University Place

Analyzing Stains in Medieval Manuscripts

– Welcome everyone to Wednesday Nite at the Lab. I’m Tom Zinnen, I would here at the UW-Madison Biotechnology Center. I also work for UW Extension, Cooperative Extension. And on behalf of those folks and our other core organizers, Wisconsin Public Television, The Wisconsin Alumni Association, and the UW-Madison Center Alliance, thanks again for coming to Wednesday Nite at the Lab. We do this every Wednesday night, 50 times a year. Tonight, it’s my pleasure to introduce to you Heather Wacha and Leah Parker. Leah was born in Seattle, Washington and graduated from Mount Lake Terrace High School north of Seattle. Then she went to American University in Washington DC, and got a degree in literature and theatre arts. Then she came here to UW-Madison and received a Master of Arts degree in English, and she’ll be finishing up her Ph. D.
in English in 2019. The parties are scheduled for May, August, and December. (laughter) Heather Wacha was born in Lincoln, Nebraska, and graduated from high school in Des Moines, Iowa. She went to Hamlin University in Saint Paul, Minnesota where she majored in French. And then she got both a Masters of Arts and a Ph. D. degree from the University of Iowa, in Medieval History. In 2016, she came to UW-Madison to talk as a post-doc with the Council of Library and Information Resources. Their topic is one near and dear to my heart, since I worked in high school as a student at the Dixon Public Library where it’s always fun to find really old stuff. Old stuff then was like, from the 1870’s.

You’re talking about old stuff. Tonight, their talk is entitled A Library of Stains, using multi-spectral imaging to analyze stains in medieval manuscripts. I’d like to point out that I usually give people bottles of water but tonight I brought coffee for these folks, since coffee makes a much better stain. (laughter) – Water leaves a mark, too. – Oh, water leaves a mark, too, that’s heavy. (laughter) Please join me in welcoming both Heather and Leah to Wednesday Nite at the Lab. (applause) – Thank you very much. First, for coming out tonight on this very beautiful evening. We’re excited to be here. We’d like to thank all of the organizers, and Tom in particular, for inviting us to present at Wednesday Nite at the Lab.

My name is Heather Wacha, I’m a CLIR Postdoctoral Fellow here at the University of Wisconsin associated with the Center for History of Print and Digital Culture in the high school and the English Department. And Leah Parker, my co-presenter, is a doctoral candidate in the English Department who’s generously contributed her time and knowledge to the data analysis part of this project. Thank you. We’re gonna talk tonight for about 50 minutes, about a project that we’ve been working on over the last year that uses multi-spectral imaging to help analyze stains that are found in medieval manuscripts. But, before we go any further, I think it’s important at this point to add a small caveat especially given the nature of this Wednesday Nite at the Lab series. Neither of us is a scientist, just putting that out there. (laughter) So, if at the end of the presentation, you have any specific questions about the scientific underpinning of this project, I will be happy to contact our specialist collaborators and try and get you the answers that you might be looking for. As an introduction to our talk, I’m going to play a 90 second video that was produced by the University of Birmingham in the UK. Mike Toth, the speaker in the video, and one of our collaborators on this project explains very concisely how multi-spectral imaging is used to investigate a possible under text or palimpsest on folios from one of the earliest surviving Qur’ans written during the lifetime of Mohammad or very shortly after his death in 632 CE. (calm classical music) – Multi-spectral imaging is when one uses the different wavelengths of light to try to reveal residues and texts and anything that’s in an object that’s not seen by the naked eye.

(calm classical music) We use light, starting in the ultraviolet, through the visible. The red, green, blue, on up to the infrared and then we combine those images, that stack of images, to better show things that are in the object that you cannot see. – [Woman] I think in this image here. – One of the strengths of multi-spectral imaging is to be able to see whether or not there is text underneath the text that we see with the naked eye. We have used multi-spectral imaging to try to reveal texts that have been scraped off and overwritten. In other words, a palimpsest. To date, we have seen no evidence of any other inks or texts underneath the text that is currently visible on the Birmingham Qur’an. (calm classical music) – Like the Birmingham Qur’an project, the Library of Stains project, also known as the Labeculae Vivae project, or the #StainAlive project (laughter) is using multi-spectral imaging to gather scientific data, but instead of looking for the undertext, we’re looking for data drawn from stains found on parchment, paper, and bindings in medieval and early modern books. Although multi-spectral imaging technologies have proven to be increasingly valuable for the study of cultural heritage objects, especially in regards to revealing palimpsests and identifying pigments, there is very little preexisting scholarship that has used imaging methodologies to try to characterize the stains found in medieval manuscripts. Our project evolved from the interests of three CLIR postdoctoral fellows in data curation for medieval studies.

