JAIC 1994, Volume 33, Number 1, Article 3 (pp. 33 to 45)
JAIC online
Journal of the American Institute for Conservation
JAIC 1994, Volume 33, Number 1, Article 3 (pp. 33 to 45)





The conservation case study is of a Taufschein in the fraktur collection of the Prints and Photographs Division of the Library of Congress (see fig. 1). The artist is unknown. The certificate is in the name of Maria Steiner, born 1800, the daughter of Michael and Lictin Steiner. Emanuel Schultz is listed on the document as the minister. The certificate was printed and written in German and executed in Dauphin County, Pennsylvania. The year 1823 is inscribed on the certificate, and it has been dated by the Library of Congress as executed between 1800 and 1823.

The artist used old-style rectangular borders for the format. The imagery is spiritual and joyous. A smiling face with rays of sunshine is centered directly above a circular wreath inside which the birth and baptismal information is recorded. On either side of the wreath is an angel dressed in deep red and blowing a horn. A background of dark blue grapes graces the scene. Below the scene, two heart shapes are inscribed with hymns and psalms. Each heart shape is crowned with a rose and a bird.

The certificate is a woodcut, with letterpress type, printed in black ink and embellished with watercolor. Iron gall ink script is also present. The deep reds and blues of the illustration appear somewhat dull and matte and were applied heavily. What appears to be gamboge (yellow) is prominent in the rays flowing from the face representing the sun and the large area around it. Gamboge was commonly used during this period on both sides of the Atlantic Ocean. What seems to be green verdigris, from its visual appearance, colors the inner border and outlines the shape of the hearts. The substrate is thin, handmade, laid paper, approximately 33 × 40 cm. A watermark is not present on the sheet.

The document had been mounted on two layers of a very modern, white, machine-made wove paper. The mount appeared to have been much too heavy for the object. The natural characteristics of the object's support had been lost because of the mount. The adhesive is a very tenacious water-soluble converted-starch paste. Several factors led the author to conclude that the adhesive used to adhere the object to the multilayered mount was most likely Yes paste. First, it tested positive for dextrine using the iodine staining test, which yielded a reddish color. Further, the makers of Yes paste, Brothers and Lane, confirmed that the company sold its product as a library paste to the U.S. Government Printing Office (GPO) during the 1960s, a period in which GPO restorators worked at the Library of Congress. The object's bond paper mount was also typical of GPO restoration. Additionally, former and senior conservation office staff members have also confirmed the use of Yes paste just prior to their arrival, when the practice was ended. It is not uncommon to encounter objects with this type of former restoration in the collections at the Library of Congress.

Brothers and Lane (1980) give this analysis of Yes: white appearance; pH 6; content a mixture of corn syrups (55%), water (32%), corn dextrins (12%), phosphoric acid (.02%), and humectants. Further investigation of the ingredients listed indicates that the corn syrup acts as a plasticizer to control brittleness and/or drying, the humectants are used to control moisture, and the phosphoric acid improves wetting and adhesion to the surface of the paper (Jarowenko 1977).


The paper was very weak and brittle. Stains and discoloration were present, as well as a number of lateral creases. Some losses had occurred along the edges of the document. Additional losses were present toward the interior. The verdigris, basic copper acetate, had caused losses around the outlines of the hearts and the inner border. Copper-based pigments have been identified as likely links to the disintegration of paper (Banik and Ponahlo 1982–83). Numerous major and minor tears ran across the substrate. The former restoration of the object left a number of tears overlapping. Some of the watercolor had suffered slight abrasion.


The conservation of the object posed several treatment problems. The iron gall ink, the printing ink, and the verdigris appeared to react only slightly adversely to water (distilled) after repeated exposure during spot testing of the pigment and ink. However, the red and blue watercolors were extremely soluble in water. Normally, a mixture of water and ethanol (absolute) is a good solution for the washing in such circumstances. However, gamboge pigment was present. This organic resin is very soluble in alcohol, as it proved to be in testing (Gettens and Stout 1966). The paper substrate was extremely weak and brittle and adhered with an extremely strong adhesive to its mount. The presence of numerous losses and tears overall complicated the backing removal of the weak paper support from the strong mount.

