Titian's technique has had a fascination for artists and connoisseurs alike from his own contemporaries down to the present day and over the centuries has been the subject of much speculation. The treatment of the ‘Bacchus and Ariadne’ in 1967–69 provided a unique opportunity for collecting data on the materials and structure of the painting and attempting to interpret such data in terms of painting practice.
It would have been impossible to make any proper study of the original paint unless the thick discoloured varnish had been removed. In 1967 when the first cleaning tests were about to be made it had been our intention to take some samples for microscopical examination and chemical analysis in order to try to discover the condition and composition of the paint beneath the old varnish. An unexpected difficulty was encountered. Not only did the thickness of the old varnish make sampling difficult, but in most areas the adhesion of the varnish to the paint layers was found to be stronger than the adhesion of the latter to each other or to the gesso ground. In the few small areas where varnish had flaked off it had probably done so taking the top surface of the paint layer with it. As a result it was almost impossible to take suitable samples for the preparation of paint cross-sections with the varnish present, and it was decided to defer sampling until after the cleaning tests. Only one of the cross-sections of samples of which the colour photomicrographs are shown in Plate 6 (p.47), Sample (h), has the old varnish still ‘in situ’ before cleaning. The sample was from the sandy shore beneath Ariadne's feet, an area which before cleaning looked brown but is now light grey. In the section it can be seen that, above a trace of gesso ground and lead-white underpaint, the thin granular, grey paint which represents the sandy beach contains a few blue (azurite) particles, a deliberate addition by the artist to impart a cool tonality. The varnish layer above, turbid yellow-brown, is approximately four times as thick as the grey paint layer beneath it. It might be said that, although patchy and uneven, the discoloured varnish present on the ‘Bacchus and Ariadne’ before its most recent cleaning was on average thicker than any previously encountered on a National Gallery picture. Some idea may be gained from the fact that the varnish layer in the sample just described has an average thickness of about 80 µm, while a single sprayed coat of a modern varnish is approximately 10 µm thick. Solubility tests on samples of the old varnish indicated that it was likely to be mastic resin (applied in a volatile solvent like turpentine distillate). Analysis of a sample by gas-chromatography carried out by J. S. Mills detected the presence of a little drying oil, probably linseed. The composition would not be inconsistent with that of the 'Gallery Varnish' (described on p.36, Note 11), but fortunately the addition of linseed oil, although sufficient to make the varnish film rather tough, and possibly increase its yellowness, did not render it appreciably less soluble in the usual cleaning solvent.
In view of the picture's history of cleanings and re-varnishings it seems almost superfluous to state that the varnish removed had nothing to do with Titian. Whether or not the picture was varnished originally by Titian, or in his lifetime, is not a question which can be answered here, or indeed perhaps at all.
In turning to the results of examination of the picture itself, it may be convenient to deal with each component of the layer structure in turn:
The original support is a linen canvas of surprisingly fine weave for the size of the picture (Fig.7). The threads are fine and smooth, the plain (tabby) weave very regular, a thread count giving average values of 24 threads/cm for the warp and 23 threads/cm for the weft. Both canvas and textile fibre seem in good condition for their age. There is a vertical seam down the centre of the picture, inclining slightly to the right near the top, the selvedges neatly overcast with small stitches and fine thread. The total width of canvas required for 1.9m painted surface plus turnover allowance (now cut off) for the original stretcher would have been about; 2m, indicating a loom width for each vertical strip of canvas of approximately 1111. Measurements on a number of very large sixteenth century Venetian canvas paintings, having several widths of canvas seamed together, indicated that the most commonly occurring loom-width was about im (plus or minus about 4cm, accurate measurements being of little significance because of variations from picture to picture in degree of stretching). It is tempting to speculate on whether the loom-widths in which canvas was available might have influenced the size and shape of paintings. It is known, however, that Alfonso d'Este sent ready-stretched canvases to Titian in Venice for the purpose of painting the ‘Bacchanals’, so it might be expected that in their case the overriding factor would have been the size and shape of wallspace allotted by Alfonso in the decorative scheme for his ‘Studio’. It should come as no surprise that a Venetian painting of the early 1520s is on canvas. Painting 011 canvas first began to predominate over painting on wood panel in Venice in the early to mid-sixteenth century, particularly for large-scale works, probably because of the impracticability of fresco-painting on damp and salt-impregnated walls. A number of quite large paintings of earlier date than this by other Venetian painters, including Carpaccio and Gentile Bellini, are on canvas, as are some of Giorgione's, and of Titian's own works more are on canvas than on wood panel.
The preparation on the canvas is a thin layer of gesso (calcium sulphate, usually as a mixture of anhydrite and gypsum, in a medium of animal glue), so thin as barely to cover the canvas grain. Its thinness is perhaps fortunate from the point of view of conservation, for a thicker layer might have cracked and disintegrated even more disastrously, particularly if in the past the canvas had been rolled for ease of transport. In paint cross-sections in which a fragment of the gesso ground is included, the latter is translucent brown, not due to any addition of brown pigment, but to a high concentration of animal glue, browned with age and probably not all original but partly the result of past lining and paint-laying operations. Gesso grounds are more often associated with Early Italian wood panel paintings, but in Venice the use of the traditional gesso ground seems to have been taken over by painters working on canvas. A reference to the practice occurs in the Volpato Manuscript, dating from the late sixteenth or early seventeenth century and deriving from Bassano. The manuscript is in the form of a dialogue between two painter's apprentices and includes a discussion of the disadvantages of the 'old-fashioned' (i.e. sixteenth century) gesso ground on canvas as compared with the 'modern', consisting of red ochre in oil medium (note 1). One apprentice remarks that old pictures on canvas are better preserved if the gesso ground is very thin, as indeed it is in the ‘Bacchus and Ariadne’. William Buchanan, before the 1853 Committee of Enquiry, displayed his awareness of the ‘Bacchus and Ariadne’ having a gesso ground (see p.26 of this ‘Bulletin’) but the knowledge seems to have been lost in more recent times and replaced with a rather generally- accepted idea that Titian painted on red-brown grounds. In the past few years, however, we have been able to examine ten paintings by Titian of widely varying dates and all proved to have gesso grounds. A late work, the ‘Madonna and Child’, in the National Gallery (N0.3948) seemed at first sight to have a red-brown ground, but examination under a binocular magnifier revealed that the colour was that of the bare canvas threads in tiny spots where both paint and ground were worn away. In a comparatively thinly-painted picture like the ‘Bacchus and Ariadne’ the use of a white or light-coloured ground would impart and help retain luminosity.
