Ariadne Kostomitsopoulou Marketou is a postdoctoral researcher in the project “EthiCodex.” Her research revolves around the materiality of the codex, viewing it as a medium to approach provenance and dating. Ariadne received her PhD in Archaeology and Conservation from the University of Oslo and holds a MSc in Protection, Conservation and Restoration of Works of Art and a BSc in Chemistry from the Aristotle University of Thessaloniki. Her broader research interests include materials analysis, the materiality of colourants, pigment manufacturing technology, and the organization of craft production in the ancient Mediterranean world.
In their recently published article, F. P. Romano and co-authors present the first experimental evidence that confirms the application of lead to delimitate the written areas and the corresponding unwritten intercolumns through in situ MA-XRF imaging.
Their research included the analysis of representative samples of unrolled Herculaneum papyri of various qualities, with dates spanning from the 2nd c. BCE to the beginning of the 1st c. CE, and the analysed samples mostly derive from Rome and Campania, while Eastern Mediterranean centres are also represented.
Figure 1 (from Romano et al. 2023). Top: Lead (Pb) distribution map obtained by MA-XRF (PHerc. 164). Bottom: NIR image of the unrolled papyrus in NIR (950 nm). Credits: Biblioteca Nazionale “Vittorio Emanuele III,” Napoli—Consiglio Nazionale delle Ricerche, Istituto di Scienze del Patrimonio Culturale and Brigham Young University, Provo. License: CC BY 4.0.
The results, illustrated by MA-XRF lead (Pb) maps (Figure 1), confirm the use of lead to mark written areas and their corresponding intercolumns.
The use of rulers and rounded leaden tools (disks) for the preparation of the papyrus rolls to delimitate the areas intended to be written or to remain blank, was so far only suggested in Roman textual sources (including Phanias, Palatine Anthology 6. 265; Catullus, carm. 22; Philip of Thessalonika, Palatine Anthology 6.62; Pliny the Elder, NH 33.60).
Interestingly, the MA-XRF maps revealed three distinct layout patterns of increasing complexity among the analysed rolls. The complexity of the ruling-lines is directly linked to the quality of the book, with the most complex, delimitating columns, intercolumns, and lines (see Figure 1), associated with the rolls of highest quality.
Finally, the inclination of the pre-defined lead-drawn parallel lines towards the left, confirms that the slanting feature of the columns was deliberate (Maas’ Law).
Did the use of leaden tools expand to early codex production?
Since lead-drawn ruling lines seem to be invisible to the naked eye, only such systematic interdisciplinary research can potentially answer such questions regarding early codex-making.
Part 3: Carbon black, red earth, and hyperspectral imaging
Carbon black and red earth
Carbon black is considered the most commonly used black ink in antiquity and the preferred ink when applied on a papyrus substrate. Essentially, carbon black ink is made of fine amorphous carbon particles dispersed in water with the help of a binding medium.
The most commonly used binding medium for ink production is gum Arabic, a natural gum produced from the trees of the Acacia species.
Gum arabic exuding from the Babhul tree (Vachellia nilotica). Photo: Ashwin Baindur, Wikipedia commons (CC BY-SA 4.0).
Even though carbon black is today considered a widely available and relatively cheap colourant, this does not seem to be the case in antiquity. Several starting materials were used to create the valuable ink (or pigment) known in Latin as atramentum, including wood, fruits, resins, and bones. The process of pyrolysis to produce amorphous carbon particles was not considered trivial and according to literary sources, specialised workshops were organised for this purpose, as described by Vitruvius (De arch. 7.10):
“A vaulted apartment is built like a sweating chamber, and is covered carefully with a marble facing and smoothed down. In front of it a small furnace is built with outlets into the chamber, and the mouth of the furnace is carefully enclosed so that the flame does not escape. Resin is placed in the furnace. Now the fiery potency burns it and compels it to emit soot through the outlets into the chamber. The soot clings round the walls and vaulting of the chamber. It is then collected and in part compounded with gum and worked up for the use of writing ink; the rest is mixed with size and used by fresco-painters for colouring walls.“
Vitruvius, On Architecture, Book 7, Chapter 10 (translated by F. Granger).
Ingredients for carbon black ink
5 g lamp black (Kremer Pigmente, n. 47250)
5 g untreated gum arabic (Kremer Pigmente, n. 63300)
30 mL distilled water
For the production of our carbon black ink, we used commercially available lamp black since we could not at this stage experiment with pyrolysis.
We worked with whole, untreated lumps of gum arabic to experience the process of grinding, taking turns until we got the desired result.
The grinding process was not particularly demanding and in about 10 minutes we were satisfied with the results.
We worked with a larger quantity than what was required for our recipe, starting with 25 g of gum which we dissolved in 150 mL of lukewarm distilled water. The dissolution of the powdered gum in water with the aid of a hotplate was a relatively easy and fast process.
