Joseph Mallord William Turner

This study on smooth white wove paper has necessarily been worked on dry paper, since the sheet is bound into a sketchbook. All the light and rapidly applied washes have a definite, slightly hard edge as a result.

The paint medium in this work was investigated. Several gums were available in Turner’s era: notably gum arabic and gum tragacanth. The former is today used alone in commercial watercolour paints. The latter could be a practical choice because it does not ‘wash up’. This nineteenth-century term means that a later wash in gum tragacanth, applied to dried paper, would not wet and re-activate an earlier wash as much as would happen when gum arabic alone were used. That makes it very useful for an artist who applies multiple washes, uses plain water for washing out, and generally works his paper extensively, all of which describe Turner’s working methods. A method of chemical analysis called gas chromatography was used, which separates out components of a complex mixture to allow later identification. This was combined with mass spectrometry for more precise identification of the separated components. The result was that the gum used here was a mixture: gum arabic as the largest component, with gum tragacanth, and quite significant amounts of sugar or honey added. Added sugar or honey give flexibility to the paint film, useful if the artist chooses to apply it thickly as Turner did often, though not in this case. This gum mixture was found in commercial watercolour blocks which Turner stuck down onto card (in fact it was the invitation to a Royal Academy dinner to mark the annual exhibition) and a piece of leather, to form a travelling watercolour palette, now in the Royal Academy of Arts, London. These blocks were of the earlier ‘hard’ type that did not include sugar, but this paint may have come from a ‘soft’ block, which did.

The blue used in the foreground is Prussian blue, while the redder pigment in the sky may be ultramarine. The other colours are so thinly applied that they cannot be identified with much certainty, though a red lake and yellow ochre are likely.