JAIC 1992, Volume 31, Number 1, Article 15 (pp. 139 to 143)
JAIC online
Journal of the American Institute for Conservation
JAIC 1992, Volume 31, Number 1, Article 15 (pp. 139 to 143)




In her article, “Abrasiveness of Certain Backing Fabrics for Supporting Historic Textiles,” LoErna Simpson suggests that duck, warp sateen, and sailcloth fabrics are more abrasive than, for example, muslin, and textile conservators might want to consider abrasive qualities as a factor in selecting fabrics to be used in backing historic textiles. Having reviewed the data as presented, I would like to raise several questions that might suggest a different interpretation of the results of this experiment.

First, and perhaps most fundamental, is there any evidence that suggests that backing fabrics used commonly as backings in textile conservation promote fiber loss in historic textiles? Certainly, in removing a textile from an old mount, there are occasions when fibers of the textile are left on the mounting fabric. In my experience, these fibers generally are grouped around stitch locations, suggesting an abrasive action by the sewing (and unsewing) process rather than as a result of friction between the textile and backing fabric.

Second, does the use of the crockmeter, as described, measure the abrasiveness of the backing fabric, or does it measure the capacity of the backing fabric to be abraded? My reading of the data presented suggests that the backing fabrics are being abraded in this test. The float structure of a warp-sateen fabric is less sturdy than a plain weave, allowing for the float threads to be easily caught and pulled away. Cotton duck and sailcloth are more coarsely woven (they have a larger yarn diameter) than muslin, and thus it is reasonable that they too would be more easily abraded in the crockmeter.

Third, while it is evidently possible to use the crockmeter to remove fibers from a fabric, does this use, for which the equipment was not originally designed, actually equate to a measure of abrasiveness?

On balance, I believe that the suggestion that duck, sailcloth, and particularly warp sateen fabrics are more abrasive to historic textiles than should be acceptable is premature based on the evidence in this article.

SARA J.WOLF, Conservator, The Textile Museum, 2320 S St., N.W., Washington, D.C. 20008


In response to the comments of Sara Wolf, I would like to make the following statements. There is evidence that the roughness of a mounting fabric can contribute to fiber removal in mounted historic textiles. I have observed it myself and have had several other conservators relay instances of its occurrence to me, in places other than where sewing stitches occurred. I would suggest that, in the situation described in the first point above, a “smoother” backing fabric would have allowed even fewer fibers to be removed; that more fibers will be removed wherever the two fabrics are held in closer contact; and that it would be interesting to study whether the fiber removal was caused by the initial stitching or by later movement. Fiber removal can occur whenever two fabrics that are held next to each other are moved by some force, whether agitation, lifting, repositioning, or gravitational forces, and in so doing slide against each other. The more fragile of the two is likely to be abraded, while the rougher, sturdier fabric (usually the mounting fabric) will be the offending fabric. I contend that common sense leads textile conservators to rule out selection of mounting fabric with that very fact in mind, along with many other considerations, certainly. Would one ever select a heavy, rough canvas to use as a mount for a fine, fragile historic silk? And why is muslin used—not canvas—to make covers for padded hangers?

The real question is not whether abrasion and fiber removal does or does not occur, but how we understand it and decide to control it. The obvious choices of extremely rough backing fabrics can easily be ruled out, but the fine gradations for many other possible choices have not been evaluated or understood. One road to understanding is to develop a way to measure the degree of abrasiveness. It is precisely for this reason that my research was undertaken.

As I pointed out in the experimental procedure, the crockmeter has been used in previous research to actually produce linting of fabrics by rubbing one fabric surface against another. Therefore, its use in this instance is logical. The description of the procedure should allow the reader to understand the testing method. “The crockmeter peg was covered in a brightly colored, napped 100% cotton fabric … and the backing fabric specimen was placed on the lower platform. The rubbing action of the peg allowed fibers from the upper colored flannel fabric surface (simulating the historic textile) to be transferred onto the white backing fabric mounted on the lower platform.”

