JAIC 1998, Volume 37, Number 1, Article 8 (pp. 111 to 116)
JAIC online
Journal of the American Institute for Conservation
JAIC 1998, Volume 37, Number 1, Article 8 (pp. 111 to 116)



ABSTRACT—The most up-to-date publication on the full range of aesthetic compensation materials for wood remains R. Barclay and D. W. Grattan's paper in the ICOM Committee for Conservation preprints, “A silicone rubber/microballoon mixture for gap-filling in wooden objects” (1988). Because of the wide range of materials covered in that paper, the issue of structural compensation was not treated in great depth. This article will focus on philosophy, materials, and techniques used today by wood conservators when structural (weight-bearing and shear-resisting) compensation is required. Aesthetic fill material is not necessarily dependent upon specific compatibility with the substrate material, so the range of choice of materials can be quite broad. However, since material and strength compatibilities are critical to a structural fill, those applied to wood objects fall within a fairly small range. Literature on the subject is limited, and practices among wood conservators are relatively uniform: wood fills, thermosetting synthetic fills, and combinations of the two.

TITRE—Mat�riaux pour la reparation structurelle d'objets en bois de grandes dimensions approches diverses et complementaires. R�SUM�—La publication la plus r�cente pr�sentant une gamme compl�te de mat�riaux pr�sent�s dans cette �tude, la question des r�parations structurelles n'y �tait pas d�velopp�e de fa�on approfondie. Le pr�sent article passe en revue la philosophie, les mat�riaux et les techniques adopt�s aujourd'hui par les restaurateurs lorsqu'ils doivent assurer des travaux de nature structurelle, c'est-�-dire obtenir un r�sultat qui offre une r�sistance aux charges et aux cisaillements. Les mat�riaux utilis�s pour la compensation de pertes uniquement dans un but esth�tique, ne doivent pas obligatoirement pr�senter de compatibilit� avec le mat�riau d'origine, ce qui offre un assex vaste choix dans ce domaine. Il n'en va pas de m�me des mat�riaux destin�s � une r�paration structurelle des objets en bois, car ils doivent correspondre � des sp�cificit�s tr�s pr�cises. Il n'existe qu'une litt�rature limit�e � ce sujet. Les restaurateurs utilisent des proc�d�s assez uniformes: mat�riaux soit � base de bois, soit synth�tiques thermodurcissables, ainsi que des associations des deux.

TITULO—Rellenos estructurales para objetos grandes de madera, enfoques constrantes y complementarios. RESUMEN—La publicaci�n m�s actualizada sobre el rango completo de materiales para compensaci�n est�tica de madera sigue siendo la de Grattan y Barclay (1988). Debido a la amplia gama de materiales que cubre ese art�culo, el tema de la compensaci�n estructural no fue tratado a fondo. Este art�culo se concentrar� en la filosof�a, materiales, y t�cnicas usados en la actualidad por los conservadores de madera cuando se requiere de una compensaci�n estructural, es decir, una compensaci�n que resista peso y desgarramiento. Aunque un material para relleno est�tico no tiene que ser necesariamente dependiente de su compatibilidad espec�f�ca con el material del sustrato, el rango de materiales a escoger puede ser bastante amplio. Sin embargo, como la compatibilidad y la resistencia de materiales si son cr�ticas en un relleno estructural, aquellos que se aplican a objetos de madera caen dentro de un par�metro bastante reducido. La literatura sobre el tema es limitada, y las pr�cticas entre los conservadores de madera son relativamente uniformes: rellenos de madera, rellenos sint�ticos de curado t�rmico, y una combinaci�n de los dos.


