The harem – painting behind the scenes

Just prior to his death, the British Victorian watercolourist and traveller, turned Royal Academician, John Frederick Lewis (1805–76) stipulated in his will that his remaining works, apart from a bequest to his wife, should be sold at Christie’s at the first relevant opportunity after his demise.1London, Somerset House, Will of John Frederick Lewis, dated 31 January 1876. Drawings and watercolours of greater or lesser elaboration bought in at the sales of 1877, reappeared in a smaller sale of 1897, and again, with ever-dwindling interest reflected in prices made, the residue from that sale came under the hammer in 1909.2The sales were held in London by Christie’s on 4–7 May 1877, 3 May 1897 and 24 May 1909. The fact that many lots were bought in and offered subsequently appears not to have been previously recognised. 

A selection of generally much slighter drawings and sketches spanning the artist’s entire working life, presumably considered too minor and valueless for inclusion in the estate sales yet scrupulously preserved by the painter’s widow Marian, more than twenty years his junior, were distributed by her executors in the years 1907–83Mrs Elizabeth Napier of Glasgow, legatee and executrix, seems to have been most directly concerned with this allocation of drawings, but papers preserved by the family of Henry Heneage Finch of London indicate that he advised and assisted her appreciably. to a number of art galleries in England and the dominions. The bulk were accepted by the South Kensington Museum and remain together in the Department of Designs, Prints and Drawings of the Victoria and Albert Museum.4These have been fully described for the first time by me, and will be published and reproduced in a catalogue of Lewis works in the Victoria and Albert Museum in 1989. A further collection is preserved in the Print Room of the Department of Art of the City Museums and Art Gallery in Birmingham. Several studies offered to the Trustees of the National Gallery were transferred immediately to the Tate Gallery, while amongst other, more token, gifts were those forwarded not only to Cape Town, Ottawa and New Zealand but also to Australia, with Sydney, Adelaide and Perth receiving minor studies through their respective Agents General.5Correspondence file in the possession of F. H. Finch, London. The Australian art museums have made no further attempts to represent the artist adequately. Amongst a small group of drawings despatched to the National Gallery of Victoria is one casually swift compositional sketch of female figures in an Eastern interior drawn in hasty sweeps of graphite outline with broad, approximate washes of transparent watercolour (fig. 1).6J. F. Lewis, Study for the harem, graphite with watercolour washes on cream wove paper, 35.7 x 49.3 cm (sheet size), uninscribed. Michael Lewis, John Frederick Lewis RA 1805–76, Leigh-on-Sea, 1978, cat. no. 491; hereafter referred to as Lewis, 1978, this monograph records previous literature on the artist.

This drawing, of course, represents the very antithesis of the concept and treatment of the artist’s famous and highly-prized watercolours of his orientalising period dating from 1850 until the year of his death, renowned for their accumulation of meticulous, circumstantial Islamic detail and the painstakingly intricate elaboration of texture in extensive bodycolour, further enriched with a full panoply of Indian ink, Chinese white, gum arabic and scraped heightening. Yet aesthetically and stylistically atypical as it may appear in general context, the sheet seems to be of some rarity and considerable procedural interest, as one of the very few direct compositional studies to have been preserved from the opulent ultima maniera of this most conscientiously craftsmanly and flamboyantly sumptuous of the British orientalists. 

Several slight drawings in the executor’s gift to the Victoria and Albert Museum (one in soft graphite, one in pen and ink) are sufficiently close to completed oils and watercolours to be considered as premières pensées at least.7D1182–1908 and D1175–1908 (Lewis, 1978, cat. no. 518). Amongst other studies, a lightly-washed watercolour of a seated Arab is not dissimilar in lassitude of handling to the Melbourne drawing, and a rapid notation of a desert procession of Bedouin and camels, executed with nervous and spasmodic dabs of the brush point in diluted oil pigment on absorbent paper, suggest early thoughts or compositional trials for paintings which appear not to have been executed.8D1184–1908 (Lewis, 1978, cat. no. 530) and D1168–1908. In the Birmingham group is another such late study in dilute oil pigment, which has long been catalogued as a preliminary drawing for the important Royal Academy exhibit of 1869, The seraff (money-changer), a doubtful coin; a scene in a Cairo bazaar – coincidentally now in the same collection.9429’07, not in Lewis, 1978; the oil is listed as cat. no. 603. It is equally, if not more closely related to the highly-finished watercolour and bodycolour variant which has been published with the title of The street and mosque of the Gouriah [sic].10ibid., cat. no. 516, illus. fig. 45; the title identification seems open to question. There are, certainly, many drawings listed in the catalogues of the posthumous Lewis sales as studies for paintings, but in most cases these would have been the brilliantly assured sketches from life made in the Near East and used subsequently as components of the sequence of monumental compositions which forged his considerable reputation. Indeed, this classic procedure was instantly discerned by contemporary critics, for Lewis, with total artistic honesty, exhibited such studies at the Royal Academy, and on at least one occasion a life drawing was recognised as the source for the protagonist in one of his exhibition machines.11The old pedagogue of A Turkish school, see Art-Journal, 1870, p. 169. 