Doctor Aaron Connolly at the Schomburg Institute for Manuscript Studies, University of Pennsylvania Libraries, who’s research interests focus on medieval medicine. Doctor Alberto Campagnolo at the Library of Congress, who researches the book in general and how to model a book as an object in the digital world. And myself, at the University of Wisconsin-Madison who’s research addresses how the materiality of manuscripts informs the historical context and content of that manuscript. Each of our institutions holds a significant manuscript and or print collection and each of us is working on independent digital projects that intersect with those collections in some way. The Library of Stains Project brings together these multiple institutions thanks to a Council of Library and Information Resources Micro Grant. And because of this generous support, we were also able to include three other very important collaborators in the project. Fanella France, Chief Preservation Officer at the Library of Congress. As well as Michael Toth, who you saw in the video. And Bill Christens-Barry. Both who provided the multi-spectral imaging equipment, as well as invaluable guidance during the imaging process at the Universities of Pennsylvania, Iowa, and Wisconsin.

Thus, from the outset, there was a natural confluence between all of our interests and backgrounds, which has culminated in a year long pilot study that we hope, as Mike described, will bring what is usually invisible to the naked eye, into the visible, and provide new insights for accessing and studying information concerning the material make up and historic uses of books. Why stains, you may ask. Well, one of the reasons is because of the intimate human connection between this, and this. (laughter) As humans, we’ve all dropped coffee on a favorite paperback, or read an old letter stained with tears, or left tracing of our own blood in a recipe book after too much excitement with a new knife. (laughter) The Library of Stains Project seeks to highlight this human experience that draws a connection between medieval and modern interactions with manuscripts. And in so doing we hope to broaden public engagement with manuscripts to include those who’s beauty may not lie in the consummate skills of an illuminator or binder, but in the human imperfections of daily life. Indeed, it’s been our experience that once you spend hours going through manuscripts looking only for stains, you can never un-see a stain again. (laughter) In addition to public engagement, we also seek to draw scholarly attention to books that have often been overlooked due to heavy soiling and damage, effects that can diminish their perceived quality and value. But the very stains that cause some scholars to dismiss a manuscript in fact carry important material information about the manuscript. For example, if a stain next to a medical recipe can be identified as an ingredient of that recipe, then we can known that the book was going used as an aide-mmoire, and that the recipes were actually being followed as written.

Or, if a book has wax stains throughout multiple folios, we may begin to surmise that the book was likely being read or written in the early morning or at night. And this would mean different things to different people, depending on where they lived. For librarians and conservators, the identification of stains such a mercury, would be very useful when deciding how much to handle a given manuscript. Drawing from the collections of our home institutions as well as two tangential institutions, Chemical Heritage Foundation in Philadelphia and the University of Iowa, we were able to gather data from the broad genre of manuscripts, and thus a broad genre of stains. At the Library of Congress we imaged some of the early printed books from the Rosenwald Collection. We also imaged medical manuscripts at the University of Pennsylvania Libraries and Chemical Heritage Foundation. And at the Universities of Iowa, and here at Wisconsin, we imaged liturgical manuscripts as well as classical works copied by University students. This aggregate collection has presented pertinent and varied questions. The goals of the Library of Stains Project were broadly conceived to address these kinds of questions, and to provide the first data set for characterized stains commonly found on manuscripts and early printed books. A sound methodology for the replication of the data gathering and analysis process, and the creation and implementation of an open source, open access database applicable to manuscript studies and librarian and conservation work.

We hope to be able to equip researchers, libraries, and conservators with additional tools for analyzing their manuscripts in relationship to provenance, use, transmission, preservation, and materiality. The project timeline highlights the four main stages of our work within the course of a year. Stage one was the very preparatory phase, where we developed our social media plan and we organized the logistics involved in coordinating and carrying out an ambitious imaging schedule in four different cities. Stage two involved imaging over 40 manuscripts at the various institutions during November and December 2017. The processing and analysis of the image data began in earnest in January 2018, and we’re continuing this analysis as we present preliminary results. The final stage we’ll focus on organizing the Library of Stains data in a repository, exhibitions of the manuscripts at the various institutions, publication, and a final assessment of the project. So this slide gives you a very general overview of our methodology and we start with a folio. And because we’re imaging it, we call it a side, and we gather the data from the stains. And that’s basically taking images of this side. We then have to process those images, and after we’re done the processing of the images, then we can begin the data analysis.

