RESTRAINT WITHOUT STRESS, HISTORY AND PROSPECTS: A LITERATURE REVIEW OF PAINTINGS AS STRUCTURES
Barbara Whitney Keyser
3 DEVELOPMENTS IN “PICTURE MECHANICS” BETWEEN 1910 and 1970
APPROACHES TO THE CARE AND THE TREATMENT of works of art became increasingly rational and scientific in the early-and mid-20th century.
One pioneer in the scientific study of works of art was A.P. Laurie, Professor of Chemistry at the Royal Academy in London. Laurie wrote extensively, thoughtfully, and perceptively for a mixed audience of art historians, curators, conservators, and artists. As early as 1914 Laurie observed artifically-aged paint samples and noted that three to four years of outdoor exposure must equal three to four centureis of museum exposure: “The impression I have formed … is that there is really very little chemical change produced by weathering, and that the destruction of the paint is very much more of the nature of a mechanical breaking of the surface than of the change in the chemical properties of the linseed oil itself.” (64, p.155)
Later, in a pamphlet written for artists (65), Laurie showed concern for permanence of paintings by proposing a “pyramid of stability” for painting pigments with the least oil content under those with the most oil content (re-interpreting the fat-overlean rule) to prevent shrinkage cracks. Laurie also intuitively sensed the importance of increasing the hygrometic half-time of fabric-supported paintings, as he suggested that artists loose-line fabric supports with waxed canvas or wood panel.
The Courtauld Institute published another landmark document in conservation literature in 1934 (1). The director of the National Gallery in London had generated this study in 1929 to provide both guidelines for museum climatology and a basis for continuing investigation of effects of various levels of relative humidity on works of art. Both aspects of the study have been long in receiving the attention they deserve, but many of the problems recently studied by such investigators as Gerry Hedley and Marion Mecklenburg are foreshadowed in the Courtauld work. For example, after describing some experiments done on exposing strips of panel paintings to varying temperatures and relative humidities, the experimenters concluded:
No similar experiments have been conducted on pictures on canvas, but it will be realized that the stress between the paint film and a stretched canvas and the subsequent strain in the film itself due to humidity changes, will depend upon the material of the canvas, the tightness of weaving, the stretch, pull, and other variables, in addition to the initial condition of the canvas and stretcher … [In changes of relative humidity] the dimensional change in the paint film will not be the same as that in the canvas, so that considerable stresses must be set up between the two, and the resultant strains must either tend to loosen the film from the canvas or produce cracks. The extent to which a paint will adapt itself to such changes without fracture will depend on its composition and age. Also, apart from hardening with age, after repetition of stresses over a long period, it may be found that cracks develop rapidly within ranges of humidity change in which there would be no apparent deterioration in the earlier life stages of the picture … it is not safe to infer that because a specimen shows no sign of deterioration after a few years under a certain range of conditions, it will not deteriorate and deteriorate rapidly a few years later though the range of conditions has not altered (1, p.14).
In noting that dimensional changes of the painted faces of panels were less than anticipated in changes of relative humidity, the experimenters also commented perceptively:
… the explanation for the small movement is that the paint film and perhaps more particularly the undercoatings restrain the shrinkage and swelling of the front panel …. If [so], they must do this at the expense of considerable stresses in themselves (1, p.28).
The report also called for conditioned storage of hygroscopic elements of paintings, such as stretchers, cradles, and panels, so that their relative movement would be lessened when they were joined together.
Second, the report stressed the need for long-term natural ageing tests under controlled climatic conditions, concluding, “Until this [sic] data has been obtained the decision of optimum conditions for any particular class of picture must be more or less arbitrary.” (1, p.16) Third, in further experiments on actual panel painting specimens, the investigators discovered the utility and effectiveness of moisture barriers twenty years before the experiments and publications of R.D. Buck. Fourth, the report deplored the lack of systematic investigation of climate control and protection of works of art through proper framing and glazing methods. Last, the report mentioned resistance to the idea of sealing paintings on the part of restorers, who insisted that paintings must “breathe.” The report tactfully commented, “No reasoned support has so far been put forward for this view, which is mainly based on the experience of picture restorers and others closely concerned with the care of works of art; but deriving from such a source it obviously calls for careful scientific investigation.” (1, p.47)
Overall, the Courtauld report is an amazingly farsighted document, so far ahead of its time (or perhaps insufficiently publicized) that fifty years later there is still a need for many of the kinds of investigation it called for.
