In all branches of learning in these enlightened times there reigns yet still among us a certain barbarity in many areas. Among other things, the unfortunate fate of most fired coloured glass panels may serve as an example, and how many, indeed how very many of these beautiful windows that were once the greatest glory of our churches have been removed from our churches in our times, on the petty pretext of allowing in more light. I say 'in our times', because the art of making this glass has been lost, and we shall never be in a position to replace these superb examples of workmanship.
The Frankfurt art connoisseur Heinrich Sebastian Hüsgen may have been only too justified in penning these lines of complaint, in view of the heavy losses to medieval window glass surviving in the late eighteenth century that could have constituted just cause for such complaint. 1 Yet without realizing it, this friend of Goethe had had the great good fortune to see many of these precious old works of art in glass before the period of major industrialization. During the course of the nineteenth and twentieth centuries, losses to medieval churches and the glass windows with which they had been furnished increased steadily on account of harmful environmental factors. Concern has often been voiced that many historical examples of stained and painted glass are today in an alarming condition, and are in need of cleaning and conservation. A particular problem furthermore is the fact that many famous panels have been restored on several occasions over the years, and often exhibit patterns of damage that can be traced back to these restoration interventions. Even stained and painted glass of the late nineteenth and early twentieth centuries is now of an age that it will barely be possible to pass it to the next generation without intensive maintenance and preventative conservation. In view of the huge onus this constitutes in the preservation of historical monuments, it is my intention here to explain some of the types of damage that occur to historical windows, and to outline some of the measures with which these types of damage can be addressed and prevented from happening in the future.
As a result of their experiences, and following the catastrophic losses of the two world wars, all European countries made particular efforts to preserve their treasure troves of historical glass panels. Since all mid-European states (primarily France of course, but also Germany and England) have large expanses of medieval stained glass, methods for its cleaning and conservation were developed everywhere. Different ways and methods of restoring historical glass arose as a result of each country’s building on its native traditions. Fortunately, over the last decade a common approach has developed, a body of principles that remained unwritten for a long time, in accordance with which restorers in many countries in Europe and overseas can deal with stained and painted glass. This has arisen not least because of the efforts made by the international Corpus Vitrearum Medii Aevi, whose committee for the restoration of stained and painted glass drew up guidelines for the conservation and restoration of stained and painted glass for the first time in 1989, in conjunction with ICOMOS. These guidelines were recently reformulated by an international working group and published in this form for the first time in German (in Die Denkmalpflege, 62. Jg. (2004), Heft 2, pp. 109-112), and in English (in S. Strobl and S. Trümpler, 'The New International Guidelines for the Conservation of Stained Glass', Journal of the British Society of Master Glass-Painters, 28 (2004), pp. 168-74; and S. Strobl and S. Trümpler, 'New Guidelines for the Conservation of Stained Glass', Conservation News, 94 (2005), pp. 34-36). This progress, welcome for addressing both the technical and the ethical aspects of restoration, was occasioned by the establishment of academic training courses for glass restorers in Antwerp, Paris and Erfurt, which are further developing research and pedagogy in the field. Yet methodological and technological differences remain still today between the different countries of Europe. Consequently, what follows should provide an insight into the restoration traditions of Germany against a background of principles that are not the subject of international controversy; mention is also made of many of the methods that differ from country to country.