Once we’ve got a set of results and we’re just starting to bring together some of our results now, we’ve got three different conferences planned, and a publication that will be coming, that will be submitted in the summer. Finally, the last piece of this is to put together and publish the data that we’ve acquired in an online, freely accessible database. And that’s going to take two different forms. One will be a reference library that will be hosted at the University of Pennsylvania, and here at the University of Wisconsin. And a second one will be visualization of the Stains Projects. To differentiate those just a little bit, the data repository, or the reference library, will be mostly the images, the TIFF files that we have taken. Because those files really are raw data, and anybody who wants to apply any kind of imaging software, imaging analysis software, can do so, they don’t necessarily have to do stains. However, our part of the project is stains, and that’s what we wanna make sure that we get out in very accessible visualizations to the public, so that the public can go back and really appreciate and love those medieval manuscripts like we do. (laughter) So this is where we started with the imaging. In a room, quite often this dark, and we have this kind of a set up.

Here at the bottom, you’ve got your object, this is your manuscript. Up here above it, you’ve got a camera, we were using an 80 millimeter lens, and we were using a camera back of 60 megapixels, achromatic. And then over here on the sides you’ve got your two lamps that are emitting specific wave lengths of light, one at a time. And these are coming in mostly, we prefer they come in sort of at a 45 degree angle, and then they’re reflected back up here and the camera takes a picture. I’m gonna take you through sort of a very quick series of what those lights might look like. We were using for our project, 10 different wave lengths going from the ultraviolet into the visible into the infrared at these values here. And it would look something like this. Each one emitting just one of the wavelengths. Right, once we had a set of 10 images per side, then we needed to process those images and for this we brought in an image processing software called Image J Software. This is completely freely accessible, you can download it after this talk tonight if you’d like.

And it represents this top bar here. However, because of the nature of our project, Bill Christens-Barry has developed two additional toolboxes to go with Image J Software that help people who want to use multi-spectral imaging for cultural heritage objects, for this kind of project in particular. So here is the first one, and it’s called a Paleo Prep Box, and it makes it very easy and streamlined to flatten our images. And basically, when we’re flattening our images, we’re just trying to create a uniform field of brightness across the entire image so that we narrow down any kind of false positives in our results. The other toolbox that Bill has created is called the Paleo Toolbox, and this adds some really nice applications in addition to the Image J Software and he also then has included this little box that is constantly running as you perform tasks and it’s a log. And really, it’s very nice if you actually want to articulate or document your methodology as you’re going through your image processing. So, at this point I’m gonna turn it over to Leah and she’s gonna take you through some specific, what it looks specifically to go through some of our particular images that we did. – Alright, thank you. So this is the point in analysis where I came in to help analyze all of the data acquired from the manuscripts of these different institutions. And I handled the data analysis for the manuscripts at the University of Wisconsin, Heather handled the manuscripts at Iowa and you’ll get to hear about some of those later.

But first I wanna set you through this process. So, in order to analyze the stains in the processed images, we continue working in Image J. You would open up 10 different image files, and you need to then create a stack. When we take, it’s fairly easy to turn them into the stack, you just open the files, click, images to stack. It’s a really cool thing, though, when you get to scroll through it. So you have your stack of images and it’s essentially, I want you to picture this like a stack of pancakes, where each image in a pancake layered on top of the others, because then what we’re gonna do it plot a Z-axis curve, and that’s by taking a sample, if I can show you right here, and then gets measured into the values of this chart. And so what this is charting are the Z-axis. And so the Z-axis is like sticking a fork through a stack of pancakes. You have your X and your Y that are flat, and then the Z-axis goes right down the middle. And so then this sample is essentially a third dimension measurement of the same spot on all 10 different images.

And that’s your Z-axis. And so that gives us a curve like this one. But the trouble is that this curve isn’t actually scaled right so that each wave length is appropriately spaced along the color spectrum. So we need to take that data, the numeric data, and transfer it out into Excel, which is where we do all of our curve generation. So we get this table of numbers, which are the mean intensity values of the pixels in the area that have been sampled, with a different value for each image in the stack and a different value for each image at different wave lengths. So that when we chart against the appropriate wavelength, they produce a spectral curve at the correct scale that we can compare against other stains in the same side or against stains in other sides. So, before we do the curves in Excel, I wanna say a few words about how we pick where the samples come from. How do we decide what to sample. And so, for starters, I can show you essentially a re-creation of what the stack looks like and how we move through it. It’s like flip book that we can move through and see the different monochromatic images of the same side with different wavelengths of light.

How cool is that? So, for example, something we can notice in this, which is a piece of music that you’ll hear more about later, in this image, at this wavelength, you have inks with the lyrics, you have inks making musical notes, and you have inks marking the stave of music. But at this wavelength, the stave starts to disappear. The stave is apparently not reflecting to the same extent that the words and the musical notes themselves are reflecting at. And so that’s an indication that perhaps they might have different spectral curves. So that’s one way of knowing we might want to sample the different inks. Of course sometimes manuscript is just really dirty and it’s easy to see that there are stains there. So again, as a reminder, we have our Z-axis going through multiple different layers of the stack measuring the reflectance at a single spot across 10 images. In order to measure exactly where we’ve taken our samples, we’ve recorder them specifically where they are on the page, not just generally this is the ink, but specifically where on the page we’re taking this because we need to find a particularly dense instance of a stain. And in some cases, a side will be very simple, like this one. It has the parchment, we always check the parchment as well as whether it’s paper or parchment, we check the main ink if there’s ink writing on the page, or different inks, and then a couple of stains.