The infra-red photographs (Fig.3) revealed practically nothing in the way of underdrawing. Some features in the infra-red photographs, such as the outlines of the hub and spokes of the wheel of Bacchus's chariot, which look like underdrawing, can be seen from comparison with the picture to be in black paint at or near the paint surface, though there does seem to be a trace of drawn outline of Ariadne's hand clutching her draperies. In only one of the paint cross-sections prepared (Plate 6c, p. 47) was there a scattering of black pigment particles which might be construed as drawing, not, as expected, on the gesso, but on top of a yellow ochre underpaint.
Many fourteenth and fifteenth century paintings with gesso or chalk grounds have preliminary under-drawings, sometimes detailed and highly finished, on the white ground. An underdrawing of this sort, provided it is done in black pigment, no matter whether with charcoal, brush or pen, can often be revealed by infra-red photography and sometimes also appears in paint cross-sections sandwiched between ground and paint layer. The Giovanni Bellini ‘Blood of the Redeemer’ (No.1233) described on pp. 11–24 of this ‘Bulletin’ is a good example of a picture of this sort with detailed underdrawing. The apparent absence of underdrawing on the gesso ground of the ‘Bacchus and Ariadne’ does seem to call for some attempt at explanation. In an early work by Titian in the National Gallery, the ‘Noli Me Tangere’ (No.270), the only drawing visible in the infra-red photographs (and that in the areas where drawing was most likely to be revealed, the figures of Christ and the Magdalen) was a rough outline of the Magdalen's left arm and hand and a zig-zag line just indicating a fold of the white sleeve. Many fourteenth and fifteenth century painters (and, of course, countless others since) have made preliminary drawings on paper for their easel paintings and it might be conjectured that such was Titian's practice, but not only do no studies by him for the ‘Bacchus and Ariadne’ survive, not more than about twenty drawings in all can be attributed with any certainty to him (note 2), and interestingly enough, drawings of Giorgione, the young Titian's contemporary and mentor, are equally rare. Most of the painters who preceded Giorgione and Titian, as well as those working in the early sixteenth century in a less 'avantgarde' style, carried on in the traditional way, starting with sketches and working up to finished drawings or even cartoons on paper, before proceeding to a detailed drawing and perhaps monochrome under-modelling on the gesso ground of the painting itself.
The actual painting process, i.e. the application of areas of local colour, was almost last in a long series of steps in the production of the finished picture. It is rather anticipating discussion of the paint layer to mention here X-radiography of paintings, but whereas infra-red photography can be used to show up underdrawings just below the paint layer, X-radiographs will show up features in the paint layers themselves. X-radiographs of paintings by Titian frequently reveal alterations made by the artist himself in the course of painting. These are not merely minor corrections, such as the slight shifting of a profile of a face, but radical changes in composition like those which have been discovered in two early works by Titian in the National Gallery, the ‘Portrait of a Lady’ (No.5385) and the ‘Noli Me Tangere’ (No.270) (note 3), and, incidentally, in Giorgione's best-known work, ‘The Tempest’ (note 4) (Accademia Galleries, Venice). The ‘Bacchus and Ariadne’ shows fewer changes in the radiograph, though Titian seems to have had a struggle with positioning Ariadne. From evidence of the X-radiographs, however, the tentative hypothesis might be advanced that Titian (and perhaps Giorgione) preferred to dispense with preliminary drawings and to work out the composition of their pictures in paint. The introduction of a new subject as in ‘The Tempest’ or the rediscovery of the world of classical mythology of ‘Bacchus and Ariadne’ would require a good deal of working out of the composition and be likely to result in ‘pentimenti’. The danger of generalizing is illustrated by the fact that when Titian's ‘Vendramin Family’ (No.4452), a work about twenty years later than the ‘Bacchus and Ariadne’, was cleaned fairly recently it was found to have a mass of black underdrawing on the gesso ground, very sketchily done, some of the drawing of the robes just hasty scribbles, and not always corresponding with the final painted image. In this picture, though, instead of having the freedom to develop a new and exotic subject, Titian was as tightly constrained as any early painter of saints and donors in that he had to fit a large number of figures into a limited space in strict order of rank, age and size, so that a preliminary sketch was a real necessity. Vasari commented that if Titian had added the art of drawing to his mastery of colour he would have been an even greater artist (note 5).
Venetian sixteenth century painting has always been famous for its colour. Venice was in an enviable position with regard to supplies of pigments, being not only the port through which exotic materials like lapis lazuli entered Europe from the East, but also a centre of the textile and dyeing industry and famous for its lake pigments which were a byproduct of the latter. Titian, as a successful painter with wealthy clients, would have had first pick of all the materials from the point of view both of variety and quality. Other than such basic requirements of the artist's palette as lead white, blacks and yellow, red and brown earth pigments, the following were identified in the main colour areas of the picture: