21. Mehra’s Eight Requirements for Linings, Revisited: Evaluation of Linings for Canvas Paintings—Then and Now

Introduction

The aim of this paper is to revisit and evaluate some early requirements for linings of canvas paintings proposed by Vishwa Mehra (Mehra 1972) and review them in a contemporary context. The requirements have been very influential within the field of conservation. They were used as guidelines for working conservators and teachers in structural conservation (see Mikkel Scharff’s paper in this publication), and the recommendations have been referenced in normative discussions about conservation praxis. Mehra’s ideas inspired a whole generation of conservators and influenced best practice considerations in structural conservation of canvas paintings for decades (Beltinger 1995; Hackney et al. 2012; Phenix 1995; van Och and Hoppenbrouwers 2003).

Vishwa Mehra is an Indian conservator of paintings and an art historian born in 1931. He studied conservation in India, Yugoslavia, and Rome and worked in Israel before he acquired a position at the Central Research Laboratory for Objects of Art and Science, in Amsterdam.¹

Through the 1960s there was an increased effort to strengthen and professionalize the field of conservation of cultural heritage. The first ethical guidelines for conservators, the Murray Pease report, were written for IIC-American group (later the American Institute for Conservation [AIC]) in 1963 (Pease 1964), and in 1964 the Venice charter expanded on these thoughts (charter and introduction can be found in Jokilehto 1998). The founding of ICOMOS (International Council on Monuments and Sites) in 1965 and of ICOM-CC in 1967 and the adoption of the 1972 UNESCO World Heritage Convention are also key events in the effort to protect cultural heritage, in light of which Mehra’s paper should be seen.

One event that also had enormous influence on this development was the catastrophic flooding of Florence on November 4, 1966. It became a milestone in the history of conservation for many reasons, one of them being that the need for education and conservation norms became evident as a result (Federspiel 2015). At the second plenary ICOM-CC meeting in Amsterdam, the Italian conservator Giovanni Urbani called for a more systematic and rational approach to lining canvas paintings (Urbani 1969). In 1972, Mehra presented his eight requirements for linings in an interim report for the ICOM-CC meeting in Madrid on behalf of the Central Research Laboratory, “Comparative Study of Conventional Relining Methods and Materials and Research towards Their Improvement” (Mehra 1972). In this report, Mehra discussed concepts that to this day remain essential to conservation. Some concepts were relevant for the general field of conservation-restoration, such as structural integrity of objects, reversibility, and compatibility of treatments. Other concepts in Mehra’s paper, including flexibility/stiffness, hygroscopicity, and permittivity, were especially important for lining paintings.

Mehra argued that traditional lining techniques using glue paste and wax resin were too invasive and nonreversible, as they caused the whole painting structure to become embedded in adhesive. Furthermore, he warned against the use of heat in linings, the forces exerted by retensioning a painting, and the extra weight of a wax-resin lining.

The Eight Requirements and the Intentions Behind Them

In his paper, Mehra proposed eight basic requirements for linings. Table 21.1 lists the requirements as they were formulated in the report fifty years ago. The questions that drive this paper are whether they are still useful and what they can tell us about our present attitude toward structural treatment of canvas paintings.

Table 21.1
Mehra’s eight requirements for lining canvas paintings

1.	Whatever the nature of the materials used, they should remain fully reversible with regard to additional relining eventually needed in the future
2.	Relining* may not in anyway cause alterations in the structural character of a painting
3.	The materials used should have passed the selection in direct reference to the specific problems of the painting involved
4.	Flexibility must be guaranteed for an unlimited period of time
5.	The use of heat should be avoided altogether or must be considerably minimized
6.	Increase of weight as a result of relining should be minimal
7.	The adhesive selected should not be allowed to permeate the canvas, ground and paintfilm alike. Instead, it should form only a film between old and new canvases
8.	It must be optional to use the selected adhesive in different degrees of cohesive strength and it is imperative that it will have proper resistance to fluctuations in temperature and humidity. It should be compatible with the other materials used for the relining which it serves

Source: Mehra 1972 (verbatim).
*Relining means lining; see Andersen 2012.