This last period of the artist’s career was inaugurated by his most famous, if not indeed, notorious watercolour, which he entitled (with obvious attempt to suggest an Arabic guttural) The hharem (fig. 2), completed to an English commission in 184912Lewis, 1978, cat. no. 413, illus. fig. 41. This is the partial replica owned by Miles Birket Foster in the Victoria and Albert Museum, which he and almost all other commentators have conflated with the original, now in the possession of the Nippon Life Insurance Company, Osaka. during the long expatriate residence in Cairo (1842–50), and which created a critical furore when shown at the Old Water-Colour Society’s rooms in Pall Mall in 1850 when the Lewises returned to their native land. It stimulated signal French interest both in itself and in the water-based medium generally when displayed at the Paris Exposition Universelle five years later, and it had a type-forming role with a considerable and varied progeny of more than twenty works within Lewis’s own oeuvre. Far from being the first imaginary depiction of a harem interior in the west (Lewis may well have seen the magisterial Delacroix Femmes d’Alger of 1834 in Paris), it is undoubtedly the most ethnically comprehensive. Despite his decade of Near-Eastern experience – he had arrived in Istanbul by October 1840 – Lewis may never have entered the occupied women’s quarters of an Oriental house which constituted a functioning harem. Considerable artifice was, however, feasible. Greek, Armenian and Jewish women were sometimes prepared to allow themselves to be depicted unveiled by visiting Europeans, as well as Egyptian servants and slaves, and even if we accept that the foreigner may have seen relatively little of the inhabited interiors of other Muslim houses, he had always available to him the essential architectural and decorative elements from which to simulate a fictitious harem, whether presented as allusive genre – a human still-life of costumed figures, or as a domestic pageant – on several occasions more or less erotically-spiced by the unveiling of a newly-purchased slave girl (in a tableau long pre-dating Hollywood) or the clumsy passing of a surreptitious love message expressed by a symbolic bouquet. Indeed, the visual premises for Lewis’s subsequent Oriental confections laboured unremittingly in his studio at Walton-on-Thames were laid down by his intimate quotidian familiarity with the various rooms and chambers of the great Cairene mansion in mamlũk style which he rented from a patrician owner in the cosmopolitan Ezbekiyeh district. His drawn records of the house, including, presumably, the women’s quarters, as well as the adjacent precincts and sukhs, served him well for a quarter of a century in that relentless production of exhibition watercolours and academy oil paintings. Above all, he recreated with the allure of idealised recollection its principal leafy courtyard and the stately mandarah in which, in upper-class Ottoman dress, he received and regaled Thackeray, who left a splendidly vivid and evocative account of the household and its bearded, lotus-eating, pipe-smoking tenant.13‘M. A. Titmarsh’ (W. M. Thackeray), Notes on a Journey from Cornhill to Grand Cairo, London, 1846. Furthermore Lewis, as we have noted, had retained those superlative life drawings, brilliantly suggestive and swaggeringly virtuosic in graphite or various chalks, with telling additions of watercolour or bodycolour wash, and which rank supreme amongst the figure studies of Victorian England; these, too, remain unrepresented in Australian collections. Along with the crisp and elegantly delineated architectural sketches, artfully incomplete, many camel studies and a few landscapes, they formed a sizeable segment of the studio sales, in company with the residuum of his records of Spanish, Italian and other Continental European excursions. Items of decorative art (a framed Koranic page turns up repeatedly, and figured in the estate auction), pieces of Chinese and Japanese export porcelain which appear scattered throughout his compositions (perhaps too liberally), and above all an important collection of Near-Eastern costumes and accoutrements, further adorned with flowers from the Home Counties garden, supply the remaining raw materials for transmutation into nearly credible fantasy. 