Which might be distinct, but that are probably representative of what’s going on. And that’s a fairly simple one. And at this point, I do want to point out that as a control we also sample this white box right here. And you’ll notice it’s labeled Macbeth, that’s a protocol that we are following from the Library of Congress’s way of doing multi-spectral imaging. So, we’re not calling out the Shakespeare play ourselves. (laughter) So the Macbeth value is basically a control, because what we’re doing in Excel, and we’ll get there in just moment, is taking a simple formula of dividing the reflectance from the places we samples over the control, the Macbeth, to produce the reflectance that is controlled across all of the different wave lengths. So, we have some very simple sides like this, where there’s just an ink, a couple of stains, and it’s fairly easy to decide what to sample. Other sides look like this. And you have a lot of stains, maybe some writing. This one actually had some writing that wasn’t visible to the naked eye, but at certain waves lengths it showed up.

And then you have stains overlapping on top of each other and it’s very difficult to tell what it similar, what is different, and how to we decide how to measure it. So, this is where the spectral curves come in. So we bring our mean intensity values into Excel, and it looks something like this, we control for Macbeth, that’s this formula that’s going on over here, each of these chunks of numbers is going to be a different sample from the side. And then we have several of those lined up, fitted into this chart, which gives us spectral curves. And from here, we can start to compare and untangle the spectral curves. These are the curves from that very complicated manuscript and if we zoom in on those, well they look kind of similar but it’s actually kind of hard to tell, so one way we go about analyzing is simply blowing up the Y-axis so that we can get a better look at the lines of the curves separately. And this super complicated manuscript actually turns out to be fairly simple. Because if you’ll notice, a lot of these curves are following a very similar path with this bump down here toward the blue end of the spectrum, and then this larger curve up over by the red end of the spectrum. And you’ll notice, of course, they do not line up perfectly. So we’re not about to start making a claim that these are the exact same compound all over this side, but we can same that it’s probably a fairly similar compound.

And there are a couple of reasons why these lines might not line up perfectly and still be either similar or the same compound. One explanation might be that they’re the same compound at different intensities, either a particularly dense spill versus a watered down spill. They might have a slightly different composition. We know that things like ink recipes weren’t always followed exactly or the measurements weren’t precise. And so it could be the same ink, but in a different batch so it’s a slightly different formulation. It could also be an older stain of a very similar ink or other compound. And another thing that might come into play is that the environment in which the stain is exposed to light, air, heat, humidity, things that could effect the aging of the stain, could effect part of a manuscript differently than other parts. So, those are some reasons why we can say that even those these have some important distinctions between the curves, the general shape being so similar suggests that they are in some way related. Or at least that some of them are related in families of stains. So another reason we can say, be comfortable with the not so perfect matches is that when we put together similar inks that we know are similar inks, or that they can be, from a bunch of different manuscripts, they do the same thing.

So these are all of the blue inks that we sampled from Wisconsin manuscripts, and you’ll notice that they have this, generally have this big spike at the beginning. That’s the blue part of the spectrum of visible light. So, it’s common sense that that would be a big part of how they are showing up in the spectral curve, but then we also have these features where a lot of them have another little bit of a bump back here. And so, we can start to characterize, like, okay, there are a couple of different ways that blue inks show up as spectral curves. And we can do the same thing with red inks. We had a lot more red inks than blue inks. We can see that they have a pretty high reflectance at the red end of the spectrum. And so seeing that across manuscripts the spectral curves for these colored inks are showing up pretty closely together, that confirms a bit our method of comparing the curves and being able to see that even if they’re not exact, they’re probably related formulas. So I wanna walk you through now a couple of the basic means by which we produce our preliminary results. So, take you through a couple a sides from the same manuscripts, these are two pages from different parts of the same manuscript codex, or book.

So if you open it up, it’s not two facing pages, but they’re from different parts of the same book. I wanna take you through how we went about sampling, identifying, and figuring out a little bit of what’s going on with these sides. So, first, let’s look at this side. What we have here is a dark ink that’s most of the text, there’s a red ink with some important text. You can also see in these little empty squares, there are, there are little letters written in you zoom in, which are indicators for someone who’s gonna come in later and draw a decorated initial, so like a large first letter. So, that hasn’t happened yet. This manuscript, you might call unfinished, but it seems to have gone into use. So those open spaces have a little letter written in them. And we don’t know, generally, if that letter was made when the manuscript was being made as a cue to the artist, or if it was added in later by a user who decided, oh this is obviously this letter, to aid himself and future readers. Another interesting this about this side, is that down here in this red box, is what’s called a manicule.