Table: Cecil Krarup Andersen

This paper was presented at the 2019 Conserving Canvas symposium held at the Institute for the Preservation of Cultural Heritage at Yale University.² The symposium title, Conserving Canvas, echoes the core of the discussions around Mehra’s paper. Mehra, in the Netherlands, and Bent Hacke (Hacke 1963–64), in Denmark, became advocates for treatment strategies that were designed to treat one problem at a time rather than the complete range of problems in a painting at once. Problems related to the canvas were now seen as a separate part of the canvas painting, and the treatments proposed (nap bond in the Netherlands and lamination in Denmark) were designed to deal with this layer. The ideas behind this concept of separation were presented in Mehra’s paper (Mehra 1972), where he warned against the long-held tradition of linings impregnating the structure. In this paper, the specific requirements written by Mehra are discussed in a contemporary context.

Reversibility

The first requirement is that the materials used should remain fully reversible (see table 21.1). According to Appelbaum, the “principle of reversibility” was written into the AIC Code of Ethics and it was advised that the conservator “should avoid the use of materials which may become so intractable that their future removal could endanger the physical safety of the object. He also should avoid the use of techniques the results of which cannot be un-done if that should become desirable” (Appelbaum 1987). However, fully reversible treatments are rarely possible, and the concept has been subject to much discussion. Muñoz Viñas has argued that the greater the reversibility, the less the conservator’s responsibility (Muñoz Viñas 2005). Concepts such as removability and retreatability (Appelbaum 1987; Charteris 1999) have been suggested as alternatives, but these still require an understanding of issues including surface tension, penetration, solubility with age, and peel strength—all topics that are still not properly investigated with respect to lining techniques. Despite the debates around the concept, reversibility is still written into the E.C.C.O. Professional Guidelines, in article IX of the Code of Ethics (E.C.C.O. 2003).

Changing the Structure

The second requirement is that lining, or relining as Mehra called it (Andersen 2012), may not in any way cause alterations to the structural character of a painting. Although it is not completely clear what he means by “the structural character,” it is known that he wanted to avoid heat and high pressure, which may change the shape of the paint. He therefore described the nap-bond treatment that targeted the needs of the canvas specifically.

The reality, now and then, is probably that most active treatments change the structure slightly, but some changes are more invasive than others, and some create more potential for future damage than others. The traditional adhesives for lining, glue paste and wax resin, were intended to penetrate the painting structure, but Mehra’s new thoughts were to avoid the penetration.

The seventh requirement—that the adhesive selected should not be allowed to permeate the canvas, ground, and paint film—is closely related to Mehra’s second requirement. Mehra wanted the adhesive to form only a film between old and new canvases, instead of penetrating the structure. This idea has prevailed in many traditions, and the nap-bond lining (or, in Scandinavia, lamination, inspired by Bent Hacke) became popular among progressive conservators of the time. The nap-bond technique was considered minimal intervention and was intended to ensure reversibility. Saving the physical integrity of an object is written into E.C.C.O.’s Professional Guidelines, in article V of the Code of Ethics (E.C.C.O. 2003) as well as AIC’s Code of Ethics, article 21 of the Guidelines for Practice (AIC n.d.). Thus, the requirement could probably still be seen as having significant weight, even though both glue paste and wax-resin lining are still practiced, as is impregnation of canvas paintings with glue and Beva.

This is much in line with the sixth requirement that increase of weight due to lining should be minimal. This, of course, was in comparison to a wax-resin lining, which was very heavy. A large format could be deformed simply by the weight of a heavy lining material if it was not supported. Lightweight treatments are still preferable, including stretchers and backing boards—especially for large formats—in light of increased loan activity between museums. In these situations, extra weight makes the painting more difficult to handle and increases the risk of high forces due to impact if the painting is dropped during transportation. Furthermore, a heavy structure can increase the risk of failure in the hanging system.

The Flexibility Requirement

The fourth requirement (see table 21.1) is that flexibility must be guaranteed for an unlimited period of time. It is not specified if Mehra means in the complete lining structure or in the canvas, the adhesive, or a possible interleaf. An argument for a flexible lining adhesive was mentioned as early as 1931, when the Canadian painter and sculptor Percyval Tudor-Hart advised that a lining adhesive should provide unlimited flexibility and recommended the “Dutch” adhesive (Tudor-Hart 1931), and that same year two American conservators remarked that a wax-lining is “the best that is yet found for relining . . . It is flexible, never becoming brittle” (Durham and Leser 1931, 38). The unlimited flexibility mentioned by Mehra should be seen in light of these sources and the long historical experience with brittle glue-paste linings. However, flexibility in lining adhesive and flexibility in lining as a whole are two very different concepts. In his requirements, Mehra mentioned the lining as a whole and not just the adhesive. In 1981, he contradicted this requirement in a paper on lining canvases, where he wrote that lining canvases should be as stiff as possible (Mehra 1981b).

