Vermilion (red mercuric sulphide, HgS) makes a single dramatic appearance as the pigment of Ariadne's scarlet sash and can be seen in Plate 6a (p.47) in a cross-section of a sample taken from where the scarf goes over her shoulder. The very thick layer of finely-ground vermilion is intensified in colour by a thin final layer of larger, darker red particles of the same pigment.
Crimson-coloured lake pigments (organic dyestuffs precipitated onto an inert, insoluble, inorganic substrate such as aluminium hydroxide or chalk) were noted, particularly in Bacchus's flying cloak and the drapery of the small faun in the foreground, in both of which they occur mixed with lead white in the light and middle tones and thickly applied as a dark crimson glaze in the shadows, the colour of which is remarkably well-preserved (see Plate 6b).
Crimson-coloured lake pigment also provides the colour in Ariadne's flesh (Plate 6a) and is a component of the pale mauve drapery of the Bacchante with the tambourine (Plate 3, p.29 and Plate 6d, p.47). It would also seem (Plate 6c) that the dark blue drapery of the Bacchante with the cymbals was initially painted in crimson or rose pink. It is a matter for regret that at the time the picture was under treatment and it was possible to take samples, the methods which have since been developed in the Scientific Department for the identification of the dyestuffs of red lake pigments (note 7) were not available.
Yellows and orange Lead-tin yellow (lead-tin oxide, Pb2SnO4) was identified as the pigment used for the pale yellow drapery beneath the urn near the bottom left corner of the picture. X-ray diffraction powder patterns of samples (note 8) confirmed that it was Type I (formula given above) of two types described by H. Kühn (note 9). The pigment seems not to occur in paintings much later than the mid-eighteenth century, and some confirmation of this is given by a remark of Sir Thomas Lawrence reported by Farington (note 10). On inspecting the picture in 1807 Lawrence praised the yellow drapery and added: 'The yellow colour used is such as we have not. Naples yellow wd. be weak to it'. A colour photograph of a cleaning test on the yellow drapery made in 1967 (Plate 2, p.29) shows how effectively the old varnish obliterated this distinctive shade of yellow. In fact in the seventeenth century copy from the Duke of Northumberland's collection the same drapery was painted in a rather dirty white, presumably because the copyist must have assumed that there was white drapery beneath discoloured varnish covering the picture at that time, a curious reversal of the usual state of affairs that a discoloured varnish makes white paint look yellow.
Orpiment (yellow arsenic sulphide, As2S3) and Realgar (orange arsenic sulphide, As2S2). These pigments occtir together in the orange drapery of the Bacchante with the cymbals, realgar in the main body of the drapery, orpiment in the lights. The pigments share with ultramarine the drawback of drying slowly when used in oil medium. During the formation of shallow drying cracks, the coarsely-ground pigment particles seem to have bunched up into small islands, producing a rather rough paint surface. Both pigments were recognized microscopically from particle characteristics and optical properties, the orpiment being notable for the laminated form and lustrous surface of the large crystalline particles. Confirmation of the presence of arsenic and sulphide was by the usual methods of chemical microscopy, and more recently arsenic was confirmed (in a small sample of realgar saved from 1967) by A. Roy by means of laser microspectrography.
Orpiment and realgar are rather rarely-occurring pigments in European easel paintings, though common in Far Eastern art. As well as the natural minerals, artificial forms, made by subliming together arsenic and sulphur, have been known since classical antiquity, probably first prepared as an alchemical exercise. Cennino Cennini, in fact, mentions only the synthetic variety (note 11).
Sixteenth-century Venetian paintings have provided the majority of occurrences noted by us. Realgar, with orpiment in the lights, was identified as the orange of S.Joseph's cloak in an early picture by Titian in the National Gallery, ‘The Holy Family and a Shepherd’ (No.4). It seems worth making the observation that Titian in the earlier part of his career, and Giorgione, seem to have had a predilection for the sort of orange drapery highlighted with yellow which is so striking a feature of the ‘Bacchus and Ariadne’. A similar colour combination occurs in the National Gallery's small panel by Giorgione, ‘The Adoration of the Magi’ (No.1160), both in S. Joseph's cloak and in the spectacular orange and yellow doublet and hose of the young man on the right leaning on a staff. The voluminous orange and yellow cloak of the grey-bearded 'philosopher' dominates Giorgione's ‘Three Philosophers’ (in Vienna), and though no record of identification exists, the appearance of the paint surface suggests orpiment and realgar. On the other hand, it seems unlikely that the two pigments, of rather doubtful durability in fresco, could have been used in Titian's frescoes in the Scuola del Santo at Padua in which the same orange and yellow colours run like a theme through the whole series.
