Study in the (mis)uses of Papyrus Cartonnage, and Recovery/Conservation of its Layers

by Robert Kraft (begun 27 Sept 2006; latest update 09 Aug 2007)

This essay describes the way in which cartonnage containing attractive color representations has been mutilated for the sake of increasing income from the sale of its pieces. This particular example is taken from the eBay sales by Bruce Ferrini and associates in 2005-2006, but it is probably typical of what has happened to cartonnage papyri through the years, when preservation of the painted surfaces was not a primary motivating factor. (In the 1960s movie on papyrology [with E. G. Turner, et al.] you can see sulphuric acid used to dissolve the plaster and painted surface to get at the inscribed papyrus!)

Here is a cartonnage fragment that has preserved both (1) a "plastered" (variously called lime plaster, gesso, stucco) and painted face surface on the left, and (2) an unpainged "gesso" surface on the other side on the right [PPennKraft 044; 83x52 mm]. Unfortunately, it has obviously been cut from the originally surrounding material, and probably broken off as well. An image of the once intact larger cartonnage piece can be seen here -- overlays of pieces known to have been sold on eBay have been made by Ernest Muro (picture layer, back layer, inbetween layers).

This fragment is made up of several layers of papyri used as a base to which the "gesso" surface has been applied on both sides. In the following view with the piece held vertically, before separation was attempted, it is possible to glimpse the painted design on the plane facing downward, and the partial forced separation of papyrus layers (gently held apart by tweezers):


Looking back at the first image, note that part of the painted surface (at the right of the lefthand image) is missing, leaving exposed a bare section of papyrus. If the partial forced separation between the two outer surfaces were continued, something like the following would happen with removal of the unpainted plastered portion (the image to the left):

Another two adjacent fragments from the same panel:

cart036-039o    cart036-039p

This joined image comes from the same original cartonnage as our first example, showing the currently exposed back side first.  In the middle of  the backside of the joined fragments (above left) is a patch of the yellowish unpainted gesso, while around it the various layers of papyri have been torn away without regard for their integrity, leaving various fiber patterns. Fortunately, in this case, the surface of the front image depicting a winged figure is relatively undamaged [PPennKraft 036+039; c 125x80 mm]. Its original position on the cartonnage panel before the panel was dismembered is visible here.

Some of the papyri layers that originally lay between the two plastered surfaces contained writing, and the splitting off of papyri layers is presumably a crude attempt to expose such saleable items. In June/July 2007, I separated the layers of 044 (pictured above), using techniques described in detail below, with the following results (note -- for help with Demotic, see also the Chicago Dictionary and Janet Johnson's Introductory Grammar):

Top layer = image and gesso/plaster

Papyrus layer #1 = immediate surface to which the top layer adhered, writing on both sides (Demotic?)

Papyrus layer #2 = writing on lower side at least (Hieratic/Demotic; blotted ink on top side?)

Papyrus layer #3 = writing on lower side (Hieratic/Demotic?)

Papyrus layer #4 = possibly blank

Papyrus layer #5 = writing on both sides (Demotic) -- on the side written with the fibers (and upside down in the composite picture, corrected here), Willy Clarysse identifies three Greek names written in Demotic: Androitas son of Philistion, Simon 30 dr., and Apollonios or Ammonios (06de2008)

Papyrus layer #6 = probably blank

Back layer = gesso/plaster

Here is an example from another cartonnage complex of a split off fragment that proved to be blank on the visible surfaces, and accordingly sold for a lesser price than fragments with writing [PPennKraft 055.2, 69x53 mm]:

cart055-2a  cart055-2b

On the left image is a relatively smooth surface, probably the "back" of a cartonnage fragment (although possibly not heavily plastered originally). On the right image are the clear evidences of forced separation of this layer or group of layers from the presumed front portion. It needed to be determined, if possible, whether any originally intact surfaces (of layered papyri) remain under the somewhat confused patterns on the right, and if so, how to remove the obstructing layer(s) as carefully as possible, with as little further damage as possible.

In the 1960s papyrology film, a steaming process was used to "relax" the papyrus pieces. My guess is that with some of these cartonnage fragments, that might not be effective, since the layers adhere so tightly to one another. It is not clear whether the layers are held together by an adhesive (natural of artificial) that is soluable in water. If so, humidification or a steaming or even a water bath treatment should work, although the danger of splitting an original layer is also present. If not, the use of liquid solvents such as alcohol, benzyne, or acetone might be effective. Experimentation is often necessary to determine the best approach to separating the original layers, while at the same time preserving those layers and any art work or plaster surfaces involved.

In the present instance, simple humidification and gentle removal of the overlaid split layers of papyrus (mainly those areas that show up in a lighter color in the right hand image) using a thin rounded blade (palette knife) proved effective, and indeed, writing was discovered on the down side of the removed layers (as seen below, with both sides of the removed layers showing). This suggests that the material split off from this piece also would have shown such writing on at least some parts.


