Process, Temporality and Material Experiment in Whistler’s Twilight Thames Nocturnes of the 1870s

From night scenes to nocturnes

Between 1871 and 1878 James McNeill Whistler painted seven atmospheric twilight scenes of the River Thames in London. These form a distinct group within his wider series of nocturnes and can be identified by titles incorporating either ‘Blue and Silver’ or ‘Blue and Gold’. They differ from Whistler’s seascapes and from Thames nocturnes described in other colours.1 Viewed chronologically, the twilight nocturnes not only map the progressive descent of dusk into night, they record a technical evolution. By capturing twilight as a gradual process, in which light and shadow fade unequally and forms slowly dissolve into darkness, Whistler captures perception unfolding over time.

Beginning with Nocturne: Blue and Silver – Chelsea 1871 (YMSM 103; Tate T01571; fig.1; hereafter Chelsea), they proceed through Nocturne in Blue and Silver 1871–2 (YMSM 113; fig.2), Nocturne: Blue and Silver – Cremorne Lights 1872 (YMSM 115; Tate N03420; fig.3; hereafter Cremorne Lights) and Nocturne: Blue and Gold – Old Battersea Bridge c.1872–5 (YMSM 140; Tate N01959; fig.4; hereafter Old Battersea Bridge). The series culminates with three paintings depicting the stretch of the Thames between Battersea Bridge and Wandsworth Bridge: Nocturne: Blue and Silver – Battersea Reach 1875–8 (YMSM 119; National Museum of Asian Art, Washington; fig.5), Nocturne in Blue and Silver - Battersea Reach 1875–8 (YMSM 151; Yale Center for British Art, New Haven; fig.6) and Nocturne: Blue and Silver – Battersea Reach 1875–8 (YMSM 152; Isabella Stewart Gardener Museum, Boston (ISG); fig.7) – hereafter distinguished by institution.

Across this series, the distinctions between water, sky, architecture and boats become increasingly compressed. In the later paintings discernible brushwork is reduced and forms blur into indistinct silhouettes, within a predominantly continuous field of colour, bringing Whistler closer to achieving the pictorial harmony he sought. The changing treatment of dusk across the series is accompanied by corresponding developments in Whistler’s technique: a sustained process of experimentation emerges in which each nocturne informs the next.

Whistler’s ambition for his twilight nocturnes can be understood in relation to an established tradition of moonlight paintings, including river scenes on panel by Aert van der Neer (c.1603–1677; fig.8), J.M.W. Turner’s Moonlight a Study at Millbank 1797 (fig.9) and the softly modulated effects over a warm coloured imprimatura – a thin coloured preparatory layer applied to the ground before painting on it – of Charles-François Daubigny (1817–1878; fig.10). Although all offered precedents for representing diminished light and atmosphere, Whistler increasingly suppressed motifs in favour of colour and tone as the unifying elements of the composition. While his ambition ‘to only bring about a certain harmony of colour’ brought him into dialogue with developing Impressionist interest in the transient effects of light,2 Whistler’s process of observation, recollection and making diverged from their emphasis on capturing the immediate impression of a given moment.3

Whistler’s familiarity with the Thames was grounded in prolonged observation. A drawing by his assistant Walter Greaves (1846–1930) depicts the artist in a boat looking out onto the river with Old Battersea Bridge in the background (fig.11). The Greaves family had longstanding ties to the river: Walter and his brother Henry (1844–1904), who assisted Whistler with domestic and studio tasks, accompanied him on the Thames where their father had once rowed Turner. Whistler’s friends and biographers Elizabeth Robins Pennell and Joseph Pennell recalled that Whistler would set out with Greaves at twilight and remain ‘on the river for hours, often all night, lingering around the lights at Cremorne, drifting into the shadows of the bridge.4 They also described the nocturnes as products of recollection: ‘it was impossible for him to work in colour at night, and he had to trust to his memory’, while ‘the long nights on the river were followed by long days in the studio’.5

This distinction between Whistler's experience of twilight, his memory and the later painting of it was reinforced by the lithographic printer Thomas Robert Way (1861–1913) who recalled how Whistler, having studied a river scene, turned away and asked, ‘Now see if I have learned it’ before reciting its details from memory.6 Such accounts suggest that memory was not merely a substitute for painting directly from nature, but a means of selecting and organising a visual experience.7 As the Pennells observed, Whistler would retain ‘the arrangement, the scheme of colour, and as much of the detail as he wanted’.8 The nocturnes thus captured a temporal process that differed from impressionist practice as they did not record a single moment but the layered duration of accumulated looking and memory.