When the perceived need for a structural fill in a wood object arises, a number of questions must be answered before choosing the appropriate material or the particular application technique:

  • Is the fill structural, or is merely aesthetic compensation necessary? If the latter, the fill may be selected from a wide range of materials commonly used by objects conservators but not discussed in this article.
  • Is the area that needs fill an actual loss of material, or is it a separation or split that will not close? In the latter case, often the result of internal vice, structural integrity may not necessarily be compromised, and a nonstructural, aesthetic fill may be appropriate.
  • Is the surrounding material sound or deteriorated? If the latter, a penetrating consolidant that will be structurally compatible with a structural fill is indicated.
  • Is the grain of the wood apparent and important to the visual presentation of the object, or is it obscured by paint, dirt, or degraded varnish? Often wood fills are necessary for visual harmony, whereas if a synthetic fill is indicated, it is more easily hidden by an opaque surface coating.
  • Is this a surface fill or one that penetrates deep into the object? Surface adhesion is much easier to deal with from the standpoint of reversibility than a deep, mechanically locked fill.


As a nonisotropic or anisotropic material, wood exhibits different properties in different orientations. Practically, wood has greater strength in the direction of the grain than across the grain, and it has dimensional response to environmental moisture variations across the grain, not in the direction of the grain. These properties present both opportunities and dangers for the craftsman and the restorer.


Damage from internal vice is by definition the result of the inherent properties of wood, often in combination with the craftsperson's fabrication choices. Nonuniform movement in response to environmental moisture variation can result in splits or breaks along the grain due to internal restraint, as in the case of solid wood, or externally applied restraint, such as cross-grain construction, in a joined wood object. In such cases, the “loss” is due to anatomical deformation under internal stress and is referred to as “compression set shrinkage” (CSS) (Hoadley 1980). Since CSS damage results in merely the separation of grain (a split) rather than breakage across the grain, there is usually minimal loss of strength in the total object.

In these cases, a strong fill is not only unnecessary, it can, under certain circumstances, propagate the split. The traditional method of inserting wood splines or hard fills into splits caused by CSS, no matter how well executed, in a sense “reloads” the system and can result in further CSS and loosening of the fill unless a stable environment can be guaranteed.

Current thinking not only perceives the danger from this type of fill but usually accepts damage from internal vice philosophically, as integral to the character of the object. However, if aesthetic concerns mandate treatment, a soft, nonstructural fill is indicated (Barclay and Grattan 1987; Barclay and Mathias 1989).

As an anisotropic material, wood can also present special compatibility problems with fills. While the traditional wood fill, properly oriented, can move sympathetically with the surrounding material, synthetic fill materials are isotropic. They respond, if at all, in a uniform manner in all directions, an action that in turn can threaten the stability of the repair.


Engineering requirements for a structural fill and the strength provided by that fill are usually subjective. Restoring adequate load-bearing strength to a large wood or joined wood object does not necessarily mean the complete restoration of original structural qualities, however. Wood as a material often exceeds the strength requirements of objects made from it (Hoadley 1985), so the conservator may choose to vary the potential strength of a fill. For example, if a damaged chair leg started life at 300% of structural engineering requirements, then a fill that restored only 50% of original strength to the damaged element nonetheless returns the object to 150% of engineering requirements. This may give the conservator an opportunity to vary the amount of adhesive contact or the strength of the adhesive itself in favor of greater overall reversibility of treatment.


Since “structural fill” as understood here implies a tenacious adhesive bond somewhere on the object, the issue of reversibility is of more than usual concern.

The AIC Guidelines for Practice recognize that complete reversibility is often an unrealizable ideal. A more realistic concept is “retreatability.” This is a useful double entendre that implies both some degree of reversibility and unhindered future treatment when and if it becomes necessary.

Although strength requirements and “retreatability” are two important characteristics of any fill, for practical purposes strength requirements must supersede retreatability if the two cannot be reconciled. Conscientious attention to actual strength requirements, without overengineering, will ensure maximum retreatability.

In some cases, maximum strength may be a necessity, perhaps sacrificing retreatability in the name of stability. For example, treating the seriously damaged leg of a table with a fill or adhesive that is not realistically reversible may be justified if it ensures against a possible failure and consequent extended damage to the entire piece or to objects that may be displayed upon it.

