Interaction of the blue pigment smalt with oil medium: The blue cobalt-glass pigment smalt has been observed to produce discolouration in paint films probably resulting from excess alkali in the glass of which it is made interacting with oil medium (note 5). Discolouration may vary from a dull greenish grey to brownish yellow and depending on whether and to what extent lead white is also present. The low refractive index of the pigment (c.1.46 – 1.55 (note 6)) is close to that of a dried oil film, and with the rise in refractive index of the oil film which usually takes place with ageing, the blue colour of the pigment, initially weak, may be all but suppressed by the discolouration of the medium. Although one of the effects is of increased translucency of the film, coupled with loss of blue colour and brownish discolouration of the medium, another type of deterioration can sometimes occur. An electron microscope study of a discoloured smalt paint revealed that the surface of the blue glassy pigment particles had in this instance become roughened and pitted from interaction with the medium (note 7). When this happens, the blue paint film is likely to become whitish and more opaque, i.e. blanched. Smalt is rarely found in easel paintings before the latter half of the sixteenth century, but becomes quite common in the seventeenth century all over Europe, though particularly in those countries farthest from Italy, like Spain and the Netherlands. That is not surprising for it was generally used as a substitute for ultramarine which, as the dry powder pigment, it closely resembles in hue, and lapis lazuli ultramarine would be likely to get scarcer and more expensive the greater the distance from Venice, its port of entry from the East. It has been noted that Claude, who for the most part worked in Italy, used a good deal of ultramarine in his pictures, particularly in skies and blue drapery, but recent examination of the National Gallery Claudes has revealed the presence of smalt in a number of samples, particularly from greens of foliage (and in one instance in the sky, of No.5, ‘A Seaport’). If used in combination with yellow pigments to make greens smalt may induce brown discolouration not dissimilar in appearance, and sometimes confused with, that which occurs in copper 'resinate' greens. Smalt can also have been introduced into ultramarine, either to act as a drier, since ultramarine itself dries slowly in oil when unmixed with lead white, or as an adulterant. It may be of interest here to mention that in samples of paint from Claude's pictures there were often seen in various layers and mixtures not only the recognizably blue glassy particles of smalt, but some apparently colourless glass particles. When interaction of smalt with oil medium occurs some cobalt combines with the fatty acids of the oil to form cobalt salts which act as drying agents. It follows therefore that some cobalt must be leached out of the pigment particles and as the concentration of cobalt in them is quite low (about 2 – 18% of cobalt oxide by weight) such loss might result in a serious reduction, or even disappearance of the blue colour which is due to the presence of cobalt. (An alternative explanation is that the colourless glassy particles are ground white glass, presumably lead glass, added as a drying agent as is sometimes recommended.) The conclusion to be drawn is that in any of the circumstances mentioned above smalt might provoke some discolouration and/or blanching of the paint film.
It can easily be appreciated that in one and the same picture, and even in the same area within that picture more than one combination of the causes and effects described above, involving either medium or pigment or both, might be operating (note 8). A full investigation is likely to be both difficult and time-consuming. In addition to the facilities already available in the Department for examination and analysis it has been proposed that a study of the topography of the surfaces of the blanched areas of Claude's paintings might provide some interesting and useful information since some relevant features may be beyond the resolving power of the optical microscope.