Untreated gum arabic solution (photo AKM).
We noticed that in contrast to the powdered material purchased from the same supplier that we used for the manufacture of iron gall ink (part 2), the untreated lumps of gum arabic enclose small pieces of bark or other foreign material, which we decided to remove using a coffee filter.
By mixing the gum arabic solution with lamp black, an intensely black ink was created.
Carbon black ink (photo AKM).Despina testing lamp black.
Finally, we used a commercially available red earth, consisting of a mixture of iron oxides (Terra Ercolano, Kremer pigmente, n. 41600) for the red ink. To produce our ink, we simply mixed the powdered mineral iron oxide with gum arabic and water.
Mixing red earth with gum arabic solution (photo: AKM).Testing the inks (photo: DW).
Hyperspectral imaging
Our mock-ups were now ready to measure!
The instrument we used was a portable hyperspectral scanner from Speccim IQ, covering the VNIR region. This was mounted on a tripod, and we made sure that the geometry was at around 45-degree angle.
Tia checking the imaging setup and Despina taking photos.
The light source was halogen lamp (200 watts) and the temperature rise was measured with the use of a thermal camera.
Seeing through the camera: the hyperspectral imaging setup with our mock-ups placed behind glass.
We are currently working on the first results and preparing the second phase of this study, which will include different types of glass.
Stay tuned for updates!
Many people contributed to this pilot study, whom we would like to thank:
Hilda Debora and Riestiya (Tia) Zain Fadillah from the Colourlab (NTNU), who set up the hyperspectral imaging experiment.
Anastasia Maravela (IFIKK, UiO) and Federico Aurora (UB, UiO), who kindly lend us spare glass frames.
David Grimaldi (who is currently completing his MA thesis at IFIKK) and Despina Wilson (artist) were extremely helpful throughout the experimental session.
Gall nuts (Aleppo galls, Kremer Pigmente, n. 37400. Photo: AKM).
Iron gall ink is made by mixing plant material containing gallotannic acid (a plant polyphenol), primarily obtained from oak gall nuts (or apples), with hydrated iron (II) sulfate (FeSO4∙7H2O), a compound also known as (green) vitriol.
Oak gall nuts are created by the female oak gall wasps (Cynips Gallae-tinctoria) as a cozy nest to lay their eggs and to provide a nutritious environment for the larvae.
The gall wasp (Image by Judy Gallagher, CC BY 2.0)
According to Kremer pigmente’s information sheet about the oak gall nuts we used for this experiment (last accessed online 09.01.2023), the main constituents of the best quality Aleppo galls, i.e., those collected before the insect escapes the gall, are 50% to 70% gallotannic acid, 2 to 4% gallic acid, mucilage, sugar, resin, and lignin (insoluble matter).
The colourant that gives the dark tones of the produced solution is ferrous gallotannate, a dark brown compound that becomes black after exposure to air with the formation of ferric tannate (oxidation from Fe2+ to Fe3+).
Before their wide use for the manufacture of ink, these tannin-rich sources found numerous applications in antiquity. Pliny the Elder mentions their medicinal properties and their use as a black hair dye (Plin. HN 24.9), dyeing wool and tanning (Plin. HN 16.26-29).
Regarding ink manufacture ancient recipes suggest that, besides galls, a range of tannin-rich plants, such as for example tree bark or fruits and alternative iron sources deriving from everyday-life materials, such as iron nails, could be used for the production (Colini et al., 2021).
For this pilot phase of our study we wanted to experience the workability of the most commonly used materials for the production of iron gall inks (galls nuts, gum arabic, green vitriol).
How easy is it to grind the gall nuts?
Is gum arabic easily dissolved in water?
Can the process be carried out by one single person?
How does the colour of the tannin extract change over time and what is the impact of time on the final product?
There are several recipes out there and we chose to use a well documented and already reproduced recipe, following closely the quantities and process as described by Sara Charles (last accessed online 19.12.2022).
For the extraction of the gallotannic acid, we used:
80 g gall nuts (Kremer pigmente, n. 37400)
300 mL tap water*
Gall nuts posing with a hammer before being crushed (photo: AKM).
* It was difficult to find distilled water at a regular store in Oslo, so we decided we could afford to use tap water since no chemical analysis is planned at the moment and contamination from the water source was not considered important.
The galls were placed in a transparent plastic bag and wrapped in paper and a cloth, to avoid losing the material through the process of crushing.
Partially crushed galls in a transparent plastic bag (photo: AKM).
And then the pieces were further processed in a marble mortar.
The colour of the crushed galls was pale yellow, pointing to the so-called ‘white galls’, i.e., galls collected after the insects have escaped. These galls contain less gallotannic acid than the best quality ones, which are of darker colour and are collected while the insects are still inside.