The relationship of the small and sketchy Melbourne compositional study to a known, coeval painting – as distinct from these Oriental notations made in situ and later quoted – is more than usually precise, and it offers an unusual glimpse of the ageing painter formulating on paper his thoughts for a complex and scrupulously elaborated oil. It provides the manifest basis of yet another work owned by the Birmingham Museums and Art Gallery,1414’49; Lewis, 1978, cat. no. 625, illus. fig. 68. In his description (p. 39), however, the author more specifically dates it to the period of The reception of 1873. I have not seen the Yale painting for a decade, and this remains a viable alternative dating to that here proposed. now entitled, in emulation of its ultimate and illustrious prototype, The harem (fig. 3). The border lines drawn in graphite suggest that in the initial stages at least, the artist envisaged a painting with all four corners rounded, but the circumstances of the commission remain, so far, quite lost. Lewis occasionally adopted the format of a rectangular panel with the upper corners shown as if arched – a commonplace of British painting over much of his century – and a number of famous works are still slip-mounted and framed in this manner. His intensely-wrought paintings were seen to suffer from being hung too high, with resultant loss of detail (specifically Waiting for the ferry boat, upper Egypt at the Royal Academy exhibition of 1859).15J. Ruskin, Notes on some of the Principal Pictures exhibited in the Rooms of the Royal Academy; Ε. T. Cook and A. Wedderburn, The Works of John Ruskin, London, 1904, XIV, pp. 218–19, hereafter referred to, respectively, as Ruskin, Academy Notes, and Cook & Wedderburn. Highly conscious of the dignity and status of the gold-framed academy piece in oil, it is hard to imagine Lewis accepting an order for a sopraporta or other decorative roundel to be incorporated, above eye-level, into an architectural or decorative scheme, which otherwise, such a shape might suggest.16One of the most blatant derivatives from Lewis’s later Harem compositions, Xarifa c.1863 by Francis John Wyburd (1826–after 1894), sale, Christie’s, London, 26 November 1982, (293), illus., is in oval format. Another little-known drawing in graphite, chalk and watercolour, formerly in the London art trade, depicting two female heads in slightly different positions against part of the same intricate background of curtain and mashrabiyyah – typical of these Cairene interior evocations – is in late style and can also be related directly (fig. 4).17With Spink & Son, London, exhibited 23 April–8 May 1979, illus. Connoisseur, April 1979, p. 72 (private collection, Switzerland). These drawings would seem to suggest that the subject was the result of considerable preliminary planning. 


The earlier provenance of the Birmingham oil painting remains to be established, and no sustained attempt has been made to rationalise a date for it within the obvious overall limits of the period 1855–76. The picture was received at the Birmingham Art Gallery on 4 July 1949 as a gift from G. T. S. Horton. It had been owned by his uncle, Joseph Horton of Moseley (who died in 1927, aged 72 or 73),18S. Wildman, in litt, Η. H. Preston, 18 June 1987, has observed that he ‘was the son of Isaac Horton, a Bull Ring pork butcher who used his profits to buy up cheap leases on city centre properties. Horton’s Estates Ltd prospered under Joseph, extending its holdings to Wolverhampton, and still very much flourishes today. Joseph, who lived at Fern Hill, School Road, Moseley, was the Chairman of Atkinson’s Brewery and a Director of the Grand and Midland Hotels (both Estate properties); he was a pigeon fancier and breeder of Berkshire pigs. No known interest in art; but I have found that he was the lender of The harem to the Glasgow Exhibition of 1901; where he (or his father, who died in 1880) got it we may never know’. Horton also lent the picture to the Royal Academy in 1907 (165). but the previous owner of The harem and its mode of acquisition has not been discovered. Arguably unfinished in minor parts, unsigned and undated, it appears not to have been exhibited during the artist’s lifetime. 

The most recent monograph on J. F. Lewis, by his great-grand-nephew, Major-General Michael Lewis (1978), lists the painting under the date heading of the last year of the artist’s life of 1876, together with other very late works which have been considered not to have been carried to the final stages of completion before the intervention of death. These include The siesta, an oil on canvas, signed and dated to that ultimate year (Tate Gallery, London) – an attenuated harem variation with a solitary, somnolent reclining female figure, encapsulated within mashrabiyyah screens and sun-drenched curtains, as if drugged by the heavy perfume of scented white lilies and giant opium poppies clustered in Imari vases. An unfinished watercolour variant, identically titled, is in the Fitzwilliam Museum, Cambridge, and a small, upright oil sketch has now re-appeared, holding a somewhat analogous position to the Melbourne drawing within the chain of creation.19Lewis, 1978, cat. no. 623, illus. fig. 67 (the oil) and cat. no. 624 (the watercolour). The oil sketch was in the sale, Sotheby’s Belgravia, London, 9 July 1974 (64), illus. The busy yet spacious The street and mosque of the Ghooreyah, Cairo20Lewis, 1978, cat. no. 626. (again, signed and dated 1876) which was exhibited posthumously at the Royal Academy in 1877, was listed there as ‘unfinished’, although to the eyes of a century later this state seems relative and hardly obtrusive. 

