Authored by Cynthia Turner Camp from data collected by Hargrett Hours 3.0 and Dr. Alice Hunt
Introduction
In Fall 2021, Hargrett Hours 3.0 students undertook a chemical analysis of the metals, inks, and pigments used to write and decorate the Hargrett Hours. The goal of this analysis was twofold:
- To determine whether the metals, inks, and pigments used in the Hargrett Hours were consistent with what is known of Parisian manuscript construction in the fifteenth century
- To look for any substantive discrepancies between the materials used to write and decorate the calendar and those used for the body of the manuscript
We did not expect any anomalies in the materials used for the Hargrett Hours; it is not a de luxe manuscript, so it was not the artistic playground of a high-end manuscript artist like Jean Bourdichon, known to have experimented with unusual pigments (Burgio et al). However, because the calendar is on a separate quire, written by a different scribe on a different grade of parchment, with a slightly larger size of text block, it is possible that the calendar originally belonged to a different Parisian manuscript and was added to the body of this one at some later point, at or before the stab-stitching phase. Additionally, the entirety of fol. 60r-v, plus two lines on fol. 52r, were written by a different scribe than the main scribe, in a hand that resembles the calendar hand. Were the calendar hand and the fol. 60r-v hand identical, the “all one manuscript” theory could be substantiated. The students therefore searched specifically for evidence that would support or counter the “all one manuscript” theory that has underpinned the research on the manuscript.
The materials used to write and decorate the Hargrett Hours were those in common use in the later Middle Ages (see Trentalman et al, 165-66 for a list). These included organics for blue and pink, perhaps indigo and brazilwood; both lead- and mercury-based reds, red lead and vermillion, respectively; a copper-based green, probably verdigris but possibly malachite; and two different shades of orange, one containing mercury and the other containing lead, suggesting shades of red mixed with an organic yellow to create orange. The artists shied away from expensive materials, using cheaper metals and pigments. The price point of the materials used aligns with our working theory that the first owner was more invested in this book as a collection of carefully chosen prayers than with it as a showpiece.
The inks and pigments used in the calendar are chemically distinct from those used in the body of the manuscript, while the ink used for fol. 60r-v aligns chemically with the ink used elsewhere in the body of the manuscript. Although these discoveries cannot prove that the calendar was created for a different manuscript, they do suggest that the calendar was created separately from the rest of the book.
Methods
The inks, pigments, and metals in the Hargrett Hours were tested with a portable X-Ray Fluorescence (p-XRF) spectrometer. A p-XRF device is a non-invasive analysis tool that can identify elements in a sample site. The p-XRF works by bombarding the elements in a small sample size with x-rays, which dislodge electrons in the elements present in the sample site; the dislodged electrons emit discrete energy levels, based on their atomic number, which are recorded and visualized as a spectrum graph. The p-XRF can only identify elements with an atomic number greater than 16, to include elements present in inks and pigments created with inorganic materials; because carbon has an atomic number of 6, it cannot detect or differentiate among organic inks and pigments. Through this process, the p-XRF solely identifies the presence of the elements, not the quantity or proportion of them. These limitations notwithstanding, the p-XRF is a useful device for differentiating between (for example) mercury-based vermillion and red lead pigments in a sample. Similarly, the absence of results in a testing spot allows one to infer that the pigment was created with an organic substance, such as indigo for blue rather than a copper-based azurite. In more robust scientific analyses of medieval manuscripts and art, the p-XRF would be used alongside other testing methods. For our purposes, the p-XRF provided sufficient information that we could infer the presence or absence of common medieval pigments.
Samples were taken with a Briker Tracer 5i portable X-Ray Fluorescence (p-XRF) spectrometer set at a voltage of 30kV with a 30μA current and a copper-tin-aluminum filter. The device was provided and run by members of UGA’s Center for Applied Isotope Studies through their STEM on the MOVE program. Sample spots were tested in 90 second intervals at 1mm away from the test surface; the test sites were approximately 6mm in diameter. Because the x-rays emitted would penetrate the parchment, testing sites were selected that did not have other writing or decoration on the reverse of the page, as the p-XRF would read the elements present on both sides of the parchment, thereby resulting in messy data. The energy spectra was output and interpreted with the software Artax, and our interpretations governed for standard spectral interference such as escape peaks and sum peaks.
Results
Trace elements
Both lead (Pb) and calcium (Ca) were regularly present across the samples, regardless of location. The calcium is present from the lime used to prepare the parchment (Trentalman et al. 162). The regular appearance of lead, even in locations where no ink or pigment possibly created from lead were present, is probably due to plummet underdrawing or simply to cross-contamination. Lead is a soft metal that easily leaves residue on surfaces, and lead also fluoresces vigorously under the p-XRF. It is therefore a comparatively “loud” element (a little bit of lead will create a dramatic spike on the spectra) even when present in trace quantities.
Inks
The Hargrett Hours is written in a rusty-grey black ink. Rubrication is executed in a bright red in the calendar and in a pinky-red color throughout the manuscript. Both calendar and body are ruled in a very faint pink. The initials in both the calendar and the body of the manuscript are either blue with bright red penwork or gold with black penwork.
