Detail from The Canal, Amsterdam, 1889, James McNeill Whistler, The Hunterian, University of Glasgow

Home > Catalogue > Browse > Pink and Grey: Three Figures <<   >>

Pink and Grey: Three Figures


The White Symphony: Three Girls, Freer Gallery of Art
The White Symphony: Three Girls, Freer Gallery of Art
Pink and Grey: Three Figures, Tate Britain
Pink and Grey: Three Figures, Tate Britain

This is, according to Whistler, a copy of The Three Girls [YMSM 088]. The artist stated, 'three Girls ... is a rough copy made from the original when that was given to the Creditors.' 1

There are many pentimenti, particularly around the heads and arms of the figures, which suggest that it was not a literal copy.

Girl with Cherry Blossom, Private Collection
Girl with Cherry Blossom, Private Collection

The Three Girls [YMSM 088] was destroyed but it is believed that a section containing the central figure was saved, which is known as Girl with Cherry Blossom [YMSM 090]. This is similar to the central figure in Pink and Grey: Three Figures, in that it is very lightly draped, whereas in an oil sketch for the composition, The White Symphony: Three Girls [YMSM 087], the model was draped in white. The figure on the left in Pink and Grey: Three Figures shows signs of having been scraped down and altered, and may also have been draped, in the original composition.


Pink and Grey: Three Figures, Tate Britain
Pink and Grey: Three Figures, Tate Britain

A survey of Pink and Grey: Three Figures by Professor Joyce H. Townsend at Tate Britain follows:

'Whistler’s very coarse and loosely-woven canvas has about 8 threads per centimetre, and it has a grey imprimatura, darker than a mid tone, very effectively covering an earlier bright greenish-blue imprimatura made from lead white, Prussian blue, a pre-mixed green (Prussian blue and cadmium yellow) and orange ochre – presumably by Whistler. He had regularly used Prussian blue for blue backgrounds in this decade. The grey imprimatura is made from lead white and bone black toned with sienna, and both are oil-based. This was not necessarily a fabric intended for artists’ use, and it does not have a conventional off-white artists’ priming beneath these layers.

Along the top edge and the right end of the lower edge, as well as on parts of the left edge, there are neatly-painted grey tick marks, evenly spaced by 7 cm (2 ½") at the figures and by 13 cm (5") elsewhere. The canvas is not actually squared up, as the tick marks are not joined up with painted or drawn lines. These are likely to be Whistler’s own marks, used to align the figures as he copied an earlier subject, possibly from memory. Being American by birth, he is highly likely to have used measurements in inches despite his youthful training in France. There are a few pencil marks near the neck of the left figure, which suggest light pencil sketching on the grey imprimatura, before painting. The paint would conceal any underdrawing used elsewhere. The canvas has not been X-rayed, but the thin paint is not at all likely to conceal an earlier image.

The rubbing down of the paint exposes the grey imprimatura but rarely exposes the bright blue beneath. Whistler worked from dark to light on this dark grey background, and at a middle stage applied paint for the grey background to the composition, lighter than the grey imprimatura, up to and around the reserved areas where the figures and other elements, such as the flower pot, had been begun. The figures were extensively rubbed down and reworked, but a final working that would have satisfied the artist is absent, giving the figures an unfinished, worn look because the last paint application was rubbed down. The eyes and the shadows under them are particularly thinly painted, which was the artist’s usual method of creating darks within a face. Some elements in the background were painted, wiped off, and never reworked.

As the canvas progressed, the red of the flower pot and the capes was made more intense, the same red colour being used each time. The shelf in the foreground, little changed from its initial painting, was painted using an optical green, made from blue and black pigments. Though the blue mountains and water of the background are a greenish blue, this is not the bright greenish blue imprimatura, but applied paint over the grey imprimatura. The flowers were painted last and must have been deemed successful, since they were not rubbed down.