When I first viewed the Birmingham Harem in store in 1971, at the moment of revival of interest in the artist and the beginnings of modern study,21The first detailed result of this was the catalogue by Richard Green, Laing Art Gallery, Newcastle-upon-Tyne, 1 October–6 November 1971, John Frederick Lewis RA 1805–76. the immediate impression gained from the rich, saturated and luminously smooth handling, lustrous, translucent yet oleaginous, was of an oil painting sufficiently close in instinctive technique to that of traditional transparent watercolour to suggest the hand of a practitioner in that medium who had just returned to oil painting after twenty-five years of abstinence, and hence a dating of c.1856. In actual fact, Lewis’s later watercolours were largely executed in heavy bodycolour, and although, to date, the small oil which has usually been identified as his second Academy exhibit after his return from the Near East (that of 1855), An Armenian lady, Cairo – the love missive, ‘The token-flowers tell what words can never speak so well’ – Byron (private collection, London)22Lewis, 1978, cat. no. 568. has not been examined, subsequent sequences of exhibited oils appear with a blond, opaque and drily-brushed facture which certainly demonstrates a gouache aesthetic. This is true of The greeting in the desert, Egypt – ‘Selamat, teiyabeen’ and Street scene in Cairo near the Babel Luk,23These two oils have subsequently been identified in sales in London, respectively Christie’s, 20 June 1986 (92) illus. in colour, and Sotheby’s, 18 October 1978 (255), illus. Of The greetings in the desert, the writer in the Art-Journal for 1856 (p. 164) noted, however, ‘In the execution of this picture there is greater degree of breadth than in the water-colour works by this painter; but everything here must be accepted as of the most unquestionable accuracy.’ paired for the 1856 season, and most readily available for close, well-lit examination, the austere The Syrian sheik, Cairo24Lewis, 1978, cat. no. 568. shown in 1857 (Fitzwilliam Museum, Cambridge) with heavy graphite underdrawing clearly discernible beneath its matt pigment layers. 

The judgements of the contemporary London critics on the success of Lewis’s crucial medium change have never been assembled. Juxtaposed samples reveal assessments, fluctuating if not entirely contradictory, of the ability of the ex-President of the Old Water-Colour Society and Associate Academician to handle the inherent textural and illusionist properties of his newly-resumed principal technique. Lewis himself would seem to have pontificated in no uncertain manner on his ideal aspirations in oil painting before the young Millais, and referring to the recent paintings he and Holman Hunt had been showing, expostulated: 

I am sure oil painting could be made more delicate than either of you make it; not sufficient pains are taken to make the surface absolutely level. Why should it be more piled up than in water-colour? 

With his advocation of the surface criteria of earlier history-painting – ‘as smooth as glass’ was his avowed desideratum – he continued assertively in his rejection of all the rich materiality of pigmentation: 

I intend to take to oil colours myself, and damme, I’ll show you how it ought to be done. The illusion of all modern painting is destroyed by its inequality of surface … Parts of your painting are level enough, I admit, but in your deep tints there is a great deal of unseemly loading.25W. Holman Hunt, Pre-Raphaelitism, London, 1905, I, p. 270. 

Old-fashioned by more advanced Continental standards at the time of this encounter (internally dateable to 1854), Lewis’s denial of the corporeality of matière was to seem increasingly retardataire in England as the century progressed. The single Academy exhibit of 1855, however found both the writer in the Art-Journal26Art-Journal, 1855, p. 171, ‘It … has all the finish of his water-colour works, but with more softness.’ and John Ruskin,27Ruskin, Academy Notes 1855; Cook & Wedderburn, XIV, p. 12. whose published opinions have obliterated him in the eyes of posterity, in initial agreement that the picture demonstrated the same laudable excellences as Lewis’s watercolours, with the corollary satisfaction of the great critic that it was in a ‘safer medium’; he had been the most vociferous of several writers who, over many years, had urged the artist to work in the less fugitive and physically vulnerable oil technique. 

In an unprecedentedly lengthy description and criticism of A frank encampment in the desert of Mount Sinai (Yale Center for British Art, New Haven),28Lewis, 1978, cat. no. 564, illus. fig. 47. The Times of 28 April 1856 had singled out those watercolourists (amongst whom Lewis was prominent) who ‘attempt by means of opaque colour to rival the effects of oil’, and while admitting ‘the beauty of the result’ it regretted that such a hypothetical artist of this ambition ‘should not have intrusted his conception to the higher and more enduring medium.’ It was, however, the art critic of the Art-Journal who concerned himself most consistently with media comparisons, and in 1857 wrote simply of The Syrian sheik that ‘the appearance of finish is not so obvious here as in the painter’s watercolour works’,29Art-Journal, 1857, p. 166. although Ruskin, by contrast, found it ‘highly-wrought’ and, indeed, Eyckian.30Ruskin, Academy Notes 1857; Cook & Wedderburn, XIV, p. 94. Of the five paintings accepted by the Academy for 1858 (as the artist became a candidate for election to his Associateship), the Art-Journal wrote of the most prominent, A kibab-shop, Scutari, Asia minor,31Lewis, 1978, cat. no. 573, illus. fig. 53. that it ‘being thin and dry, bears allusion to water-colour practice; it is finished with that stipple in which the author realises all his objects.’32Art-Journal, 1858, p. 163. Of course such ‘water-colour practice’, that is to say that firm adherence to a gouache manner, was fresh in everyone’s mind, for the artist had just shown his last major drawing to be widely exhibited in London, Hhareem life – Constantinople33Lewis, 1978, cat. no. 572, illus. fig. 52. (Laing Art Gallery, Newcastle) – rich in bodycolour – as his farewell contribution at the ‘Old’ Society’s rooms the year before. He continued to make watercolour and bodycolour replicas, versions or variants of his oils to the end of his life, as we have observed, but these were not so accessible to public view. 

