The Collation

Research and Exploration at the Folger

A Conservation Intern’s Observations on STC 2608

A guest post by Kevin Cilurzo (with particular thanks to Adrienne Bell)

For a conservator, to disbind and rebind a book is a rare chance to study and understand its binding structure. With broken sewing and loose detached leaves, Folger STC 2608 needed major conservation treatment before it could be safely handled by readers.

Folger STC 2608 is a hybrid book combining a manuscript and a printed text. The manuscript portion was written by the Calvinist minister Francis Frampton (ca. 1607-1650) and contains his sermons and notes. The printed portion is Thomas Sternhold’s edition of The whole Book of Psalmes published in Cambridge in 1628.

Completing thorough documentation of the binding, especially of the sewing structure, was the first step before disbinding the book. In this post I will discuss some of the features of the binding and discoveries made during the documentation and the treatment of the book.

A full leather binding with a combination of blind tooling and gilt decoration. The binding includes two parts: the manuscript from 1632 and the printed part from 1628.

The sewing structure

The binding of STC 2608 is sewn over four recessed single supports of twisted alum-tawed skin. Two types of sewing patterns were used on this binding: one for the printed text and another for the manuscript.

A two-on sewing pattern was used to sew the printed text. Two-on sewing patterns allow the binder to sew a textblock much faster and to minimize the swelling of the spine because there is less thread in the centerfolds. It was the perfect pattern for the printed text, which is on a very thin paper, but less so for the manuscript on a heavy paper. This must be the reason why a different technique, the bypass sewing pattern, was used for the manuscript. Clearly the bookbinder was totally aware of what she or he was doing and why. In the graphic below you can see how the sewing pattern changes from “bypass” to “two-on” sewing. The thread is shown in red. Sewing supports are represented by dots:

Section 27 shows the main sewing pattern of the manuscript portion. The thread passes completely through each section even if it doesn’t catch every sewing support. In this example, the thread catches supports #1, #2, and #4, but #3 is bypassed. Sections 28 and 29 show the sewing pattern of the printed text. This pattern is called a “two-on” sewing because sections are always sewn in twos and treated as a pair. This is indicated by the large bracket in gray. It is wonderful to see how a single pass of thread links two sections.

Let’s take a closer look at the bypass sewing used on the manuscript part. It is a pattern designed to speed up the sewing process. In bypass sewing, the binder skips certain sewing supports rather than catching each of them. Usually this is done in a somewhat regular pattern, but here this is not the case!

The first sections were sewn all-along (and each sewing support was caught), but the binder switched to a bypass sewing pattern for the next 5 sections. Another two sections were sewn with a two-on pattern, and then again back to bypass sewing for the remainder of the manuscript part. For the printed text, the binder used a two-on sewing pattern. The last sections, like the first, are sewn all-along. Sewing the first and last sections all-along makes the sewing strongest in the areas that are used the most.

The animation shows the bypass sewing. The thread is not catching each sewing support and results in visible pattern at the spine. The sewing support is highlighted in blue. Sewing thread that catches the sewing support is highlighted in green. Orange highlights the sections where thread bypasses the support.

In the animated image below, do you see the tear in the center fold, indicated by the blue bracket? This gives us a clue about something that happened in the bindery and explains an irregularity in the sewing pattern.

The binder over tensioned the sewing thread and tore through the center of the section, indicated with the blue bracket. The original thread left an impression in the paper, which is perfectly visible as one straight line, indicated in red. The new thread is already in the centerfold and indicated in green.

Here, the binder over tensioned the sewing thread and tore through the center of the section, indicated with the blue bracket. If you have ever bound a book yourself, you know how easily this can happen. Instead of repairing it, the bookbinder changed the pattern and skipped the damaged part. How can we be sure of that? The original thread left an impression in the paper, which is perfectly visible as one straight line, indicated in red. An imperfection with a story to tell! The photograph was shot during the resewing of the text block. The new thread is already in the centerfold and indicated in green.


The binding is composed of 34 folded sheets, 27 for the manuscript, and 7 for the print. The manuscript portion is made up of two sheets folded simultaneously to make a quarto: i.e. the two sheets were placed on top of each other and then folded twice to make 8 leaves or 4 bifolia. The printed portion is is also made up of two sheets folded simultaneously; here, each set of two sheets were folded three times to make 16 leaves, or 8 bifolia.1 The folding pattern is very consistent, but six leaves are missing. In this simplified representation of the gatherings, the missing leaves are noted with -1.

34 gatherings: 8 IV + IV-1 + 10 IV + IV-1 + IV-1 + 2 IV + IV-1 + 2 IV + IV-1 | 6 IV + II-1

When a leaf is missing there is always the question of why. Sometimes leaves are lost accidentally and sometimes they are removed deliberately. Regardless of the cause, it is usually possible to see the evidence of their removal: a stub near the gutter, or a damage to the conjoined leaf on the opposite side of the center fold.

