Miniatures on ivory first appeared in the eighteenth century. While they present stylistic variations depending on where and when they were executed, in general the technique is relatively uniform. Old treatises describe them as painting in gouache, a term that covers water-based pigments on ivory. The surface tension of the water creates a droplet and prevents the pigment from spreading uniformly across the ivory, which is why old treatises incorrectly described ivory as a “greasy and slippery” material which was de-greased using ox gall. In fact, the ox gall functioned to reduce the surface tension of the water, thus allowing the paint to spread more easily across the ivory. To ensure better adhesion to the support, artists sometimes created a slightly rough surface with polishing agents such as ground pumice stone.
The small panels of ivory were made by cutting the tusk longitudinally, for which reason its width determined the diameter of the sheet. The largest ivory panel in the Prado's collection is the Portrait of the Infanta Luisa Carlota by Luis de la Cruz y Ríos (O-669), which measures 12.5cm wide. For very large-format works the outside layers of the tusk were sometimes used, which is a rougher, yellower area that can be seen at the edges of some miniatures.
The ivory panels are extremely fine, approximately 0.4 to 0.5mm thick, although in works of lesser quality they can be more than 1mm thick. The use of such fine panels was not due to a lack of raw material but to the fact that the translucent, whiteish tone of the ivory was used for the flesh tones. The ivory tone was used as the base while on some occasions a sheet of silver leaf, silver-plated copper or gilt metal was placed behind the miniature as this lit up the flesh tones from behind and increased the depth of the shadows.
This thin metal sheet was generally only located behind the flesh areas although we have encountered examples in which it was placed behind the entire image. In other cases artists used what was called tiza roja, a piece of red card that “warmed up” the flesh tones and gave the sitter a lively, healthy appearance. When these backing sheets have been lost or have deteriorated, the miniature looses light and the sitters' appearance becomes rather dull and pallid. In the case of the Portrait of the Marchioness of Lazán (O-701), José Alonso del Rivero not only used a sheet of silver leaf but also painted on the back of the ivory panel in order to make the background darker.
The backs of miniatures were generally covered with a piece of protective paper attached with a few spots of glue, frequently with a piece of vellum on top of it. In other cases, the ivory panel was stuck down onto a thick piece of rigid card that made it safer to handle and to mount in a frame. This system was used in the Portrait of the Infanta Luisa Carlota by Cruz y Ríos (O-669). When the ivory panel is uniformly stuck to the card this provides excellent stability but unfortunately, on some occasions, particularly when later restoration has been carried out, only one side is stuck down, as a result of which the ivory has become distorted in the area around the adhesive.
Miniaturists began their portraits with a preparatory drawing. Various examples in pencil have been encountered but it is not easy to access them as on some occasions they are concealed beneath the paint while in other cases they were executed with a brush in the same technique and using the same pigments as the miniature itself. Dechateaubourg's Portrait of a Girl (O-675) reveals a particularly precise under-drawing of the dress and the bow of the edging of the dress.
Another interesting example is Juan Pérez de Villamayor's Portrait of Isabel II (O-698), which reveals an area of squaring-up behind the dress and various lines that establish the position of the mouth. This was an institutional portrait that may have been made on the basis of a larger preparatory sketch that was reduced using the squaring-up technique.
For the paint layer artists used pigments with water-soluble agglutinants. It is difficult to specify exactly which agglutinants were used due to the small scale of these works and the thinness of the pictorial layer, which has prevented us from taking representative micro-samples for analysis. In general, therefore, we have referred to old treatises that discuss the use of tempera and gouache techniques in general terms. However, during the restoration of Portrait of a Girl (O-675), a fragment of the paint layer became detached and could therefore be analysed2 using gas chromatography and mass spectrometry. These tests revealed that the agglutinant was gum arabic, which is typically used with watercolour. When used, the amount of pigment could be varied or other elements added, allowing the artist to deploy both translucent brushstrokes that exploited the transparent nature of the ivory, and dense, opaque ones that concealed the support.
Three principal pictorial techniques have been identified in the Prado's miniatures: transparencies in the manner of watercolour, stippling, and hatching over superimposed, opaque strokes. There are miniatures in which only one of the three is used, but the most common practice was to combine them.
In general, the faces and flesh areas were painted using stippling and hatching, as in Portrait of an elderly Lady by Antonio Tomasich, which reveals a large number of tiny dots, or in the above-mentioned Portrait of a Girl by Dechateaubourg, which uses a lesser density of dots. In some cases stippling was painted over glazes applied with the brush and which functioned as the initial lines of the composition. Only in a very few miniatures in the collection are the faces painted with transparencies characteristic of watercolour technique, of which the most spectacular example is Portrait of a Boy in which the anonymous artist used a very limited number of brushstrokes over the translucent tone of the ivory to create an impressionistic effect of volume.
In the grisaille of Profile Portrait of two young Women attributed to Bourgeois de la Richardière the brushstrokes are executed in black with a few white highlights.
To depict the hair artists generally applied long, fine brushstrokes that overlapped the stippled area of the face, while for the clothes a wide range of techniques were deployed: the interplay of transparencies was exploited for areas of thin, light-coloured material while for the other colours a thicker, more opaque layer of pigment was applied, starting with the darkest under-layers, which were then graduated with increasingly pale toned strokes. For areas of embroidery, mantillas and above all jewels such as brooches and necklaces, a sense of volume was achieved through the use of heavily charged spots of white pigment that stood out in relief. In some cases we also encountered the use of gold dust to complete this effect.
Scraping was also employed, both to define an outline and to lighten the effect of an area of colour, while another technique occasionally encountered, particularly in the backgrounds, is that of the application of drops of water to create a pale spot with a dark outline.
Very few pentimenti or errors were encountered in these miniatures due to the fact that they were based on carefully conceived preparatory studies, and the few corrections that were found were carried out by scraping out rather than painting over.
The backgrounds tend to be thick and opaque and in miniatures that used a sheet of silver leaf the edges of the sheet had to be concealed from the front given it would otherwise have showed through. This is evident in the unfinished miniature Portrait of a Lady in a Black Dress (O-715), in which the sheet of silver shows through around the head.
The frame is important for the preservation of a miniature as it acts as a protective and stable environment. All the Prado's miniatures are framed although many of the frames are not original and a large number have been subsequently altered. It is common to find pieces of paper, card or cloth that fill in the space at the back and fix the miniature into the frame. In the case of works from the Perera collection the pieces of paper and card found during restoration were prescriptions or from boxes of medicines. Arturo Perera was a doctor and thus used materials from his everyday surroundings for his miniatures.
An interesting case is that of the miniature of the Duchess of Alba (O-694) by Guillermo Ducker. When the back of the frame was removed a carefully folded sheet of paper was found, which was a page from a book with a detailed description of an aphrodisiac called satirión or 'dog's testicle' (orchis mascula). Examples of this type raise the question of whether the piece of paper simply acted as filling or whether it was placed there for another purpose.
Some frames take the form of lockets for hair. In commemorative pieces of this type the hair is arranged decoratively in curls or is woven or plaited. In both cases the hair was prepared in the same way: it was soaked in an aqueous adhesive3 and then woven into the desired pattern while still damp. When the adhesive dried the woven hair was placed in the locket.
The sheets of glass in old frames were blown by hand and are unique to each frame. They were known as vidrios lupa [magnifying glass] although this effect is only evident at the edges as the centre is flat. This subtle effect functions to enlarge the miniature. Many of these sheets of glass have deteriorated, resulting in the hydration of the glass and the appearance of tiny surface drops that crystallize in relation to atmospheric conditions.
A raking light study revealed that their internal structure was micro-fractured, which is an associated type of damage that occurs when hydrated glass dries out.
An analysis of the “drops” revealed that they were deliquescent salts of sodium acetate4, for which reason the compound that destabilised and started to corrode the glass was acetic acid. This type of deterioration was frequently triggered off by inappropriate wrapping material and damp combined with gas emissions from unstable materials such as wood, resulting in the onset of the chemical process of crizzling.
In some cases these 'drops' only appear on the inside face of the glass and seem to have spread out as if wiped with a cloth. This suggests that a product containing acetic acid such as vinegar was used when handling the miniatures and that this substance had not been dried or completely removed before the frames were sealed up, leading to the deterioration of the glass5.
