The Grau Framing Campaign and the Finish of Whistler’s Late Work

In 1892, in preparation for his retrospective exhibition at the Goupil Gallery, London, James McNeill Whistler undertook a systematic campaign to reframe as many of his canvases as possible. Working with the frame maker Frederick Henry Grau (1859–1895), Whistler replaced his bespoke, hand-painted frames of the 1870s with a uniform architectural design – deeply reeded, broadly moulded frames that converted the frame from an individual artistic gesture into a standardised mark of authorship. Reeding, a series of narrow, parallel convex mouldings resembling bundled reeds, gives these frames their distinctive ribbed profile. The campaign Whistler conducted with Grau between 1890 and 1895 – the ‘Grau campaign’ – was not incidental refurbishment. It was an attempt to fix, once and for all, the ideal finished state of each work: not merely how a painting should appear at Goupil in 1892, but how it should be seen in perpetuity.

The conviction behind the campaign had already been stated plainly a year earlier. In March 1891 Whistler sold Arrangement in Grey and Black, No. 2: Portrait of Thomas Carlyle 1872–3 (YMSM 137; Kelvingrove Art Gallery and Museum, Glasgow; fig.1), hereafter Carlyle, to the Corporation of Glasgow.1 Concluding the sale, Whistler assured the purchasing committee that the painting had been newly framed to his ‘own design’ and that he trusted it would remain in that frame ‘always’.2 He was doing more than negotiating a sale; he was articulating a state of completion. The Carlyle was, in this sense, a rehearsal. Within a year, Whistler would extend the same conviction to an entire exhibition. The Grau campaign was a systematic effort to create this ideal finish at precisely the moment his work began passing from private collections into public ones. While scholarly inquiry into Whistler’s ‘finish’ has traditionally remained focused on the painted surface, this article reveals that for Whistler, a painting did not reach its finished state when the paint dried. It reached its final state only when it was fixed, permanently, within a specific frame of his own design.

The frames Whistler left behind

To understand what the Grau frame replaced, one must first take the earlier frames seriously – as Whistler himself did. From the early 1870s his paintings were characteristically housed in what is defined as the ‘flat Whistler’: a variant of the cassetta, an Italian Renaissance frame type in which a wide, flat frieze is bordered by raised inner and outer mouldings. In Whistler’s version the profile was shallow, the mouldings reeded and the frieze decorated by hand in colours drawn from the enclosed painting.3 The frame surrounding Nocturne: Blue and Silver – Chelsea 1871 (YMSM 103; Tate T01571; fig.2), hereafter Chelsea, is the type’s exemplar. The frame was likely constructed by the London maker Frederick Fox and decorated by Whistler’s studio assistant Walter Greaves (1846–1930), under the artist’s supervision. Across its frieze runs a painted seigaiha (lit. ‘blue sea wave’) – a traditional Japanese pattern of overlapping concentric arcs – that continues the visual rhythm of the blue-toned twilit river Thames scene it surrounds. On the frieze, amid the waves, sits Whistler’s painted butterfly: the same monogram he placed at the base of his canvases, now applied to the frame (fig.3).

These frames were not accessories. Whistler said so explicitly, and repeatedly. Sending paintings to Paris in 1873, he warned the dealer George Aloysius Lucas (1824–1909): ‘You will notice and perhaps meet with opposition that my frames I have designed as carefully as my pictures – and thus they form as important a part as any of the rest of the work – carrying on the particular harmony throughout.’4 Five years later, he said it under oath. During his libel trial against the critic John Ruskin (1819–1900) Whistler was to explain a ‘peculiar dark mark’ on the frame of Nocturne: Blue and Gold – Old Battersea Bridge 1872–5 (YMSM 140; Tate N01959; fig.4) hereafter Old Battersea Bridge. Whistler told the court that the mark was his butterfly, and that ‘The frame and the picture together are a work of art.’5 The painted frames of the 1870s – bespoke, hand-worked and individually signed – were, by the artist’s own testimony, constituent parts of the works they enclosed.