Which is a pointing hand saying this is an important piece of text. We see them in a lot of late medieval and early modern manuscripts. So, we have these two additional inks and then we have a couple of stains. In this, it’s lower here, but it’s the outer margin, in the outer margin of the book, we have a couple of stains. So, when we put these into a chart of curves, we’ve got a couple of pretty expected things. The parchment is this higher gray curve. Then we notice that these two curves, which are the initial notation in those little, the first letter of those blocks, and then the manicule are actually in very similar inks. So perhaps those happened at the same time. We also see that these two stains are fairly similar as well, except that they spike up at the beginning and that’s a pretty common thing we’re seeing when something is a fairly thin or light stain, it’s basically the parchment showing through. So what we can see here is that these stains are basically inks, probably.

There’s a strong likelihood given how close they are, especially at the high end of the spectrum, that those stains were made by the same person who drew the manicule, and possible also filled in the first letter of those text blocks. So that’s exciting, we’re seeing a person making a stain. But sometimes we have more exciting stains than that even. (laughter) This one, we use it in a lot of our process presentations, because it’s just such an exciting side in that, like, what could possibly have been spilled on this page? Spoiler, we don’t know, but we could tell you some stories about what we think it might be. So, this page had a bit more going on, but it’s again from the same manuscript. We have our dark ink, we have our red ink, we have another bit of marginalia in a similar ink to the ones from the last side. There are also some parchment distortions that I measured. They weren’t anything exciting, but I wanted to explain those for you. So, we sampled this big ring stain, and stain one is the ring stain. There’s also another one, but it’s not as interesting.

So, what we get in our spectral curve, again is fairly expected, here’s our parchment curve all up there, and then down in the curves in here we can see that our stain one doesn’t really match anything else. It kind of matches stain two, so okay that’s probably a splatter of the same thing making the ring, but really stain one stands alone. It’s not anywhere near a match for the inks on the page. And that maybe isn’t surprising, because that would be a very strange way to spill your ink. But, it tells us something very interesting in that we can now move ahead and think about what else it could be. We’ll say more about our control samples that we’re putting together, but we can test whether it’s maybe tea or oil or wine, just to see what might have discolored the parchment in that way. And that’s something that we can take in future directions. So these kinds of ways of thinking about individual sides of manuscripts, we’re calling Stain Stories. And this is something that I inherited from a journalism and geography scholar studying soil stories, so I’m stealing it. But we’re calling these Stain Stories because what we can do is create fragmentary narratives of a manuscript’s history and it’s use based on the spectral signatures of it’s curves.

So by looking at those graphs, and looking at the pictures, and looking at where we took our samples, and what the stains look like, we can start to tell stories about these stains. And for now, of course, these are also hypotheses, these are hypothetical stories of what might have happened. Because there’s always a sense of uncertainty with this kind of history, and we also don’t have quite the large enough catalog of spectral curves to make certain identification. But for now, we have relative data, relational data, that’s not necessarily hard and fast identification, so it’s giving us goals for future inquiry at this stage. So to recap, from the sides I just talked about. The first one, with the manicule, we can think of the person going through and writing in the letters of the initials to be drawn later by an artist, they were maybe gonna have someone do it after they’ve owned the book. They were kind of careless with their ink and smudged it in the margin, perhaps. For the larger ring stain we can say that this is almost certainly not ink, and the possibility remains open that it could be any of a number of characteristic stains like wax, wine, oils, chemicals, or other compounds that we have not even dreamed of finding in medieval manuscripts. But now that we have this base line, we can go find out more about that particular side’s Stain Story. And now I’d like to hand it over to Heather to talk a little bit about wax.

– Sorry. So my first Stain Story is a story about wax. And, it’s based on this question you see here. Can we use the clear identification of wax in a spectral curve on one folio to match stains on other folios and other manuscripts. And I start with these two very heavy, dark lines down here at the bottom of the spectral curve. These two come from a folio, they’re two stains on the same folio, in a psalter held at the University of Iowa. And they are wax. And I know this because you see the wax on the folio. It’s still there, and it’s thick. And I think that’s probably why, down here, there’s a bit of gap here because what you’re seeing here in the real wax, the very intense wax, and what happened then is I took these curves and I thought to myself, well I went through all the other spectral curves and I said are there any curves in other manuscripts or on other sides that might match these wax curves that I’m pretty certain are wax.