Given that the sheets of glass are concave the drops have only been in contact with the miniatures at the edges, although in some cases we encountered small surface marks only visible with a binocular lens. However, the acidification of the atmosphere can affect the paint layer. The most damaged work in this respect was Portrait of a Girl (O-675) by Dechateaubourg in which some white brushstrokes in the area of the dress had blackened on the surface and acquired an iridescent effect. When analysed, the pigment was found to be lead white, which oxidises in an acidic environment and takes on the dark grey tone found in the miniature6.
In addition, the pictorial layer was fragile and brittle, for which reason various micro-fragments had become detached. Their analysis revealed that the above-mentioned agglutinant, gum arabic, was damaged, possibly due to hydrolysis characteristic of an acidic incident.
The miniatures' state of preservation varied considerably and depended on the work in question. Deterioration partly derived from natural aging but the most serious damage was the result of inappropriate restoration in the past and poor handling.
Changes of relative humidity and temperature influence the natural aging of a miniature given that ivory is a hygroscopic material and atmospheric changes can thus result in warping and twisting.
When the ivory twists out of shape it is counter-productive to attempt to return it to its original form given that its composition involves an inorganic, rigid part called cementum that may crack or split if forced back into shape. In numerous cases the twisting and splits in the miniatures resulted from tensions produced by frames or support systems that prevented the natural movement of the ivory or from the fact that the works were originally stuck down onto thick sheets of card.
Another intrinsic problem is the sulphuration of the silver leaf, which is a corrosion that changes the colour of the metal, making it irregularly darker. These darkened spots show through in the flesh tones and produce shadows that disfigure the face. In the present day when this problem is encountered with miniatures that retain their original sheet of silver leaf, the sheet is not removed as restorers apply the criterion of conserving the work intact and hence the materials from which it is made. In addition, no miniature in the Prado's collection had darkened patches that were so serious as to disfigure the fact and thus raise the issue of removing the metal sheet. However, old restoration procedures did involve removing these sheets and in order to do so the pieces of paper and other types of protective backing were also removed. In such cases the modern response has been to insert a loose sheet of aluminium, as this is a more stable metal and produces the same optical effect as the old silver sheets. Juan Pérez de Villamayor's Portrait of a Lady with Flowers in her Hair and a green Dress (O-699) has a silver-plated copper sheet in which the edges had become sulphurated. As the metal sheet was loose it had moved and the darkened edge was now located behind the sitter’s forehead, producing a dark patch that disfigured the image. In order to correct this it was only necessary to move the sheet back to its original position.
Miniatures were also encountered in which the silver leaf had been completely stuck down with a gum adhesive or protein-based glue that resulted in more significant damage as the silver in contact with the glue oxidised far more and resulted in larger dark areas on the face.
Among inappropriate, earlier restoration methods the most significant was the incorrect application of adhesives both to stick down pieces of paper or card on the back and to support the miniature in the frame. We encountered two types of adhesive that were analysed in the laboratory7. The first was a stable, transparent one that was found both in original joins and in subsequent interventions and which was identified as a gum. The second was an animal-type glue that was only found in later restorations. It is grey-brown in colour and in some cases has left patches on the ivory. In cases when these adhesives derive from subsequent intervention and in general whenever possible, they were removed as they can result in tensions and stains.
In addition, when the adhesive were applied in these old restorations the paint layer was sometimes stained, particularly at the edges.
The most serious damage of this type was found in Portrait of a Lady with Flowers in her Hair and a green Dress (O-699), as this work may have suffered an accident in which adhesive spilled onto the area of the dress. Fortunately, we were able to remove this patch of animal glue by mechanical means.
With regard to the paint layer, we found few cases of old repainting and these were only located in thickly painted background areas, possibly because repainting intended to cover over damage would be too thick and opaque over flesh tones and transparencies. Interestingly, we found miniatures that had been repainted from the back: using the transparency of the ivory some elements were reinforced in this way, for example, Portrait of an Officer in uniform Jacket (O-717) and Venus and a Satyr (O-680), in which the officer's face and the flesh tones of the bodies had been strengthened respectively.
In other cases varnishes or adhesives had been applied to intensify or 'fix' the colours, as in Portrait of Isabel II (O-698). Unfortunately, the application of this project resulted in craquelure and losses to the paint surface.
Another problem was that of abrasions to the paint. The most serious cases were found in miniatures that had lost the glass of the frame, such as Akbar the Great (O-745) in which almost all the polychromy had been lost due to abrasions. In other cases, such as Portrait of a Knight of the Order of Calatrava (O-703), the abrasions were the result of a new sheet of glass that was flat and had sharp edges.