To begin, we should perhaps remind ourselves of the complex processes involved in the production of stained-glass windows. Because the techniques used for nineteenth-century and 'classic' stained and painted glass, which are still in use even today, are those that were already being used in the Middle Ages, our description of the process must begin with that period. The production of a stained and painted glass panel in medieval workshops has been described in detail frequently. 2 Although in the most famous art-technological handbook of the Middle Ages, the Diversarum Artium Schedula by Theophilus Presbyter dating to the start of the twelfth century, 3 the first step - the production of the glass itself - is described as if it were part of the glass-painter's job, we should probably proceed on the assumption that in reality glassworks were not in the same locations as the glass-painters' studios, and that they employed different personnel. If nothing else, this was necessary to satisfy the glassworks' enormous need for firewood. 4 It is remarkable that the condition of medieval glasses today can vary greatly, even within the same window. Even though the corrosion of glass is affected to a variety of factors, we may conclude from this that as early as the thirteenth and fourteenth centuries, workshops were simultaneously using differing base materials from different, sometimes very distant glassworks. 5
The starting point for the execution of the panel itself was a cartoon. This is necessary from a purely technical point of view and is corroborated by Theophilus. This sketch was very probably drawn onto a wooden panel covered with a chalk whitewash, and an example of one such panel from mid-fourteenth-century Gerona in Spain has survived. 6 At the same time, sketches made on parchment were in use: these were more easily transportable. 7 Designs and cartoons on paper survive in increasing numbers from the late fifteenth and early sixteenth centuries. 8
The sketch constituted the basis on which the pieces of glass were cut. For each colour desired, a piece of glass was cut out to the correct shape according to its planned outline by means of a hot cutting-iron; the edges were then shaped more precisely with a grozing iron. At later dates, diamond glass-cutters were increasingly used for this purpose. The design for the interior surface was then applied in a dark glass-paint and with washes of varying thicknesses; at later dates silver stain 9 and enamels, 10 with which images could be modelled, were used to create the design. In addition, the use of etching and cutting techniques, such as were applied on occasion in the late middle ages and the nineteenth century, allowed the layering of colours in multi-layered flashed glasses. 11 After the painted glass had been fired in a wood-fired oven, the pieces of glass were assembled by means of lead cames. These cames were cast to be H-shaped in section, and in the medieval period they were shaved and cut with a knife to the desired width and height. At later dates, the use of lead strips became the norm, whereby the raw material could be milled cold into the desired profile. 12 When the leading was complete, the joints between the cames were soldered with a lead and tin solder. It remains unclear whether the panels were then further insulated against wind and rain with putty (as was normal in post-medieval times), as there is no mention of putty in contemporary art-technological handbooks, and demonstrating the presence of medieval putty on specific objects is exceptionally difficult. 13 The panels of glass assembled in this manner were usually installed directly into the locations for which they were intended, in a groove in the stone window-frame; in the Romanesque period, wooden framing systems were probably used on occasion. 14 Very often, however, there were ironwork framing systems, consisting of wrought-iron flat bars, that subdivided the stone-framed openings in the window. Panels were then secured in their stone frames with mortar and to the ironwork with iron T-bars, against which the panels were placed and secured with wedges. In later times, a whole variety of systems for holding glass panels in place developed. 15
Some panels produced in this way have been able to last centuries in good condition, as is demonstrated by the panels that have been kept in museums since the beginning of the nineteenth century. However, those panels that have remained in their original locations and been exposed to the weather and perhaps suffered from inadequate maintenance today show signs of severe damage. Particularly noticeable in the first instance are the thick layers of dirt that cover the interior and exterior surfaces of nearly all historical panels. These encrustations of dirt usually constitute a firmly attached stratum of dust, soot (from candles and heating systems), remains from applications of putty, and other noxious substances. The exterior surfaces can often be observed to have been dirtied with pigeon droppings, which have an etching effect on the surface of the glass. It is not just dirt that damages the panels: very often it can be seen that the painting on the interior and exterior surfaces of the glass has gone missing. The layers of dirt and the corroded paint layers are usually so completely intermixed, that they can no longer be distinguished from each other. These problems are not just found on finely painted figurative panels: they are also found in ornamental glazing schemes, which today constitute the larger part of those historical windows that survive.