In the mid-nineteenth century the Thames was not considered a place of poetry or unnoticed beauty. Even upriver near Chelsea, where Whistler first lived in 1866, the view would have often been obscured by smoke from coal-powered factories and chemical works on the south bank (fig.12). The river also carried factory effluent and sewage downstream towards the centre of the city and became noxious in summer.9 In the City of London, smoke from coal fires in houses and offices thickened the air every winter, creating a fog that could be seen on still days. Whistler described the river being perpetually obscured by ‘the blurring of distant objects, and the complete veiling on nine days out of ten of everything beyond two miles’, an atmospheric quality that would come to define his Thames nocturnes.10 Before the construction of the Chelsea Embankment in 1874 the river was wider than it is now, its flow more sluggish and it lacked the tidal range seen today. Old Battersea Bridge, a reoccurring subject for Whistler, was eventually demolished in 1885 to make way for a new bridge that would meet the needs of the growing population and increased traffic.

By twilight, the same conditions that made the river an unpromising subject offered Whistler a setting that could transform under its changing conditions. The Thames seen from Chelsea has been described as a dark, still expanse on which the intermittent illumination of the gas lamps was reflected, with the river itself appearing to absorb the light (fig.13).11 Whistler preferred the calm of the river at night to the bustle of the Thames by day, describing in his ‘Ten O’Clock’ lecture how ‘The evening mist clothes the riverside with poetry, as with a veil, and the poor buildings lose themselves in the dim sky’.12 One century earlier the painter Joshua Reynolds described the challenge of painting indistinct forms and optical effects seen at twilight, describing how the eye perceives a scene as a whole ‘without seeing any one of the parts distinctly’.13 Scientists in Victorian England increasingly sought to explain human perception of twilight’s optical phenomena – in an 1869 lecture the physicist John Tyndall demonstrated that finely dispersed particles could scatter light to produce a blue appearance, contributing to a growing understanding of atmospheric colour.14 For Whistler, the instability of twilight offered a pictorial opportunity, as forms weakened and colours shifted, tone could unify the composition, while perception dissolved into imagination.

Throughout the nineteenth century, paintings of reduced light were commonly identified through descriptive terms such as ‘night piece’ denoting a subject depicted after dark, and ‘moonlight’ referring to a naturally lit evening.15 Whistler’s earliest Thames paintings were initially presented in this inherited language, for example, Chelsea was exhibited at the Dudley Gallery in 1871 as Harmony in Blue & Green – Moonlight.16 The title retained a conventional description of the lighting effect, while ‘harmony’ shifted attention towards the organisation of colour.17 Whistler’s mother, Anna Matilda (1804–1881), referred to Chelsea simply as a ‘moonlight picture of the Thames’.18 As Chelsea, the first of Whistler’s nocturnes, was neither conceived or initially exhibited under the title, it marks a point of transition, preceding his adoption of the term ‘nocturne’ in public from 1872.19 The name was inspired by his patron Frederick Richards Leyland (1831–1892), whom Whistler thanked that year ‘for the name “Nocturne” as a title for my moonlights’, remarking that it ‘does so poetically say all I want to say and no more than I wish’.20 Whistler drew upon the term’s established musical associations with night, lyrical expression and tonal unity. Contemporary response confirms this was perceived as novel. The Echo remarked upon the use of ‘symphony’ or ‘nocturne’ as titles ‘more usually applied to musical pieces than to works of pictorial art’.21