Other circumstances also highlight the need for flexibility in applying the concept of retreatability. For example, a solid piece of wood that has split or broken has never before been apart, nor was it ever intended to be by the craftsman. Therefore, it could arguably be rejoined with less reversible materials. However, a loss or break in the vicinity of a joint impinges on an area that bears information about fabrication techniques such as tool marks or layout lines. Because of cultural or craft information, a “worked” area or surface is more significant than an “unworked” area and thus requires heightened regard for retreatability. This statement in no way endorses gratuitously irreversible methods, since any object may need to be retreated for a variety of reasons, and previous treatments should have minimal impact on any new work.


Generally, fills are two-part systems: an adhesive or mechanical interface with the object and a sympathetic mass of cohesive material that replaces the loss. These characteristics can literally be separate, as in the case of a well-fitting wood fill affixed with an appropriate adhesive. Or the fill and its adhesive can be combined into a single component, such as bulked epoxy paste. The appropriate fill balances characteristics of both parts of this system to attain the best combination of strength and retreatability.


Although wood fills are structurally and aesthetically desirable, they do pose some obstacles. Matching grain pattern and color demands a skilled and experienced eye, and fitting a solid fill to an irregular or deep area of loss requires a high but by no means unattainable level of skill. In the past, their high skill level notwithstanding, traditional restorers did not apply the concept of integrity to the original material of an object. In the interest of saving time and effort, a wood fill would often be treated much like inlay. It would be cut to a convenient shape, i.e., slightly larger than the loss, straight-edged, sharp-cornered and slightly taper-sided for a tight fit. It would then be laid over the area of loss and traced out with a sharp knife or pointed scribe; the area so outlined would be relieved with a knife and chisel. Glued into place and leveled with the surrounding area, this method produces a tidy appearance but results in the loss of original material.

With heightened regard for original material, the contemporary conservator carves the fill to fit the loss instead. This approach does indeed require skill, and many conservators are able to execute the task by eye. However, it can be facilitated with carbon paper. The paper is laid in the interface, ink toward fill, so that when the fill-in-progress is temporarily pressed into place, high spots are registered on the surface. These spots in turn are progressively carved away until a good fit is achieved and carbon registers overall.

It is important not only to choose but also to align the grain of wood fills to closely approximate the surrounding wood. When the fill must match visually as well, grain orientation must be exact. Grain orientation includes the degree of tangential vs. radial exposure as well as the angle and direction of the pores relative to the surface. Otherwise, light reflected from the surface will belie even the best-fitting fill. One of the benefits of a wood fill fitted to a odd-shaped loss is that the irregular interface at the surface is less eye-catching than the straight lines of the traditional fill.

It is accepted, particularly in architectural conservation, that for ethical identification of compensation, wood of a different species from the original should be used. However, conservators of decorative objects produced with a heavy reliance on the figure and color of the wood accept that wood fills are virtually never completely undetectable, so they generally use the same species whatever the circumstances.

Although thermoplastic wood “glues” such as PVA and aliphatic resin emulsions produce excellent bonds, reversibility is problematic for wood fills. For strength, versatility, and retreatability, traditional hot animal-hide glue, produced in a wide range of gelling characteristics and tensile strengths, remains the best adhesive for today's conservator. (Fish glue, which has the same general characteristics as hot animal glue, is traditionally preferred when working in the mixed medium of Boulle work, i.e., tortoise, brass, and pewter.) So-called liquid hide glue is adulterated with a gel suppressant such as urea. The experience of many conservators is that while such glue is sometimes desirable for nonstructural repairs, it may fail at elevated temperature or humidity. Liquid hide glue also has a shelf life, and manufacturers do not always provide an expiration date on the package.