Black ink. The black ink of the calendar contained iron (Fe), indicating that it is iron-gall ink. It also contained higher than anticipated peaks of calcium (Ca). Iron-gall ink is typically created from oak galls, which contain gallotannin and gallic acid, and iron sulfate (FeSO4·7H2O), often called vitriol or copperas in medieval sources. The black ink used in the body of the manuscript contained iron, copper, and traces of zinc, although the spectra from fol. 60 (written by a different scribe) differed significantly in the excitation of elements, if not their presence. The zinc and copper commonly appearing in iron gall ink samples can perhaps be traced to the brass vessels in which the ink would be created (Trentelman et al. 163), or they may be contaminants in the iron sulfate (Hahn et al. 234).
Rubrication. The calendar rubrication contained mercury (Hg), indicating that it was probably vermillion (HgS), a synthetic form of cinnabar. The “red” rubrication in the rest of the manuscript is actually a deep magenta color. Sampling suggested that this ink is made from an organic colorant, such as brazilwood. Brazilwood, or Caesalpina sappan, is a tree found in India and Southeast Asia, and its heartwood was used by Parisian illuminators to create reds and pinks of different shades, depending on the exact recipe (Melo et al). The pink ruling was not visible to the sampling process, suggesting that it too is an organic.
Penwork initials. The red penwork in both calendar quire and body of the manuscript contained mercury (Hg), pointing to cinnabar (HgS). Blue initials in both sections contained copper (Cu), indicating that it was formed from azurite (Cu3(CO3)2(OH)2). The black penwork could not be sampled cleanly enough for conclusive results to be obtained. Although it is a darker black than the regular ink, hinting that it may be a carbon-based black, such as lampblack, the relevant spectra consistently contained iron and copper, as for the regular writing ink. Whether this is a true result from the black penwork ink or contamination from the adjacent writing ink could not be determined.
Metals
Both gold leaf and shell gold were used in the manuscript. Gold leaf was used for the initials throughout the book, while shell gold (powdered gold leaf mixed with gum and applied with a brush) (Panayotova 131) was used in the rinceaux borders and line fills. The shell gold was applied after the black vinework was laid down. Some gold leaf initials are more severely worn than others.
The sampling process revealed that both the shell gold and the gold leaf contained (unsurprisingly) gold (Au). Unlike other examples these classes have encountered, where p-XRF testing revealed the “gold leaf” to be a copper and gold alloy (Payatova 132), neither shell gold nor gold leaf appeared to be alloyed with another metal. Notably, however, all gold leaf samples except those taken from the calendar exhibited marked traces of calcium. Gold leaf is frequently burnished with an animal tooth (Clemens and Graham 33), a possible source of the trace calcium there; alternately, it may be present in a gypsum-based gesso, used to adhere the gold leaf to the parchment (Trentalman et al 158). For the shell gold, a similar explanation suggested for the trace calcium in the ink might obtain: that the shell gold was mixed within a shell or horn, and the calcium leached into the shell gold from the container (Clemens and Graham 33).
Pigments
Painted decoration in the Hargrett Hours is limited to the rinceaux borders, embellished with acanthus leaves and floral motifs, which accompany the decorated initials that open each of the Long Hours of the Passion. Students identified five distinct colors used for the acanthus and floral elements in addition to the ubiquitous white (invariably lead white) and black. As Trentalman et al 168 reminds us, however, the presence of a inorganic pigment doesn’t mean that the artist didn’t also mix that colorant with an organic substance to enhance the color.
Blue: The blue pigment is evidently an organic, like indigo, because no copper was present (which would have signaled azurite, Cu3(CO3)2(OH)2).
Green: The green pigment is copper-based, either verdigris (a copper acetate) or malachite (Cu2(CO3)(OH)2).
Orange: Lead was present in the orange sample; this could be evidence of an orange formed partially from red lead, orange formed from mixing an organic with lead white to lighten it, or simply trace lead from the plummet underdrawing. We did not identify arsenic (As) in the sample, which would have indicated the presence of realgar (As4S4).
Red-Orange: The occasionally used red-orange color includes mercury (Hg), which indicates that vermillion (HgS) was used, perhaps mixed with an organic, to create the color.
Pink: The pink pigment is an organic like brazilwood, rather than a red pigment lightened with lead white (Pb)
Interpretation
The scribes and artists employed in the Hargrett Hours’ creation use standard materials commonly identified in fifteenth-century manuscripts, and most of the surprises in our testing results can be explained as the result of the production process. The scribes unsurprisingly used iron gall ink, although the rubrication inks differ between calendar and manuscript body. All the pigments used in the manuscript are economical choices, especially when compared with the more expansive palettes used by other Parisian artists of the same time period (Melo et al 440-41). The artist used an organic blue rather than the pricier azurite or lapis lazuli ((Na,Ca)8(AlSiO4)6(S,SO4,Cl)1–2); vermillion was used sparingly. Given this evidence, it is likely that the artist also used the cheaper verdigris rather than the mineral malachite for green, although the two are chemically indistinguishable through this testing method. These cheaper pigment choices align with other aspects of the Hargrett Hours that suggest the first owner was not invested in the book as a form of social display, but instead on the book’s textual contents. The use of gold throughout the manuscript does not undermine this inference. Gold leaf is not as expensive as its name suggests, while the heavy wearing of the gold leaf in many places hints at a less-than-perfect application of the leaf to the underlying ground, a point that substantiates the book’s overall appearance of having been decorated on a budget.