After thorough examination, a treatment plan was devised that took into consideration the problematic characteristics of the pigments, support, and mount. The plan involved light dry cleaning; facing the recto of the object to stabilize the weak substrate during removal of the backing; removal of the backing via Gore-Tex humidification or damp packs; reduction of adhesive residue; washing on the suction table with combinations of ethanol and water; and a two-step lining process. The primary goal of the plan was to minimize the amount of water used in the treatment to reduce the risk of altering the water-soluble pigments and damaging the fragile paper support.

There were several ways in which the presence of water was minimized during the treatment:

  1. The Gore-Tex damp pack allows the adhesive of the mount to swell in a controlled manner. Moisture enters the interstices of the paper without fully wetting the object, therefore posing less risk to the water-sensitive colors and the fragile paper substrate.
  2. Residual adhesive must be removed following the removal of the mount. The local use of ammoniated water (pH 9) via swabbing and 4% Dow A4M premium methyl cellulose poultices during the adhesive residue removal treatment reduces the threat to the watercolor.1 This precaution controls the amount of moisture present at the verso of the object, preventing moisture penetration through to the recto.
  3. Ethanol and water solutions used on the suction table during the washing phase maximize washing without endangering sensitive watercolor.2 Various solutions of ethanol and water were used, depending upon the solubility of a particular area. Also, testing showed that the watercolor exposed to a 70:30 solution of ethanol and water swelled only slightly and did not move laterally across or absorb into the paper. The gamboge was found to be stable in a 35:65 ethanol and water solution. Aqueous-sensitive watercolor swelled more in this solution but did not run or absorb into the substrate and was not otherwise endangered. The 35:65 ethanol and water solution safely treats adjacent alcohol-sensitive and water-sensitive areas. During the long washing procedure on the suction table, a 50:50 ethanol and water solution would be used sometimes to lightly rehumidify the object to compensate for evaporation in areas near ethanol or water-sensitive pigments. The suction table confines the various ethanol and water solutions to desired areas. Additional control of the washing solutions is attained by spraying and blocking areas of the object with blotters held above.
  4. Brittle paper has very little water content, and water rapidly evaporates from it once it is exposed to normal room relative humidity and temperature. In this respect, ethanol and water solutions have additional advantages. Application of these solutions controls the amount of paper expansion and seems to provide moisture for a longer period of time than simple humidification.
  5. Wheat starch paste saturated with ethanol minimizes the amount of water in the lining process, which may affect the aqueous-sensitive watercolor.3 It also controls unwanted further expansion of the object once the substrate is in contact with the lining adhesive.



After the media-free borders were lightly dry cleaned with ground Staedtler Mars plastic erasers, the object was prepared for the backing removal by facing large areas of the recto with heat-set tissue. Library of Congress heat-set tissue adhesive (acrylic emulsions) specifications are as follows: 1 part Rhoplex AC-73; 1 part Rhoplex AC-234; and 1 part water. The adhesive is applied to lens tissue by brushing. The heat-set tissue was activated by applying toluene with a brush. A spatula gently tamped the heat-set tissue to the surface as the toluene was applied to assure adhesion. Four percent Acryloid B-72 (ethyl methacrylate copolymer) in toluene was used as a facing adhesive in those areas that needed additional strength and reinforcement. The B-72 was applied over the heat-set tissue with a brush and allowed to saturate through the heat-set tissue that was already attached to the surface. The use of toluene was very important since some of the weakened areas were underneath ethanol-sensitive gamboge. Gamboge was found to be stable in the presence of toluene. In addition, B-72 has been found to be extremely stable and reversible in toluene (Feller 1975). Since one can never be sure of the complete 100% removal of adhesive from paper, B-72 is a good choice as any extremely small trace residue of B-72 would most likely not harm the paper.