In 1859 the Art-Journal recorded the opinion that Waiting for the ferry boat, upper Egypt ‘proclaims the painter not so entirely at home with oil as a medium as water’,34Art-Journal, 1859, pp. 163–64. with no further analysis or explanation of the painting hanging on high, but Ruskin, who saw it closer at Lewis’s dealers, was bored by the repetitious subject matter rather than by any technical ambivalence.35Ruskin, Academy Notes, 1859; Cook & Wedderburn, XIV, pp. 218–19; and in a letter to Lewis of 1 May 1859, in the possession of Major-General Lewis (Lewis, 1978, p. 28). Reversing his usual pattern of procedure in 1863, Lewis displayed an oil version of his great watercolour ‘which men will come to England from far away to see’, famous from Ruskin’s supersaturated panegyric of 1856,36Ruskin, Academy Notes, 1856, pp. 73–78. A frank encampment, which the Athenaeum found to be ‘gaining in some respects from its translation into the new material. It is better in colouring, free from chalkiness, and less scattered and glittering in effect.’37Athenaeum, 9 May 1863.

Writing in 1865 of A Turkish school in the vicinity of Cairo, Egypt38Lewis, 1978, cat. no. 595. the Athenaeum of 6 May found the work was treated with breadth but the figures were lacking in substance, ‘their execution is rather thin’ (Lewis had used an old drawing for the dominant bearded pedagogue), an opinion shared by The Times in its more appreciative notice concentrating on a ‘new quality … in its breadth and unity of effect, till now imperfectly attained by this most minutely elaborating painter!’39The Times, 29 April 1865. The writer in the Art-Journal, in linking the picture with In the Bey’s Garden, Asia minor (Harris Art Gallery, Preston)40Lewis, 1978, cat. no. 598, illus. fig. 59; a replica is recorded (Witt Library, London) as having been in the possession of the Maharaja Gaekwar of Baroda in 1920. pin-pointed his reasons for satisfaction in categorical terms: 

It strikes us in these two works Mr. Lewis has overcome the loaded opacity which used to prejudice his technical practice. His colours are now liquid and transparent, and he thus adds to his feats in drawing and composition the full advantages of oil medium’.41Art-Journal, 1865, p. 169. 

By 1869, however, the same periodical in a somewhat tetchy dismissal of The seraff, observed a reversal: ‘the mode of painting is that of opaque watercolours transferred to oils’, it reiterated,42ibid., 1869, p. 164. and a similar delicately-touched, brilliantly-hued chalkiness may be seen in the major masterpiece, An intercepted correspondence, Cairo, which accompanied it.43Lewis, 1978, cat. no. 602, illus. fig. 61. As late as 1874 The Bezestein bazaar of El Khan Khali44ibid., 617, illus. fig. 65. was widely admired, above all by the sculptor Thomas Woolner who subsequently owned it (‘so entirely unique and original, and of such quality in art … of such superb excellence’),45In a letter to Lewis of 3 May 1874, in the possession of Major-General Lewis (Lewis, 1978, p. 32). and A lady receiving visitors; the apartment is the mandarah, the lower floor of the house, Cairo (Yale Center for British Art, New Haven)46ibid., cat. no. 611, illus. fig. 64, the watercolour variant is cat. no. 612. induced the writer in the Illustrated London News to continue insisting on technical incongruities in Lewis’s 

miracles of patient elaboration, in which, however, the sense of atmosphere is more than ever wanting, owing, we think, to a method of hatching and stippling (appropriate only to water colours) with pigments, especially blues, too dilute with oil and varnish.47Illustrated London News, 9 May 1874.