This piece of papyrus was sold in the same lot with another on which no significant writing was visible, PPennKraft 055.1. Since the latter clearly showed at least two adhering pieces, and since the smaller piece (5.5 x 6 cm after flattening) was already partly loosened from the larger (12 x 4 cm), I decided to attempt separation by using alchohol, spot blotted on the adhering areas with a "Q tip" swab rather than soaking the entire configuration in water. This alchohol approach proved to be largely effictive, although in the most stubborn area involving a split layer (see definitions below), some damage occurred that might possibly have been avoided with a water immersion approach.


Vocabulary: here are the terms I propose to use in describing the following processes:

"layer" = an original papyrus structure, with horizontal and vertical surfaces on the respective planes

"split layer" = the result of separation of the original planes of a "layer"

"fibers" = the visible remains of a damaged "layer" or "split layer"

"surface scraps" = smaller pieces adhering to a surface, obscuring what lies beneath

"painted plaster" = decorated surface with paint applied over gesso/plaster

"unpainted plaster" = surface that has been plastered, but not decorated (also called "gesso")

"shiny surface" = unpainted surface to which some sort of hard, shiny substance has been applied

"peanut brittle" = unpainted surface covered with a light brown grainy substance

"panel" = the original complex from which pieces have been broken and/or split


  Materials and Equipment: here are the basic items to be used (ideally) in the following processes:

digital camera and adequate lighting for documenting stages

centimeter ruler or grid for image scale (width and height)

note making equipment (paper, pencil, computer)

containers for the humidifying process -- preferably "see through" and heat resistant

distilled water if immersion will be necessary

surfaces for flattening and drying (e.g. blotters, flexible mylar, etc.)

tools for separating, flattening, moving (e.g. palette knife, tweezers, etc.)

cleaning solution (benzyne, alcohol) -- note dangers of toxicity, etc.

tools for cleaning (camel hair brush, "Q tips")

temporary and permanent storage material (mylar, plexiglass, glass)

Main Challenges: here are some of the difficulties to be anticipated:

how to preserve plastered surfaces, especially those with painted designs

how to distinguish integral layers from split layers, especially when no ink appears

documenting each significant stage for later reference

preserving and labelling each significant stage


Test #1. Jumbled mass attached to modern cloth base [PPennKraft 035; 92x70 mm]

cart035+  cart035o-1

On the left is the eBay image. Traces of ink are visible at the top of the right side cut (almost half way up the image) and a bit farther up and to the left. On the right is the breakdown after removing the mass from the cloth base (using acetone, in this instance). The fragment with the fine black writing at the right has survived ("group C"), and the other fragment just above it, with traces of at least one letter ("group B").

The other side, which was firmly attached to the cloth by a very tenacious adhesive, proved to be covered with the hard, peanut-brittle resembling coating, possibly also a modern feature (if so, it is unlikely to have covered any writing). Here is how it looked (with lots of fragmentation caused by the removal process); no single continuous piece of papyrus seems to have covered the entire area (the four larger segments will be labelled groups A-D below):


In the right image are the inscribed fragments that are barely visible in the first image above. The two on the left ("group C" and a separate small fragment) might be from the same plane (written along the fibers), but the one on the right ("group B") is written against the fibers and thus not continuous with the others.

Disassembly of the largest piece proved difficult, and was finally accomplished by soaking it for several hours in water that had originally been heated to the boiling point. The main results appear below, with the larger recovered pieces in approximately their original positions in relation to the original lump (which actually divided itself into three smaller configurations along the weak lines visible in the left hand images). Writing appeared on at least three of the layers, including one small fragment that appears to be hieratic or hieroglyphic. In the process, the "peanut brittle" layer was destroyed except for some small chips.

For convenience and clarity, we need to label the items processed. Since the above section is the largest unit to survive removal from the cloth backing, it will be referred to as "group A" composed of what proved to be three separable parts, "A-1" (top, when viewed in the original position imaged above, with the exposed papyri facing up), "A-2" (middle section), and "A-3" (lower and smallest of these sections). The pieces rescued from group A will be numbered 035.01, 035.02, etc. (there are 15 showing in the above images).

"Group B" is the piece showing what looks like the Greek letter "T" or possibly "P"; "group C" is the piece with parts of two Greek (?) lines in small ligatured letters; "group C" is the next smaller piece. All of these groups had the rough "peanut brittle" backing that needed removal to facilitate flattening, possible exposure of writing (there was none) or other layers, etc. The process used here was humidification, plus dabbing on of alcohol, which softened the peanut-brittle layer and made it possible to scrape it off carefully. Aside from straightening the edges and removing small surface split layering, groups B and C needed no extensive treatment. There were no significant additional layers.

Group D, however, proved to be a small piece that had been rolled and folded, and after the backing had been removed, opened up into a larger piece with some writing on it. No attempt was made to remove the backing from the smaller scraps.