Have a look at the digitized opening that shows pages 126 and 129 in Luna. You can clearly see the jagged stub of leaf that would have formed pages 127/128, which seems to have been cut away with a pair of scissors. The stub of leaf for pages 458/459 also shows a straight and clean cut. If you compare these examples to the remnant of leaf before page 302, you will see a difference. Here, the traces near the gutter suggest that this leaf was crudely ripped out. Most of the missing leaves in the book show similar evidence of rough removal (for example the remnants of the leaves before pages 314 and 360).

When we take a closer look at the pagination of the text penciled in at the corner of each leaf in the manuscript portion of the item, we can extract some useful information. For example, where there is a remnant of leaf for pages 126/127, there is a corresponding gap in the foliation. This detail tells us that this leaf was removed after the book was bound. And this allows us to understand a small part of the history of the book. The pagination and foliation of this binding must be deciphered in different stages.


Watermarks can be viewed through transmitted light. A disbound book provides a great opportunity to study such features. Examining STC 2608, I found at least four different watermarks. Those in the printed text are sometimes incomplete and hard to identify due to the trimmed edges of the paper. In the manuscript, by contrast, they are very well-preserved.

The bifolia in the manuscript have a watermark of two columns with a pinecone in between. Looking carefully at the two photographs below you will see that the watermarks are similar but not identical. The reason for this is that the papermakers worked using a matched pair of two molds to make each stock of paper.

A pair of watermarks in the manuscript: two columns or towers with a Zirbelnuss (pinecone) in between. Both are about 55mm long. The Zirbelnuss may indicate an origin in southern Germany, particularly the city of Augsburg. To this day, the pinecone is the city emblem. The watermark database Wasserzeichen Online attributes several papers with the Zirbelnuss to Augsburg.

What is interesting in this case is that the two watermarks appear in a regular repeating pattern throughout the manuscript. This alternating order probably results from the process of the paper production. As mentioned earlier, paper was made alternating between two near-identical molds. A diagram of the sewing structure helps us to visualize this pattern.

This diagram shows the regular distribution of watermarks in four sections from the manuscript portion. The different colors indicate the watermarks of one pair of molds (watermark shown on the left, above, in cyan and watermark shown on the right, above, in magenta).

The location of the watermarks depends upon how many times a sheet has been folded to create a section. In this case of the manuscript part, the sheets were folded twice to create quarto format bifolia. A second pair of watermarks found in the manuscript depicts a design in the shape of a grape (or perhaps an abstract pinecone form?). The second watermark occurs in the same position as the first—at the spine-fold near the center of the page.2

Examining and Executing

The text block is separated from the leather cover and boards during the conservation treatment.

The treatment of the book only began after a thorough examination of the item and a conversation with a curator. Constant evaluation and experience lead to a successful treatment. After STC 2608 was disbound, the textblock was sent to Photography and Digital Imaging, where high quality images were created and made available online for researchers.

The treatment included the mending of tears in the paper, reinforcement of the sections at the center fold, repair of the original sewing supports, resewing of the textblock, and connecting it again to the boards and the leather cover. The textblock was resewn using  new thread, but following the original sewing pattern. The original thread was removed from the binding and preserved for future researchers to understand the past.

Images show the binding before and after the treatment. The loose sections and sheets put back in place. With careful handling, the treatment will allow use for years to come.

It is crucial to study and preserve the bindings of books and extract as much information from them as possible—before, during, and after treatment. Such work is always an interesting journey of discovery about the production of the book.

As an upcoming book conservator, my examination and treatment of STC 2608 truly underscored the importance of looking carefully. I think it is essential to share my personal observations with interested researchers and to put the knowledge of historical bookbinding practices to fertile ground.


Kevin Cilurzo is a book and paper conservator with training from the University of Arts, Bern, Switzerland. He was an intern with the Folger’s conservation team in 2017 and is now a paper restorer in the Graphic Arts Collection at the ETH Zürich.

  1. When preparing sheets for binding, the bindery will fold every sheet required to make up a gathering simultaneously.  By doing it this way, the folds sit inside each other with no air in between the bifolia which is important when the book block is sewn.
  2. Examination of the full sewing diagram shows that the watermarks do not have a regular order that they appear in.  The pattern seemingly established by  16IV – 18IV is not at all consistent throughout the book block.


  • Thank you Kevin!
    That was interesting! I had never seen the two-signature sewing before! Also interesting to hear about the various molds for paper production too! Very thorough explanations and interesting graphics for all the visuals too! Thanks again!

  • That is the most detailed binding analysis I’ve ever read! This was quite educational.
    Thank you, Kevin.

Leave a Reply

  • (will not be published)