A year later, Mecklenburg summed up his approach to providing stability in a painting: “to support a painting, the material chosen must: 1) be capable of developing and maintaining a large percentage of the total force of a stretched painting; 2) be able to restrain in-plane displacement as developed by desiccated glue and paint films; 3) restrain out-of-plane distortion such as cupping and buckling” (Mecklenburg 1982, 27). He furthermore advised that a high modulus³ of the lining material was preferable and that thickness would provide support against out-of-plane distortions.

Thus, research has shown that stiff linings can reduce the risk of further cracks and cupping in a painting (Mecklenburg 1982; Hedley 1988; Andersen 2013), which contradicts Mehra’s 1972 report. However, it is worth remarking that Mehra had been measuring the bending stiffness of the linings he tested, as opposed to the later studies, which measured stiffness in tension. For paintings, this means the difference between in-plane and out-of-plane forces.

With this development in mind, it is interesting to observe how Mehra’s early thoughts on flexibility have inspired conservators to require flexible linings (Heiber 1987; van Och and Hoppenbrouwers 2003). Thus, the properties of lining canvases remained a cause for disagreement within the conservation profession for many years after Mehra’s report, despite the fact that he was the first to change his statement.

Degradation Agents—What Is Worse?

The fifth requirement (see table 21.1) was that the use of heat should be avoided altogether, or at least must be considerably minimized. Heat could deform the painting and increase the chemical degradation rate. While this is true, high moisture/water, pressure, and solvents can also have a lasting negative effect on paint layers, and all are used in paintings conservation. The question of what is causing more degradation remains unanswered and relies heavily on the materials and properties of each individual painting. Solvents can cause swelling and leaching in paint and ground layers (Phenix 2002), and the development of metal soaps can be accelerated by moisture (Garrappa et al. 2020). However, all of this is a question of the extent and duration of exposure to these factors. The discussion of what is worse for a paint layer—exposure to fluctuating climate or exposure to water or solvents while cleaning, lining, or other conservation treatments—is still unresolved. New research is, however, starting to tackle these questions with, for example, nanoindentation, which allows for the study of the mechanical properties of paint surfaces (Andersen et al. 2019).

Choice of Materials for Specific Purposes

Mehra’s requirements for linings were widely seen as an argument for nap-bond linings only, but the third and the eighth requirements are actually pointing to the need to adapt each treatment to the specific problem at hand. Thus, the last requirement is that it must be optional to use a selected adhesive, in different degrees of cohesive strength, and that it is imperative that it has the proper resistance to fluctuations in temperature and humidity.

Mehra advised that the materials used in conservation should have passed selection in direct reference to the specific problems of the painting involved. While this is a nice principle, more specific failure/success criteria are needed in order to make such decisions. Very little advice is provided for choosing the amount of strength, stiffness, moisture, solvent, pressure, and so forth needed; thus, the choice of materials and methods largely remains a matter of practical experience and tacit knowledge.

Consider the case of a very torn canvas painting with frayed tacking edges. Different conservation traditions would imply different solutions to this problem. Everything from tear mending to glue-paste, wax-resin, and synthetic adhesive linings would be used and justified as the right treatment for this specific problem, depending on the tradition in which the conservator was trained. This is a testament to the reality that conservation is still a craft when it comes to structural treatments and that very little science-based consensus is found.

Compatibility

In the eighth and last requirement, Mehra wrote that the lining adhesive should “be compatible with the other materials used for the relining which it serves.” The meaning of “compatible” is not clear, and Mehra does not define it. It remains an open question whether poor compatibility means that the material does not bond well, that it is historically or ethically wrong, or perhaps that it has different properties from the original materials. It is also not clear how compatibility may be judged in a scientific manner. And yet the word is frequently used.