Nor was the practice of reframing new to Whistler. Throughout his career, he habitually altered or entirely replaced the frames surrounding his own canvases, sometimes decades after a painting was first exhibited. Multiple instances can be documented across his correspondence and in the online catalogue raisonné, clustering in two distinct periods: the first in the mid- to late 1870s, and the second – the subject of this paper – beginning in 1890.6 The frame for Symphony in White, No. 1: The White Girl 1861–3 (YMSM 038; National Gallery of Art, Washington) was reworked and redecorated well after it was fitted in 1862; Harmony in Blue and Silver: Trouville 1865 (YMSM 064; Isabella Stewart Gardner Museum, Boston) passed through at least two frame states before its probable alteration under Grau in 1892.7

The earlier alterations had generally been prompted by external circumstances: a change of owner, a new setting, a shift in fashion or in Whistler’s artistic vision. What distinguishes the campaign that began in late 1890 is its motive. For the first time, Whistler set out to reframe not in response to circumstance but in an explicit bid to fix each work’s appearance permanently – and to do so uniformly. The butterfly had made each early frame an individual signature. The Grau frame would make every frame the same signature.

The campaign begins, with Valparaiso

This shift is best illustrated by the history of the Tate’s Crepuscule in Flesh Colour and Green: Valparaiso 1866 (YMSM 073; Tate N05065; fig.5), hereafter Valparaiso. The frame surrounding the painting today is the work’s third. Of the first, nothing certain is known, neither of its style nor its whereabouts. Its replacement, however, can be reconstructed with some confidence. In 1879 Whistler wrote to the painting’s owner, the dealer Charles Augustus Howell (1840–1890), insisting that ‘Valparaiso ought to be framed without a moments [sic] delay … I should never forgive you if it didn’t.’8 The frame that may have resulted was made by the London makers Foord & Dickinson in preparation for the work’s exhibition at the Grosvenor Gallery, and based on the other frames made at this time it most likely carried a hand-painted seigaiha pattern of the kind that still survives on Chelsea and Old Battersea Bridge.9

Almost immediately after the painter and collector W. Graham Robertson (1866–1948) purchased Valparaiso from Howell in November 1890, Whistler intervened. In a letter postmarked 26 November, he issued a directive that sought to strip the work of its 1870s identity:

The frame Whistler now dismissed as ‘rickety’ was the very Foord & Dickinson design he had demanded from Howell in 1879. In condemning it, Whistler was repudiating his own earlier taste. And in insisting on Grau’s precision – the sight size reduced by ‘an eighth of an inch … all round’ – he was not repairing a structural flaw so much as recalibrating the painting’s visual impact. The sight edge is the frame’s inner lip, the point at which it meets and slightly overlaps the canvas; narrowing the opening by even a fraction crops the margins of the composition and pulls the eye inward, a decision of pictorial rather than structural consequence. The letter demonstrates that the campaign of standardisation associated with the Goupil exhibition was already under way, painting by painting, nearly two years earlier.

The Valparaiso reframing signals a fundamental shift in the function Whistler assigned to his frames. Where the painted frames of the 1870s had operated as atmospheric extensions of the canvas – each surface individually decorated in sympathy with its picture – the Grau frames were instruments of standardisation and status. In abandoning the hand-painted designs of his middle career, Whistler was converting the frame from a unique artistic gesture into a uniform mark of authorship.

The Robertson reframe was not an isolated request but the opening move in a calculated campaign of persuasion. Whistler used the newly framed Valparaiso as a prototype to pressure other collectors. In February 1892 he urged his patron William Cleverly Alexander (1840–1916) to visit Robertson’s home at 23 Rutland Gate specifically to ‘see the beautiful effect of my new frame’, promising that a similar Grau commission for Alexander’s Chelsea would yield a ‘gain three times in stateliness and charm’.11 When Alexander proved resistant, Whistler appealed to the lady of the house, urging Rachel Agnes Alexander (1837–1900) to ‘persuade’ her husband and touting Grau’s ‘moderate’ charges.12 The Alexanders remained unmoved. Their refusal inadvertently preserved a rare historical record: the Alexander nocturne, now hanging in Tate Britain, is one of only three works to survive the Goupil preparations in an unaltered 1870s state – a silent witness to the style Whistler was so keen to erase.13

The accessions of 1891

The campaign acquired urgency in 1891, when two of Whistler’s most significant paintings entered public collections in quick succession. That spring, the Corporation of Glasgow purchased the Carlyle for one thousand guineas; in November, Arrangement in Grey and Black: Portrait of the Artist’s Mother 1871 (YMSM 101; Musée d’Orsay, Paris; fig.6) was acquired by the Musée du Luxembourg in Paris for four thousand francs.14 Both works were reframed by Grau in the months surrounding their sale (the Paris work ‘lined and reframed’ in direct anticipation of its accession), their earlier surrounds exchanged for the new architectural profile.

