Tag Archives: codicology

Hair and Flesh

After giving a lecture on codicology in Stuttgart last month, I got a follow up question sent via the organiser (Dr Anja Thaler) regarding differentiating between the hair-side and flesh side of parchment. Rather than just replying, I thought it would make a good subject for a blog post (especially as the month is almost at an end and I’ve not written one yet). So, without further ado hair and flesh.

Before musing on my own experiences, it may first be useful to outline what I mean by ‘flesh side’ and ‘hair side’. First, though, we need to step back and consider parchment (sometimes called membrane, sometimes vellum if derived from cows or bulls). Parchment is the de-haired skin of an animal (particularly cows, sheep or goats, but hey, go wild), which has been soaked, scraped and most importantly dried under tension. The modern definition that leather is prepared through soaking in tannic acid while parchment is not, do not hold for the medieval period, the defining point is the drying under tension. I digress. Imagine the skin while it is still on the source animal, it has two sides: the side facing out from the animal, which has hairs on it, and the side facing in towards the animal in questions flesh and muscles and bones. Flay that skin in one big piece from the (dead, please!) animal and it still has two sides, the ‘flesh’ side that originally faced inwards and the ‘hair’ side that faced out.

Soaking the skin in water, perhaps with some appropriate plant materials added to make a dilute solution of tannic acid or some quicklime, helps loosen the hairs. Scraping the hair side of the parchment helps remove the loosened hairs. More thorough scraping removes the physical structures of the follicles themselves and helps to thin the parchment. Scraping the other side of the skin is also beneficial, any fat left clinging to the flesh side can be removed. Ideally, after its been stretched and dried (and scraped some more during this process) the skin will end up smooth and thin, without too much distinction between the two sides. The practice of pouncing, that is applying finely powdered chalk to the surface, is not always employed, but when it does it makes for a smoother writing surface and causes the hue of the parchment to become paler and more homogenous. Which can add further complications to the job of identifying which side is which.



This is a manuscript I’ve not yet had the pleasure of looking at in person: Madid, Biblioteca Nacionale MS 413 (I chose it because the images are freely available and hosted by the library on a creative commons license). It’s a tenth- or early eleventh-century century copy of the Lombard laws originally produced in southern Italy, an overview of the manuscript can be found here. The image here shows an opening with the hair side of two facing folios (30v and 31r). The parchment, from the photo at least, seems relatively pale, but look in the right hand margin of the folio on the right, and a clear area of follicle patterning can be seen. Contrast that with the next opening (fols 31v and 32r), and the relative paleness of the flesh side (even in photo) becomes apparent.



When I first began doing my own manuscript research in the second year of my PhD, I remember sometimes struggling with identifying which side of a folio was the hair side, which the flesh. For every folio where it was instantly apparent, there were others which caused me to scratch my head in confusion. It got easier with time and practice, most things do of course. Knowing that, as a general rule the flesh side tends to be paler and whiter in hue helps, especially when the more starkly coloured hair side has not been pounced, is a useful hint.

During my doctoral research I gathered some useful secondary resources on the subject, and can highly recommend both Robin Reed’s Ancient Skins, Parchments and Leathers and Christopher Clarkson’s ‘Rediscovering Parchment’ (full bibliographical details at the bottom of this post). The majority of the information in this post (and in my brain on the subject) has been assimilated from these. It’s worth noting some of the distinctive points:

  • Parchment from kid skin tends to be white on the flesh side and grey on the hair side with older goats often having grey-black regions on the hair-side.
  • Parchment from lamb skin tends to be yellow on both sides, while sheep skin tends to have a paler flesh side, and
  • Parchment/vellum from calf skin tends to be creamier in colour, but again with a darker colour on the hair side for mature animals.


Reed notes that distinguishing between goat and sheep can be almost impossible (even with living animals, and includes a cute photo to prove it), while Clarkson suggests that goats differ and can often be distinguished from sheep or calf on account of having softer and more flexible texture.

Both also note that older animals tend to have more established networks of veins and arteries, the presence of which can sometimes be seen or felt in the texture of the parchment even if remaining iron in the blood has not reacted with tannic acid to leave a dark stain in the shape of the vein. Which apparently can happen, although I have not yet seen it in person. Similarly, older animals are more likely to have suffered injuries and scar tissue or (vastly) stretched holes from tick bites, etc., can frequently be spotted on the surface.

Sometimes the hair side of the folio would have been so well scraped that details such as follicle patterns and bits of skin colouring are not really apparent on the surface. Sometimes later damage, abrasion and the such like to the flesh side can have darkened it so much it looks like hair side. Sometimes the parchment has been laterally split to make two sheets, one of which has a hair side and an ‘inside’, as it were, the other a flesh side and an ‘inside’. I’ve not seen this in any manuscripts I’ve studied myself, as it is a later practice not yet developed or employed in the tenth to twelfth century law-books that my research focuses on. It’s a detail worth remembering though, especially as Reed comments that many extra thin parchments assumed to have been produced from uterine vellum (that is parchment produced from calf foetuses) has probably actually been produced as a split skin instead.

Needless to say, in my early days in the archives (still less than a decade ago) I frequently confused hair and flesh, or spent what felt like hours pondering a folio, flipping back and forth. Sometimes I’d just put a question and come back to it. (Hands in the air honesty, it still happens sometimes). On more than one occasion I’d spend ages pondering my way through quire, finally decide on some tricky instances and step back to look at a quire diagram – that gave the same face of two halves of a single bifolium hair on one part, flesh on the other. It happens in recording, but clearly not in the manuscript. The bifolium though sometimes provides an easy solution, if one half is tricky to the point of frustration and the other has a nice, diagnostic feature such as a patch of follicles then the information can just be transferred across the quire. Tricks of knowing where else you can (should) look become a part of the working repertoire alongside an ever attuning eye and holes in the memory that let you slowly forget the difficulties you once had.



‘Madrid, Biblioteca Nacional 413’, in Biblioteca Legum: A Database on Secular Carolingian Law Texts, ed. by Karl Ubl <http://www.leges.uni-koeln.de/en/mss/ [Accessed 28 July 2016]

Clarkson, Christopher, ‘Rediscovering Parchment: The Nature of the Beast’, in Conservation and Preservation in Small Libraries, ed. by Nicholas Hadgraft and Katherine Swift (Cambridge: Parker Library, 1994), pp. 75-96

Reed, R., Ancient Skins, Parchment and Leathers (London: Academic, 1973)


On Ruling grids

Having just spent a month and a half travelling around France and Northern Italy looking at some of the manuscripts of the Liber legis langobardorum for my research project, I am now back in the office and writing up copious amounts of notes. One of the main things I have to be working on is the descriptions of the ruling grids, as used by the scribes to lay out the page and position the main text-block, decorative initials, glosses in the margins and the such like.

I’ve previously found analysis of these to be highly informative when considering the general ways in which a given manuscript was physically produced, and long hours spent considering the minutae of exactly which lines are extended to the outer edge of the folio or across the centrefold of the bifolium often prove to be worthwhile. Even if at the time, barely perceptible scratched lines, damaged parchment and dozens upon dozens of folios can cause the eyes to cross and the attention to wander.

At the moment I am writing up the details for Padua, Biblioteca del Seminario Vescoville, MS DXXVIII, one of the latest copies of the Liber legis langobardorum dating to the first quarter of the twelfth century. It is a plain manuscript overall, written in a single column with simple, pen-drawn initials for each clause offset from the text-block in the left margin, made in what appears to be the same ink as used for copying the text itself. There are glosses copied in the outer margin of many of the folios (they appear to be in the same hand as the main text-block, but detailed palaeographic analysis is currently a little further down my to do list), and running headings in the upper margin of each folio, recto and verso.

Here and now, however, it is the ruling grid that has my attention. My focus in this post is to consider the methodology of describing and analysing these, the implications for the production of the book (skirted around here somewhat, but nonetheless present) must wait for a fuller article on the manuscript itself. A summary glance at the data could easily be used to say that only two grids are used: Grid A has 38 long lines to the column and is used on the vast majority of the folios, while Grid B has only 37 lines and is used only on folios 31-33 and 60-61.

Looking at which of the long lines the scribe chose to extend to the outer margins of the page and into the centrefold (close analysis shows that here the ‘through lines’ do not actually go through, but instead stop just shy. Each folio was ruled individually, not as an open bifolium), gives a more detailed picture. Between variations in extension of the first two to four lines, the last one to three lines, and any of none to three lines around the middle of the page (between lines 16 and 23), a much larger range of ruling grids appear. three for type B, and at least 24 for type A.

In fact, there may be more than 24 of type A, as there are some subtle variations on the given theme, usually where only one line is apparently missing from one side of the grid. As the lines are often faint and difficult to discern, I have given the grid the benefit of the doubt and labelled these variants with an additional code e.g. Grid AD2 or AD3, rather than giving them a whole new code to themselves. If the variants are included, a further eight ruling grids can be added to the total – making for thirty-two of type A (38 lines) and thirty-five in all. A sizeable number for a manuscript of only eighty-three folios, averaging at one for every 2.4 folios.

In all those grids, the duplicates are generally adjacent. Despite the overall similarity of the grids, there are surprisingly few repeats of exactly the same pattern. Grid AP first appears on fols 46-49 and again on fols 6-67, while Grid AF appears on fols. 14-15 and 22-24, then perhaps in a variant form (AF2, which includes line 36 as a ‘throughline’, but not as an extender) on fols 76-79.

My initial instincts are to draw a diagram to summarise each one, and it is an approach I have used in the past on other manuscripts (both the collection of Anglo-Saxon laws in Cambridge, Corpus Christi College MS 383 on which my doctoral project was based, and on the copy of the Liber legis lagobardorum in Vienna, Österreichische Nationalbibliothek MS 471, that formed the foundation for my current postdoctoral research project). It may still prove to be a useful tool here, but the variations in the ruling grids are subtle and I am not sure how much information the reader might glean from comparing to ostensibly similar grids. Take for example grids AA and AB below, where the only variation is that mid-lines 18-20 are extended in AA, and 19-21 in AB.

Ruling Grid AA (fols 2-5)

Ruling Grid AB (fols 2-5)

Instead, I have focused on tabulating the information, a PDF of the draft file can be found here. This provides a very dense appendix to an article or manuscript description, which is nonetheless far more approachable than the initial extended table of page-by-page data, painstakingly gathered in the library direct from the manuscript. Editing the data into this tabulated form is as much a process of analysis as presentation, and as actively constructed data (is there any other kind?) belays interesting suggestions on how the manuscript was produced. As such, there is a balance between the density of the information provided and its usability.

I wonder (idly, not as a challenge!) how many people will check the tabulated data in detail, preferring instead to refer to the key points that I draw attention to. This, then, like any research, is a case of forming, analysing and curating data. And as such, I will keep sifting through it partly to see what patterns and suggestions fall out for myself, but also to see if better ways can be found to present it.