If a wood fill is not practical or desirable, a gap-filling adhesive paste or bulked liquid can be used. As a general rule, thermoplastic fills such as bulked acrylic, PVA or PVOH resins, or even the traditional sawdust and glue fill, are less structural than thermosetting fills. These thermoplastic polymers have adequate theoretical cohesive strength, but they are generally unsuitable for gap-filling due to shrinkage during solvent release curing. Bulking with glass microspheres, glass microballoons, phenolic microballoons, or microfibers can increase the viscosity of these adhesives, but enough bulking to create a thick paste can result in reduced cohesive strength (Grattan and Barclay 1988).

Thermosetting resins are irreversible with standard solvents, and although gap-filling properties of catalyzed epoxy and polyester resins are excellent, the bond with a wood substrate that is mostly mechanical is also virtually impossible to reverse without damage to the object.

It is most practical to utilize the desirable characteristics of each of the resin types by using a thermosetting material for the gap-filling, structural fill but introducing a thermoplastic adhesive “bridge” at the interface for retreatability. There are a number of techniques for this system. Some conservators cast the fill in place with a nonadhesive barrier at the interface. The fill is then removed, trimmed, perhaps even colored, and then re-adhered. Complex or deeply undercutting fills may have to be built up in separate pieces so that they can be individually removed, worked on, and readhered with a skim coat of filler to blend the surface. This method is desirable if the mechanical undercutting of the loss could make later removal difficult.

Others prefer to accomplish the fill in a single step. First, the surface of the loss is coated with the appropriate thermoplastic resin or adhesive. The thermosetting gap-filler is then poured, spread, or otherwise molded in place, adhering to the bridge. These fills must be trimmed and colored in place (Anderson and Podmaniczky 1990). In both cases, practical reversal usually requires destructive removal of the mass of the fill in order to deliver the solvent to the barrier.

For maximum structural qualities, epoxies remain the favorite over polyesters for the majority of conservators polled for this article. Ciba-Geigy epoxies such as Araldite 1253 (paste) and Renweld 306 (thick liquid) are reported as widely used. Although technical representatives of epoxy manufacturers are hesitant to comment on the strength of bonds that are, in a very real sense, disrupted from the start, initial testing by conservators has provided useful results and the technique has been widely accepted. Some barrier adhesives that have been used include Paraloid B-67, Paraloid B-72, “Butvar” B-98, and hot hide glue.

Hide glue is quite popular as a bridge or barrier. If it is allowed to dry completely to a hard gloss, the surface of the glue must be roughened for good bonding of the epoxy fill. It has been found that the epoxy bond is enhanced if applied before the hide glue is completely dry. This technique avoids the need for roughening the surface. Although there has been little scientific testing of this specific system, conversations with technical representatives suggest that there may be some degree of primary bonding between epoxy catalyzed in direct contact with collagen.

In many cases, even fills that are required to provide structural support do not have to fill losses completely. This is particularly important when considering retreatability factors. For example, an epoxy fill on a barrier resin may not need to adhere to the entire surface of a deep void to provide adequate strength (Podmaniczky 1988). One way to reduce the adhesive contact area is to apply paste wax to inner surfaces so that there is only structural bonding at the perimeter despite complete filling of the loss. (Dental crowns exhibit this principle, adhering only to the outer layer of a tooth stub.) The author has also incorporated a void in the interior of a fill to provide an access reservoir into which solvent could be injected to saturate the barrier if removal was ever necessary. On one occasion this procedure was accomplished by packing cotton balls into the center of the loss. A mass of microcrystalline wax has also been used for the same purpose.

To reduce penetration of the epoxy paste on the surface of a loss that may have many voids or undercuts, a two-dimensional barrier, such as Japanese tissue, fine-wove polyester, or another sheer fabric can be incorporated into the bridge adhesive in order to restrict penetration of the fill material (Anderson and Podmaniczky 1990). In the absence of specific engineering data on the strength of restricted bonding, consulting colleagues or one's own past work experience usually provides guidance in these cases.


Often it is desirable to combine wood with synthetic resin fills. In these cases, the wood fill is fitted well, but only to the perimeter of the loss, saving time by not going through the laborious work of fitting to the entire interface. A bridge adhesive is applied, and the fill is then adhered with epoxy paste.