An even more serious problem is the corrosion within the glass itself. This starts as point-sized pits and spreads outwards from different centres (figs 1 and 2). The surface of the glass breaks up, and the pock-marked and damaged core glass underneath becomes visible. This type of corrosion can cover the whole surface of the glass and make it totally opaque. How can this type of damage be explained in more detail? Images produced with a scanning electron microscope show that the glass is divided up into strata (fig. 3). In glass of this type and date, a zone builds up over the core glass where, over the course of time, silicate molecules originating in the centre of the glass gather. 16 The effect of this is to render the zone harder and more chemically stable than the core glass. Unfortunately, one characteristic of this so-called 'gel layer' is the fact that it is riddled with tiny fissures and cracks. Water gets into the cracks and lead to the formation of syngenite and gypsum, causing further damage. Both syngenite and gypsum are opaque, which is why the glass seems to the viewer to become gradually darker and eventually black; both are also precipitated in crystal form. The crystals gain in volume, thereby breaking open the surface of the glass. On top of this package of corrosion is the original, endangered paint layer, as well as dirt and, very often, micro-organisms, such as mould or algae. 17
Another phenomenon causing serious damage is the browning of medieval glass at the centre of the core glass, which can also lead to the glass's becoming totally opaque (fig. 2). This phenomenon, usually caused by the oxidation of the manganese present in the glass, is unfortunately quite common in glass of the thirteenth and fourteenth centuries. 18 In addition to the patterns of damage inscribed onto the glass by corrosion, other forms of damage should be mentioned, such as the crazing of glass (as is often seen in panels from the nineteenth century), as well as the many almost banal-seeming cracks and gaps in historical windows, which are nevertheless very expensive to deal with from a restoration point of view.
How should one proceed with historical stained and painted glass in light of this complex damage situation? Before anything else, an absolutely essential requirement for these valuable works of art is research by art historians and restorers, which could include preparatory chemical analysis if need be. Even taking this first step is usually beyond the capabilities of a traditional glaziers' workshop, with the result that it becomes absolutely essential, right at an early stage of such measures, to take on as project planners experts or glass-restorers with specialist qualifications. While the various types of browning that occur in medieval glass cannot be reversed in the present state of restoration technology, 19 in the case of the corrosion encrustations, appropriate interventions may be made. It is normally useful to thin out the corrosion encrustations, as these are hygroscopic, absorbing moisture, which can lead to further damage. It is advisable initially to try to expose the various layers of corrosion: increasingly intensive cleaning should be carried out, and samples investigated with a scanning electron microscope, in order to establish the boundary between the corrosion layer and the 'gel layer' (fig. 3). Ideally - though in practice almost completely impracticable - the corrosion layer should be cleaned off without damaging the underlying gel layer. This layer is chemically more stable than the core glass, so less susceptible to corrosion. However, cleaning a piece of glass too hard or too deeply will probably result in increased corrosion in the future. In order to avoid this, restorers have gone over to leaving a thin layer of corrosion material in situ. This uncontroversial, common-sense approach has been adopted by all conscientious restorers and experts in Europe. However, the ways in which people try to achieve the desired level of controlled removal of corrosion encrustations are diverse. In view of the problems associated with the glass surface for most stained and painted glass, traditional cleaning methods are out of the question. Attempts to clean glass with sandpaper, steel wool, steel brushes, sponges, dipping troughs, etching acids or strong soapy water, which were still occurring in inexperienced workshops some years ago, are thankfully a thing of the past. Instead of this, different countries have developed different restorative cleaning methods. For instance, in England work is often carried out with a micro-jet process, by which the corrosion encrustations can be removed with exactly regulated pressure. 