Conservation Report
The Conservation of the Egyptian Polychrome Coffin of Djheutymose
Kelsey Museum Accession Number 89.3.1
Alan J. Hogg, Jr. & Geoffrey I. Brown
Left side of the coffin lid, after conservation
Introduction
In early 1995, coinciding with the exhibition Preserving Eternity: Modern Goals, Ancient Intentions at the Kelsey Museum of Archaeology, conservation treatment was completed on the polychrome wood coffin of Djheutymose. Djheutymose, a priest of Horus and Hathor, is believed to have lived near Edfu in southern Egypt. Based upon stylistic and textual criteria, the polychrome coffin is dated to the Saite Period (685-525 BC). The coffin was donated to the University of Michigan in the 1930s, a gift from Mr. Albert M. Todd, a resident of Kalamazoo, MI. On long-term loan since the 1930s to the Kalamazoo (MI) Public Museum, the coffin was returned to the University in 1990 when it became part of the Kelsey Museum's collections [1].
Description of the condition of the Djheutymose coffin
Photographs of the coffin top half before conservation.
Photographs of the coffin bottom half before conservation.
The Wood
The anthropomorphic coffin is carved from two large pieces of wood which form the lower half and most of the top half. A small third piece forms the projecting 'toe and foot' of the top. A very small wood sample, taken from the open area of the left side of the coffin's face, was analyzed by McCrone Associates, Inc. [2] and tentatively identified as ash, based on wood fiber morphology (the sample was too small for a cross sectional examination).
Along the mating surfaces of both top and bottom are six sets of mortise holes for joining the two halves. The wood varies in condition, but most of it has apparently retained its original strength. Notable exceptions are the areas underlying the red and green surface paints, where the paint has decomposed and darkened the wood. There is also evidence of biological damage, especially along the edges, and some abrasion. The shrinkage of the wood over the years has resulted in numerous cracks, some which penetrate through the thickness of the wood. Various modern metal nails and pins were found embedded in the coffin, apparently from previous display supports.
Ground Layer
A ground layer (CaCO3), dirty white in color, and of fairly grainy texture covers the wood on all surfaces. This ground material, presumably mixed with a gum or adhesive binder, was used as a primer wash, filler, modeling plaster, and adhesive plaster. It was applied as a very thin wash on all surfaces except the exterior of the top, where it is significantly thicker and contributes to the modeling of the shape. The ground material was also used to build up areas such as the nose and knees, which may have been too detailed to carve directly in the wood. Shrinkage of the wood beneath this ground layer has resulted in many cavities and detached areas between the wood and ground layer shell. Consequently, the fragile shell is extremely susceptible to further flaking and chipping whenever the coffin is handled, moved, or exposed to fluctuations in humidity.
Paint Layers
Samples of paint were selected from chipped and flaking areas, and identified by McCrone Associates, Inc. [2]:
Pigment Identification Summary
| Color |
Pigment |
| Green |
Unidentified yellow and blue |
| Blue |
Egyptian blue (Ca, Cu, Si) |
| Black |
Antwerp brown (bitumen) |
| Red |
Red ochre (Fe2O3) |
| White |
Whiting (CaCO3) |
| Yellow |
Orpiment (As2S3) |
Although areas of the paint layer were still well attached to the ground layer, others were flaking and crumbling. Areas of paint above cavities caused by wood shrinkage were in especially poor condition due to the disruption of the ground layer. The areas of green and red paint were often detached and highly friable, due to their corrosive effect on the wood support which was not mitigated by the typically thin ground wash beneath these pigments.
All surfaces of the coffin are painted, inside and out, with the exception of the contact surfaces between the two halves. These are ground coated only.
The Resinous Coating
An uneven, transparent, orange-colored resin coats most of the surface. This layer is apparently a vegetable gum and is soluble variably in acetone or ethanol. It is unclear whether this coating is of ancient or modern origin, or partially of both.
Areas along the foot of both halves were stained with a red paint, perhaps from a display case, as the stains are consistent with the contact pattern that might be made if the coffin were displayed 'on its feet,' an orientation that was used for many years.
The Connecting Areas
The joint between the top and bottom of the coffin was sealed in preparation for burial with a linen and plaster band around the perimeter. This linen (approximately 10 cm wide) is both bonded and coated with the plaster. The outer surface is painted to continue the iconography on the rest of the coffin. This linen composite provides a strong support for the paint layer, but it tends to detach more easily from the wood than the ground layer alone. Large portions of the linen, ground, and paint were completely detached from the coffin (most notably on the proper left side of the top, near the shoulder).
