New and Old Data

This project uses data I collected while working with the Harvard Art Museums to catalogue the scientific analyses undertaken on the marble sculptures in the Ancient Art Department. The first phase of the process studied over 175 artworks, mostly through the files held by the department or at the Harvard Art Museums Archive, which I compiled in a spreadsheet that contains notes on each object. Eugene Farrell from the Harvard Art Museums undertook the majority of these analyses beginning at latest in 1978 and lasting throughout the 1990s. Most of these analyses focused on stable isotope ratios because, at that time, they produced adequate data for differentiating between quarries (see Basic Chemistry and Terminology).

I first took the subset of objects that had quantitative results from their analyses: 38 data points for 25 artworks, primarily focusing on stable isotope ratios. (The dataset of all 178 objects and the artworks with stable isotope analyses can be accessed on the Harvard Dataverse.)Three artworks had multiple samples that produced different results because they are pastiches (they are assembled from pieces that came from multiple objects). Some analyses produced qualitative data that cannot be compared against other observations, for example looking at the stone under a raking light. Would we still consider these analyses scientific? The use of specialized instruments makes a case for calling this scientific, or at least empirical, in its goals, and new methods seek to quantify characteristics that were previously identified visually, such as color.[1] I am currently working on a database that will present the full catalogue of objects studied during the course of this project.

In order to study the quantitative information, a reference dataset must be used. When I began this project, I was unaware of any online or otherwise fully published datasets, but I found the Miss Marble database online. After downloading the program, however, I could not open it due to a connection error, so I opened it in a hex editor to determine whether the data was stored in the program or on a server. Unfortunately, it seems the data was stored in a server that is no longer connected to the program.

I then contacted Walter Prochaska, an Austrian geologist whom I interviewed in summer 2017, and he alterted me to Donato Attanasio’s multivariate database. Attanasio's 2003 and 2006 books on marble respectively focus on the use of EPR spectroscopy and stable isotope ratios. The main parameters I could analyze were stable isotope ratios, dolomite proportion, maximum grain size, and whiteness.[2] As noted in Basic Chemistry and Terminology, results from EPR are not comparable between different laboratories, while other data types are. (Stable isotope ratios use Pee Dee Belminite or NBS-19 limestone as a universal reference and whiteness is measured against an 8-bit grayscale.) This source is particularly important because it comes with a disc containing an Excel file with the dataset, making it accessible so long as copies of the book are available. Other more recent efforts to create digital resources for the study of marble, like the Miss Marble database, have fallen through.[3]

See Methods and Preliminary Results for the applications of the quantitative data.