Some remarks on Titian as a colourist will be deferred until the layer structure has been described.
Preliminary solubility and combustion tests on one or two paint samples, combined with the limited range of biological staining tests then in use indicated that the medium had the characteristics of drying oil. There was nothing in the results to suggest the presence of natural resins, and certainly not of natural resins soluble in the range of organic solvents used for picture cleaning. In 1967 the system evolved in the Scientific Department by J. S. Mills and first published in 1966 (note 12) for the identification of painting media by means of gas-chromatography was just beginning to be applied to samples of paint from actual pictures. One sample from the ‘Bacchus and Ariadne’,of white paint, was analysed by this method and the ratio of fatty acids present showed that the medium was essentially a drying oil, and specifically linseed oil. It has to be admitted that as yet no analytical method is available for the detection of minor components, particularly natural resins, in the oil, nor does the method used throw any light on its initial preparation, whether for example it was boiled or sun-thickened. Nevertheless, it was a considerable step forward at the time to have ascertained that the medium of the ‘Bacchus and Ariadne’ was essentially linseed oil. Since then the media of several samples of three other major works by Titian have been analysed by gas-chromatography, one turning out to be linseed, the other two walnut oil ((note 13) and see p. 58 of the 1977 ‘Bulletin’). Over the centuries the idea has been current that the special quality of Venetian painting in general and of Titian's painting in particular resided in the use of some special and secret medium. In the eighteenth century a variety of recipes and nostrums changed hands, sometimes for large sums of money, which purported to be the 'Venetian Secret', including such unlikely candidates as pastel colours (note 14).
The situation was further confused from Titian's own time right up to the present century by attempts to copy paintings by Titian and other sixteenth century Masters, often, of course, from originals half-obscured by discoloured varnishes. At the beginning of this century, Max Doerner, the authority at that time on techniques of the Old Masters, discovered that he could make plausible copies of old pictures, including those of Titian, Tintoretto and Rembrandt, using an egg and oil emulsion for the impasto whites and an oil medium with a high proportion of a fairly easily- soluble resin such as mastic (or even an entirely resinous medium) for the translucent darks and glazes (note 15). He therefore concluded that because he could copy accurately the surface appearance of, say, a picture by Titian using a particular range of materials and techniques, the original artist must also have used the asme materials and techniques. The fallacy of this argument is demonstrated daily in the National Gallery's and other conservation departments where paint losses are inpainted to match the surrounding original paint using modern pigments and synthetic resin media. The fact that the translucency which oil paint develops with great age can sometimes by achieved at once by the use of a medium with a high proportion of natural resins or balsams led to the belief that the media of the originals, and particularly of the glazes, must contain a high proportion of 'soft' or easily-soluble resinous material and hence be vulnerable to solvents used in varnish removal. The argument was very persuasive. The results of the gas- chromatographic analyses cited above constitute the first objective and non-speculative evidence of the composition of Titian's painting medium.
The layer structure
The layer structure provides some information about the painting procedure of the artist and also about alterations made in the course of painting by the artist himself (‘pentimenti’) or by a later hand. X-radiographs are useful for gaining an overall concept of the layer structure; the microscopical examination of cross-sections of paint samples for gaining precise information about very small specific areas. The two are complementary. A composite photograph of the X-radiographs is shown in Fig.2. Plate 6 (p.47) shows photomicrographs of some of the paint cross-sections of samples from the ‘Bacchus and Ariadne’ photographed at c.150 x magnification, with, on the facing page, captions giving details of layers present in each, so that it is unnecessary to repeat all that information here. The layer structure of the ‘Bacchus and Ariadne’ sections is more complex and at the same time less regular than that which has been observed in samples from fifteenth century paintings, as for example some by Crivelli and Giovanni Bellini. Also, the layer structure of the Titian ‘Bacchus and Ariadne’ samples cannot be interpreted in terms of the orderly and systematic technique of earlier painters in which underdrawing, underpainting, main paint layer and final glaze or scumble follow each other logically and may be read off, as it were, from the paint cross-section. From the information derived from the paint cross-sections and detailed in the captions on p.46, it is possible to deduce from the cross-sections (a–f) respectively that:
(a) Ariadne's scarlet scarf is painted over the flesh of her shoulder and in turn the flesh of her shoulder over the blue paint of the sea.