20 Comparatively soft materials are used for the micro-jet, such as crushed walnut shells, sodium hydrogen carbonate, and wheat flour. In this way, the corrosion layers can be thinned out in a focused manner. Many French restorers, in agreement with the Laboratoires de recherche des Monuments historiques in Champs-sur-Marne, prefer to remove corrosion chemically, by means of poultices or gel pads. 21 With this type of intervention, a cotton-wool poultice steeped in EDTA (ethylenediaminetetraacetic acid, a loosening agent for chalky encrustations), or a EDTA thickened to gel form, is placed directly onto the glass. After a while, the poultice or pad can be removed and renewed if necessary. In the hands of a specialist, this method can be used for deep, focused cleaning. However, the suspicion remains that chemically cleaned surfaces corrode more swiftly, as has been shown by experiments in Vienna, where sample areas of glass that had been cleaned with EDTA developed new corrosion encrustations after only eight years. 22 In addition, the chemical method has the disadvantage with fragile painted glass of endangering such broken remains of the paint as are found on the glass. In such cases, work can be undertaken in a more controlled manner by using mechanical means. Consequently, in Germany (and often in Belgium and the Netherlands as well), there is a preference for working with mechanical means (though sometimes with the help of chemical cleaning agents, such as ethanol poultices) - with cotton buds and/or scalpels, with which the corrosion encrustations are removed in layers. 23 In most cases, this type of cleaning takes place under the microscope. Many mechanical means are available for further cleaning of the surface, such as different varieties of hard brush, fat-absorbing powder, etc. Fortunately, opinion as to these cleaning methods is no longer divided into camps. It is also agreed that all three methods, in the hands of an inexperienced restorer, can cause extensive damage. It is therefore very important only to employ specialized and experienced restorers on a contract. 24 Thankfully, in France, Belgium and Germany, glass restorers qualified to diploma level are emerging in increasing numbers from the academic training institutions mentioned above.
As already intimated, restoration work is particularly difficult when original paint lines sit on top of the corrosion layers. These are often fragile and prone to become detached (fig. 4). Cleaning should not take place as a first step in such cases, and it may be necessary first of all to secure the original paint outlines. 25 This is done at specific points with small drops of synthetic resin, which are applied under the microscope. Under no circumstances does it make any sense to cover the whole piece of glass with synthetic resin. Such layers of resin are in danger of becoming detached in a short space of time and tearing off the painting underneath them, as has been shown by investigations on the windows of the Church of St Martha in Nuremberg, which were completely covered in the synthetic resin Paraloid B72 about twenty-five years ago. Here, the layer of synthetic resin has developed large bubbles and is becoming detached from the layer beneath over large areas. Layers of paint can also detach themselves from the glass to which they are applied, even when there is no corrosion of the glass. This often occurs in cases where enamel colours have been used. This technique, developed in the fifteenth century, involved firing ground, coloured glass onto support glass. 26 Frequently, stresses arise in the paint structure, on account of the differing expansion coefficients of the support glass and the paint layer; the brittle support glass inevitably suffers the effect of these stresses. The result is flakes that look like crystals, which are very hard to secure in place and conserve.
A further significant problem is overpainting by earlier restorers. Oil- and tar-based paints were used for overpainting on many windows, particularly in the nineteenth and early twentieth centuries. Over the course of time, the overpainting has become black, and is today working its way loose in the form of large clumps, often taking large parts of the original painting with it. Although there is often a desire to remove the overpainting on aesthetic grounds, the decision requires careful consideration. In the first instance it should always be clarified whether the 'cold', unfired painting layers are not in fact retouchings by the artist; these should of course never be removed under any circumstances. 27 The dangers associated with the removal of overpainting mean that such serious interventions should only be made when there are important conservation grounds for doing so. For example, if the overpainting is endangering the historical paint layers underneath, one should ascertain what the binding agent in the overpainting paint is through chemical research. The layers of overpainting should then be removed layer by layer with poultices, without damaging the original layers of paint underneath by rubbing too hard or by over-wetting