The Glasgow purchase shows how institutional expectations and Whistler’s ambitions had begun to converge. Confirming the sale on 24 March 1891, the museum’s curator, James W. Paton (1843–1921), stipulated that the work be delivered ‘in good condition, in a frame suitable for the work, and for a public gallery’.15 The institution, in other words, invited precisely the permanence Whistler wished to impose. His reply, two days later, was a declaration of the work’s final state:

Every clause performs an act of control. ‘[T]his I have seen to myself’: Whistler is his own conservator, certifying the painting’s condition. ‘I have had the picture newly framed in the frame of my own design’: the surround is claimed as a gesture of authorship, not trade work. ‘In which I trust it may always remain’: a bid to bind the museum’s conduct after the artist’s authority has formally ended. Whistler understood that accession meant the loss of his sovereignty over display; once inside a public gallery, the work would hang, travel and be treated at the institution’s discretion. The Grau frame was his answer; framing became his final act of governance.

Whistler’s contemporaries registered the change. Before the Carlyle travelled to Scotland, it was shown briefly at the Goupil Gallery, where a correspondent for the Pall Mall Gazette observed that Whistler would no longer exhibit a picture ‘until a year or more after it has been painted – until it has been varnished, in fact, and entirely completed’, noting that the Carlyle had ‘lately been varnished, put behind plate glass, and sealed up in one of the artist’s most recently-designed frames’.17 Varnished, glazed, framed, sealed: the vocabulary is that of closure. The painting had made its transition from studio object to fixed public artefact – and the frame was the instrument of that fixing.18

The man at 570 Fulham Road

The new frame was inseparable from the man who made it. The painter Mortimer Menpes (1855–1936), an assistant to Whistler in the 1880s, observed that Whistler’s tradesmen were drawn into his aesthetic as thoroughly as his pupils. Menpes noted that Whistler’s newly employed frame maker ‘very soon, fell under the influence of the Master, became an impressionist’. In Grau’s case, the transformation is not mere anecdote; it is etched into the archival record of his workshop at 570 Fulham Road.

Frederick Henry Grau was born in 1859 into a family of German cabinetmakers, and his early career was rooted in the traditional London trade. His first work for Whistler dates from 1887, and the frame on Chelsea in Ice 1864 (YMSM 053; Colby College Museum of Art, Waterville; fig.7) may be the earliest surviving example of the partnership. This frame is a wide flat Whistler, similar in design to those Whistler used during the late 1870s but devoid of any painted surface pattern. Unusually, it also retains a paper label advertising Grau’s services as a ‘cabinet maker, upholsterer, and decorator’ (fig.8), the only such label this author has encountered on a Whistler frame. The label is as telling as it is rare: it presents the frame not as part of an artistic collaboration but as the product of a general furnishing business. Over the next four years Grau’s account of himself changed. By the 1891 census, he recorded not the trade term of his family’s business but a new self-description as ‘artist in furniture’ (fig.9).19

The same shift is legible on the frames themselves. Around 1891 the commercial paper label disappears from Grau’s frames for Whistler, replaced by a hand-written pencil signature on the verso of the mouldings. The inscription on the Valparaiso frame – ‘F.H. Grau, London, 1891’ (fig.10) – suggests that Grau, like Whistler, had come to regard these frames as works of art: a tradesman labels his product, but an artist signs his work. The two men’s marks, however, now occupied different places. The painted frames of the 1870s carried Whistler’s butterfly on their public face; the Grau frames dispensed with it. Whistler’s authorship did not disappear but changed form. In place of an individual signature, the standardised profile and consistent gold of the frame itself became his mark, recognisable at a glance as a Whistler surround. Grau’s authorship, by contrast, remained on the verso, visible only to those who handle the frame. His new identity was short-lived: he died in 1895, and with him the ‘true pattern’ Whistler had entrusted to his workshop.