And, yes I found at least four in the University of Iowa collections and I’m gonna take you through those a little bit. But these, you will see, are residues, the wax in no longer on the actual page, all that’s left in the residue of the wax. And I think that’s probably why we’re seeing this discrepancy at the lower end of the spectrum. But these here are the exact folio. You can’t see it super well here, but what’s circled in blue, those are the wax stains, the wax is still there if you had it in front of you, you would see the wax, you could feel the wax, and it’s definitely wax. And those two, again, are the two stains that create this very bottom dark curve here. This is a folio from the same manuscript. And this stain, too, which is right here, is one of those other lines just above it. This is a folio from the same manuscript, and this stain here, right on the end of the page, also matches that curve, and it is not the same as this stain over here. This stain is from a completely different manuscript.

And it’s right here, near the gutter of the manuscript, it matches with the others, it is not a match with this stain or this stain. And this stain here, also is matches with the others, as does this one, I didn’t circle it but it does. They are the same stain, and they match this similar curve that I’ve just been talking about of wax. And I think what’s really interesting here in this wax story is that you can start to see that a lot of these stains are on the outskirts or in the margins of the folio, and now you can start to see the manuscript in front of a person who’s holding a candle, either at the top or the side or the bottom, but certainly around the margins of the manuscript, not over the manuscript. And this in particular, if this is your manuscript, and your candle’s right here, and perhaps we’re turning a page and we accidentally knock the candle and it splatters onto the manuscript. And that’s my wax stain story, I’m now gonna pass it over to Leah for another kind of Stain Story. – Thank you First, I get to do my favorite part of the talk. So, some implication of Heather’s wax story. From my field, which is Anglo-Saxon studies, the study of English literature before the Norman primarily in Old English, I wanna introduce you to the Exeter Book, which is one of the four major codices surviving containing Old English poetry. It contains a lot of major poems including The Sea Farer, The Wanderer.

Many of, well all of the Exeter Book riddles, which are fabulous, you should read them. It also contains two poems that I write about in my dissertation, so it has a special place in my heart. And it’s also marvelously stained. (laughter) So, zooming in just on this, sort of what’s the front page now, we have one of those ring stains. We have this huge spill of something dark. And if you also look at these markings here, at some point it’s been used as a cutting board. (laughter) I wanna point out that these are poems about the birth of Christ. And for most of it’s history, this book was held in a cathedral. So I have a lot of questions about what circumstances led to it being used as a cutting board. (laughter) But, speaking of wax, there’s also this damage to the manuscript.

This is the back of the book. At some point, something burned it. We know that these are burns because of singeing. But several poems on these last pages are damaged. The ones in this opening are The Husband’s Message and The appropriately named Ruin. And here’s the thing, we don’t really know how it was burned. It’s not like most manuscripts that are burned that are all singed around the edges because the library caught on fire. Some theories are, though, that either a hot poker was dropped on it or that a candle fell on it. And I believe if we were able to get multi-spectral images of the Exeter Book, and recognize the spectral curves of wax around the edges of these holes in the manuscript, it’s not like we’re bringing back additional letters or words or poems, but we can give this damaged book just a little bit of history in terms of the exciting moment in which a substantial chunk of Old English literature was destroyed. So, that’s that Stain Story, but let’s talk about music.

We saw a bit of this manuscript before. I stepped you through a sample stack of one of these sides. So, this is a manuscript that is held in the University of Wisconsin collections over in Memorial Library. We took three sides from it, but it’s all one piece. So you can see that it sort of curves around and there’s some overlap here. The interesting thing about this manuscript is that it’s a single piece of parchment that at some point was part of a book of music. And then at some point was taken apart and used as the cover of a book, a codex. So you can see that here in this middle side, that’s the spine of a book with very characteristic wear. And naturally as the cover of a book it also got stained. So we have some questions about this.

We wanna know, are these stains from when this piece of parchment was the cover of a book, are some of them from when it was a page in a music book, and potentially what was getting spilled on this book. So, we took a lot of samples. Quite a few because there a couple of different inks on each side, there are a couple of different stains of each side, and the spine is particularly exciting because it’s just dark everywhere. And when you put all of these different samples together, we end up with this chaos. (laughter) These are all of the samples from all three sides put together into one chart. Don’t worry about interpreting it, because I’m gonna break it out for you. So let’s start with the inks. We have here these two higher curves, these are samples from the music staves. And remember back in the stack, when we were flipping through it, the musical staves disappeared a bit, where the ink doing the lyrics didn’t? That’s what’s happening here. These have such a different curve that they disappear in some wavelengths.

