archaeological obsidians from the Mediterranean area: A comparison with traditional XRF methods, Archaeometry. 50, 337–350.

Provenance from M. Arci should be investigated further, for examples with a specific in-detail scatterplot, reporting only the four sub sources from M. Arci, as used in Freund, Tykot 2011, Lithic technology and obsidian exchange networks in Bronze, Archaeol Anthropol Sci (2011) 3:151–164. I think that this important evidence should be emphasized and dissected more carefully. Some other recent works on the topic, such as that of Mameli et al. 2023, Intra-source provenance study on Monte Arci (Sardinia) obsidian by pXRF: Role of the data acquisition and analysis tools, Helion, 9, 3, e13958, could also be considered and cited.

Some other minor aspects

-          Chapter 3 in my opinion is somewhat redundant. I suggest to summarize some parts, avoiding some digressions, as in the sentences reported at lines 103-118. Some sentences are actually duplicated, as the ones at lines 129-131 and 165-166.

-          Figure 4 is not clear: are these the samples that were selected to be studied? If this were the case, I think that this image, reporting the resulting provenances, should be postponed at the end of paragraph 5, where the objects were reconnected to sources/sub-sources, eventually creating the “group d” comprising the main cluster of the tools from – I think - Lipari.

-          I think that the paragraph at lines 253-265 could be transposed in the beginning of the paragraph 5.

-          The reference apparatus is detailed, but perhaps too self-referential. The impact and substantial contribution of the authors, especially the first one, in the study of prehistoric obsidian artifacts is certainly very important and prominent in the field of studies, but it is recommended to reconsider some of the cited entries, for example by removing some of the older studies and adding more recent research, such as the already mentioned Mameli et al. 2023 and De Francesco, Crisci, Bocci 2008, or other studies such as:

M. Orange, F.X. Le Bourdonnec, L. Bellot-Gurlet, C. Lugliè, S. Dubernet, C. Bressy-Leandri, A. Scheffers, R. Joannes-Boyau, On sourcing obsidian assemblages from the Mediterranean area: analytical strategies for their exhaustive geochemical characterisation, J. Archaeol. Sci. Reports. 12 (2017) 834–844. https://doi.org/10.1016/j.jasrep.2016.06.002

Z. Kasztovszky, B. Maróti, I. Harsányi, D. Párkányi, V. Szilágyi, A comparative study of PGAA and portable XRF used for non-destructive provenancing archaeological obsidian, Quat. Int. 468 (2018) 179–189. https://doi.org/10.1016/j.quaint.2017.08.004.

For these reasons, I suggest a thorough revision, mainly related to the improvement of the graphical apparatus, increasing the number of discriminating scatterplots and effectively reporting the distribution of the analyzed artifacts within these diagrams. I think this can greatly enhance the quality of the data, which are already very good.

Author Response

The paper “Obsidian artifacts from Multiple sources and sub-sources at Valdesi in Western Sicily” reports an intriguing analysis of provenance of obsidian Prehistoric artifacts, highlighting the incredible economical network and circulation of these materials in the central Mediterranean area during Prehistory.

The topic is mastered by the authors as demonstrated by the long tradition of studies on provenance recognition of obsidian artifacts. The non-invasive technique for collecting geochemical data is one of the most appreciable aspects of the methodological approach. Calibrated pXRF investigations of already-analyzed obsidian sources by XRF and ICP-MS demonstrate the possibility of collecting valuable trace elements without damaging the tools. This is a great advantage as it allows to analyzed small finds stored in private collections or Museums, usually inaccessible and extremely safeguarded by owners and conservators.

The structure of the paper is good and the provenance results extremely interesting and intriguing in particular regarding the tool from Sardinia (mt. Arci), that testifies forms of open-water connections and commercial networks between Sardinia and Sicily since Prehistory.

In fact, even if some of the principal discriminant scatterplots adopted for provenance recognition of obsidians are reported in the draft, they just present the distribution of the geological markers, but I do not see the actual plot of the obsidian tools under investigation, that is the most important part of the research. I think that these graph were just edited from figures 3-4 of Tykot et al. 2017, Obsidian Studies in the Prehistoric Central Mediterranean: After 50 Years, What Have We Learned and What Still Needs to Be Done?, Open Archaeology 2017; 3: 264–278.

Therefore, I suggest to edit these images with new ones where the archaeological tools are displayed together with the obsidian source markers. Moreover, the graphical layout of the scatterplots should be improved, as they are not coherent: one report Sr over Nb vs Rb over Nb (that should be Sr/Nb vs Rb/Nb, in order to match with the one at figure 6). For the sake of clarity, the labelling of these figures should be explained in a legend clearly reporting the full description of, for example, BdT (Balata dei Turchi I suppose), LdV1 and LdV2. The same should be done for table 1.

Graph for Fig. 5 replaced with appropriate version

Graphs 6-7 replaced with those showing the artifacts in this study.

Having considered this, I also think that the provenance determination needs to be cross-correlated with other types of discriminant graphs, such as Y vs Rb/Nb, adopted in Tykot, Foresta Martin 2020, Analysis by pXRF of Prehistoric Obsidian Artifacts From Several Sites on Ustica (Italy): Long-Distance Open-Water Distribution From Multiple Island Sources During the Neolithic and Bronze Ages, Open archaeology Open Archaeology 2020; 6: 348–392. I suggest to use this scatterplot as the authors stated that Y was acquired from archaeological tools and that this element was used for XY graphs (line 240), but actually it was not used to plot any of the figures. Other useful provenancing graphs could be based on non-factionary elements such as Nb vs Zr or Zr vs Sr, as reported in A.M. De Francesco, G.M. Crisci, M. Bocci 2008, Non-destructive analytic method using XRF for determination of provenance of archaeological obsidians from the Mediterranean area: A comparison with traditional XRF methods, Archaeometry. 50, 337–350.

Ratios are best for overall use since many of the artifacts are fairly thin and some of the Xrays will not be detected (and effecting the totals but not the ratios).

Provenance from M. Arci should be investigated further, for examples with a specific in-detail scatterplot, reporting only the four sub sources from M. Arci, as used in Freund, Tykot 2011, Lithic technology and obsidian exchange networks in Bronze, Archaeol Anthropol Sci (2011) 3:151–164. I think that this important evidence should be emphasized and dissected more carefully. Some other recent works on the topic, such as that of Mameli et al. 2023, Intra-source provenance study on Monte Arci (Sardinia) obsidian by pXRF: Role of the data acquisition and analysis tools, Helion, 9, 3, e13958, could also be considered and cited.

Freund & Tykot 2011 cited

Mameli et al. cited

Some other minor aspects

Chapter 3 in my opinion is somewhat redundant. I suggest to summarize some parts, avoiding some digressions, as in the sentences reported at lines 103-118. Some sentences are actually duplicated, as the ones at lines 129-131 and 165-166.

Deleted duplicate part of sentence

- Figure 4 is not clear: are these the samples that were selected to be studied? If this were the case, I think that this image, reporting the resulting provenances, should be postponed at the end of paragraph 5, where the objects were reconnected to sources/sub-sources, eventually creating the “group d” comprising the main cluster of the tools from – I think - Lipari.

While the objects are arranged in Figure 4 by source, these are pictures of the artifacts while the assignment to these groups is based on the analysis and graphs in the following section.

- I think that the paragraph at lines 253-265 could be transposed in the beginning of the paragraph 5.

Visual assessment comes first, before pXRF analyses

- The reference apparatus is detailed, but perhaps too self-referential. The impact and substantial contribution of the authors, especially the first one, in the study of prehistoric obsidian artifacts is certainly very important and prominent in the field of studies, but it is recommended to reconsider some of the cited entries, for example by removing some of the older studies and adding more recent research, such as the already mentioned Mameli et al. 2023 and De Francesco, Crisci, Bocci 2008, or other studies such as: 1. Orange, F.X. Le Bourdonnec, L. Bellot-Gurlet, C. Lugliè, S. Dubernet, C. Bressy-Leandri, A. Scheffers, R. Joannes-Boyau, On sourcing obsidian assemblages from the Mediterranean area: analytical strategies for their exhaustive geochemical characterisation, J. Archaeol. Sci. Reports. 12 (2017) 834–844. https://doi.org/10.1016/j.jasrep.2016.06.002. Kasztovszky, B. Maróti, I. Harsányi, D. Párkányi, V. Szilágyi, A comparative study of PGAA and portable XRF used for non-destructive provenancing archaeological obsidian, Quat. Int. 468 (2018) 179–189. https://doi.org/10.1016/j.quaint.2017.08.004.

added all four refs

For these reasons, I suggest a thorough revision, mainly related to the improvement of the graphical apparatus, increasing the number of discriminating scatterplots and effectively reporting the distribution of the analyzed artifacts within these diagrams. I think this can greatly enhance the quality of the data, which are already very good.