The campaign at scale

The campaign reached its full scale in March 1892, when the Goupil Gallery opened Nocturnes, Marines, and Chevalet Pieces.20 Organised with the gallery’s director, David Croal Thomson (1855–1930), the retrospective gathered forty-three paintings spanning nearly thirty years of Whistler’s career. The artist treated it not as a historical survey of his development but as an opportunity to present the work as a unified whole – and the frames bore the burden of the unification. During the preparations, thirty-three frames – roughly three-quarters of the exhibition – were replaced or significantly altered, an undertaking involving the restorer Stephen Richards (1844–1900) alongside Grau.21 By setting paintings of the 1860s and 1870s into the architectural profile of the 1890s, Whistler superseded his earlier – although no less intentional – frame designs with a single, uniform mode of presentation.

Whistler regarded the result as definitive. Writing in March 1892 to John Gerald Potter (1828–1908), one of his earliest patrons, who had lent paintings to the exhibition, he declared: ‘I hope you are as pleased as I am with my new frames – at last the pictures have a dress worthy their own dignity and stateliness, Wherefore you may thank me for finally inventing them – You see it takes years to know these things.’22 The word ‘inventing’ is significant, and so is the language around it. ‘At last’, ‘finally’ and ‘it takes years’ all describe a long search brought to an end. Whistler presented the Grau frame not as one design among many but as the solution to a problem he had been working at for twenty-five years, after which no further design was needed.

Potter’s pictures were not unusual: most of the works at Goupil belonged to other people, not Whistler, and the letter to Potter reveals how the campaign actually proceeded. Works borrowed for the exhibition were cleaned by Richards and reframed by Grau, and the owners were presented with the bill afterwards. Whistler’s letter to Potter does exactly this. ‘What an execrable knobly [sic] horror was round your “Blue Wave, Biarritz!”’ he wrote of the frame Grau had replaced, adding that The Little White Girl ‘also ought to have a new frame. The old one is quite too weak for the picture … Do order one from Mr Grau. 570. Fulham Road. His prices are very little.’23 He pressed the point: ‘I do trust you will keep the works in the frames I have put upon them – and so pay the little man Grau.’24

Potter refused. In a letter of 4 April 1892, written while the exhibition was still open, his son, John Charles Potter (1854–1920), objected on his father’s behalf that the pictures had been cleaned and varnished without the family’s sanction – although he conceded that they appeared ‘immensely improved’ for it.25 When the paintings were returned, the Potters declined to pay for the Grau frames.26 They were not alone. A list of lenders’ accounts sent to Whistler by the Goupil Gallery in May 1892 records six owners, among them Frederick Leyland and Louis Huth, who refused either a new frame or payment for work already done; only a handful had paid. When Potter began selling the pictures on in 1893–4, Whistler drafted a letter of magnificent sarcasm, opening with ‘Hurrah Potter! – Well done Potter!’ and accusing his ‘dear Patron, friend & speculator’ of having rejected ‘the one or two frames in which they looked their best’ and of having ‘shoved the beautiful “Blue wave” … back into the mean old abomination of years ago’.27

It is tempting to read Potter’s decision simply as an obstacle to the artist’s vision, but the conflict is better understood as a collision between two legitimate forms of authority. Three parties may claim the right to determine how a painting is framed: the artist, the owner and, eventually, the museum.28 Potter, as an owner, was under no obligation to involve Whistler in decisions about the display of his work, and his property had been altered without his consent. His refusal, moreover, had the same unintended consequence as the Alexanders’: by reverting his pictures, Potter preserved earlier states of Whistler’s framing that the campaign was designed to erase. The owners who resisted Whistler became, inadvertently, the conservators of his earlier style.

The Tate collection preserves both outcomes of this negotiation. Valparaiso, bearing Grau’s 1891 signature, represents the campaign fulfilled: a work sealed in the state Whistler intended as final. Chelsea, still in its painted frame of the 1870s, represents the campaign refused. For the modern scholar, these disparate frames are not merely decorative borders but historical documents of a negotiation over who determines an artwork’s final state.

Grau Whistler versus Grau-style

The collaborative era of the standardised frame was as brief as it was consequential. Grau died on 11 March 1895, at the age of only thirty-six, and with him died the campaign’s London operations. On learning of his death, Whistler wrote to Thomson ‘of course nothing can be done with his successor – so that’s all over’.29 His lament is revealing. If the Grau frame had been merely a design, a pattern of reeds and gilding, a successor could have inherited it. Whistler’s verdict shows that it was something else: a specific standard of execution, resident in one workshop and one pair of hands, and not transferable.

