UNIVERSITY of GLASGOW

While this copper plate was taken through many states, the image itself varied little from Whistler's initial conception. After he cleaned the abraded surface of the plate and reduced the scale of the figures, most additional work - in both etching and drypoint - involved adjustments to the size of the shadow below the awning and relatively minute changes to the two figures within the doorway.

Fruit Stall was published by Messrs Dowdeswell and Thibaudeau with A Set of Twenty-six Etchings in 1886. The complete edition was to consist of 30 impressions. A deceptively simple plate as originally composed in 1880, this was radically altered by Whistler as he worked on it between 1886-1887. He printed only a few impressions at relatively long intervals in 1886, sending solitary impressions to Messrs Dowdeswell on 2 April, 31 July, 20 August, 21 September and 2 October. 6 These were probably from the first five states of the etching. He may at this time have been printing complete sets rather than completing the edition of any one etching.

One document records impressions of Fruit Stall over a year and lists single impressions delivered on 31 July and 20 August 1886, two on 2 October 1886, as mentioned above, plus two on 2 April 1886, twenty-nine on 8 February 1887, a total of 35. 7 This does not exactly correspond to the individual receipts, and it may be that it simplifies the numbers printed, particularly during February 1887.

Since Whistler kept one or two proofs of each new state, as printer's perquisites, there are more impressions than might be expected of this etching.

Only one impression of the first state has been located. 8 It is a working proof, in black ink on light-weight cream laid paper, and shows vertical scratches overall, where the plate has not been cleaned and polished, so that it looks as if it is pouring with rain. The first owner Henry Studdy Theobald (1847-1934) was collecting Whistler's work in the mid-1880s from Messrs Dowdeswell, so it is hardly surprising that he should have acquired a unique early proof ().

Two impressions of the second state have been recorded, of which one has been located (). The impression noted by Kennedy as owned by Thomas Jefferson Coolidge jr (1863-1912) in 1910 was printed when the 'Second Venice Set' was published, in 1886 (). This suggests that most work on the plate was done back in London. All impressions from this time on appear to be trimmed to the platemark and signed on a tab with a butterfly and 'imp.' to show that Whistler had printed them (or at least supervised and approved the printing). Of the third state, there is only one impression, printed in very dark brown ink on ivory laid paper with the watermark 'PG' and a trefoil over an ogee arch (). This was printed with very deep plate tone at the bottom of the plate, which formed into uneven waves under the stresses created when the plate went through the press: Mansfield commented, 'The dark water is an effect of printing.' 9

An impression of the rare fourth state is in black ink, again on ivory laid paper, this time with a posthorn watermark (). This state shows for the first time the conspicuous mark of the vice used to hold the plate at the bottom edge, when it was being reheated and reworked. Three impressions of the third state are recorded, one being printed in dark brown ink on ivory laid paper with a 'WW' countermark ().Only one or two impressions of several subsequent states have been located.

Finally, in the ninth state Whistler was sufficiently satisfied with the plate to allow it to go to Dowdeswell's, who passed it to Émile Frédéric Salmon (1840-1913) to print. He printed an impression in black ink on dark cream Japan paper, which is reproduced below.

However, it appears that Whistler did not particularly like this impression, despite the soft, rich effect of the shadows, and he decided to print it himself. He also made small but extensive alterations to the figures and shading in the tenth state, and it was printed in dark brown ink on ivory 'antique' (pre-1800) laid paper, and is reproduced below.

Impressions were uniformly trimmed to the platemark and signed on the tab with a butterfly and 'imp.' to show that - in theory - Whistler had printed them; in practise, he had both volunteer and professional printers to help him. Some are, indeed not signed by Whistler: the proof of the tenth state, for instance, reproduced above, was signed with a slightly squashed butterfly and 'ip.' - rather than 'imp.' as shown below.

Whistler continued to change details of the figures and shading and print one or two impressions, as before, of each state. Several impressions of late states were kept in Whistler's collection, and are now in the Hunterian, showing that he kept a proof of each state for his own use or resale. Some of these impressions - presumably 'proofs', which he was entitled to keep - were kept by the artist. Impressions in brown ink on cream laid paper,such as the 14th state reproduced below, look more delicate and atmospheric, as well as warmer in colour, than earlier impressions in black on ivory paper.

Eventually, under the necessity of completing the edition, Whistler printed a larger print-run of the 20th state.

Some impressions are in black ink, one, for instance on ivory laid paper with an elaborate watermark of a crown () and two on cream Asian paper (, ). However, it was usually printed in brown or dark brown ink on laid paper. A few watermarks are seen on these later impressions, including 'Pro Patria' () and posthorn (). In addition some are on Asian laid paper (, ).

Finally, four impressions are recorded of the cancelled state. One cancelled impression has a 'Marlowe' watermark (). The edition was completed by 8 February 1887 when Whistler sent Dowdeswell's 28 impressions and one of the cancelled state. 10