On the joint surfaces of both the cover and lower section, there are deposits (patches) of a black, organic material (possibly bitumen) mixed with plant fibers. The composition and purpose of this material is unknown, but it may have been part of the bonding process used to join the two halves.
Previous Restorations
An earlier, incomplete attempt at restoration in 1990 intruded modern, white fills and cosmetic in-painting on the coffin's surface. The PVA emulsion/CaCO3 filler, watercolors, and acrylics were used to fill losses in the original paint and ground coat. These were found covering raw wood, original plaster, and sometimes original paint. Beva 371 Film had been used to reattach sections of paint, ground, or linen to the body of the coffin. Japanese tissue was used to face the underside of the lower half of the coffin, in an attempt to stabilize the flaking paint from stresses induced from the coffin resting on that surface. The facing was apparently attached with a starch paste. When this tissue shrank as it dried after application, it pulled up entire sections of paint and ground from depressed areas of the surface. These detached surfaces were shattered, but still properly oriented, as they remained attached to the tissue.
The 'toe' of the coffin had been reattached (probably by a dealer in the 1930s) at an incorrect angle with two wood tenons using plaster as an adhesive. One of these tenons is of modern wood, and the other appears to be of re-used ancient wood.
Treatment
Solubility tests were performed on all surface materials in preparation for formulating the treatment plan.
Solubility tests indicated that ethanol or acetone would be the best solvent to remove both the modern in-painting and the modern filler, while leaving the original materials unaffected. Acetone was selected for use because of its greater effect on the modern filler.
The initial step was to eliminate, if possible, any paint or plaster that was not original. The top of the coffin was treated first, since this was the portion (predominantly the exterior) which had the majority of the modern restoration work. The modern paint was easily removed with cotton swabs and acetone, and a similar procedure was used along with mechanical cleaning to remove the modern filler. In some areas, heavier mechanical cleaning with dental tools was required to detach the filler. All of the in-fills were successfully removed, uncovering the grain of the wood as well as some small areas of painted decoration. The newly revealed areas were thoroughly cleaned with acetone, in order to remove remaining filler residue in an attempt to reduce the 'halo' effect of embedded filler particles. Since the filler is abrasive, the areas of decoration that were covered with modern fill suffered minor abrasion which occurred either when the filler was initially applied and worked, or as it was removed.
After the exterior was cleaned of intrusive materials and grime, the stabilization process began. The voids between the painted plaster and the wood underneath required filling to reestablish both structure and adhesion. Initial consideration was given to using resin or glue injection, but the size of the cavities ruled out this procedure; the voids were simply too large to reasonably expect a glue or resin solution to fill and strengthen without a filler or bulking agent. Therefore, an adhesive fill treatment was considered. A primary criterion for this fill was that it would flow easily enough to be pumped into the cavities with hypodermic syringe and needle and fill the entire cavity from a single entry point. Another criterion was that the fill would have minimal shrinkage upon drying, to prevent the formation of shrinkage voids within a sealed off cavity. Such voids would require the drilling of access holes for further filling.
As a consolidant for wood, polyvinyl butyral (PVB) has been shown to be a highly compatible material. Furthermore, when dissolved in ethanol, the least problematic solvent, and used as a binder, PVB does not coat the particles in a matrix, but collects and bonds the particles only at points of contact between them. This results in a consolidant which will not intensify color, or make the surface shiny, unless relatively high concentrations are reached. Additionally, PVB is effective at low concentrations, so a relatively small amount of resin is added to the artifact. All of these factors suggested that PVB was a better choice of resin than Acryloid B-72, since B-72 saturates the surface, intensifies color, was likely to stain, and requires solvents that are more toxic than ethanol.
The density of the glass microbubbles was chosen after testing various densities in sample filler matrices to determine shrinkage, best resin percentage, strength of bond, and crush resistance. Lower density bubbles crushed too easily after the fill had solidified, and higher densities are essentially solid glass spheres, losing the weight advantage of hollow bubbles. Therefore, the optimal density would be the lowest density that would have adequate crush resistance in the filler matrix.
A consolidant/fill solution consisting of 35% w/v Scotchlite K37 glass microbubbles in 10% PVB in ethanol, tinted with dry pigments, met the necessary criteria.