There are often cracks and gaps in historical panels, and the way in which these are treated from a restoration point of view throws up a large number of questions. While the decision-making process for cracks in glass is relatively simply, since repairing the crack with glue is only really necessary if there is danger of further loss, 28 the problems associated with missing areas of an image are harder to solve. A fundamental principle that must be observed when restoring any object, is that the work's authenticity should not in any way be compromised. 29 Any works undertaken must respect what is preserved of the original state. In the case of severely deteriorated stained and painted glass, there would be no technical problem with creating an 'as new' version of what the glass may have looked like - an extremely satisfactory solution for the lay public. However, restoring severely deteriorated glass in this state would involve destroying large parts of the object, and of course its historicity. The huge controversy as to whether partly or totally destroyed works of arts should be reconstructed has caused quite a stir in recent years. 30 Unlike as in wall-painting (for example), the introduction of 'neutral' stop-gaps of white or grey glass when restoring glass panels can lead to excessive levels of brightness in these areas (on account of the sunlight passing through them) and render the original composition much less legible. Thus in each individual case a plan of action needs to be researched and thoroughly reviewed. Results that are acceptable from an aesthetic point of view as well as from a restoration-ethics standpoint can often be achieved with suitably coloured and toned stop-gaps without any painted design on them. 31
Because of the many different materials involved, restoration systems for stained and pained glass are invariably highly complex. In addition to the different types of glass and the glass surfaces, the leading and the putty sealing usually also require restorative treatment. 32 Often the joints between the leads have broken and the leadwork is bowing out, which leads to multiple breakages in the glass and compromises its water-tightness. Many lead cames suffer from corrosion particularly around the solder points. It is often suggested that such areas should be releaded, which is usually not necessary, and is also not acceptable. Leadwork form part of the object's historicity, and are an integral component of these works of art. For this reason, the leads should also be cleaned and undergo restoration. Broken solder points can be resoldered, and if necessary, the leadwork can be carefully reputtied on the unpainted side of the panel. Even very damaged leadwork can be sufficiently restabilized and preserved in this way. Mention should also be made of the stonework and ironwork that frame the window. There are also important restoration decisions to be made here, on which specialists should be consulted. 33
It is self-evident that there should be written and visual documentation of all measures adopted during restoration that may be handed down to future generations. Standards for such restoration documentation have developed over recent years. 34 In all cases, consideration should be given to the suitability for archiving of all materials relating to such documentation. 35 Even in financially constrained projects, documentation is non-negotiable, as later damage can often only be explained and correctly treated through knowledge of a window's earlier history. Documentation is an integral part of the measures taken and constitutes the concluding stage of every single restoration project.
Many of the types of damage that occur to historical stained-glass windows can be traced back to the effects of moisture over a long period of time, either in the form of rain water, or condensation on the interior surface of the glass. Moisture can cause and aggravate not just corrosion to the glass, but weathering of paint layers and the corrosion of leadwork. In addition to their artistic function, all windows were also intended to act as protection against the weather and as space enclosers, and they were usually able to fulfil these architectonic functions for centuries. However, if we wish today to preserve these important works of art in the long term, the urgent necessity arises of preventing the further effects of moisture on these sensitive surfaces. This normally requires the installation of protective glazing, which prevents rain and polluted outside air reaching the precious panels (fig. 5). 36 If the decision is made to embark on a conservation measure of this nature, a protective panel is installed in place of the original leaded glazing. On the original ironwork support system is mounted a supporting structure that extends into the interior of the building, on which the original leaded glazing is placed, at a distance from the protective panel. 37 It is absolutely essential to ventilate the space between the two layers of glazing sufficiently, in order to minimize the occurrence of moisture in the ventilation gap. It is recommended that the frequency of condensation on the panels be checked with atmospheric measuring equipment in the interior of the church and in the space between the original glass and the protective layer. If protective glazing is installed, the various types of damage described above can be greatly slowed, or even prevented. In addition, restorative intervention to the historical panels can be limited to a conservatory minimum, as the painted glass is relieved of its function of enclosing the space by the protective glazing.