Therefore, the procedure is not measuring the “capacity of the backing fabric to be abraded” but rather the capacity of the backing fabric to be an abradant, i.e. to abrade. Webster's New International Dictionary (1961) defines abrade as “to rub or wear away, especially by friction,” which is what was happening when the red fibers from the flannel were removed by friction and transferred onto the backing fabric on the platform. (The backing fabric was not being abraded by the red flannel; no white fibers were found on the red flannel.) Note that the linen tester was placed on the backing fabric and the number of red fibers removed by abrasive action was determined by counting fibers on the backing fabric, as clearly shown in figure 1 in the article.

Equipment certainly can be, and frequently is, adapted for use in other situations. Similarity in the underlying principles of the equipment and test are the items to be examined. Is the force of action the same as that force needing to be investigated (in this case, friction)? Can the equipment securely hold the fabric pieces being investigated? Does the equipment provide a repeatable, reliable procedure? In all cases the answer is yes for using the crockmeter to study abrasiveness of backing fabrics.

In the article my conclusions simply indicate that duck, sailcloth, and warp sateen fabrics are more abrasive in nature than is muslin. I clearly state the “abrasive action is only one of many factors to be considered when appropriate backing fabrics are selected. However, conservators may find this information useful in selecting backing fabrics which are less abrasive in nature, when concerned about fiber loss with fragile historic textiles.” It is up to the conservator to decide when abrasiveness of mounting fabric should be a decision–making factor. Then the results of this testing procedure simply provide them with facts about the abrasive characteristics of fabrics. Readers may be interested in my additional research on abrasiveness of eleven backing fabrics forthcoming in the Clothing and Textiles Research Journal in 1993.

LOERNASIMPSON, Assistant Professor of Textiles, Department of Apparel, Interiors, Housing, and Merchandising, Oregon State University, Milam Hall 224, Corvallis, Oreg. 97331-5101



In the article “The Identification of Blue Pigments in Early Sienese Paintings by Color Infrared Photography” by Cathleen Hoeniger (JAIC, vol. 30, no. 2), Hoeniger indicates (p. 117) that a yellow filter is used in conjunction with black–and–white infrared film to record differences in infrared absorption by artists' pigments. For this purpose, either red (e.g. Kodak Wratten #25) or visually-opaque infrared-transmitting filters (e.g. Kodak Wratten #89B, 88A, or 87 series) are proper. Although surprisingly it is not mentioned specifically in the article, we can assume the color infrared film used for the study is Kodak Ektachrome Infrared Film (IE135–36, cat. no. 168 5718). A yellow filter (specifically a Kodak Wratten #12) is suggested by Kodak for use with this film. This is because the film's blue sensitive emulsion layer is also the one make sensitive to infrared; the purpose of the yellow filter is to absorb blue wavelengths so that this emulsion layer will thus respond only to infrared wavelengths.

Hoeniger also suggests that black-and-white infrared film (we assume she is referring to Kodak High Speed Infrared Film) has a slower speed rating than Ektachrome infrared film and for this reason is an easier film with which to work. Actually the suggested speed ratings (exposure indices) of both these films are relatively similar to each other and both vary greatly depending on the light source and filtration used; in fact, when working under tungsten illumination, most of the suggested exposure indices for the black–and–white infrared film are higher than that for the Ektachrome.

It should also be noted that there are a number of factors, some discussed below, that make use of Ektachrome infrared film more difficult, not “easier,” to use than black–and–white infrared film. We found it problematic that none of these were addressed in the article. One difficulty not mentioned is processing of the film. Kodak processing plants (now Kodalux) have not offered the E–4 processing required for this film since 1984 (all other Ektachromes are processed in E–6 chemistry). While E–4 chemistry can still be purchased and used by the film user, there are very few commercial labs that offer E–4 processing services.

Of greater concern, however, are those factors which are necessary to insure consistency in the “false color” rendering of the pigments. The character of the light illuminating the subject must be controlled rigorously. The inclusion of daylight as a suggested light source is particularly inappropriate because direction, weather, time of day, and time of year can vary the spectral distribution of the illumination and result in variation in color renderings.