The question of whether the calendar was created to accompany the rest of the manuscript remains open, although the chemical evidence reinforces the visual evidence that they are distinct productions. As discussed on the Codicology page, the parchment used in the calendar is of a different quality and texture than the parchment used elsewhere in the manuscript, and the size of the text blocks differs. The rubrication in the calendar and the body of the manuscript were executed in visibly and chemically distinct inks, vermillion red and an organic pinky-red respectively. Similarly, although neither calendar nor manuscript body gold leaf appeared to be an alloy, the calendar gold leaf’s ground did not contain the trace calcium present in the gold leaf elsewhere. Both sections used azurite for their blue inks.
The evidence of the ink is the most diagnostic. The black ink used in the calendar is chemically distinct from the ink used in the body of the manuscript. The calendar ink is pure iron-gall ink, while the ink in the body of the manuscript contains copper in addition to iron. Moreover, the calendar ink also contains unexpectedly high levels of calcium. Dr. Hunt suggested that the calcium may have leeched into the ink from its inkwell, which were sometimes made from animal horn. Under this theory, not only are the inks different, but they were stored in different types of inkwells. Importantly, the ink on fol. 60r-v, which was written by a different hand than the main scribe, perhaps by the calendar scribe, contains the same elements as the ink of the main scribe, albeit with a somewhat different spectra, suggesting slight differences in chemical makeup. It does not contain the excess of calcium present in the calendar ink. The ink on fol. 60r-v is therefore distinct from the calendar ink, and more like the rest of the ink in the body of the manuscript.
Although this evidence proves different production processes, and probably different production locations, for calendar and manuscript body, it cannot prove that the calendar was originally made for a different manuscript altogether. The difference in the size of the text blocks between calendar and body (75mm x 102mm for the calendar, 58mm x 92mm for the body) is real but not extreme; in another Books of Hours in the Hargrett Library, the calendar is also ruled slightly larger than the text block in the body of the manuscript (54 x 94mm for the calendar, 54 x 88mm for the body), suggesting this difference may not be anomalous for fifteenth-century Books of Hours. Moreover, the Hargrett Hours’ level of decoration in the calendar quire is consistent with the modest decoration in the rest of the book, while the evidence of the textual contents – the idiosyncratic anthology of Passion material, the presence of Parisian saints in the suffrages – aligns with the calendar evidence that points to a liturgically savvy Parisian owner. Books of Hours could be produced piecemeal, as Kathryn Rudy has demonstrated of mass-production workshops, so it is equally possible that the compilation of the calendar (itself seemingly a more complex process than usual, as discussed in the Calendar interpretive essay) occurred in a different workshop, with the calendar always intended to accompany the rest of this manuscript.
Bibliography
Burgio, Lucia, et al. “Spectroscopic Investigations of Bourdichon Miniatures: Masterpieces of Light and Color.” Applied Spectroscopy vol. 63, 2009, pp. 611-20. http://dx.doi.org/10.1366/000370209788559593
Clemens, Raymond, and Timothy Graham. Introduction to Manuscript Studies. Cornell University Press, Ithaca, 2007.
Hahn, O., W. Malzer, B. Kanngiesser, and B. Beckhoff. “Characterization of iron-gall inks in historical manuscripts and music compositions using x-ray Fluorescence Spectrometry.” X-Ray Spectrometry vol. 33, 2004, pp. 234-39. https://doi.org/10.1002/xrs.677
Melo, Maria João, et al. “A Spectroscopic Study of Brazilwood Paints in Medieval Books of Hours.” Applied Spectroscopy, vol. 68, no. 4, 2014, pp. 434-43. https://journals.sagepub.com/doi/10.1366/13-07253
Paynayotova, Stella. “Painting Materials and Techniques in Western Illuminated Manuscripts c. 600-c. 1600.” The Art and Science of Illuminated Manuscripts: A Handbook, edited by Panayotova, Harvey Miller Publishers, Turnhout, Brepols, 2020, 126-70
Rudy, Kathryn M. Piety in Pieces: How Medieval Readers Customized their Manuscripts. Open Book Publishers, 2017, https://doi.org/10.11647/OBP.0094
Trentelman, K., C. Schmidt Patterson and N. Turner. “XRF Analysis of Manuscript Illuminations.” Handheld XRF for Art and Archaeology, edited by Aaron N. Schugar and Jennifer L. Mass, Leuven University Press, Leuven, 2012, 159-89.