When dry, the PVB formed a sufficiently strong bond with the glass bubbles (and the wood or plaster), but when in solution it had a low enough effective viscosity to allow easy injection. When needed, the solution was cut 7:1 with ethanol to allow it to be easily injected with needles as thin as #19 or #22. Once injected, this solution worked as an adhesive filler that supported and bonded the plaster, without affecting the original paints. Usually the solution could be injected behind detached sections of paint and plaster, but occasionally a tiny perforation was drilled through the plaster to provide access to fill an inaccessible cavity behind it.
The consolidant/fill solution also worked well as an adhesive to reattach flakes of paint and plaster. After the surface layers had been made relatively strong and continuous, a 7.5% PVB/EtOH solution was brushed over the entire surface as a consolidant, to strengthen it, to prevent further paint flaking, and to slightly darken 'halo' areas caused by the modern plaster residue or abrasion. A 1% PVB/EtOH solution, tinted with a small amount of raw umber pigment, was used to further blend in light haloes around previously filled areas.
The interior of the top received the same treatment. The interior of the bottom of the coffin was treated next. First the figure on the interior was consolidated with 7.5% PVB, to stabilize the green and red paint which were extremely friable to the touch. The remainder of the interior was very dusty and dirty, and it was cleaned with an 80% ethanol/water solution rolled on gently with cotton swabs. This mixture was chosen after testing various proportions of water and ethanol on small areas at the foot of the coffin. Too high a percentage of water leached out staining materials from the ground layer (especially in the areas covered with yellow pigment) and too low water content did not effectively clean the surface. The green/red painted figure was then cleaned with 100% water, which was very effective since the figure had been consolidated with PVB. This initial treatment with PVB prevented stain leaching, but allowed the water soluble dirt to be removed with the water. After cleaning, voids between the layers were filled with the consolidant/fill solution used on the upper section. The entire interior and joint surfaces were then similarly consolidated with an application of 7.5% PVB in ethanol.
The underside of the bottom of the coffin presented the final challenge, as the paints and plaster were somewhat sensitive to water. The tissue paper facing and starch paste were carefully removed by saturating the local area with water/ethanol. The change in cleaning material from ethanol to water/ethanol was dictated by the presence of the starch paste, which is removable with water. Since many loose flakes were released by the removal of the tissue paper, an adhesive was selected that could be used while the surface was still moist. Loose flakes, as well as detached paint which adhered to the tissue, were reattached with PVA/PVOH emulsion adhesive (9% w/w PVOH/Jade 403 PVA). The plaster layer on the bottom is very thin and tended to crumble into dust when PVB/bubble solution was injected. The consolidant/fill solution was thus simply unusable. The PVA/PVOH adhesive, diluted in 1:1 H2O/EtOH, readhered the paint fragments to the wood with a minimum of cracking and flaking. The starch paste adhesive's solvents simultaneously dissolved and softened and detached the paint layer from the tissue paper. After cleaning and stabilizing the surfaces the entire interior surface was consolidated with 7.5% PVB in EtOH.
Conclusion
Stabilization of the coffin's surface has been accomplished. Continued chipping of the surface has been slowed or stopped, and the integrity of the paint/plaster/wood structure has been reestablished.
Photographs of the coffin top half after conservation.
Photographs of the coffin bottom half after conservation.
Acknowledgments
A generous contribution from Todd and Linda Herrick made the conservation and display of the coffin of Djheutymose possible.
Materials and suppliers
Scotchlite K37 Glass Microbubbles, 3M Industrial Tapes and Specialty Additives Division, Product Information: Scotchlite Glass Bubbles: General Purpose Products, K Series, 3M Center, Building 220-8E-04, St. Paul, MN 55144-1000
Monsanto Butvar B-73 (polyvinyl butyral)
Jade 403 PVA
References
1 RICHARDS, J. E., and WILFONG, T. G., Preserving Eternity: Modern Goals, Ancient Intentions, Kelsey Museum of Archaeology, Ann Arbor, MI (1995) 50-55.
2 McCRONE ASSOCIATES, INC., Analysis report, Kelsey Museum of Archaeology Archive, Ann Arbor, MI.
Related Material
JOHNSON, C., HEAD, K., GREEN, L., 'The Conservation of a Polychrome Egyptian Coffin,' Studies in Conservation, 40 (1995) 73-81.