However, the installation of external protective glazing represents a major intervention in the church's architectural structure, and must be weighed against the desire to preserve historical stained and painted glass. One particular problem that should not be glossed over is the fact that changes arise from the repositioning of the historical leaded panels, both inside, where the profiles of the stone framework are obscured, and outside, where the view of the window is altered on account of the flatness of the protective glass panels. While the change to the interior view is normally tolerated, huge efforts are made in many European countries to mitigate the problem of the exterior view. Many historical windows in Germany are fitted with protective glazing consisting of hand-blown window glass (known as 'Goethe glass'). This glass contains small irregularities, which enliven the exterior view of the building (cf. fig. 5). Another way has been adopted in the new German states, where, in the context of a project set up by the Deutsche Bundesstiftung Umwelt, many churches have been fitted with protective glazing panels that have leadwork with rectilinear interstices; this is intended to subdivide the large expanses of reflective protective glazing. This solution does not seem very useful, because the leadwork of the protective glazing cast shadows on the internal, original glazing, thereby compromising its aesthetic impact. 38 Many attempts have been made to use glass that has been given a matt finish. From some viewpoints, these matt-rendered panels of glass are almost invisible, but they have the significant disadvantage that from other standpoints they give the effect of a completely opaque surface. A technically interesting solution was developed by French restorers for Tours Cathedral. 39 A mould was made of the external surface of the original leaded glazing, and a clear glass panel was subsequently produced from this mould. This glass panel reproduces all the irregularities of the medieval panel, including the leadwork. At the next stage of the work, the external surface of the protective panel was painted with various colours in imitation of the corrosion encrustations of the original. As a result, the external appearance of the protective glazing can hardly be distinguished from that of the original panels. However, there is a problem with this system, in that the painted on colours slightly reduce the amount of light in the interior, and distort such light as does shine through. It would therefore be sensible not to proceed with any painting of moulded panels. Yet the use of such heat-moulded panels poses the basic question of the authenticity of the exterior view, because in the system described, large areas of architecture are filled with 'imitation' leaded glazing. 40 One very simple and often very useful method of improving the visual impression of protective glazing is the installation of woven-wire grilles, such as have been used for centuries as protection against stones thrown at windows. The grating disguises the flat protective glazing behind it and can be a good solution, particularly for buildings where it can be proved that historically such gratings were fitted. If windows that require work are masked by nearby buildings, it is perfectly possible to dispense with expensive, specially produced glass, and to use cheaper window glass or Goethe glass to make up the protective glazing; this has the advantage of allowing the maximum possible transmission of light. As can be deduced from the variety of possible solutions, it is always absolutely essential to test the effect of the different systems on the building structure itself with different models, before embarking on the reconstruction of whole groups of windows. Solutions that may have provided very satisfactory results in one case cannot be transferred wholesale to another building.
It remains only to emphasize that in the restoration of historical stained and painted glass there are no 'patent solutions' that could have been detailed in the course of the present article. Both the complex nature of the damage to historical windows and the aesthetic problems of protective glazing demand intensive preparation and research in each individual case. A plan for the restoration of a particular object can only be drawn up on the basis of technical, art-historical and scientific research. Only then is it possible to put out to tender a detailed, object-specific list of the works to be undertaken by qualified glass restorers. 41 In view of the difficult questions posed by works of art to those working on them, it is imperative that specialists from different areas work together while the restoration is being carried out. At least three partners should be closely involved: the client; the expert and/or architect contracted by the client; and the restoration workshop carrying out the work. Responsibility also falls on the appropriate building authorities and monument curators. With objects of such importance it is also sensible to involve an expert from the relevant national committee of the Corpus Vitrearum Medii Aevi. In addition to expertise in the field of glass restoration, an important prerequisite is therefore collegiate co-operation in working groups, so that our precious stained-glass heritage may be conserved and preserved for the future.
This article originally appeared as 'Konservierung und Restaurierung historischer Glasmalereien. Eine Einführung in die Problematik' in Die Denkmalpflege, 62. Jg., 2004, Heft 2, pp. 141-50 and 154 (colour plates).
Translator: Joseph Spooner.
Publisher: Deutsche Kunstverlag.
Images: Ivo Rauch.