One of the Courtauld report's most telling observations was its recognition of the active structural role played by the paint/ground layers in works of art. What conservators lacked was a conceptual framework, not to mention technical means, for isolating and measuring the forces acting on and in paintings as they age and respond to environmental influences.
An early, and probably the first, attempt to quantify the mechanical behaviour of paintings, as well as to apply the recently-developed science of rheology to works of art, was made in 1942 by F.I.G. Rawlins, Scientific Officer at the National Gallery in London.81 Rawlins' stated aim was admittedly suggestive: “… to bring into focus the part played by flow and deformation in the science of picture mechanics.” (81, p.71)
Rawlins analysed paintings' mechanical behaviour by considering each material and/or layer of a painting in terms of eight rheological properties, including elastic modulus and tensile strength. He then assigned each material a 0–5 rating for each property to indicate the extent to which the material possessed or displayed the property. He thus established a “rheological profile” for each material. Rawlins then added up and compared the overall rheological properties of fabric-supported and panel paintings. When he compared his results with the hypotheses of the Courtauld report, he found a striking correspondence: the layers of the painting above the support made a much higher than anticipated contribution to the rheological/mechanical behaviour of the painting as a whole for both fabric-supported paintings and panels. Thus Rawlins concluded: “It would seem as if these layers tend to take upon themselves an Atlas-like burden which may, in certain circumstances, exceed their powers of endurance. (81, p.71) But no one took up this promising line of research, to locate and quantify the forces acting in paintings, for almost three decades.
The next investigator to search for a rational approach to the structural problems of paintings was R.D. Buck (23). In the 1950's and the 1960's Buck attempted to apply principles of mechanics and rheology in his search for an alternative to cradling panels, and he drew heavily on existing wood technology literature (4, 57, 75) for information leading to a rational treatment of panel paintings.
Some of the phenomena Buck studied also apply to fabric-supported paintings:
- He used moisture as a plasticizer to relieve stress in wood while slow-drying from a high, uniform moisture content (8, 55, 56). (There may be an analogy here in using moisture to plasticize cupped, brittle paint films in fabric-supported paintings in pretreatments before lining.)
- He extensively investigated moisture barriers to lengthen the hygrometric half-time of panels, thereby lessening destructive responses to daily and even seasonal relative humidity fluctuations.
- His stated ideal for the structural security of panels embodied the concept of restraint without stress: the panel would be plasticized, flattened slowly, and held flat by passive restraint.
While stressing the importance of R.D. Buck's work, it is nevertheless not unfair to point out some of its omissions and shortcomings:
- One senses that he was so fascinated with the behaviour of wood that he regarded the paint film as passive—a passenger on the active wood support.
- He did not explicitly recognize the role of moisture in plasticizing the paint film.
- He did not explicitly apply his insights to the care and treatment of fabric-supported paintings.
Nevertheless, Buck deserves great credit for his positive contributions to the understanding and the treatment of panel paintings.
1969 proved a watershed year in the study of paintings as structures, as two important papers appeared. The first was “Mechanical Alteration of the Paint Film” by Sheldon Keck (60). At the beginning of the article, Keck rightly pointed out that the study of the mechanics of paintings is an example of statics, that is, the physical behaviour of matter under mechanical stress which usually involves properties pertaining to flow and deformation: strength, elasticity, plasticity, ductility, malleability, creep, relaxation, fatigue and hardness.
Keck went on to cite paint and coatings literature (16, 20, 34, 38, 54) which describes the mechanical behaviour of paint films under varying loads, loading rates, times, and at different ages. He thus began to add rigour to the common empirical observation that paint loses elasticity and plasticity as it ages, while it simultaneously gains hardness and brittleness. Keck then described the factors involved in the development of “age” cracks, which are caused, he said both by forces within the paint and external to it. Keck noted that most paintings in North America over one hundred years old display a continuous crack network: since brittle fracture occurs perpendicular to stress, patterns of cracks indicate patterns of stresses—most notably the classic pattern further explained by the work of Marion Mecklenburg (26, 70).