However firm the artist’s hand remained, evidently little affected by the fatal onslaught of locomotor ataxia during these last years, his creative imagination meandered in a haze of past recollections, his paintings becoming more and more quintessential fantasies distilled from memory in the long afterglow of his Oriental idyll, poignantly projected in lush and refulgent colouristic sonorities. The question yet remains as to when, during his last quarter-century, Lewis painted the Birmingham Harem from the Melbourne drawing and perhaps other sources.48With rare exception, securely documentable cases of the use of Near-Eastern figure drawings in subsequent paintings involve male figures, and no studies from the decade 1840–50 which might have been used for the Birmingham oil have come to light to date, although each season brings to the London art market, in particular, unpublished drawings originating, ultimately, in the Lewis estate sales. Demonstrating very strikingly the ‘liquid and transparent’ handling of the group of 1865 Academy exhibits, the painting has something technically in common with other contemporary oils, such as the two versions of The hósh (courtyard) of the house of the Coptic patriarch, Cairo; the patriarch is dictating to his secretary despatches to a convent in the desert to be conveyed by the Arabs in waiting, the larger being that shown at the Royal Academy in 1864 and, prior to recent rehabilitation, ending as decoration for the foyer of a Brooklyn cinema,49Lewis, 1978, cat. no. 590, reproduced in colour in its post-1926 truncated form in the sale catalogue, Sotheby’s, London, 24 November 1981, (307). the better-preserved being the smaller sketch (or replica) in the Tate Gallery.50Lewis, 1978, cat. no. 591, illus. fig. 56. The most notable similarity in handling, however, may be observed in the very small panel of 1863, shown with The hósh, and entitled A startling account, Constantinople, in the Joshua Dixon Collection at the Bethnal Green Museum, London (fig. 5),51ibid, cat. no. 588; the watercolour version is cat. no. 589. one of the most immediate oil equivalents to what might be considered normative water-based technique. It is this period around the early to mid-1860s, and certainly prior to the major display of paintings in 1869, to which I would now, tentatively, ascribe the Birmingham picture. 

The mahogany panel bears the label of the London manufacturer, Charles Roberson & Co., yet it has been pieced out and is now composed of three elements all of similar timber, whose grains, however, do not marry. This is perhaps not entirely consistent with the notion of a structure ordered from a commercial artist’s supplier, although it is just conceivable that a completed support of fine wood might have been hastily enlarged in the workshop in response to an order with precisely stipulated dimensions.52The mahogany support measures 88.7 x 111.7 cm overall, the main panel, strengthened with chamfered battens, measures 81.0 x 101.6 cm, the horizontal member along the top edge being 7.7 mm high, and the vertical extension to the left 10.0 cm in width. No longer possessing my original notes on the picture, I am most grateful to Stephen Wildman for having the support remeasured. Despite the somewhat atypical preparation of the subject with preliminary drawings, which implies a very concisely-focused mental image of the composition fixed in its essential disposition of forms, Lewis may have felt the need to extend the field of the work in progress, and it could perhaps have been that the original, partially-painted, panel was returned to the supplier during the course of painting for extension at the top and left, either through this exigency or because of the development of some flaw, fault or damage. The tale of analogous procedure with the initial paper sheet of the first Cairo Hhareem, as recorded in a French review,53Edmond About, Voyage à travers l’Exposition des Beaux-Arts, Paris, 1855, p. 32: ‘Lorsqu ’il avait manqué un morceau, il l’enlevait proprement, remplaçait le papier par une pâte de fabrique anglaise, et recommençait la besogne. Un jour, il trouva l’aquarelle fendue en deux. Il l’envoya en Angleterre, fit recoller soigneusement les morceaux et reprit la suite de son travail. Vous voyez que la fabrication des aquarelles est chose compliquée …’ sounds somewhat improbable, but may be confirmed when the recently rediscovered chef d’oeuvre becomes available for professional scrutiny.54Permission has so far been refused by the owners for the watercolour to leave Japan. The timber joints of its Birmingham ‘descendant’ have long since opened up, giving the picture an unsightly appearance around the area of the upper left, which has meant that it has not been shown alongside other important works in recent years nor has it been available for loan. Full conservation treatment would be lengthy, laborious, difficult and costly. As the cracking and various infills show some signs of age, it might be suggested, with great caution, that the cleavage of the original members could date back to the period when the panel was in Lewis’s own Surrey studio – given his very slow rate of production – and that the work might have been abandoned when virtually complete through this incipient disruption to the painter’s criterion of mirror-smooth surface. The picture cannot be identified with any of the lots in the studio sales, although it is not impossible that it was one of the items remaining in family ownership. 