Even original pieces of glass with curved edges can damage the paint surface and produce shiny areas and scratches if the glass has swollen or moved.
Finally, damage was encountered resulting from poor handling. Given that miniatures are extremely small, delicate items the way in which they are treated can leave permanent marks. The most habitual type of damage is that of fingerprints on the paint: sweat from hands is damp and greasy and varies in acidity depending on each person. These factors affect the paint layer and while deterioration is not immediately evident, dark fingerprints become visible with the passing of time dark.
Sneezing also leaves its mark on the paint surface. Another problem is the practice of cutting down the ivory support, which was generally done to eliminate damaged areas, however, in the case of Portrait of a Man with long Hair and Moustache the miniature was cut down with scissors to adapt it to the frame.
When deciding to restore the Museum's collection of miniatures, the aim was that of conservation and for this reason no repainting or replacement of materials was carried out. Instead, the restoration process focused on stabilising the works through the elimination of products and materials that affected or destabilised their state of conservation. In most cases this related to adhesives, pieces of card or frames that had produced deterioration or tensions.
The restoration of the Portrait of a Man (O-813) offers an example of the excessive presence of adhesives and poor handling. The work had dark patches on the face and adhesive tape stuck on the paint surface and on the back.
When removed from the frame it was observed that it had two pieces of yellowed paper stuck on with old, brittle glue. These pieces of paper were removed, revealing behind them the silver leaf crudely stuck to the ivory. The silver had corroded due to contact with the adhesive and was brittle and dark, resulting in the dark areas visible on the face.
The silver leaf was removed and traces of adhesive removed from the back of it, at which point the face recovered its original flesh tones. In the case of the mark left by adhesive tape on the paint, the remains of the adhesive were removed mechanically with a binocular lens in order to clean up the paint layer but this zone lost some of its precision.
In order to recover the effect of the sheen of the silver leaf a piece of aluminium film was inserted as this is more stable and has the same metallic sheen. It was attached to an acid-free piece of card that was used as a backing when the miniature was replaced in its frame.
A comparable case was that of Portrait of a Man (O-799) by Antoinette Brunet. The sitter's face had become dull and yellowish.
As in the previous case, it was found that the silver leaf and pieces of paper had been stuck to the ivory. They were removed and the adhesives cleaned off but in this case they left the ivory with a yellowish tone in the area of the face, as a result of which the sitter still had a sickly appearance.
Removing the stain on the inside of the ivory would have required an aggressive intervention and it was therefore decided to make use of its transparency. A piece of tinted red card was inserted as a backing that counteracted the yellowish tone. As a result, the sitter recovered his healthy, pink appearance with a minimum degree of intervention.
Another interesting procedure, in this case an old one, had been used on Portrait of a Lady in black (O-690). The portrait was extremely delicate but the dress and background were very thickly painted to the extent that the paint layer had craquelure. It was also curious that this miniature had silver leaf in the zone of the bust when this was covered over by a black shawl. An x-ray of the work and subsequent analysis revealed that this miniature had been 'dressed in mourning' as the sitter originally wore a lighter dress, but a family death or the fact that she had been widowed had resulted in mourning dress being painted onto her, concealing the original paint with this thick black garment.
The restoration of the Prado's miniatures collection and the study of their state of preservation have revealed the stability of these works when they are kept in stable environments and have not been treated with inappropriate materials, given that the damages found were generally due to old manipulation or restoration that altered the state of the materials.
Restoration criteria for works of this type should aim to preserve the original technique and methods, involving the minimum degree of intervention necessary with the aim of recuperating and preserving the original material. For this reason older treatments such as sanding down to remove stains, whitening the ivory or filling in paint losses were considered inappropriate as procedures of this type on objects of such delicacy conceals the works themselves, making it difficult to see the image and preventing the recovery of the original painting.
During the restoration of the collection a specific study was undertaken on each work's state of preservation. While the techniques were comparable, each miniature was a separate case with specific issues that required a precise and particular approach to restoration with the aim of avoiding both invasive treatments and those that could result in new conservation problems in the long term.
Elena Arias Riera. Restorer. Museo Nacional del Prado