Test #2: Very small painted cartonnage scraps [PPennKraft "cart114.1-6"]

These six small pieces were purchased together and all display a painted surface of some sort, sometimes carelessly oversmeared with a white paint. Two of them also have layers of material adhering to the painted plaster. The other four show evidence of having been united to papyri, but now have only the indentations or at best, half of a split layer embedded. Here is the original image supplied on eBay, followed by basic descriptions:

[insert image of cart114-6p+.jpg]

114.1 is the largest, and some writing is visible on the back layer. It will be treated in more detail below.
114.2 (1.6 x 1.0 cm)  shows dark blue-black and white paint, and has a split layer adhering to the back. I did not attempt to remove it.
114.3 (1.6 x 0.9 cm) shows traces of red and blue paint, oversmeared with white,  and a complex small layer-lump adhering to the back. The lump was removed, but is held together by a shiny hard material which resisted treatment.
114.4 (1.2 x 0.8 cm) shows red, pink, and traces of white, with some fibers from a split layer embedded on the back.
114.5 (1.0 x 0.8 cm) shows red, tan, and a hint of white, also with fibers embedded on the back. It is almost certainly from the same larger piece as 114.6.
114.6 (1.1 x 1.0 cm)  is red and tan almost button shaped, also with fibers embedded on the back. It is almost certainly from the same larger piece as 114.5.

[insert image of 114-1all-scan.jpg]

Test #3: Small painted cartonnage group [PPennKraft 078]

Here we have a layered piece of cartonnage with painted greenish-blue stripes against a tan background. The problem is how to preserve the painted surface while also removing the papyri layers. The original images are on the right side below, and the progressive layering progresses in order to the left. The topmost images show the layers as you proceed down from the painted surface; the lower images proceed up from the back of the clump -- that is, the topmost layers when looking from the back are to the left, matching the top row of images.

At the start of the separation process, small traces of ink were visible at the edges of the lower image, which turned out to be a nicely written documentary fragment, the first layer up from the gesso/plaster. The written surface was covered by a small rectangular piece on one side, and over that a rough triangular piece and some stray fibers (unless they originally were part of the triangular piece).

The separation process consisted of the following steps:

(1) Three coats of "Sally Hansen Hard as Nails" clear nail polish was spread carefully over the painted surface, trying not to let it flow over to the papyri but at the same time making sure it covered the cracks in the gesso/plaster in order to hold everything together (bleed-through from the cracks presented a minor problem in the separation process). As discovered later, it would have been better to use clear paraloid or acryloid B-72 lacquer at this stage, removeable with acetone.

(2) Standard humidification did not penetrate the lump sufficiently to permit separation of the layers, so the entire piece was immersed in heated water for about two hours.

(3) This softened the side of the plaster that adhered to the papyri sufficiently that a thin hobby-knife blade could be gently worked between plaster and papyri to separate the layers from the plaster. The coated painted surface remained intact through the process. Tape could have been used over the coated surface (and removed later with acetone or alcohol) to insure that it would remain intact, but that was not necessary in this instance.

(4) With the papyri layers separate from the plaster, the process of separating the layers proceeded normally, with careful use of the thin blade sliding between layers.

(5) Cleanup of stray fibers, pieces of dirt, etc., came next,

(6) Then the pieces were dried between glass, to insure flattening.

(7) Finally, three coats of the clear nail polish were applied to the back side of the painted gesso/plaster, to strengthen that material, which was then mounted between glass and a mylar cover over the painted surface (to reduce chances of crushing the plaster, or cracking it).

Before, during and after the process, digital images and (after drying) scans were made, as well as written notes on the various stages. Probably the nailpolish coating on the painted surface can now be removed by blotting (not rubbing) with acetone, but I would like to experiment first with some less interesting pieces before attempting that. I don't consider it an important step at this point, but there might be reason for it in the future.

Finally, here is an enlarged image of the written fragment (3.7 cm high x 6.0 cm wide), ready for us to attempt transcription (with some help also from Willy Clarysse, email 06de2008, probably from a contract)!

1. ] . . Φ [

2.  ] Δ H N A R I A . [   ==  δηνάρια

3.   ] . Λ Ο Υ . . . Κ A Q H Μ A Τ Ο Σ . [

4.  ] A Θ Ο N T I N Ο Υ Ν Τ R Ο P [     ==  κ]αθ᾽ ὁντινοῦν τρόπ[ον

5.] . . . .I  Ε P I  Τ Ο Υ T O [   ==  ἐπὶ τούτο[ις  (or the like)

Test #4 (PPennKraft cart053)

This piece presents fewer problems in that it seems to be folded and crushed, but without needing extra effort to remove adhesive or gesso/plaster covering. The evidence of cartonnage paint covering is very limited, with some small red paint patches and remnants of gesso, but not much else.

(1) photographed and folds noted, along with evidences of ink

(2) placed in humidification tank, some boiling water added to speed the process.

(3) immersed the piece when humidification was not effective.

(4) flattened and separated various sections, put them between glass to dry.

(5) photographed and [in process] reconstructed original locations of the resulting pieces.

[to be continued]


[prepared 27 Sept 2006; update 24oc06, 09au07, 15ap09]