(b) The (azurite) blue paint of the distant landscape, together with its lead white underpaint, goes just over the edge of the crimson-glazed shadow of Bacchus's cloak near his right arm (a section of a second sample, not illustrated, of the main area of the crimson cloak showed a pink opaque underpaint, followed by crimson-coloured glaze, painted directly on the gesso ground).
(c) The dark blue drapery of the Bacchante with the cymbals was initially coloured crimson or deep rose- pink, but was subsequently obliterated with a layer of lead white and then repainted in several layers of deep blue. The crimson colour can, in fact, be seen at the bottom of some of the cracks in the blue drapery on the picture itself.
(d) No changes have occurred in the mauve drapery of the figure with the tambourine, but it can be seen that there is a thin brown layer (of undermodelling) between the mauve paint and the gesso ground.
(e) The sequence of green and blue layers representing foliage on landscape has beneath it two yellowish opaque layers which might be from a ‘pentimento’ of Bacchus's golden chariot.
(f) The dark brown (once green) feathery foliage has been painted over blue sky, but beneath that pair of paint layers is another almost identical sequence of green, not-discoloured, glaze over blue sky. The artist must have repainted sky and foliage in this area. Therefore of the six samples of which the paint cross-sections are illustrated in Plate 6a-f inclusive, five might be termed ‘pentimenti’ either in the sense of deliberate alterations (c and f), or as one feature of design or area of colour overlapping another (a and b). The occurrence of even one of these changes in an earlier and more traditional picture would excite comment. The evidence from the paint cross-sections does seem to point to Titian having worked out the composition of the picture as he proceeded with the painting, using either the minimum of underdrawing of the design or even none at all.
At this stage it is useful to consult the X-radiographs (Fig.2), although at first sight they seem to present a rather confused image on account of the numerous small losses and puttyings. The gesso ground, containing only the light element calcium, would contribute virtually nothing to the impedence of X-rays, particularly by contrast with any paint layer containing lead. Now the blue paint of both sea and sky contains a high proportion of lead white, as does the pale grey and fawn of the sandy beach, hence these background areas look white in X-radiographs. Since the total absorption of X-rays in a particular area is a sum of the absorption of each of the layers, it should easily be possible to see which features of the design have been painted directly on the gesso ground and which on top of the blue sky. Clearly some feature which is painted directly on the gesso ground is more likely to have been conceived early on in the production of the painting than one applied to existing paint. This principle having been accepted, an interpretation of the X-radiographs in the light of the paint cross-sections can be made:
Bacchus's cloak, particularly in the deep crimson-glazed part streaming out to the right, appears in the X-radiograph as a dark silhouette against the light background of the lead-containing paint of sea and sky. This shows that it was painted directly on the gesso ground and not over the blue paint of sea or sky and agrees with the paint cross-section in Plate 6b, which shows a little of the blue-green paint of the sea over the crimson glaze of Bacchus's cloak. Around the figure of Bacchus, which appears beautifully modelled in the radiograph in carefully modulated thicknesses of lead white, there is a strong black contour, though discontinuous in parts, as if the surrounding sky, land- and seascapes had been painted not quite up to the edges of the flesh.
By contrast, the figure of Ariadne appears in the radiograph, apart from her legs and feet, as a swirling mass of lead white, denser than anything else in the picture. Within the area all colours visible on the picture's surface, flesh, blue and white drapery and scarlet sash, seem equally opaque to X-rays. The contours of the cloak and sash or the outline of the hand can be discerned only by the direction of the brush strokes. Ariadne's blue cloak, so similar in appearance to the deep blue drapery of the Bacchante with the cymbals, looks disproportionately denser in the radiograph. Part of this thickness of lead-containing pigment could be accounted for if the figure of Ariadne had been painted on top of the blue sea and sky, but would not fully explain the dense, shapeless, yet disturbed nature of the mass. A possible explanation is that Titian, having at any rate positioned the figure of Bacchus in the early stages of painting had some difficulty in getting the right pose for Ariadne in relation to him and had to resort to repainting her. It also looks from the radiograph that at least some of the figures in Bacchus's train were laid in on the gesso ground at an early stage in the composition, as was his chariot, the cheetahs which draw it, and the little dog, since these last three all appear in the radiographs as black silhouettes surrounded by the light opaque paint of the shore and landscape. The enchanting little faun in the centre foreground must belong to a later stage of painting for even on the picture itself a ‘pentimento’ of the chariot wheel is visible under his leg. On the upper right the trunks of the tall trees can be seen to have been laid in on the gesso ground before the sky was painted, but the foliage applied on top of the blue paint of the sky, as in the cross-section of browned foliage seen in Plate 6f. Titian then added a few patches of blue paint to indicate blue sky seen through the trunks of the trees, and these show up strongly in the radiographs. It might therefore be tentatively proposed that the artist had from the very beginning a fairly clear idea of how he wanted to dispose most of the group on the right, and seems, from the radiograph to have painted Bacchus with an unhesitating brush. He may have delayed painting the figure of Ariadne until later, or else been dissatisfied with his first attempts and done some extensive repainting of it before arriving at the present happy solution.
With regard to documentary sources on Titian's technique, the only nearly-contemporary one is a description given by his pupil Palma Giovane (1544 – 1628) to Marco Boschini (note 16), the historian and biographer of Venetian art and artists, some years after Titian's death. His account of how Titian laid in the composition with areas of solid colour, and of his ruthless alteration and revision of his half-finished pictures shows fair agreement with what has been observed in the ‘Bacchus and Ariadne’ and other of Titian's pictures we have examined. His report of Titian rubbing glazes and scumbles onto the surface of his pictures with the finger rather than the brush seems more in accordance with the very late works, such as the National Gallery ‘Madonna and Child’ (No.3948), or the ‘Pieta’ (Accademia Galleries, Venice) which Palma Giovane in fact finished after Titian's death. It is these late works which would probably have been fresh in his mind in his reminiscences to Boschini. In any case, ‘Bacchus and Ariadne’ was painted twenty years before Palma's birth.
Titian as a colourist
The reputation of Venetian sixteenth century painting for excellence of colouring rests firmly on Titian. In the mid-sixteenth century there were rival advocates of Florentine art, as personified by Michelangelo and his mastery of drawing, and those of Venetian art, as personified by Titian and his mastery of colouring, ‘disegno’ versus ‘colore’. This led eventually to the idea that perfection in art would be achieved by a combination of both excellences, as advocated for example by Paolo Pino in his ‘Dialogues of Painting’ of 1548 (note 17). It seemed worthwhile, as a conclusion, or rather, perhaps, corollary to this article, to attempt to assess, from the standpoint of materials and techniques, and in the light of the results of examination of the ‘Bacchus and Ariadne’, what physical factors contributed to Titian's enormous reputation as a colourist. It should be pointed out that in order to attempt this there is no need to survey the entire ‘oeuvre’ of his very long life. By the time Pino was writing in 1548 it is clear that Titian's supremacy as a colourist was already well-established and indeed he may have gained a reputation as such as early as 1518 with the completion of the great altarpiece of the ‘Assumption of the Virgin’ in the church of Santa Maria Gloriosa dei Frari in Venice, which since its recent cleaning looks at least as colourful as ‘Bacchus and Ariadne’. In any case, in some of his late works Titian is no longer interested in the juxtaposition of large areas of pure colour, but more in the flickering effect of light over surfaces, as in the National Gallery's ‘Death of Actaeon’ (No.6420). Since the ‘Bacchus and Ariadne’ seems always to have been admired for its colouring, it seems fair that it should serve as an exemplar.
The principal physical characteristics which contribute to the 'colourfulness' of the ‘Bacchus and Ariadne’ seem to be (not necessarily all in order of importance):
1. The inclusion ‘within a single picture’ of all the most colourful pigments available (comprising, as we have seen, some rather uncommon ones like malachite and realgar).
2. The choice of the finest quality of the most colourful pigments (the ultramarine and vermilion are of outstanding purity and richness of colour).
3. The employment of each of these pigments separately in sizeable areas of colour, whether sky, drapery or foliage), well-spaced within the picture area (N.B. for this purpose orpiment and realgar may be regarded as a pair, one being used for lights the other for shadow within the same area).
4. The juxtaposition of areas of strong contrasting colours.
5. The use of each of these colourful pigments at ‘full strength’, i.e. unmixed with another pigment, or mixed only with white to produce a brighter colour of the same hue. In the colourful areas of sky, sea, landscape and drapery physical mixtures of pigments, other than with white, seem to have been avoided, presumably because all such mixtures tend towards grey. (The only example of a 'colourful' area which has been produced by mixtures containing two different coloured pigments in the same paint layer is the mauve of the drapery of the Bacchante with the tambourine.)
6. The intensification of the colour in some areas by means of glazes; an example is the final layer of large deep blue particles of ultramarine on top of the blue body colour of the drapery of the Bacchante with the cymbals. The use of a dark, transparent glaze over a lighter underlayer produces an increase in saturation (i.e. apparent depth of colour) without as great a loss of brightness as would occur if the two pigments were mixed instead of being used in separate layers. The use of a green 'copper resinate' glaze over green or yellow underpaint is a further example, but the intended effect has in some areas been destroyed by the browning of the 'copper resinate'.
This is not, of course, the whole story, for what was most admired in Titian's paintings was his skill in unifying such strong contrasting colours to produce 'harmony'. The meaning of the term as applied to art is less precise than in music, but we could do worse than settle for: 'agreeable effect of apt arrangement of parts' (O.E.D.) getting close to the idea of tone.
In fact, other painters may employ the same formula but with not quite the same results. It so happens that the identical range of pigments occurs in Sebastiano del Piombo's ‘The Raising of Lazarus’ (No.1), and there is even repeated a touch of the lilac colour which appears in the drapery of the Bacchante with the tambourine in the ‘Bacchus and Ariadne’. Although Sebastiano began his career in Venice, he settled in Rome in 1511 and this picture was painted some years later, although he still signs himself on it as 'SEBASTIANUS VENETUS'. Cecil Gould made the point (on the occasion when the colours of ‘that’ particular picture were revealed after cleaning) that 'The colours are not tonally related, as in Venetian painting [...]' (note 18). In addition the many small bright areas of contrasting colour in the middle distance give a rather jumbled effect. The Florentine Mannerists sometimes employ a not very different range of colours, but used in a rather arbitrary or 'non-naturalistic' way, if one dare use such an equivocal term. In addition, many of their pictures of this period, for example those of Bronzino who is an almost exact contemporary of Titian, are painted on very smoothly prepared wood panels and have an enamel-like 'unpainterly' surface. Although the ‘Bacchus and Ariadne’ is a comparatively thinly-painted picture, Titian has exploited to the maximum textural effects (they may have been reduced in the course of the many past lining operations) which help to give variety to the areas of pure colour. An example is the pale lead-tin yellow drapery beneath the urn in the bottom left corner (Plate 2, p.29) where almost all the modelling of the folds is in the form of the brushstrokes, rather than painted shadows.
The large separate areas of pure colour in the ‘Bacchus and Ariadne’ are also skilfully knit together by the quieter earth-colours of the foreground landscape. In this connection it might be remarked that the picture would probably seem much less pleasing or 'harmonized' were it not for the comparatively good state of preservation of the green areas. The flesh tones also provide a link right across the picture and it is interesting to note how subtly they are varied. The very pale flesh paint of Ariadne is faintly tinged with a crimson-coloured lake pigment, with, as we have seen, ultramarine blue in the shadows, while the shadows of Bacchus's flesh contain green earth, and the flesh of the Bacchantes and satyrs is coloured with a variety of red and brown ochres. In general the flesh tone increases in warmth and depth of colour from left to right of the picture, culminating in the Laocöon figure with the snakes, but with the proviso that female figures are depicted lighter-skinned than males.
No amount of analysis, chemical or otherwise, can explain the particular magic of the picture, but the results of technical examination may have given some small glimpse of the mechanism of what goes on behind the scenes. If nothing else, it will have served to refute the statement (meant as complementary rather than the reverse!) by the eighteenth century historian and critic Antonio Zanetti that 'the pigments on Titian's palette were ordinary and few' (note 19).
Notes and references
1. Merrifield, M., ‘Original Treatises in the Arts of Painting’, Vol.11 (London 1849), p.821.