Whistler had said as much while Grau lived. Urging the dealer Edward Guthrie Kennedy (1849–1932) of Wunderlich & Co., New York, to reframe a group of newly purchased paintings in June 1892, he insisted that Kennedy go to Grau because ‘he is the only one who has the true pattern of my frame.’30 The artist’s own criterion, then, entails a distinction that the study of his frames has not always drawn. A ‘Grau Whistler’ frame is a documented product of the Fulham Road workshop – often signed or labelled – reflecting the architectural sobriety of the 1890s campaign. A ‘Grau-style’ frame follows the same deep-reeded pattern but was produced by another maker, frequently on the instruction of dealers or collectors seeking to replicate the Grau Whistler look.

The Grau pattern can be described precisely, and it is the profile, rather than the surface, that defines it. A frame’s surface is its least stable feature: gilding darkens and wears with age and pollution, and may be cleaned, regilt or toned by well-intentioned restorers and conservators. Whistler cared intensely about the colour of his gold – Grau was to use ‘exactly the same gold’ for the Valparaiso – but the gilding that survives on a frame today may reflect more than a century of alteration. The profile, the cross-sectional shape of the moulding, is far harder to change, and it is here that the ‘true pattern’ resides. The surface is fickle; the profile remains constant.

Read from the outer edge inward, the Grau profile descends in three registers from its highest point. The outer section is a broad cushion (a wide, gently rounded moulding), itself reeded, rising to a fillet (a narrow flat band) at its crest. The middle section comprises two large reeds flanked by small fillets, stepping down through graduated reeds. The inner section is a run of large reeds alternating with fillets, ending at the sight edge – the inner edge, beside the picture – in a small step or bevelled flat (fig.11).

The substitution began even while Grau lived, as the case of Variations in Flesh Colour and Green: The Balcony 1864–73 (YMSM 056; Freer Gallery of Art, Washington; fig.12) shows. The painting was among a group formerly owned by the collector George John Cavafy (1805–1891) that Kennedy acquired from his estate in June 1892.Whistler, assuming that the pictures ‘must be in hideous old things’, urged Kennedy to have them reframed by Grau.31 Kennedy ignored the instruction and shipped the canvases to New York unframed. When Whistler’s wife, Beatrice (1857–1896), pressed him – ‘Why didn’t you have new frames for them all[?]’ – Kennedy explained that his firm made its own frames, saving on import duty and producing ‘a better article, or rather one which won’t split or crack in our climate’.32 He then asked, with a dealer’s bafflement at the artist’s priorities: ‘But why this anxiety as to frames and not a single word as to the paintings themselves?’33 The question shows that Kennedy did not share Whistler’s conviction that the frame was part of the work. Yet he did not disregard Whistler’s wishes entirely: The Balcony was given a Grau-style frame, made in New York by Kennedy’s own framers rather than by Grau. The painting remains in this frame, correct in pattern but not from Grau’s workshop. And because Grau’s mark, when it exists at all, appears in pencil (or marker) on the verso, the difference is invisible from the gallery floor.

Authentic Grau frames are remarkably rare. The reeded gold surround has become the global standard for displaying Whistler – yet the great majority appear to be Grau-style frames or modern reproductions. The signed survivors, such as Valparaiso, are a finite corpus, and a precisely datable one: a documented Grau frame can date only from the window between the campaign’s beginning in 1890 and the workshop’s dissolution in 1895. In a field where, as we have seen, the date of a frame can sit decades from the date of its painting, Grau’s short life supplies what Whistler scholarship frequently lacks: a fixed point.