In later life Lewis appears to have worked entirely to commission, and to have sold his oils either to William Vokins or specific patrons, such as the prosperous and loyal collector Charles R. Matthews. Perhaps The harem, marginally incomplete, and not to be displayed, was yielded to a client – possibly at a lower price. Assuming the habitual exhibition of commissioned works, it would not be too simplistic to suggest that this particular variation on Lewis’s most famous theme had been intended initially for the Royal Academy walls, hypothetically for one of the years when the painter had nothing else to show, although, despite some new clues which have recently (1986) come to light, the private life of the artist still remains obscured in mystery and conjecture. Lewis had just one oil at the Academy for the years 1855, 1857, 1859, 1863 and 1866 (his Diploma Piece, the little-admired The door of a cafe in Cairo).55Royal Academy of Arts, London, Lewis, 1978, cat. no. 597, illus. fig. 58; the watercolour version is cat. no. 599, and a smaller replica from the T. S. Kennedy collection, perhaps commissioned, was in the sale at Sotheby’s, London, 17 June 1968, (24), repr. in colour. There were no works of any kind sent up for the seasons of 1860, 1867, 1871, 1873 and 1875. In 1862, 1868 and 1870 he appears to have exhibited earlier watercolour studies made abroad to represent his name in the catalogues, and although it would be exceedingly tempting to postulate that the Birmingham Harem was intended as an exhibit for one of these years of absence or stop-gap display, particularly around the seasons between 1862 and 1868, the old Academician on rare occasions sent in four and even five paintings at a time, often of widely varying dimensions, and it is an equally plausible speculation that the work in question might have been commenced to form part of a larger showing of that period to bolster up a single contribution or a small group of paintings. 

Harley Preston 


I first became aware of the relationship of this, perhaps less spectacular, item in the Melbourne collection of British drawings with the Birmingham painting, in 1962, when perusing the files on individual drawings in the National Gallery of Victoria, initiated by Ursula Hoff. 



1          London, Somerset House, Will of John Frederick Lewis, dated 31 January 1876. 

2          The sales were held in London by Christie’s on 4–7 May 1877, 3 May 1897 and 24 May 1909. The fact that many lots were bought in and offered subsequently appears not to have been previously recognised. 

3          Mrs Elizabeth Napier of Glasgow, legatee and executrix, seems to have been most directly concerned with this allocation of drawings, but papers preserved by the family of Henry Heneage Finch of London indicate that he advised and assisted her appreciably. 

4          These have been fully described for the first time by me, and will be published and reproduced in a catalogue of Lewis works in the Victoria and Albert Museum in 1989. 

5          Correspondence file in the possession of F. H. Finch, London. 

6          J. F. Lewis, Study for the harem, graphite with watercolour washes on cream wove paper, 35.7 x 49.3 cm (sheet size), uninscribed. Michael Lewis, John Frederick Lewis RA 1805–76, Leigh-on-Sea, 1978, cat. no. 491; hereafter referred to as Lewis, 1978, this monograph records previous literature on the artist. 

7          D1182–1908 and D1175–1908 (Lewis, 1978, cat. no. 518). 

8          D1184–1908 (Lewis, 1978, cat. no. 530) and D1168–1908. 

9          429’07, not in Lewis, 1978; the oil is listed as cat. no. 603. 

10        ibid., cat. no. 516, illus. fig. 45; the title identification seems open to question. 

11        The old pedagogue of A Turkish school, see Art-Journal, 1870, p. 169. 

12        Lewis, 1978, cat. no. 413, illus. fig. 41. This is the partial replica owned by Miles Birket Foster in the Victoria and Albert Museum, which he and almost all other commentators have conflated with the original, now in the possession of the Nippon Life Insurance Company, Osaka. 

13        ‘M. A. Titmarsh’ (W. M. Thackeray), Notes on a Journey from Cornhill to Grand Cairo, London, 1846.

14        14’49; Lewis, 1978, cat. no. 625, illus. fig. 68. In his description (p. 39), however, the author more specifically dates it to the period of The reception of 1873. I have not seen the Yale painting for a decade, and this remains a viable alternative dating to that here proposed. 

15        J. Ruskin, Notes on some of the Principal Pictures exhibited in the Rooms of the Royal Academy; Ε. T. Cook and A. Wedderburn, The Works of John Ruskin, London, 1904, XIV, pp. 218–19, hereafter referred to, respectively, as Ruskin, Academy Notes, and Cook & Wedderburn. 

16        One of the most blatant derivatives from Lewis’s later Harem compositions, Xarifa c.1863 by Francis John Wyburd (1826–after 1894), sale, Christie’s, London, 26 November 1982, (293), illus., is in oval format. 

17        With Spink & Son, London, exhibited 23 April–8 May 1979, illus. Connoisseur, April 1979, p. 72 (private collection, Switzerland). 

18        S. Wildman, in litt, Η. H. Preston, 18 June 1987, has observed that he ‘was the son of Isaac Horton, a Bull Ring pork butcher who used his profits to buy up cheap leases on city centre properties. Horton’s Estates Ltd prospered under Joseph, extending its holdings to Wolverhampton, and still very much flourishes today. Joseph, who lived at Fern Hill, School Road, Moseley, was the Chairman of Atkinson’s Brewery and a Director of the Grand and Midland Hotels (both Estate properties); he was a pigeon fancier and breeder of Berkshire pigs. No known interest in art; but I have found that he was the lender of The harem to the Glasgow Exhibition of 1901; where he (or his father, who died in 1880) got it we may never know’. Horton also lent the picture to the Royal Academy in 1907 (165). 