The object was then placed in a Gore-Tex damp pack. The arrangement of the pack placed the object between two sheets of Gore-Tex barrier laminate with the membrane side against the object. The barrier laminate consists of a Gore-Tex membrane (polytetrafluoro-ethylene) laminated to a 1/16 in polyester felt. The Gore-Tex barrier laminate and object were then placed between a set of lightly dampened blotters. Sheets of Mylar (polyester film) covered the damp blotters, allowing the moisture to be forced through the Gore-Tex barrier laminate and into the object. The slight weight of a ¼ in felt was added.

The mount was peeled back very slowly after sufficient moisture had swelled the adhesive, but at the same time leaving the sensitive pigments unaffected. This degree of humidification took about 30 minutes. During the backing removal procedure only very small portions of the mount were removed at a time because of the fragility of the object's weak paper substrate. Small reinforcements of tengujo Japanese paper were adhered with wheat starch paste to the verso of the object as the backing was peeled back. The reinforcements secured and supported prior tears and weakened areas, thus helping to protect the substrate from undue stress as the backing procedure progressed. Often the object and mount became too dry after a short period of time and had to be rehumidified. Moisture had to be reintroduced to swell the adhesive via the humidification damp pack after a small area of the mount had been removed. The procedure had to be repeated numerous times. The complete backing removal treatment was accomplished over several days of tedious work. The mount was successfully removed intact from the object and without damage to the support, inks, or watercolor.

Examination of the object after removal of the backing revealed several former reinforcements as well as a thick adhesive residue layer. The adhesive layer caused the paper to be tight and cockled. The extent of the damage caused by the iron gall ink and the copper-based green pigment was evident by the losses in areas of the paper where the corrosive media were present. In addition, the verdigris was somewhat visible on the verso of the object. From the author's past experiences with the behavior of verdigris, two conclusions were drawn about its appearance. First, the appearance was most likely due to the prior restoration attempts. Such attempts may have exposed the support and pigment to an overabundance of moisture during treatment. The verdigris was then absorbed further into the paper. But it may also have been caused by the natural migration process of the corrosive pigment into paper over time. The slight amount of moisture used in the damp pack in the presence of a thick adhesive layer would not have caused the absorption of the verdigris to the verso in the manner that was evidenced.


Ammonia water (ammonia hydroxide) swells starch and starch-based adhesives, making them easier to reduce and remove. Ammonia water at pH 9 and cotton swabs and balls were used locally to remove the thick layer of adhesive residue. Dull knives and ammoniated methyl cellulose poultices applied directly on the adhesive also helped remove the residue. The dull knives removed the thick mass of the adhesive residue after it had been swelled by the poultices. Cotton swabs and balls slightly dampened with ammoniated water removed the remaining thin film of residue. Former reinforcements as well as some of my own reinforcements were removed from the verso of the object. The absorbance of moisture to the recto did not occur during the treatment. Overall, the object appeared much less cockled and tight as result of the treatment.


The heat-set tissue facings and B-72 adhesive were easily removed on the suction table by spraying with toluene and lifting the carrier with a pair of tweezers. A Crown Spra-tool spray gun operated by a pressurized can was used to apply the toluene. Examination after treatment revealed no damage or discoloration to the media or paper. Despite the fact that B-72 is chemically very stable, it is important to note that media and papers may darken when the B-72 is first applied, and the darkening will remain unless the adhesive residue is thoroughly removed.


Washing the object on a suction table proved a good method for cleaning the object without subjecting the pigments to undue risk. A Maxwell suction table was prepared by masking off the top with polyester film. An opening was left at the center for the blotter support. A thin blotter was humidified with a 50:50 ethanol and water solution and placed in the opening.4

To expand and relax the object before it was placed on the suction table, it was sprayed with various solutions of 35:65, 70:30, and 50:50 ethanol and water solutions. In this way the sensitive pigments would not be affected during the procedure. While supported face up on a sheet of Hollytex (spun-bonded polyester web), the object was placed on top of a blotter resting on the unrestricted portion of the suction table. The suction table motor (explosion-proof) was first run at very slight vacuum pressure. Then the vacuum pressure was increased very slowly as the object was delicately manipulated to avoid creases as it flattened out.