Freer, Hanna and the Little Blue Girl

Grau’s death ended the campaign’s London operations, but not the campaign. After Whistler’s own death in July 1903, it continued across the Atlantic, in Detroit, where the collector Charles Lang Freer (1854–1919) had assembled what would become the largest Whistler collection in the world. In September 1903, two months after Whistler’s death, Freer reported on its condition to Rosalind Birnie Philip (1873–1958), Beatrice’s younger sister and now the artist’s executor:

The ‘capital workman’ was the Detroit frame maker James E. Hanna (1858–1920), who reframed nineteen of Freer’s Whistlers to the Grau pattern.35 The arrangement deserves a moment’s scrutiny, because materially it repeats exactly what Kennedy had done a decade earlier: Freer had frames produced in the United States to replicate the London product. The difference lay in intent. Where Kennedy had treated the frame as a practical matter of cost and climate, Freer treated it as a condition of the work’s integrity, reproducing Whistler’s ‘standard form and colour’ under his personal supervision. The Freer collection’s celebrated uniformity is, in the strict sense established above, Grau-style rather than Grau, with the pattern faithfully continued after the workshop Whistler had trusted to execute it was gone. Whistler’s ‘finish’ survived in America as a practice of stewardship rather than a product of the workshop.

Amid Hanna’s reframing, one painting remained untouched: Harmony in Blue and Gold: The Little Blue Girl 1894–1903 (YMSM 241; National Museum of Asian Art, Washington; fig.13). It is the campaign’s definitive exception. Built by an unknown maker, the frame is a standard Grau-style reeded cushion, to which Whistler did something he had long since abandoned: he painted the frame, adding a blue and gold chequerboard pattern and his butterfly. The Whistler scholar David Park Curry observed that the chequerboard echoes the rug beneath the model’s feet, creating a ‘carefully integrated pair’.36 But a purely aesthetic reading misses the chronological weight of the decision. Why, after years of insisting on the sober architecture of plain gold, did Whistler return to the painted frames of the 1870s for this work?

The answer lies with Beatrice. She died in 1896, and the painting, which Freer had commissioned from Whistler in 1894, became a vessel for his grief. ‘I write to you many letters on your canvas!’ he told Freer in March 1897, ‘and one of these days, you will, by degrees, read them all, as you sit before your picture.’37 If the canvas was a series of letters, the frame became the envelope. Throughout his widowerhood, Whistler’s correspondence was edged with a thick black mourning border (fig.14), and the painted frame of The Little Blue Girl may be read as that border’s translation into blue and gold. In this final frame Whistler set aside his own rules, joining the stateliness of the 1890s profile to the painted surface of the 1870s. Freer understood. Having once asked Whistler to ‘hurry the picture’ to save on duty, he withdrew the request in 1897 and promised that the painting would ‘rest with its own beautiful kind’ – and when Hanna reframed the collection, The Little Blue Girl was left as Whistler had made it.

Conclusion

The framing history of Whistler’s late career shows that his pursuit of ‘finish’ did not end at the edge of the canvas; for Whistler the frame was not applied to a completed work but was itself the act of completion, the moment a studio object became a fixed and public one. The campaign he conducted with Grau pursued that fixity systematically: painting by painting, with Robertson and Valparaiso; institution by institution, with Glasgow and the Luxembourg; and wholesale at Goupil in 1892, where nearly the entire retrospective was reframed to a single standard. The bespoke, painted, individually signed surrounds of the 1870s gave way to a uniform architectural profile, each frame no longer an individual signature but the same signature, repeated.

Yet permanence proved to be the one thing the campaign could not manufacture. Potter reverted his pictures to their older frames; the Alexanders never yielded theirs – and the resisting owners thereby became the accidental conservators of the very styles Whistler sought to erase. Institutions, too, altered the ‘sealed’ surrounds: at the Tate, glazing doors were fitted to Whistler’s frames within years of accession, conversions the gallery has since revisited. And even the campaign’s most faithful continuation was a substitution: the celebrated uniformity of the Freer collection is the result of Hanna’s Grau-style work, the pattern perpetuated after the workshop that defined its standard had closed. Grau’s death in 1895 fixed the authentic corpus forever; of this, Valparaiso at Tate is a rare survivor.

The reeded gold frame that Whistler designed with Grau succeeded absolutely as an image: it remains, worldwide, the definitive dress of a Whistler. But nearly everywhere it survives as reproduction – Grau-style, not a ‘true’ Grau. It is a version of what Whistler wanted for posterity, rather than the product he certified ‘in absolutely perfect condition’ and trusted would ‘always remain’. The frames themselves, in all their states – signed by Grau, faithful copy, reverted ‘abomination’, painted exception – record something the pattern alone cannot: a negotiation over authority in determining an artwork’s final state, in which the artist’s control was exerted, resisted and ultimately outlived by the stewardship of owners and museums. For Whistler, a work was finished only when finally, properly enclosed.