19        Lewis, 1978, cat. no. 623, illus. fig. 67 (the oil) and cat. no. 624 (the watercolour). The oil sketch was in the sale, Sotheby’s Belgravia, London, 9 July 1974 (64), illus. 

20        Lewis, 1978, cat. no. 626. 

21        The first detailed result of this was the catalogue by Richard Green, Laing Art Gallery, Newcastle-upon-Tyne, 1 October–6 November 1971, John Frederick Lewis RA 1805–76

22        Lewis, 1978, cat. no. 568. 

23        These two oils have subsequently been identified in sales in London, respectively Christie’s, 20 June 1986 (92) illus. in colour, and Sotheby’s, 18 October 1978 (255), illus. Of The greetings in the desert, the writer in the Art-Journal for 1856 (p. 164) noted, however, ‘In the execution of this picture there is greater degree of breadth than in the water-colour works by this painter; but everything here must be accepted as of the most unquestionable accuracy.’ 

24        Lewis, 1978, cat. no. 568. 

25        W. Holman Hunt, Pre-Raphaelitism, London, 1905, I, p. 270. 

26        Art-Journal, 1855, p. 171, ‘It … has all the finish of his water-colour works, but with more softness.’

27        Ruskin, Academy Notes 1855; Cook & Wedderburn, XIV, p. 12. 

28        Lewis, 1978, cat. no. 564, illus. fig. 47.

29        Art-Journal, 1857, p. 166. 

30        Ruskin, Academy Notes 1857; Cook & Wedderburn, XIV, p. 94. 

31        Lewis, 1978, cat. no. 573, illus. fig. 53. 

32        Art-Journal, 1858, p. 163. 

33        Lewis, 1978, cat. no. 572, illus. fig. 52. 

34        Art-Journal, 1859, pp. 163–64. 

35        Ruskin, Academy Notes, 1859; Cook & Wedderburn, XIV, pp. 218–19; and in a letter to Lewis of 1 May 1859, in the possession of Major-General Lewis (Lewis, 1978, p. 28). 

36        Ruskin, Academy Notes, 1856, pp. 73–78. 

37        Athenaeum, 9 May 1863. 

38        Lewis, 1978, cat. no. 595. 

39        The Times, 29 April 1865. 

40        Lewis, 1978, cat. no. 598, illus. fig. 59; a replica is recorded (Witt Library, London) as having been in the possession of the Maharaja Gaekwar of Baroda in 1920. 

41        Art-Journal, 1865, p. 169. 

42        ibid., 1869, p. 164. 

43        Lewis, 1978, cat. no. 602, illus. fig. 61. 

44        ibid., 617, illus. fig. 65. 

45        In a letter to Lewis of 3 May 1874, in the possession of Major-General Lewis (Lewis, 1978, p. 32). 

46        ibid., cat. no. 611, illus. fig. 64, the watercolour variant is cat. no. 612. 

47        Illustrated London News, 9 May 1874. 

48        With rare exception, securely documentable cases of the use of Near-Eastern figure drawings in subsequent paintings involve male figures, and no studies from the decade 1840–50 which might have been used for the Birmingham oil have come to light to date, although each season brings to the London art market, in particular, unpublished drawings originating, ultimately, in the Lewis estate sales. 

49        Lewis, 1978, cat. no. 590, reproduced in colour in its post-1926 truncated form in the sale catalogue, Sotheby’s, London, 24 November 1981, (307). 

50        Lewis, 1978, cat. no. 591, illus. fig. 56. 

51        ibid, cat. no. 588; the watercolour version is cat. no. 589. 

52        The mahogany support measures 88.7 x 111.7 cm overall, the main panel, strengthened with chamfered battens, measures 81.0 x 101.6 cm, the horizontal member along the top edge being 7.7 mm high, and the vertical extension to the left 10.0 cm in width. No longer possessing my original notes on the picture, I am most grateful to Stephen Wildman for having the support remeasured. 

53        Edmond About, Voyage à travers l’Exposition des Beaux-Arts, Paris, 1855, p. 32: ‘Lorsqu ’il avait manqué un morceau, il l’enlevait proprement, remplaçait le papier par une pâte de fabrique anglaise, et recommençait la besogne. Un jour, il trouva l’aquarelle fendue en deux. Il l’envoya en Angleterre, fit recoller soigneusement les morceaux et reprit la suite de son travail. Vous voyez que la fabrication des aquarelles est chose compliquée …’

54        Permission has so far been refused by the owners for the watercolour to leave Japan. 

55        Royal Academy of Arts, London, Lewis, 1978, cat. no. 597, illus. fig. 58; the watercolour version is cat. no. 599, and a smaller replica from the T. S. Kennedy collection, perhaps commissioned, was in the sale at Sotheby’s, London, 17 June 1968, (24), repr. in colour.