Washing began with the suction table running at full vacuum pressure. Large bottles of ethanol and water solutions were prepared in advance so that they could be readily used to refill exhausted container bottles of the spray gun. The washing procedure used the same ethanol and water solutions used to humidify the object. Controlled spraying and the blocking of sensitive areas of the object with a blotter held 4–5 in above the object's surface helped to keep the pigments stable during the treatment. The blotter paper underneath the Hollytex was changed and examined several times during the treatment to evaluate the results of the washing. Washing was completed when the blotters indicated that no more discoloration could be removed from the object. The object was then air-dried and inspected. The watercolors and inks appeared unaffected, with no apparent lateral movement or absorption into the paper other than a very slight movement of the verdigris onto the blotter. Overall, staining and discoloration were diminished, and the object appeared brighter and fresher looking when it was visually examined.


A two-step lining process was used to stabilize the fragile paper support. The first step involved performing a preliminary lining on the suction table. Using the suction table allowed good contact between the lining papers and the object and eliminated the need for manual pressure on the surface of the object to create a uniform bond between the primary support and the lining paper. In this way, contact with the delicate watercolor was avoided.

The lining began by setting up the suction table with areas of the surface masked off with polyester film in the same manner that was used to wash the object. A sheet of Hollytex was placed on the unrestricted area of the suction table. A thin, lightweight sheet of usui-usu-minogami Japanese paper was pasted with a Japanese paste brush on top of the Hollytex using a thin solution of ethanol-saturated wheat starch paste. Light vacuum pressure from the suction table ensured constant overall contact between the Hollytex and minogami surfaces.

Ethanol in the paste reduced the amount of water in the system that could affect the water-soluble pigments. Not enough ethanol was present, however, to harm the ethanol-sensitive gamboge. The procedure for making ethanol-saturated paste begins by mixing cooked and strained wheat starch paste with water using a Japanese paste brush until the desired consistency is found. A small amount of ethanol (5–10 ml) is added to about 1 liter of thinly blended paste. Some of the paste will congeal on contact with the ethanol. The ethanol is reducing the amount of water in the paste mixture through dissociation of the water molecule, thereby slightly solidifying the starch particles. The paste is then blended until it is smooth. A small amount of ethanol is again added and blended. The steps are repeated until the ethanol no longer blends into the paste. The paste is now saturated. A small amount of water is then added to finish blending the paste back to a smooth consistency.

Using the ethanol-saturated paste, a sheet of tengujo Japanese paper, which is very lightweight and thin, was then pasted on top of the minogami paper. The grain pattern of the minogami paper ensured that the lining and the object would contract in a similar dimensional pattern during the drying process. The tengujo paper provides a cushion that allowed the tears to remain seated properly during the contraction and drying phase. Both Japanese papers are thin and supple so that they did not overwhelm the natural characteristics of the object's support.

The object was placed on the lining after having been humidified and expanded with the same 35:65 and 70:30 ethanol and water solutions that were used to wash the object on the suction table. The ethanol and water solutions were sprayed lightly on the object. The object was gradually flattened out as the vacuum pressure was slowly raised. The lined object remained on the suction table under full suction for about 20 minutes, giving the lining the opportunity to achieve full contact with the object and a small amount of drying to occur.

The second stage of the lining procedure involved transferring the object to sanded 5 mil polyester and a sheet of glass for the stretch-drying phase. Sanded polyester film (Mylar) was used to avoid ferrotyping of the lining paper and to provide a bit of grab. A medium-grade sandpaper can be used to sand the polyester film, which should be sanded well enough to produce a completely opaque sheet. After sanding, the Mylar should be thoroughly washed to rinse away the fine polyester particles.