The paint for the flowerpot was extremely thinned, and had to be wiped to retain the shape of the pot, as drips ran off. The flesh paint was less thinned and remains thicker, even lightly impasted, in places. Some of the paint fluoresces strongly in ultraviolet light, which could indicate an added medium modifier incorporating natural resin, which would confer a similar appearance. Certainly the paint medium is darker today than originally. 2

Lead white, bone black, vermilion, Prussian blue and some synthetic ultramarine form the paint for the laying-in of the composition before the grey background was applied, and the same colours can be seen in the later paint, augmented with madder and yellow ochre for the flowers, and a purplish-red ochre for the caps worn by the women. The red paint is almost pure vermilion, and Prussian blue is common in the later applications of paint. The range of colours employed is wider than in a 'Nocturne' or 'Harmony', yet narrow in that no green at all, nor bright yellows nor browns, are included. The colours themselves have not altered with time. 3

Conservation History

Professor Joyce H. Townsend, Tate Britain, reports:

'The painting was lined and re-attached to a new and rather flimsy stretcher at an unknown date, rather unevenly stretched at the top left and bottom right corners. Prior to this lining the canvas had suffered some damage, and during the lining it was crumpled in places. Cusping on all four sides indicates the points of attachment of the canvas to the original stretcher, and the depth of the cusping makes it clear that the format was slightly reduced at top and bottom during the lining, by less than one centimetre. Both the lining canvas, medium weight and therefore finer than the original, with a more conventional thread count, and the new stretcher bear export stamps from France where the work was likely carried out. This lining could date from Whistler’s lifetime.' 4

Professor Townsend adds:

'In 2001 two thick and yellowed varnishes that filled the hollows of the coarse canvas were removed by Stephen Hackney at Tate. Analysis of the combined varnishes suggested oil as well as natural resin, and the difficulty in removing them is consistent with this finding. Whistler’s own varnishes – or more likely those applied for him – were generally based on natural resin alone, which suggests these might have been applied at the time of the lining, or else later to ‘refresh’ the surface. Before the 2001 treatment, the surface appeared very dull: it became evident during cleaning that the paint medium has discoloured towards brown, making the surface permanently darker and more dull.' 5


1891: When it was on exhibition with Messrs Dowdeswell, London, Whistler wrote, 'I am in no way responsible for the taste of the frame with its astonishments of plush and varied gildings.' 6 The whereabouts of this frame is unknown.

1997: Flat Whistler frame with incised frieze, gold leaf and silver powder (oil and water gilding) on wood (pine), made at the Tate in 1997, based on Whistler designs. 7

This frame replicates that on Whistler’s Girl with Cherry Blossom [YMSM 090], which itself was cut down from a larger format. The inner frieze is covered with rows of small whorls incised into the whitening layer of powdered chalk and protein glue. This was a pattern used by Whistler in 1864 on Symphony in White, No. 2: The Little White Girl [YMSM 052] and reflects his interest in blue and white Chinese porcelain.

Whistler did not develop the reeded outer-frame until the 1870s. Therefore the combination of elements replicated on this 1997 frame is possibly an early twentieth century interpretation of a ‘Whistler’ frame. 8


1: Whistler to H. E. Whistler, [22 March 1880], GUW #06688

2: The paint medium was not investigated further than by heating a fragment of it, which yellowed and softened at a low temperature. These observations suggest added natural resin-based medium modifier, and also imply that early darkening of the paint with age is likely.

3: Report by Prof. Joyce H. Townsend, Tate Britain, November-December 2017.

4: Townsend 2017, op. cit.

5: 'At that time, the painting was marouflaged to an aluminium honeycomb panel, without having been X-rayed first. The metal and the honeycomb effect would together conceal any underlying image on the coloured imprimatura, so only four digital X-ray plates exist from 2017, giving very little information.' Townsend 2017, op. cit.

6: Whistler, J. McN., letter to the Editor, The Morning Post, 27 July 1891; reprinted in Whistler 1892 [more], p. 288; GUW #13566 and #05842.

7: Dr Sarah L. Parkerson Day, Report on frames, 2017; see also Parkerson 2007 [more].

8: Ibid.

Last updated: 25th November 2020 by Margaret