Composition

A photograph in Glasgow University shows that the cloak over the woman's shoulders was originally much darker, with a glossy surface like satin, and her back was silhouetted against a light area. 1 A photograph in the Lucas Collection, Baltimore, signed with a butterfly that can be dated about 1881, and inscribed by Whistler 'Arrangement in Grey & Yellow', shows the painting in a later state. A dark wash had been added down the right side of the picture and had dripped over the hem of the dress and down to the bottom of the picture.

From the present state of Arrangement in Yellow and Grey: Effie Deans it appears that Whistler worked on it after the Lucas photograph was taken, darkening some of the folds of the dress, and the shadow cast to the right, with very wet paint which has dripped and run in several places. He also lightened the cloak and added a flower pattern to it, although the head, hand and skirt do not appear to have been touched.

The photograph in Glasgow shows the painting before the signature and inscription were added. Walter Richard Sickert (1860-1942) told the Pennells (1911) that the inscription was added by someone other than Whistler, and Elizabeth Robins Pennell (1855-1936) was not sure that it was in Whistler's writing. However the Pennells also point out that Whistler made no objection to the inscription when he saw the painting in 1902. 2

Technique

It is painted very thinly, with the paint dripping from the long, broad brushstrokes. Erma Hermens and Arie Wallert, who conducted extensive technical research on the painting, note that Whistler regarded it as one of several 'studio experiments', in which he employed 'fluid and transparent paints … with extensive dripping':

'In an early photograph, a series of long narrow drips can be seen at the bottom left, dripping from the dark, shadowy background that surrounds the cloaked figure. A later photograph shows a lighter tone on the model’s bodice and conspicuous black drips on the right side of the figure. Apparently Whistler was darkening the background with extremely thin washes of paint, to enhance the contrast with the figure while retaining the soft outlines at the edges of the dress. The photographs show several areas, particularly in the folds of the dress, where the shadows have been darkened and the highlights accentuated.' 3

A summary of their research findings on the paint and medium employed follows:

'The drips of carbon black-containing paint mixtures seem to make furrows in underlying paint, … Staining tests on cross-sections suggested an oil medium, and the presence of regular linseed oil was confirmed ... A small addition of beeswax was also identified. The extreme fluidity must have been achieved by greatly diluting the paint mixture with volatile solvents such as oil of turpentine or spike oil. After evaporation of the solvent the paint surface would have had a matt velvety appearance, enhanced by the beeswax.

… [In] ultraviolet light, … the painting shows an array of fluorescent tones, indicating a repeated application of partially overlapping thin layers. …

Elemental analysis of the greyish ground layer confirmed the presence of strontium, barium, zinc, lead and small amounts of iron oxide and ivory black (phosphorus was identified in the black particles ...). Three paint layers follow: a dark grey black, then a thin blackish layer that caused some dripping, followed by an even thinner black wash which added more dripping especially on the right. On top there is a highly fluorescent transparent layer, probably an oiling out layer, which he applied before adding the signature in 1889. ... In most cases the paint stratigraphy consists of just two or three thin paint layers using a fairly limited palette: white, yellow ochres, red and brown earth pigments and black. [A] cross-section taken from a light area on the fold of the skirt show[s] the presence of lead, iron, barium, strontium, calcium and zinc.

Interestingly, most of the samples showed the presence of barium (Ba) ... Examination of pigment particles … gave a strong indication of the presence of natural mineral barites. ... The barium sulphate pigment consists of rather large particles and seems to be mixed with conventional lead white. … The particulate characteristics of this paint mixture allowed it to be thinned to a very large extent without becoming slick and smooth, and without losing its matt appearance.

... X-ray fluorescence analysis also indicated the recurrent presence of strontium (Sr), … this pigment could be identified as celestine (strontium sulphate, SrSO4). … [T]he paint … seems to be a mixture of large celestine crystals and smaller amounts of tiny barite particles. … The increased production of this luminescent and gritty pigment around the 1870s and 1880s, and the appearance of the almost ephemeral, hazy and extremely thinly ‘washed’ full-length portraits by Whistler at this time can hardly have been coincidental. … The fairly large celestine particles, however, would have given his paint mixtures a rather open texture, while its often slightly bluish colour would impart a subtle cool tonality to Whistler’s silvery greys used both in the ground layer and the light areas in the skirt of the Effie Deans portrait. …

… There may be another explanation for the use of the pigments barite and celestine in the paint. Whistler was not only meticulous when choosing his painting materials, but also in the decoration of his houses and exhibition spaces, which he personally coordinated in every detail. [For instance, using] a first layer of grey to tone down bright hues …

[In] the 1870s, John Bryson Orr ... founded a factory for the production of lithopone, a white paint containing zinc sulphide and barium sulphate, which he patented in 1875, [and he] developed another white paint by adding celestine to the lithopone mix; this Duresco paint was patented in 1884 but Orr started production in the late 1870s. As all the cross-sections of paint samples from the Effie Deans portrait contain small amounts of zinc, barium sulphate and mineral strontium sulphate, it is possible that Whistler used oil-based house paint from Orr containing this mixture, with an addition of some lead white tube paint.

Conclusions. The group of full-length portraits from the late 1870s introduce a period in which Whistler experimented with strongly diluted paints, applied in thin washes. It also seems he used unusual pigment combinations in his white paints either by mixing tube paints such as blanc fixe (lead white and barium sulphate) and zinc white (zinc oxide), and adding mineral strontium sulphate, as its characteristics suited the desired translucent effect. However, it is also possible that he used oil-based house paint such as the Duresco produced by Orr to which he added some lead white tube paint. ' 4

Conservation History

Unknown.

Frame

Portrait Whistler frame, dating from 1876/1879. 5