Whistler’s adoption of ‘nocturne’ gave expression to ideals central to the emerging Aesthetic movement, in which relationships between colour, tone and form were increasingly valued as the basis of pictorial harmony rather than of narrative subject. Reflecting on the name ‘nocturne’ during his 1878 libel trial against the art critic John Ruskin, Whistler explained that he wished ‘to indicate an artistic interest alone’, thereby ‘divesting the picture of any outside anecdotal interest which might have been otherwise attached to it’.22 He defined a nocturne principally as ‘an arrangement of line, form and colour first. The picture is throughout a problem that I attempt to solve’.23 This concern with pictorial unity was echoed in his assertion that ‘as music is the poetry of sound, so is painting the poetry of sight’ and that ‘the subject-matter has nothing to do with harmony of sound or of colour’.24 Reviewing Whistler’s nocturnes in 1872, the Athenaeum recognised this shift, reporting: ‘the deep, warm twilight, which obscures all things here’ while attributing the pictures strength ‘not in its verisimilitude, but in the super-subtle arrangement and exquisitely delicate nature of its colour’.25

Envisaging a twilight nocturne

Before Chelsea, Whistler had already explored the atmospheric possibilities of the Thames in his etchings. In Battersea Dawn (Cadogan Pier) 1863 (fig.14), we see the beginnings of subdued tonal contrasts and later in Chelsea and Church 1871 (fig.15) the use of the tone of the paper as a unifying device and softened distinctions between river and sky, all of which anticipate aspects of his 1870s nocturnes and other works on paper. Whistler’s earlier technical explorations informed a developing approach to painting. This is evident in Symphony in Grey: Early Morning, Thames 1871 (YMSM 098; figs.16a–c), which presents a muted view across the Thames towards Battersea at dawn. Painted in shades of grey, it is characterised by thin paint films through which the canvas texture remains visible and the reddish preparatory layer begins to contribute to the overall tonal effect (fig.16b). Broad lateral strokes extend across the river and sky, while smaller touches articulate buildings, boats and smoke (see fig.16c).

This marks a technical change from many of Whistler’s paintings of the 1860s, in which opaque layers or impasted passages coexist with thinner applications, creating a more varied surface texture, seen in Blue and Silver – Trouville (fig.17) and Harmony Blue and Silver – Trouville (YMSM 064; Isabella Stewart Gardner Museum), both painted in 1865. In these earlier works, sky and water are distinctly separated by both line and tone, and the widespread network of drying cracks reflects Whistler’s experimentation with paint applications. Selective wiping back reveals glimpses of the coloured preparation and canvas weave. In Early Morning, Thames, the more even application of thin paint gives greater prominence to the imprimatura, which together with the restrained tonal range establishes a subtle continuity between the sky and water, contributing to a more unified pictorial surface. Whistler would develop this approach to painting further in his twilight nocturnes.

The earliest nocturnes do not appear to have been painted in response to commissions. Anna Whistler’s account of Chelsea describes a spontaneous beginning, arising from an evening spent on the Thames. In pursuing his own pictorial interests, Whistler had scope to experiment with his materials and methods. In the 1878 trial, when asked to describe making a nocturne, he stated that he ‘completed the mass of it in one day after hav[in]g arranged it in [his] mind’.26 The ‘mass’ here describes the establishment of a continuous tonal envelope in which sky and water were fused, conceived before the painting began. His pupil Mortimer Menpes later recalled Whistler studying nocturne scenes until their colours and forms were fixed in his memory, before painting ‘without a single note’, suggesting that he approached the canvas with the composition already established in his mind.27

Technical evidence shows that, although sometimes disparaging, contemporary critics characterised his approach with a degree of accuracy. In October 1872 the London Echo reported that a nocturne was made by ‘one or two broad sweeps of monotint cover[ing] the canvas from end to end’, followed by ‘small streaks of light and dark paint … touched on here and there’.28 Although probably intended satirically, the observation corresponds closely to the broad tonal fields and sparing touches of paint revealed by technical examination. Reviews in the Athenaeum, the Graphic and regional newspapers likewise observed that detail appeared only ‘by suggestion’ and that the image only fully emerged after prolonged looking.29 Yet these accounts reveal little of the preparation, drying intervals and successive working sessions that shaped the paintings’ surface finish.