And these again are not very close together. They’re probably the same ink because they’re the staves, the lines that the music is written on, ultimately the same page just on different sides of what became a cover of a book. So, a hypothesis for why they’re so far apart is that maybe one of them got a lot more wear than the other, maybe one was exposed to light more, maybe on got dirtier, based on which side of the book it was on. So that’s the staves. We have another ink here which is actually a fairly typical black ink curve, it’s a bit of writing on the manuscript. But then we have all of these guys. All of these dark inks down here, which are actually peculiarly low for even black inks. And so, I came up with this hypothesis. That these dark inks, which I started calling the low riders, that they might help us establish or develop some information about this manuscript. Because currently, right now, we don’t know exactly when this manuscript was made, when it was turned into a book, or where either of those things happened.

I though maybe the low rider inks, if we can connect them to other manuscripts that have similar inks, that might start to develop a pattern and terms of geographical areas or moments in time where this ink formula is in use. So, I assembled all of the dark black inks, the potential low riders from across the Wisconsin and Iowa manuscripts that we sampled. And again, because they’re very dark, we need to blow up the Y-axis, but looking at them very closely, they’re actually not very similar after all. The up side is, there’s no low rider really, but sort of a family of low riders that have some similar characteristics. And it’s actually by putting these inks from the music manuscript in the context of other dark inks that we can see they’re actually very similar after all. So the music manuscript’s dark inks are these yellowish green and blue curves that compared to the other crazy curves, are actually fairly similar to one another. These curves going all over the place, they are not probably the same ink formulation, they’re just also very dark. One interesting thing I do want to point out is that these two are a curious, very close match, and you’ve been noticing, we keep talking about why they’re not perfect matches. This is a suspiciously perfect match. And the darker of these two is actually something that is in an Iowa manuscript that Heather, from examining it in person, has good reason to believe is rust.

And so, if this, the lighter orange, is also kind of rust, we might wonder, okay it has this bump down here like the inks do, is this perhaps an ink that is composed partially of iron which is not an uncommon thing, but a particularly rusty iron? That’s an interesting thing we might follow up on. But let’s turn back to the stains. Cause remember, there were a lot of them. So, the first stain we can talk about it one that turns out to match the ink of the stave. So, it’s this stain here. You can see extending sort of from the stave, looks like it might have even been a mistake before or after the lyrics were written and it’s hard to tell. But it’s got a very similar curve to that of the staves. So that’s probably the same ink. And then we have a slightly more exciting one. Over around the spine, we have these, it’s duplicated here because it’s two different angles, but there’s this big stain that respects the line of the spine.

And that means to me that it was stained after this was a book cover. Because that would be a very weird coincidence for it be stained along those lines when it’s just a flat piece of parchment and then it’s, that’s just where the spine it. So it’s probably after it became a book. And then, if we go back to those curves, these two close ones in the middle, those are two different samples of that definitely the same stain, because it’s the same stain from different angles. These two are two other stains. So if we’ve got those two, there are two more stains that have very similar curves, though different intensities. So, related but not necessarily, they don’t have the same curve in the same way that the two that are the same stain from different angles do. And what I find interesting about this is that if you think about this being folded as a book, because it was the cover of a book, you have a stain along the spine, you have another stain on what we might think of as the front cover, but then there’s also this other stain on what would them be the back cover. And so, somehow whatever this is stained both the front and the back of the book. And, my inclination is to assume that someone has their book folded open and plopped down table the way we try not to treat our books today.

But then it wouldn’t have that really neat line right along the spine of where the stain wasn’t really able to seep. There’s something interesting going on there, and finding out what that might be could tell us lot, I think, about what this manuscript has gone through, what it’s seen in it’s many centuries. So for comparison, we have this other stain. The one that’s probably the same ink as the staves, this is probably a stain when the manuscript was being written. But the other stains, all of these ones in red circles, those were probably made when it was a book, when it was the cover of a book. And for future consideration of this Stain Story, we might think more about how to match spectral signatures of different kinds of inks and stains to give us information about date and place of origin for manuscripts. That’s something that may be able to get developed in the future. But we also wanna think about whether these non ink stains can be determined to identified so we can determine something about how this book was used, how the parchment was used when it was the cover of the book. So, would you like to do one more Stain Story? – Sure. If you will indulge me.

So, this is the last Stain Story that we have here. It comes from my colleague Alberto Campagnolo at the Library of Congress. The manuscript itself comes from the Chemical Heritage Foundation, Othmer one. Early 15 century recipes, textural extracts of alchemy, medicine, medal working, and cosmetics. So, at first glance, you’re looking at this stain and you’re thinking that could be some really cool thing, right? Spoiler alert. Okay, so we image it and we image the stain as an entire unit, and we get these spectral curves. So here is the ink, pretty standard spectral curve for ink. This is the paper, the purple. And then we get, sorry, the blue is the paper, and then the purple is the stain. And it’s a little bit erratic and it doesn’t really say much other than it’s not paper and it’s not ink.

