JAIC 1999, Volume 38, Number 2, Article 5 (pp. 176 to 185)
JAIC online
Journal of the American Institute for Conservation
JAIC 1999, Volume 38, Number 2, Article 5 (pp. 176 to 185)





Samples of unsensitized albumen-coated papers were obtained from the Image Permanence Institute (IPI) and the Chicago Albumen Works (CAW). Both fabricators coated the sheets by hand, using the same general procedures (Reilly 1980) and the same paper base (Munson 1996). The papers were coated in the late 1970s to early 1980s. Descriptions of their procedures are presented in appendix A.

Thirty-two samples of the albumenized paper, each approximately 5 by 8 in. (12.7 20.3 cm), were sensitized with silver nitrate, exposed, and processed in our laboratory (Reilly 1980; Munson 1996). The sensitized papers were printed out to maximum density without using negatives, as a dark surface was expected to make any cracks that formed easier to observe. Processing involved the following steps: wash, gold tone, wash, fix, wash, hypo clearing treatment, final wash, and drying. Details of the sensitizing, exposure, and processing procedures are described in appendix B.

After processing, it became apparent that there were three types of prints, distinguished by their average gloss and thickness. The prints from IPI tended to be thinner and less glossy than the ones from CAW.

  • Type 1 (16 prints): thin (∼0.09 mm), low gloss (12 GU)
  • Type 2 (8 prints): thick (∼0.10 mm), low gloss (12 GU)
  • Type 3 (8 prints): thick (∼0.10 mm), high gloss (24 GU)

Because the two manufacturers used the same paper base and the same procedures to coat the sheets, the observed variations in initial gloss and thickness can be attributed to differences in the albumen viscosity and flow during application of the coating. Both manufacturers felt that variables inherent in the process such as temperature, relative humidity, age of egg, breed of chicken, and quality of the chicken life (free range or not) affected the final product. Both also stated that drying was important and that higher drying temperatures produced glossier emulsions. Aging of the solution can also be a factor. A 19th-century photographer working in Paris observed: “The older the albumen becomes, the more brilliant the coating will be, but there is some danger of yellowing of the proofs after the hyposulfite bath” (Ogonowski 1891, 17).

The focus of this study was to observe the relative effects of the treatment baths on a given set of available materials. As will be shown below, the relative effects of the baths were the same for all print types, so these variations in initial gloss and thickness, although interesting, did not affect the overall results of this study.


The 32 prints were divided into 4 treatment groups (A, B, C, D) of 8 samples each. Each group contained 4 prints from each fabricator. The print types based on initial gloss and thickness were also distributed equally among the treatment groups: each treatment group contained 4 prints with a thin albumen coating and a low initial gloss, 2 prints with a thick coating and a low initial gloss, and 2 prints with a thick coating and a high initial gloss. An identification mark was applied to each sample in pencil on the reverse. The samples were repeatedly subjected to the four wet treatments listed in table 1.


Baths A, B, and C were chosen as examples of baths that may be used in the treatment of photographs. Bath D was chosen because it was thought there might be some benefit to replacing the water in the saturated prints with ethanol, causing the sample to dry more quickly while still retaining its enlarged dimension. A similar procedure is used in the preparation of biological samples for scanning electron microscopy/transmission electron microscopy (SEM/TEM) analysis (Harkins 1996). All solutions were used at room temperature.


After each bath treatment, the samples were drained and blotted to remove standing water. The drying then proceeded under restraint, usually for five to seven days. The prints, with lens tissue against the recto and verso, were placed face down between layers of four-ply mat board, 3/8 in. (.95 cm) wool felt, and 1/2 in. (1.3 cm) Plexiglas. Once dry, the prints showed a strong tendency to curl, and so were kept flat in Mylar envelopes until the next treatment cycle. For the purposes of this study, it is only important that the drying technique was consistent throughout the project, as our objective was to study the effect of the treatment baths. The effects of different drying techniques will be the subject of a future study.

The wet-to-dry treatment cycles were repeated until obvious differences were seen when comparing the treatment groups. Messier and Vitale (1994; Vitale and Messier 1994), working with more fragile vintage albumen prints, observed significant changes after a single wet treatment, but this was not the case with the newly manufactured prints in this study. After 10 cycles, there were only slight differences among the groups. After 20 treatment cycles, clear differences among the groups had emerged, making it possible to compare the data and draw conclusions.


Changes in the surface of the prints were monitored by measuring the reduction in gloss. As roughness and cracking develop, the amount of scattered light increases and the gloss is reduced. The samples were measured with a Macbeth Novo-Gloss 20/60/85 Statistical glossmeter. This device measures gloss in incremental units of 0.1 GU (gloss units) and has a specified accuracy of 0.5 GU. The Novo-Gloss meter is capable of measuring gloss at 20, 60 and 85 with respect to the surface normal (perpendicular to the plane of the sheet), and the selection of angle is determined by the smoothness of the surface being measured. For surfaces of low gloss, measurement at a larger angle will maximize the collection of the reflected light, thereby producing a more reliable result. The average thickness of the prints was measured with a Mitutoyo digital micrometer.

The gloss of the samples was recorded before the initial treatment bath and after every second treatment cycle. To reduce planar distortion, the samples were held flat on a paper suction table during measurement. The initial gloss of all the prints was relatively low, and therefore the surface was measured at both 60 and 85. At 60, the glossmeter measures an area approximately 2 cm2 in size, and this area increases to approximately 3 cm2 for measurements taken at 85. Five spots were measured on each sheet, using a template as a guide to ensure that the same area was measured each time. The gloss was measured parallel to and across the grain (machine direction of the paper substrate) of the print at each spot. In all, 10 measurements were made at each angle after every second treatment cycle for each of the 32 samples in this study.

The data obtained at 60 and 85 exhibited similar trends, but the data obtained at 85 appeared to be more sensitive to smaller changes in gloss. For the sake of simplicity, only the data collected at 85 are presented here. As the bath treatments were repeated, the paper substrates showed increasing distortion, especially across the grain direction. As a result, gloss measurements taken perpendicular to the grain direction of the sheets exhibited a large degree of variation, while the measurements taken parallel to the grain showed a consistent decrease in gloss. For this reason, only the measurements taken parallel to the grain direction were used to assess the treatment results.

No cracks were apparent prior to the prints having been subjected to the treatment baths. During the course of the study, visual examination in normal, raking, and reflected light revealed that the prints appeared to become rougher and more matte. At the end of the study, cracks were clearly visible to the unaided eye, although not all prints exhibited the same degree of cracking. Stereo-microscopic examination (10 to 40x) and photomicrography were used to observe and record the relative quantities and size of the cracks that had developed as well as their preferred orientation, if any. At several locations on each print, the number of cracks in the field of view (at 40x) was estimated, and the results were averaged for the print. Similarly, the average length of the cracks was estimated.

Copyright 1999 American Institute for Conservation of Historic and Artistic Works