Keck also cited a useful paper by F. duPont Cornelius which had appeared in 1967 (28). Cornelius had been one of the first individuals actually to measure the dimensional change of raw, sized, and primed cotton and linen canvas in response to cycles of high and low relative humidity. Cornelius's measurements confirmed the common empirical observation that sized canvas behaves differently than raw canvas. Noting that Cornelius recorded dimensional changes up to 5% in canvas cycled between 95% and 20% relative humidity, Keck hypothesized that even when restrained by the stretcher, the canvas could readily elongate more than the 1% required to rupture an aged oil paint film.
Having considered some possible mechanisms for age cracking, Keck made far-sighted recommendations for preventing such cracking. First, he stressed the importance of environmental control in the care of fabric-supported paintings. He also perceived the need for further research to measure and correlate tensile strength of young and old paint films, the stresses exerted by stretching a canvas, and dimensional changes in supports caused by variations in temperature and relative humidity. In addition, he called for development of lining supports which would be tensionless and dimensionally stable in the presence of moisture, and would have a coefficient of thermal expansion similar to that of paint. Keck also saw the necessity for reexamining the possibilities of marouflage.
Keck concluded by quoting R.D. Buck, who had asked, “What are the plastic and elastic limitations of old paint? What other rheological factors are involved in defining the properties of paint, and how may they be estimated quantitatively?” (23, p.39) Twenty-five years later, these questions have still not been adequately addressed by paintings conservators. Keck's own conclusion also remains true today:
… many of the current practices in the treatment of defects caused by mechanical stress are not based on an understanding of the forces involved. Consequently, some of the treatments performed today will only tend either to aggravate the stresses which they are attempting to overcome, or, inadvertently, to create new ones. (60, p.26)
In fact, the art of lining paintings was in a state of crisis in 1969 (50, 77). During the 1930's and 1940's wax had replaced glue as the dominant lining adhesive in northern Europe and North America, and it was generally perceived to be an improvement (79, 85). The vacuum hot table came into use in the 1950's and seemed at the time to be the ultimate solution for remedying the structural problems of fabric-supported paintings (86). However, by the 1960's indiscriminate application of vacuum waxlining onto linen was being questioned by several eminent conservators, and the search for alternative methods and materials was on (8, 13, 18).
The roots of these efforts, especially those of the ICOM working group, can be found in Giovanni Urbani's presentation to the second triennial meeting of the ICOM Committee for Conservation in Amsterdam in 1969 (91). Urbani perceived paradox in the lining process: while conservators attempt in lining to re-establish the characteristics of a new fabric-supported painting, that is, its flatness, surface character, tension, and elasticity, their methods involving heat and pressure can also be destructive to the original canvas and structure of the design layers. He also pointed out the absurdity of using as remedial treatment the same system of linen fabric on a wooden stretcher which caused trouble in the first place. He emphasized that structural treatments of fabric-supported paintings lacked a rational basis because the causes of deterioration and effects of remedial treatments had never been systematically studied. As a corollary, Urbani called for exploration of factors determining how long a painting could be made to last without being lined. He also called for exploration of means to measure the forces acting in and on fabric-supported paintings, and for functional analysis of technical solutions adopted.
To provide answers to these fundamental questions, Urbani proposed the following lines of research:
- Systematic empirical observation of naturally aged canvases and paint films, comparing lined and unlined examples;
- Chemical and physical testing of lining fabrics, original fabrics, and lining adhesives;
- Monitoring development and use of synthetic lining canvases;
- Proper definition of the adhesive problem. Urbani believed confusion in this area would continue as long as consolidation of the painting and adhesion of the lining canvas were attempted in the same operation;
- Application of adhesive technology to paintings conservation, including systematic studies of phenomena such as adhesion, cohesion, and creep.
Urbani concluded by stating that to understand properly internal tensions, deformations, and structural deterioration of fabric-supported paintings, the entire structure must be schematized mathematically in a continuous system. Like R.D. Buck, Urbani saw rheology as the branch of science which could be fruitfully applied to this problem: the result could be an exact model of the movements of a painted textile constrained by a stretcher.