From sawdust to gold dust: The conservation of a 16th century birch bark book

Shelfmark: OMS/Or 13300
Curator: Pasquale Manzo
Treatment Time: 113 hrs
Estimated time: 92 hrs

Introduction:
Late 2016 a black acidic shoe box with a note was transferred to the British Library Centre for Conservation for treatment. The note was dated from 1972, reading ‘object is extremely fragile - do not touch’. Inside the shoe box was a mass of tissue, which when carefully lifted out, had thousands of tiny entangled bark fragments entwined in its fibres. Beneath the tissue was the birch bark book.

The following blog is about how the manuscript was conserved so that it could once again be safely requested and handled by the general public.

The Manuscript:
The manuscript is part of the British Library Asian and African collection. It was originally made in Kashmir, is written in Śāradā script and dates to the 16-17th century. It includes three different texts: (A) Nirṇayāmṛta by Allāḍanātha, a work in 4 chapters concerning suitable times for various Hindu religious ceremonies. (B) Narasiṃhaparicayā by Kṛṣṇdāsa son of Rāmācārya, a text on Vaiṣṇava ritual and (C) a fragment of the Padmapurāṇa.

Binding Structure:
The manuscript was formed of 10 sections, each with 8 folios to the centre. The manuscript was sewn with a thick hemp chord in an unsupported Coptic-style. The sewing had a knotted incongruous double loop centrally on the 3rd section.

Figure 1: Spine edge of the text block showing the 2 central sewing stations and the headband tie-downs at head and tail.

 

Both the head and tail had headbands, similar perhaps in style to Monastic headbands. However the cores were made of a Mahogany-type wood dowel, looped several times over with chord, then wrapped in a layer of alum-tawed skin and finally a turned-in leather flap extending up from the cover. The headbands were attached to the text block via tie-downs on each section.

Figures 2-4: Wood-core headband with alum-tawed and leather wraps secured to text block with chord tie-downs.

The manuscript was covered in thick brown goat leather with an inner parchment wrapper. The limp leather case was attached to the text block by the headband tie-downs.

Figures 5 & 6: Limp goat leather cover with parchment wrapper below.

 

Dimensions: (L x W x H) 190 x 189 x 70 mm

The Text block:
The text block had 231 folios, foliated 27-258. The folios were transcribed with a carbon-based ink. The media was stable. The folios themselves were made from very fine layers of bark from the outer periderm of the deciduous birch tree. The cork cells of the birch bark are compacted in radial rows according to seasonal growth, and the periderm layers were adhered together by pectin as well as physical knots and streaks.

Figure 7 (Left): Tree anatomy diagram showing the periderm layer of the birch tree (Wojtech 2013). Figure 8 (Right): Peeling birch bark off a birch tree to use as a substrate. Image shows inherent knots and streaks (Wojtech 2011).

 

Condition:
Unfit: Significant risk of damage even under controlled display conditions due to existing damage or extreme sensitivity, inherent vice.

Principal Substrate

In response to fluctuating environmental parameters, the inherently weak pectin adhesive in the birch bark had failed, causing the periderm layers to delaminate. It appeared that each folio was made up of around 7 periderm layers. Each folio was in a different state of delamination. Similarly, as a result environmental fluctuation the natural resins in the birch bark had been drawn to the surface, causing a gentle white blooming on the majority of the birch bark folios.

Changes in the birch bark’s moisture equilibrium over time had caused dimensional changes, forcing tangential curling and making the bark stiff and extremely brittle. This had resulted in each folio having significant tears extending from the foredge. Some of the folios had even degenerated into piles of fragmentary leaves. There were vertical cracks on most folios towards the spine, due to the stress induced by turning the pages.

Figure 9 (Left): White blooming from natural resins on the surface of the birch bark. Figure 10 (Right): Delaminating periderm layers of embrittled birch bark.

Binding Condition

The leather and the parchment wrapper beneath were stiff and distorted and subsequently did not effectively cover or protect the text block. The cover was only partially attached to the text block by the headband tie-downs at the head of the spine. As such, when fully opened some of the text block spine was exposed. The sewing had broken at multiple points, however disparately the sewing structure was relatively stable. The opening angle of the manuscript had been compromised by the brittle substrate, so could open to around 60 degrees in a section or 160 degrees between the sections. Two sections at the back of the manuscript had detached and four pages were sitting loose in the back cover.

Figure 11: Distorted leather no longer covers and protects the text block.

Figure 12: Showing detached tie-downs, exposed spine and broken sewing at the spine tail.

Evidence of Previous Repairs:

Previous repairs were carried out in 1972. There appeared to be a white bloom on the surface of the leather cover, suggesting the possible previous use of wax or oil based leather coatings.

Two folios, f.211 and f.212, were coated on both sides with a texacryl 13-002 adhesive (at various strengths) and nylon tissue, with Japanese paper borders. This conservation process was discontinued as the result is visually distracting (sharp contrast in tone and light refraction) and had irreversibly changed the nature of the original birch bark.

The areas of loss, of around twenty of the most fragile folios at the front of the volume, had been infilled with hand-transcribed western paper. Similarly heavy-weight cream paper repairs had also been crudely adhered to multiple folios along edge tears.

Figure 13: Showing the crude paper repairs at the foredge of multiple folios

Figure 14: Showing the crude texacryl-adhered nylon lining on 2 birch bark folios.

Figure 15: The fragmented first two sections with annotated western paper infills.

 

Analytical Adhesive Testing

The decision was made to identify the adhesive used to adhere the paper patch repairs and the annotated infills. This would not only provide more information about the history of the object, but would inform the treatment decision-making process.

A microsample of adhesive residue was removed from beneath the distorted and lifting paper infills. The microsample was then analysed via use of FTIR-ATR to identify and characterise the present adhesive. The results suggested the adhesive used in previous repairs was a protein-based animal glue.

Figure 16: Image of the subtraction spectrum suggesting the adhesive residue was a protein based animal glue.

Figure 17: Adhesive residue visible on birch bark folio. Highlighted area shows micro-sampling location.

Conservation Treatment:

The choice of treatment for any object of historical or cultural significance must reflect its artifactual value, uniqueness and the accessibility of the information it holds. It was decided a minimally interventive treatment would be carried out with the aim of preserving the original and rare binding structure, whilst stabilising the folios for digitisation.

Repair Method Selection

Due to the laminar structure of the birch bark, it was decided that traditional paper-patch repairs would not suffice, as they would only encourage the delamination of the top layer. After experimentation, it was decided that thin strips of toned Japanese tissue, woven in between the stratified layers of each folio would act as an effective repair method.

Figure 18 & 19: An inter-woven toned Japanese tissue repair.

 

Paper and Adhesive Selection

Kozo 2 Japanese tissue was selected as the repair material. It was chosen as it was semi-opaque and thus not visually obstructive; weaker than the primary substrate; and had a degree of stiffness that enabled it to be inserted and woven between bark layers. The tissue strips were adhered with a low concentration of methylcellulose (2%) in water. Methylcellulose was chosen as adhesive firstly because of its cellulosic similarity to birch bark, secondly its refractive index was similar to birch bark, and lastly it had a greater water retention capacity and flexibility than wheat starch paste.

Figure 20 & 21: Tears and areas of delamination were repaired and reinforced with toned Kozo 2 tissue adhered with 2% methylcellulose.

 

Tissue Preparation

The Kozo 2 tissue was toned using an airbrush, with dilute burnt umber and raw sienna Golden heavy body acrylic paints. The tissue was toned in a colour-range of tones so that the repairs match the multi-tonal folios. The tissue was then cut into thin strips using a scalpel. The widths of the strips varied from about 3-6 mm.

Figure 22: Tonal variation of different birch bark folios.

Repair

A strip was selected, woven through the delaminated layers using tweezers, adhesive applied with a paintbrush (size 001), and the joint gently pressed into place using a cotton swab. The swab also removed any excess adhesive. The repair was then pressed gently under weights. Pressure-light weights were used due to the brittle nature of the substrate. Large areas of delaminated layers were re-adhered to their folio similarly, using a thin application of methylcellulose and gentle pressure.

Figure 23-25 (Left): Applying toned Kozo 2 to delaminated corner. Centre: applying methylcellulose with a fine brush. Right: Applying gentle pressure to the join using a cotton swab.

 

The folios were conserved systematically, one by one, starting from the back of the volume (as these were in better condition than those at the front). Each folio took around 15-20 minutes to repair depending on the extent of its damage.

The crude previous paper repairs and the annotated infills were lifted from the birch bark using tweezers, and the dry powdery adhesive carefully scraped off the surface of the folio using a scalpel. The repair’s locations were documented prior to removal.

The four loose pages at the back of the manuscript were stripped up with Japanese tissue and adhered into the two detached sections according to their foliation. The texacryl-coated pages were trimmed to remove the Japanese paper border, and likewise stripped up and attached into these sections.

The final repairs were made to the outer spine folds of the sections. These were carefully repaired in-situ using slightly wider strips of toned tissue.

Figure 26: The trimmed texacryl 13-002 and nylon coated folios, stripped up and sewn into their section.

Re-sewing Loose Sections

The two detached sections at the back of the volume were attached at the head-edge sewing station via Coptic chain stitch. An extra Coptic stitch was attached to the 3rd section to reinforce the attachment. The decision to not attach the section at two sewing stations was due firstly to the fragmented state of the sewing on the tail-edge station, secondly because the minimal sewing sufficed, and lastly because the purpose of the attachment was solely to prevent loss and disassociation.

Figure 27 & 28: Showing the repaired spine edge of the text block and the three chain stitches at the head of the volume.

Cover Decisions

SC6000 leather treatment was rubbed lightly into the leather cover to reduce the white blooming.

The sewing, despite being broken, was remarkably stable post treatment, . This is perhaps because the chords were consolidated in place by the Japanese tissue repairs. The relative stability of the binding enabled the decision to not interfere with the original sewing, and to leave the binding, as well as the cover, as it was. The object post treatment was stable enough to digitise and even stable enough to be handled and viewed by researchers.

It was noted also that by leaving the leather cover attached only at the head, it facilitated the viewing of the sewing structure and spine. As the rare binding style is as much of cultural value as the textual content of the object, the fact that it was left exposed was regarded positively.

Figure 29 & 30: Showing volume post treatment with limp leather cover left in its original state.

Re-housing Fragments

Figure 31 & 32: Showing adhesive labels and previous condition reports re-housed in Melinex sheaths.

Figure 33 & 34: Showing annotated infills and birch bark fragments spot-welded into Melinex sheaths.

After treatment there was varied ephemera to re-house and to keep with the object:

1. The old adhesive labels on the old box
2. The previous condition reports
3. The annotated infills
4. Three unplaceable birch bark fragments.

The old adhesive labels and the condition reports were placed in independent inert polyester Melinex © sheaths. The top edges of the sheaths were left open so the items could be removed, unfolded and read in the future.

The annotated infills and the birch bark fragments were secured in place in their independent Melinex sheaths using an ultrasonic spot-welder. Both sets of sheaths were hole-punched and secured together in the top left hand corner using an archival snap-ring.

Re-housing the Manuscript

To impede the likelihood of potentially harmful physical or environmental damage, it was decided that a custom made drop-back box would be made. It was decided that a shelved compartment would be included in the design in which to store the sheathed ephemera. A four flap wrapper was also made for the volume from Kraft paper, to further impede the potential of damage and to facilitate future handling.

Finally the shelfmark of the item was gold tooled onto the spine of the box.

Figure 35 & 36: Showing Kraft paper 4 flap folder and buckram-covered drop-back box with shelf compartment.

Figure 37 & 38: Gold tooling the manuscript shelfmark onto the box.

 

Before and After Treatment Photographs:

By Daisy Todd

Image References

Wojtech, M 2013, The Language of Bark, American Forests, article: http://www.americanforests.org/magazine/article/the-language-of-bark

Wojtech, M 2011, Getting to know bark, Northern Woodlands, article: http://northernwoodlands.org/articles/article/getting-to-know-bark

Further Reading

Agrawal, OP & Bhatia, SK 1981, Investigations for preservation of birch bark manuscripts, ICOM committee for conservation, 6th triennial meeting, Ottawa, September, pp. 21 – 25.

Batton, S 2000, Seperation Anxiety: The Conservation of a 5th Century Buddhist Gandharan Manuscript, WAAC Newsletter, Vol. 2, No.2

Florian, ML, Kronkright, DP, Norton, RE 1991, The Conservation of Artifacts Made from Plant Materials, Getty Trust Publications, Getty Conservation Institute.

Gilberg, MR 1986, Plasticization and forming of misshapen birch-bark artifacts using solvent vapours, Studies in conservation, Vol. 31, No. 4, pp. 177-184.

Gilroy, N 2008, The Stein birch-bark collection in Oxford: Thirty years of developing treatment options for our most fragile manuscripts, ICOM Committee for Conservation:15th triennial meeting, New Delhi, 22-26 September, Delhi: Allied Publishers, Vol. 1, pp. 264-269.

Suryawanshi, DG, 2000, An ancient writing material: Birch-bark and its need of conservation, Restaurator: International journal for the preservation of library and archival material, Vol. 21, No. 1, pp. 1-8.