Author Response File: Author Response.pdf

Reviewer 3 Report

This is a useful contribution to obsidian characterization research in the Central Mediterranean region, building on the senior author's longstanding contributions to this topic. The text is generally clear, well-written, and well-illustrated.

I have only minor criticisms of the presentation, as follows:

1. In the download PDF copy for review, there is no inset map of Italy in Figure 1. It is essential that this is provided.

2. There are some issues relating to the use of English, word choice, spelling, grammar, capitalization, etc. (see WORD file attached).

Comments for author File: Comments.pdf

My comments and editorial suggestions are marked in the accompanying WORD file (<TykotTEXT_edited>) using track changes.

Author Response

This is a useful contribution to obsidian characterization research in the Central Mediterranean region, building on the senior author's longstanding contributions to this topic. The text is generally clear, well-written, and well-illustrated.
I have only minor criticisms of the presentation, as follows:
1. In the download PDF copy for review, there is no inset map of Italy in Figure 1. It is essential that this is provided.
Added
2. There are some issues relating to the use of English, word choice, spelling, grammar, capitalization, etc. (see WORD file attached).
All corrected

Round 2

Reviewer 2 Report

The paper is substantially improved in the edited version as the authors accepted and integrated most of the observations of the first round of revisions.

The only point that could be further corroborated is the layout of the images.

In particular, figure 5 is not clear as both the sub-figures report the same data: the lower one just reports the legend of the provenance but actually it reports the same distribution of the sourcing obsidians of the upper one. It would be important to plot in this figure the actual distribution of the analyzed Valdesi obsidian tools, in order to have a clear representation of their overall distribution, as it has been done for figures 6 and 7. For example, in figure 5 the tool from Sardinia should plot in the related provenance interval, as I do not detect it, as well as the other analyzed ones.

 Moreover, for figures 5, 6 and 7, the two subfigures should be reported, named with letters (a, b) and accordingly described in the captions. For the sake of clarity, I also strongly suggest to use the same range for x and y axis, in order to make the comparison between sourcing obsidian and Valdesi tools more evident. For example, in figure 6 (up), x ranges from 0 to 25 and y from 0 to 1200; in the lower one x ranges from 0 to 24 and y from 0 to 1000. A similar situation affects figure 7 too.

Author Response

IN HIGHLIGHTED TEXT

The paper is substantially improved in the edited version as the authors accepted and integrated most of the observations of the first round of revisions.

The only point that could be further corroborated is the layout of the images.

In particular, figure 5 is not clear as both the sub-figures report the same data: the lower one just reports the legend of the provenance but actually it reports the same distribution of the sourcing obsidians of the upper one. It would be important to plot in this figure the actual distribution of the analyzed Valdesi obsidian tools, in order to have a clear representation of their overall distribution, as it has been done for figures 6 and 7. For example, in figure 5 the tool from Sardinia should plot in the related provenance interval, as I do not detect it, as well as the other analyzed ones.

Figure 5 is cited directly after the “Mediterranean sources”, with the purpose of showing that distributing the geological sources is fairly straightforward with just three elements. Other element ratios may be used as alternatives.

Figs. 6-7 are specifically for distinguishing Lipari and Pantelleria subsources for the Valdesi artifacts.

For the Sardinia sample, how it is determined is described now in the text (and the compositional data that stands out is in the Table).

Moreover, for figures 5, 6 and 7, the two subfigures should be reported, named with letters (a, b) and accordingly described in the captions. For the sake of clarity, I also strongly suggest to use the same range for x and y axis, in order to make the comparison between sourcing obsidian and Valdesi tools more evident. For example, in figure 6 (up), x ranges from 0 to 25 and y from 0 to 1200; in the lower one x ranges from 0 to 24 and y from 0 to 1000. A similar situation affects figure 7 too.

Descriptions in the captions adjusted.

The axes each have different ratios/elements and thus different scales.

Revision attached, shown as markup

Author Response File: Author Response.pdf