Walter Greaves recalled that he and his brother assisted Whistler in the studio, including preparing colours and canvases while leaving the finishing of the pictures to the artist.30 The Pennells also recorded that ‘some [nocturnes] were laid aside to dry slowly in the studio, some were put in the garden or on the roof to dry quickly’.31 This division of labour provides context for Whistler’s testimony at the libel trial when he described completing a nocturne in ‘a couple of days … one day to do the work and another to finish it’32 while acknowledging ‘I do put my paintings in the open air so they may dry well as I go on with my work.’33 While the principal applications may indeed have been rapid, they formed part of a longer sequence of preparation, drying and adjustment. Studio trials using the same materials as those found by analysis in the nocturnes demonstrated that preparation and drying could extend over over several days, while the subsequent paint layers could be applied relatively swiftly. The ‘finish’ of the nocturnes, perceived as ease and immediacy, concealed a more sustained and carefully managed process of making.

Painting a twilight nocturne

The support

Whistler painted his first nocturne, Chelsea, on panel. Its rigid, comparatively smooth surface provided a stable support that enabled the thinned paint to flow evenly, its fluidity sometimes requiring the support to be laid ‘flat’.34 The panel’s smoothness, however, imposed limitations: long, horizontal sweeps of paint remain visible across the surface, leaving ribboning marks and limiting their optical fusion.35 After selecting panel supports for both Chelsea and Blue and Silver (Harvard Art Museums; fig.18), Whistler painted the remaining Thames twilight nocturnes almost exclusively on canvases with a plain weave, ranging principally from fine to medium weave densities (fig.19).36 He also adopted fairly consistent dimensions for the Thames twilight nocturnes, reserving larger formats for seascapes.37 The transition from panel to canvas altered the possibilities for manipulating the painted surface without fundamentally changing the underlying approach to its preparation. Canvas presented a textured network of raised threads and intervening spaces that could be exploited by applying paint thinly enough to preserve the topography of the weave or by locally removing paint to reveal the weave tops, creating a broken paint film that diffused light in a way not possible on a panel support.38

Light grounds

Whistler’s use of light preparatory layers, followed by coloured imprimature, was equally strategic. The first four nocturnes began with a lead white-based ground in oil, applied with a broad brush, leaving pronounced brush striations visible in X-radiographs.39 In the landscape orientation of the nocturnes these strokes appear predominantly vertical (fig.20).40

In the case of Chelsea, cross-sections indicate two preparatory layers consisting of a white commercial priming and an artist-applied imprimatura with a small addition of bone black producing a pale off-white or very light grey layer (fig.21a). This pale grey imprimatura is consistent with the grounds found on Blue and Silver (Harvard Art Museums; fig.21b), where coarse lead white is extended with chalk and barium sulphate, Cremorne Lights (fig.21c) and Old Battersea Bridge (fig.21d), the latter of which in addition to traces of bone black, contains brown and yellow ochres. Rather than indicating identical formulations, these preparations point to a preference for pale, oil-bound surfaces as the starting point for subsequent paint layers.

Coloured preparation layers

In addition to the light ground, Whistler applied a thinner, coloured imprimatura to establish the tonal and chromatic field of the painting before the principal image was developed. In Chelsea, the imprimatura extends across the panel, providing a continuous mid-tone beneath the overlying blue paint. Its role was not simply preparatory; its warmer colour intensified the cooler passages, preventing the surface from appearing tonally flat or uniform. The first two nocturnes on panel have a rich, deep grey-brown imprimatura (fig.21a and b), while the remaining five paintings, all on canvas, have exclusively dark grey tones (fig.21c and d). Walter Greaves later recalled that, for Whistler’s blue nocturnes ‘the canvas was covered with a red ground, or the panel was of mahogany … the red forc[es] up the blues laid on it’ (fig.22).41 The Pennells likewise recorded Whistler’s insistence that the tone of the preparation should correspond to that of the subject, noting that ‘if the night was grey … the sky is grey, and the water is grey, and, therefore, the canvas must be grey.’42