Consistency in filtration is also essential. We were puzzled by the wide disparity between the “false color” rendering of ultramarine obtained by Hoeniger (“cherry red”) and that obtained by Olin, and Carter, Matteini, et al., and by our laboratory as well (“deep” or “dark” blue or purple). We and the others did observe, however, that cobalt, smalt, and cerulean blues were rendered as reds; the former two “intense” or “bright” reds, the latter a “magenta” or “reddish lavender.” It is possible that this inconsistency can by linked to the specific combination of light source and filtration used in the present study, which, although apparently useful, is probably at variance with those suggested by Kodak or used in the other studies. It is extremely unfortunate that the author does not indicate the light source and filter combination used and more troubling does not detail whatever testing, if any, might have been done to corroborate the conclusions drawn (e.g. optical or chemical microscopy) in each example.

Besides lighting and filtration, so many other variables can affect the “false color” rendering of any given pigment (e.g. underlying colors, variations in manufacture or source, even the age of the infrared film itself) that use of color infrared photography is best suited to indicate the presence of different, but visually similar pigments in a given passage rather than to identify the specific pigments themselves. Indeed, the apparently singular reliance on the use of color infrared film for pigment identification as that undertaken by the author in most of the examples presented is extremely risky.

DANKUSHEL, Associate Professor, Art Conservation Department, Buffalo State College, 1300 Elmwood Avenue, Buffalo, N.Y. 14222–1095


In response to the letter from Dan Kushel, I would like to say that perhaps when more art historians and conservators are working together in the field of artists' techniques, conditions will become more ideal for this kind of research. But as it stands, in most European galleries it is very difficult even to get permission to photograph works of art, let alone for detailed examinations in a laboratory setting. Furthermore, the type of survey I was interested in entailed looking at a large number of paintings. As a result, I required a method that would yield information about the two blue pigments–ultramarine and azurite–without the necessity of microscopy or sample analysis. Color infrared film had been used successfully by Norman Muller of The Art Museum, Princeton University, in the mid–1970s under similar circumstances.

In my research I used Kodak Ektachrome Infrared Film 2236–IE 135–36, which as I mentioned, was introduced by Kodak in the 1960s. This is the only color infrared film listed in Kodak's film catalog and widely available. The film can be processed at the Rocky Mt. Film Lab, 145 Madison St., Denver, Colo. 80206; (303) 399–6444. With the film I used a Kodak Wratten #12 filter (yellow) as suggested by Kodak in the packaging instructions. (I am grateful to Kushel for pointing out my mistake in referring to a yellow filter in connection with black–and–white rather than color infrared film, an error that found its way into the manuscript during revisions.) I worked in museum settings under mixed lighting conditions: hand–held flash supplemented by whatever light was available. Before photographing each painting, I examined the surface carefully to identify restored from original paint areas. A color chart consisting of early Italian pigments of graduated strengths mixed with egg yoke, was photographed with color infrared and used as a guide for comparing colors in the paintings.

My investigation was limited to distinguishing between ultramarine and azurite, the two blue pigments used in early Italian panel-paintings of the late-13th and 14th centuries (with the rare exception of indigo, a visually darker pigment). Smalt and cobalt blues had not yet been introduced. The results of the study were very clear—the presence of ultramarine was recorded consistently, as “cherry red” and azurite, in contrast appeared as “dark blue” on the false color film. The test was more that adequate for distinguishing the two pigments involved.

I cannot explain why the findings for the false color appearance of natural ultramarine in the work of Kushel (dark blue or purple) and Matteini, Moles, and Tiano, (1980) (“wine-colored”) differ from my results and those of Norman Muller (cherry red). In the 1970s, Norman Muller used color infrared film with tungsten lighting in a dark room and had it processed by Kodak. I used hand–held flash with some daylight, had the film processed by the standard E–4 method at the University of Guelph's Photographic Services and obtained the same results. Experts at Kodak have suggested tentatively that the color variations may be related to differences in lighting or the nature of the pigment. Hopefully further research will clarify the applications and limitations of color infrared film in art historical and conservation research.

CATHLEENHOENIGER, Assistant Professor, Department of Art, Queen's University, Kingston, ON, K7L 3N6, Canada

Copyright 1992 American Institute for Conservation of Historic and Artistic Works