The extraction of the tannins without heating takes a few days, so the process had started already at home before we all met at the laboratory of the Department of Archaeology, Conservation and History of the University of Oslo.
At first, the mixture is yellowish:
According to the recipe we followed, the container should be placed by the window for three days. Oslo in December is rather dark, and the window could not help with the maintenance of a warm temperature. The jar therefore changed several locations during these three days, according to which place was the warmest and not the sunniest.
After a couple of hours, the mixture was already darker.
The glass jar containing the mixture is placed next to the window. The sun is relatively low since the experiment was carried out in December in Oslo (Norway).
The glass jar changed several positions during the day trying to find the highest temperature.
Three days later, and we were already in the lab setting up the hyperspectral imaging study that aimed to explore the impact of the glass frames that house papyri on the imaging process (part 3).
The mixture had a dark brown colour, indicating the successful extraction of the tannins in the water medium. We strained the solution through a coffee filter and since we wanted a small quantity of ink, we used 70 mL of the dark brown liquid to produce our iron gall ink, mixing it with 40 g of green vitriol and 70 g of gum arabic.
Ingredients for iron gall ink:
70 mL extract from the galls
40 g vitriol (ferrous sulfate FeSO4∙7H2O, from Kremer pigmente, n. 64200)
70 g of powdered gum arabic (Kremer pigmente, n. 63330)
With the addition of vitriol, the solution became intensely black et voilà!
Our ink was ready to use.
Many people contributed to this pilot study, whom we would like to thank:
Hilda Debora and Riestiya (Tia) Zain Fadillah from the Colourlab (NTNU), who set up the hyperspectral imaging experiment.
Anastasia Maravela (IFIKK, UiO) and Federico Aurora (UB, UiO), who kindly lend us spare glass frames.
David Grimaldi (who is currently completing his MA thesis at IFIKK) and Despina Wilson (artist) were extremely helpful throughout the experimental session.
Ancient papyri are most commonly stored and displayed in glass frames, which protect the papyrus and at the same time permit the display of the written surface. The papyrus fragments are basically sandwiched between two glass sheets held together by tape.
P.Oslo I 4 (inv. 434): 4th century CE Egyptian magical papyrus (spell of attraction) from the Oslo Papyrus Collection (image under CC-BY-NC 3.0). The papyrus is sandwiched between two thin glass sheets, which are held together with adhesive tape.
Various types of glass and adhesive tape are used across different collections, with the materials being based on a series of factors, including cost, the display conditions and environment, access to suppliers, etc. (Kaye 2019). Window glass (soda lime glass) has been used traditionally, while the more chemically stable borosilicate glass is gaining ground in recent times and UV radiation protective filters are necessary for papyri that are often on display.
Given the fragile state of surviving papyri, curators are reasonably hesitant to open the glass frames for materials analysis. So, a common technique for preliminary analysis of mounted papyri involves imaging with a portable microscope (namely using a DinoLite microscope with UV and IR light sources) above the glass frame.
This methodology, developed by the research team of Ira Rabin and colleagues (BAM), makes it possible to rapidly scan the manuscript and get qualitative information about the inks based on a simple principle: carbon inks remain black across the spectrum, iron gall inks become less intense (the lines become thinner) under IR illumination (set to 940 nm in the commercially available model of the DinoLite), and plant-based inks tend to disappear at 740 nm (Colini et al. 2021). Moreover, for plant based inks, the contract between the ink and the substrate is expected to increase (ibid.). By comparing the opacity and thickness of the lines, information about the makeup inks can be extracted.
However, this technique relies on the (rather subjective) ability of the researcher to recognise these variations between the three predefined modes of illumination (Vis, IR, UV) and the results are qualitative and descriptive. Hilda Debora and I were wondering if we can use the equipment of the Colourlab (NTNU, Gjovik) for the quantitative measurement of glassed papyrus with hyperspectral imaging.
What would the impact of the glass be on our measurements?
We also decided to make our own inks and apply them on both papyrus and parchment.
The next post (part 2) describes the process of making iron gall ink from commercially available starting materials. The last part of this series (part 3) describes the process of making carbon black ink, starting from commercially available carbon particles and gum arabic and finally the setup of the hyperspectral camera.
Many people contributed to this pilot study, whom we would like to thank:
Hilda Debora and Riestiya (Tia) Zain Fadillah from the Colourlab (NTNU), who set up the hyperspectral imaging experiment.
Anastasia Maravela (IFIKK, UiO) and Federico Aurora (UB, UiO), who kindly lend us spare glass frames.
David Grimaldi (who is currently completing his MA thesis at IFIKK) and Despina Wilson (artist) were extremely helpful throughout the experimental session.
The Early History of the Codex 