The technique is a modification of a lining method that is used at the conservation laboratory of the Bibliothèque Nationale in Paris. At the time the technique was introduced to the author in 1984, the staff usually employed Tylose (sodium hydroxymethyl cellulose) at approximately 2% as an adhesive. In this technique sanded polyester film is also placed on wet glass. The Japanese lining paper is then pasted on the sanded polyester (sanded side up). Then the object is placed on the lining paper and gently manipulated to achieve good overall contact. Linen cloth strips adhere the edges of the lining and the sanded polyester film. The object is then allowed to air dry.

The technique has a number of advantages. Contact with the object's surface can be avoided, which is a crucial point when lining objects with delicate media. Tears do not open or tent during the drying phase. Papers of the early and mid-1800s such as Whatman, which have severely disintegrated, are especially difficult to control and line when they have been poorly restored in the past. Tears that were overlapping and set during the lining procedure have a strong “memory” of their previously overlapped positions. A conventional lining, other stretch lining techniques, or even a suction table lining alone cannot always prevent the overlapping of tears from recurring. Stretch lining on sanded Mylar with multiple layers of Japanese paper prevents the phenomenon from occurring.

In general, it is a good idea to use multiple layers of Japanese papers when lining an object that has massive tears or losses no matter which lining system is used. Multiple layers of Japanese papers seem to give much better overall support to a heavily damaged object than a single layer. Two or three thin sheets of Japanese paper sometimes give better overall support than a single heavy sheet. As there are many Japanese papers with various characterist-ics and qualities, it is important to choose papers that are sympathetic to the characteristics of the object. In general, the lining promotes initial rapid drying that prevents tears from gradually moving and opening. It effectively complements the setting characteristics of the thin but supple tengujo Japanese paper.

The next step involved preparing the stretch-lining system. Sanded polyester film was placed on wet glass with the sanded side up. Water was sprayed on the glass to adhere the polyester film. As many of the air bubbles as possible were removed between the polyester film and glass to ensure overall contact. The lined object was removed from the suction table while it was still damp and placed on the sanded polyester film. Both the lining paper and the sanded polyester film were the same size. No pressure of any kind was used when the object was placed on the polyester. The lining paper was tacky enough with adhesive to allow for sufficient contact between the two surfaces without using mechanical pressure. The object was sprayed lightly with an ethanol and water solution to keep it from drying too rapidly at that point. Linen cloth restraining strips were liberally pasted with wheat starch paste and placed along the edge of the lining paper with wheat starch paste to restrain the object on the glass. The width of the cloth-restraining strips depend on the size and expansion-contraction characteristics of the object. A 30 × 20 in object with great expansion-contraction characteristics, for example, would most likely need a 3 in wide strip of cloth to prevent the lining from pulling itself off the glass or table during the drying period. Half of the restraining strip covers the lining paper.

The object was then air dried over several days. The drying time will depend on the relative humidity of the environment. It is important not to remove the lining from the restraints too early. As with any lining technique, the object will curl if it has not fully dried when it is removed from its restraint.

While the object was still attached to the glass, the losses were filled with gouache-toned kizukishi Japanese paper and retouched with fixed pastel. Wheat starch paste adhered the fills. Blotters and weights were not needed for fills to dry flat because the overall tension of the lining automatically flattened out the insert paper or pulp fill. The glass also made it convenient to work directly on a light table while filling the losses.

The object was removed from the glass by first peeling away the linen strips. A spatula was inserted under the edge of the lining paper and polyester film support, and the linen strips were slit away. The attached polyester film support was easily removed by turning the object face down on a clean blotter and peeling the polyester film away. The object's paper support appeared very relaxed and flat. To complete the treatment, the excess lining paper was sanded away using a sanding stick to create a softer appearance to the edge. The completed treatment revealed no change or disturbance in the appearance or characteristics of the media (fig. 6). The aesthetic integrity of the object was left intact while a stable support was provided.

Copyright © 1994 American Institute for Conservation of Historic and Artistic Works