Unlike the wide and varied selection of materials available for merely aesthetic compensation of losses, materials for strength-providing fills in wood objects fall within a narrow range of options, defined by cohesive strength and adhesive compatibility with each other, and ultimately with the wood itself. The favored system for restoring strength to a loss or break while maximizing retreatability is the use of epoxy or wood-epoxy combination to fill the gap and a thermoplastic resin or glue to act as barrier and adhesive bridge.


Anderson, M. J., and M. S.Podmaniczky. 1990. Preserving the artifact: Minimally intrusive conservation treatment at the Winterthur Museum. Wooden Artifact Group preprints, American Institute for Conservation 18th Annual Meeting, Richmond, Va.Washington, D.C.: AIC. 5–18.

Barclay, R., and D. W.Grattan. 1987. A silicone rubber/microballoon mixture for gap filling in wooden objects. International Council of Museums Committee for Conservation preprints, 8th Triennial Meeting, Sydney. Paris: ICOM. 1:183–87.

Barclay, R., and C.Mathias. 1989. An epoxy/microballoon mixture for gap filling in wooden objects. Journal of the American Institute for Conservation28:31–42.

Grattan, D. W., and R. L.Barclay. 1988. A study of gap-fillers for wooden objects. Studies in Conservation33:71–86.

Hoadley, R. B.1980. Understanding wood. Newton, Conn.: Taunton Press. 114.

Hoadley, R. B.1985. Lecture notes. Conservation Analytical Laboratory, Museum Support Center, Smithsonian Institution, Washington, D.C.

Podmaniczky, M. S.1988. Conservation of a degraded late 18th century Windsor settee. Wooden Artifact Group preprints, American Institute for Conservation 16th Annual Meeting, New Orleans, La.Washington, D.C.: AIC. 111–25.


Barclay, R.1981. Wood consolidation on an eighteenth-century English fire engine. Studies in Conservation26:133–39.

Fuller, R. D.1985. An investigation of the physical and tensile properties of selected elastomeric gap fillers for wood. Master's thesis, Queen's University, Kingston, Ontario.

Grattan, D., and R.Barclay. 1984. Report on the use of silicones and glass spheres as a wood filler. IIC-CG Newsletter (December).

Hatchfield, P.1986. Note on a fill material for water-sensitive objects. Journal of the American Institute for Conservation25:93–96.

Nakhla, S. M.1986. A comparative study of resins for the consolidation of wooden objects. Studies in Conservation31:38–44.

Philips, M., and J. E.Selwin. 1978. Epoxies for wood repairs in historic buildings. Washington, D.C.: U. S. Department of the Interior.

Sakuno, T., and A. P.Schniewind. 1990. Adhesive qualities of consolidants for deteriorated wood. Journal of the American Institute for Conservation29:33–44.

Schniewind, A. P. and D. P.Kronkright. 1984. Strength evaluation of deteriorated wood tested with consolidatns, In Adhesives and consolidants, ed.N.Bromelle et al. London: International Institute for the Conservation of Historic and Artistic Works. 227–316.


MICHAEL S. PODMANICZKY is head of furniture conservation at Winterthur Museum and adjunct associate professor in the Winterthur/University of Delaware Program in Art Conservation. He has an extensive craft background as carver/cabinetmaker, boat builder, modelmaker, pattern maker, and machinist. He received his B.A. in fine arts from Kenyon College, 1970, Certificate in Furniture Conservation from the Smithsonian Institution, and congruent master's degree in decorative arts from Antioch University in 1990. He is a member of the class of 1989 of the Attingham Summer School on the English country house. He has responsibility for accessioned woodworking tool collections at Winterthur and maintains active research of related historic trade practices. He is a Professional Associate of AIC, a member of the Objects Specialty Group, and past chairman of the Wooden Artifacts Group. Address: Winterthur Museum, Winterthur, DE 19735.

Section Index

Copyright � 1998 American Institute for Conservation of Historic and Artistic Works