In E.C.C.O.’s Code of Ethics article IX, compatibility is included as a requirement, but it remains undefined. The European Committee for Standardization (CEN) standard defining the most important general terms used in conservation of cultural heritage, EN 15898:2011, defined compatibility as the “extent to which one material can be used with another material without putting significance or stability at risk” (CEN 2011, 10). Acknowledging that this was still a rather vague definition, Revez proposed the following definition: “Compatibility corresponds to the extent to which a product, method or action may be used upon a heritage object without putting its present or future significance at risk” (Revez 2016). In the updated version of the CEN standard from 2019 (EN 15898:2019), the definition was clarified further: “extent to which one material can be used with another material without compromising significance or stability of the object. Note 1 to entry: this relates to conservation materials or to components of the object” (CEN 2019, 25).

Useful compatibility indicators have been proposed for stone conservation and the conservation of archaeological sites by Rodrigues and Grossi (Rodrigues and Grossi 2007), but due to the difference in materials, these are not easily translated to paintings conservation. One indicator that might be useful, however, is physical content, which represents “intrinsic properties responsible for the performance of the intervention action, considered in its material side” (Rodrigues and Grossi 2007). The other indicators are the operational background, the sociocultural context, and the environmental constraints. Assuming that compatibility is a multifaceted concept that can be broken down into a number of simpler categories, Rodrigues and Grossi propose a rating system (from 1 to 10) for each action taken, which would then be translated to an incompatibility degree for a proposed treatment of stone and buildings. The question remains whether this concept works for structural paintings conservation.

It seems that in the context of Mehra’s paper, the physical indicator was probably what was more on his mind. If for the sake of simplicity one stays with this concept, it may be possible to propose compatibility indicators for mechanical and structural properties of conservation materials. Problems with the concept of compatibility in the context of canvas paintings include that the materials of canvas paintings (canvas, glue, ground, paint, etc.) have very different properties, and research suggests that lining material should not have the same mechanical and structural properties as the original canvas, as described above. Stiffer and less hygroscopic materials have been found to provide better supports, so perhaps compatibility of structure should be considered in relation to the paint and ground layers. Another, and possibly more comprehensive, approach to assessing treatment options is risk assessment, which is considered in the next section.

Evaluation of Risk for Linings

In structural conservation of canvas paintings, very little research progress has been made in the last decades, especially when it comes to predicting the risks of structural conservation treatments. Instead, the tendency has been to avoid treatments. However, one tool that can help us move toward rational risk assessment is computational finite element modeling to predict stresses and the resultant strain and failure in the painting structure with fluctuating ambient climate conditions. A comprehensive computational model and subsequent parametric studies are missing for lined paintings, as is validation of the results through epidemiological paintings collection studies.⁴ Computer-generated prediction requires not only detailed model design but also reliable experimental data on material properties that can help answer specific questions about structure and mechanical properties. The potential is promising and can point the way to more solid guides for conservators, but until such a tool exists—with validated results—the best option we have is to rely on risk assessment for guidance.

In cleaning literature, the concept of star diagrams for evaluating specific parameters was developed and used repeatedly (Daudin-Schotte et al. 2013; Chung et al. 2017; Ormsby et al. 2019; Bartoletti et al. 2020). This approach could be useful for other conservation treatments as well, and would follow the trend of preventive conservation that moves away from requirements and recommendations or guidelines and toward risk assessments.

Linings and related structural treatments are still performed widely, and the different alternatives should be evaluated when choosing a treatment option or considering paintings that have already been lined. A star diagram shows risk factors situated around the center (fig. 21.1). A value from 1 through 5 is given for each risk factor, with 1 as low risk and 5 as high risk. This is different from the approaches proposed in the articles referenced above in that it is intuitive that the higher the rating number, the higher the risk, and vice versa. Thus, a smaller shape means lower risk, as shown in figure 21.1. Of course, the scale is a matter of interpretation and must be discussed for each factor.

The first step is to identify the risks. Using Mehra’s guidance and some updated information, the list in table 21.2 could be proposed as examples of eight factors that may imply an increased risk of change in a lined canvas painting. The chosen parameters are inspired by current discussions in the field, such as stresses, color change, change in properties, and so forth.

Table 21.2
Eight lining risk factors and example ratings

Change in paint properties or structure—due to solvent exposure (including water) or temperature change No risk or very low risk. Small risk of subtle change. Medium risk of flattening or stiffening the paint due to the use of solvents or temperature. High risk of visible flattening areas with impasto and changing paint properties permanently. Very high risk of flat impasto or paint loss and permanent change of properties.