So, when Alberto then went back through the images, he realized that the images, something that is not visible when you just look at the manuscript, there’s actually ruling on these images. And the ruling underneath, which is very usual for medieval manuscripts, but it’s not necessarily visible to the eye. And, again, this is probably not unusual, in the sense that especially if a medieval manuscript is passed through a book dealer in the 20th Century at some point they may have erased the ruling or tried to get rid of the ruling, as if it was something that was detracting from the value of the book. But, once he realized that there was ruling as well, he went back and redid his methodology and instead then decided to image the center part of the stain, the darkest part, as well as an outside part of the stain. And there, you start to see very interesting results. Because, the center green almost follows identically the ink that is used for the writing. And the outside of the stain, which is this light blue, almost follows identically a different, slightly different ink that was used for the ruling. And so basically, in this Stain Story, what we get, two different stains, what we assumed at one point was one stain, it’s two and the one on the top would have been made, likely have been made, when the ruling was taking place which is always before the writing. So, two different stains, two different times, two different moments in this manuscript’s Stain Story. – So I want to say a few words as we move toward concluding about why this matters, why this is important to us as scholars of various different fields.

For me, as an Anglo-Saxonist, as a Medievalist, even these preliminary results give us so many new directions. So, they’re all preliminary results that we don’t have any hard and fast answers yet and the goal here is to develop a methodology, right? The beginning of a methodology. We can already affirm though, from our work over the last several months, the value of these methods for the humanities as a whole, and especially for medieval studies and manuscript studies. Because the study of medieval manuscripts is not just about language and literature, even for someone like me, who’s housed in an English department. We’re also investigating the cultural and historical contexts in which texts are created and distributed and received. We’re investigating where a text was made, or used, or read, and when was it made, used, or read. Who was doing the making, the using, or the reading. More abstractly, we care about the vehicle of language, and literature in the middle ages. We care about the material cultural objects, the inks, the parchment, the pages, and the books themselves, that are bearers of ideas, tellers of stories, and repositories of knowledge. And so, we care how these manuscripts were treated as carriers of texts, and we want to investigate further how these manuscripts and texts were valued, or in some cases not.

For example, how did the Exeter Book, which spent most of it’s history in the possession of a cathedral library come to be used as a coaster and a cutting board? (laughter) In short, we want to know how people used their books. We want to learn more not just about the literature and the text in these books, but also about the people and the culture and the history to which these books belonged. And Heather is going to share a bit about the next steps in that process. – So, in the course of this year, we’ve been able to get this far. We still have a few next steps that we are waiting for. And we’ll be doing before we finish up. One is actually imaging samples of known stains. So, part of this entire project and part of the methodology is to create a sample database of known stains. And thus, we actually stained pieces of handmade paper and parchment with certain stains that we thought would probably be common and might arise in the manuscripts here, that we are imaging actually. So things like water, olive oil, red wine, iron-gall ink, black tea, mold, and ammonia.

The ammonia solution is out effort to replicate urine and that’s not as disgusting as it sounds. It just so happens that one of the manuscripts that we imaged is actually called The Color of Urine. In the middle ages, doctors would look at the color of urine and they would gage it according to a scale or a spectrum and do diagnoses of your illness as it were. And in this manuscript that we imaged, the entire thing is covered with yellow stains. (laughter) And all we can image is some doctor accidentally spilling the vial onto the manuscript. So we just put that in there, we have to know, in the end we have to know. These have not been imaged yet, they’re waiting to be imaged. As soon as they are, they will give us a set of spectral curve against which we can compare several of the curves that we have already imaged. Next step is to get our data organized. Because we are absolutely committed to making it open access and sharing it with other scholars.

As I talked about earlier, this will be a data repository of images, raw images, TIFF files, that show the layers basically for each side. And then scholars can take that and they can use any analytical tools they want on that, doesn’t have to be stains, they might wanna look at pigments, they may wanna look at inks, they may wanna look at parchment or the substrate. And then, in this we will also include a methodology of what we did so that others can follow it. But also important of this idea of a data visualization and for this we’re using a different environment and this is the Digital Mappa environment that Martin Fois is gonna be talking about on May 16th. It’s been recently released, designed and developed by Martin in the English Department here at the University of Wisconsin. And this works for us absolutely perfectly, because in it, we can upload the TIFF, the image as a peg, we can create a highlight around the stain or the area that we actually imaged, and then we can link that, create annotations and link that to various different pieces of data that are associated with that highlight. This is just a mock up of a testing kind of model, it’ll probably be a little bit more robust than what I’ve just got on this slide. But this will be able to be done for every side that we imaged at Iowa and Wisconsin and will be freely accessible and open to the public as well. That’s it. We wanna thank you so much for coming, and listening to our Stain Stories tonight.

(applause)