Change in canvas properties or structure—due to impregnation of adhesive or moisture/water exposure No change or slight effect of adhesives bonding to surface. Slight change due to adhesives penetrating the surface. Decrease in strength or changed response to RH due to impregnation. Significant brittleness or weakness and/or change in response to RH due to the effect of the adhesive, or risk of shrinkage. Serious chemical and/or physical weakening and deformation of canvas due to impregnation with adhesive, or high risk of shrinkage.

Unstable environment (long exposure)—creep due to lack of support No change with change in RH or temperature. Risk of increased cracks over 50 years. Risk of increased cracks and cupping over 10 years. Risk of bulging, cupping, and delamination over 5 years. Risk of delamination after 1 year of climate exposure.

Short exposure to forces—lack of support with fast climate events, keying out or restretching Support is as stiff as the paint layer—fully supported. Support is stiff, but not as stiff as the stiffer paint layers. Lining provides some support, but not enough to prevent damage with significant use of force (e.g., keying out in corners). Lining provides very little support, and there is risk of cracks and cupping. Treatment provides no support, so any slight or significant increase in forces will affect the paint layer directly.

Shear forces between canvas and lining—due to stiffness in lining and shrinkage forces in canvas No forces. Slight forces with changes in RH. Some shear forces with change in RH in case of canvas contraction. Local delamination of lining with change in RH if canvas contracts and the binder is not strong. High forces with delamination of lining if canvas contracts and the binder is not strong.

Shear forces between canvas and paint—due to free movement of the canvas and stiffness in paint No forces. Slight forces with changes in RH. Some forces and risk of long-term development of cracks in paint layer due to compression or tension. High forces and risk of cupping and delamination. Very high forces and risk of tenting and delamination of paint in high RH situations.

Visual change—due to impregnation and subsequent change of refraction index No change. ΔE = below 1: Change cannot be seen with naked eye. ΔE = 1: Change is just noticeable. ΔE = more than 1: Change is obvious, but the artwork can still be enjoyed. Significant change and seriously compromised appreciation of the artwork.

Reversibility—due to materials used or construction of structural treatment Easy to remove both canvas and adhesive (if present). Can be removed with little effort. Canvas is easy to remove while adhesive is more difficult, but low risk for original materials if care is taken. Canvas and adhesive can be removed only with some risk of damaging original materials. Canvas and adhesive cannot be removed without causing significant damage to the original painting.

In the example of star diagrams for a lining in figure 21.1, the risk for each parameter for a particular painting has been assessed. The nature of specific paintings and treatments will change the nature of the risk, of course. Risk factors can be added or subtracted as appropriate, and different options can then be compared. Other risks, such as weight change, out-of-plane forces, infestation, and mold growth can also be assessed as needed. In some cases, the values for various parameters can be measured in a lab, and the outcome can then be adapted to the same scale, as seen in Fuster-López et al. in this publication.

Conclusion

Fifty years after Mehra formulated his requirements, the concepts he discusses largely remain relevant, even though the context and the understanding of some of them have changed. The eight requirements became very influential, but they must be seen in the context of the lining techniques in use at that time, when wax-resin and glue-paste linings dominated. However, the imposition of requirements for conservation treatment is not a contemporary way of approaching conservation—partly because each painting is different, as Mehra acknowledges, and partly because best practice requirements call for a common understanding of the concepts they codify.

Many of the concepts used in conservation are elastic. They can be interpreted in various ways and have multiple meanings, and this can prompt endless discussions. The specific interpretation in each case can be attributed to differences in culture, social parameters, stakeholders, technical knowledge, and practical indicators. This means they can be bent and understood differently depending on the context. In case of misinterpretations of the concepts used, requirements may be worthless or even misleading. Tacit knowledge is deeply embedded in local conservation traditions, which, of course, complicates the matter further.

Until a comprehensive prediction tool is created, assessment of risks remains the best option. An update of Mehra’s approach was proposed in this paper, in which risks are assessed from certain criteria for each case individually by the conservator who needs to decide between different options.

1. https://www.mehra.nl/vishwa-raj-mehra/. ↩︎

2. Funded by the Getty Foundation. ↩︎

3. E-modulus is derived from a stress/strain test representing change of stress divided by the change in strain in the elastic (Hookean) region. High modulus can be understood as high stiffness. ↩︎

4. Since this paper was written, we have proposed a computer model for untreated paintings; see Lee et al. 2022. ↩︎