Owing to the importance of this tone and its dominant role in the painting, Whistler took care in choosing the colour, even if it was applied by Walter Greaves, who recalled ‘the grey distemper ground which [Whistler] so liked working upon’.43 However, use of distemper, which is water-based, is not supported by materials analysis of the Tate nocturnes. Technical examination of the three Tate twilight nocturnes identified oil-bound grounds containing predominantly lead white, with coloured imprimature applied on top.44 Greaves’s recollection is evidence of Whistler’s preference for grey-toned preparatory surfaces, rather than a guide to their exact material composition.

Although consistent in stratigraphy, the preparation layers of the seven twilight nocturnes vary in tone and application, evolving over the series. The two panel paintings have oil-bound coloured imprimature made from oil tube paint containing finely ground lead white, fine bone black and vermilion. In Chelsea Whistler adjusted the tone with red lake, while in Blue and Silver (Harvard Art Museums) he used duller red iron oxides, the latter also containing chalk and barium sulphate in oil.45 These preparatory layers were applied thinly but are opaque and relatively thicker than the upper layers of paint. There are no notable drying cracks present on these twilight nocturnes, indicating that the preparatory layers had dried sufficiently before subsequent layers were applied, thus providing stable surfaces for further work and perhaps evidence of a technical lesson learned from earlier paintings.

Studio trials investigated how the materials and layer structures identified through technical analysis might behave in practice. For the tests on panel, a wooden panel was sanded, sized with rabbit skin glue and primed with a lead white oil tube paint.46 After the ground had become touch-dry within a couple of days, the pigments identified in the imprimature were mixed in different ratios to produce a spectrum of rich grey-brown tones.47 Even the slightest addition of red lake, red oxide or vermilion caused marked shifts in the warmth of the mixture, with the final colour resembling the depth and tonality of a mahogany panel (fig.23). When covered by the thin blue paint, these variations altered the depth and tone of the surface colour. Differences in viscosity and opacity of the imprimatura also affected the handling and resulting surface finish.

Whistler adjusted his preparation for his first nocturne on canvas, Cremorne Lights 1872. Over its white ground, a thin ‘wash’ of medium-rich oil containing solely coarse bone black particles forms a translucent dark grey imprimatura (fig.21c).48 In the corresponding studio trials, a medium-weight linen canvas was stretched, sized with rabbit-skin glue and primed with lead white oil tube paint. Ivory black oil tube paint diluted with a mixture of turpentine and linseed oil created a comparable medium-dark grey preparation (fig.24).49 The other Thames twilight nocturnes on canvas consistently have dark grey imprimature over white grounds.50 In Old Battersea Bridge, the coloured imprimatura contains lead white and bone black in oil, producing a dark grey-brown. Its tone is reminiscent of the ‘chocolate brown’ colour (fig.25) that the diplomat and collector John Chandler Bancroft (1835–1901) described to Thomas Armstrong for Battersea Reach (Smithsonian).51 In these later works, the darker imprimatura assumes a more prominent role within the composition, extending beyond its function to bring tonal unity and contributing to the final image found in silhouettes, foreground or motifs such as boats and buildings (fig.26).

Equally crucial to Tate’s twilight nocturnes was Whistler’s increasingly sophisticated use of localised coloured underlayers, applied thinly over the imprimatura to selected areas of the composition to create a distinct optical contribution to the surface colour. Absent on Chelsea, this practice evolved in complexity from a single, pale grey tone applied on Blue and Silver (Harvard Art Museums) and Cremorne Lights, used in both to establish a lighter river foreground extending to the horizon, to the multiple and tonally distinct coloured underlayers discovered in Old Battersea Bridge, indicating a clear planning stage and compositional development (fig.27). High-resolution digital microscopy, supported by cross-sections of samples from corresponding areas (fig.28), identifies these layers beneath the final image: the dark grey imprimatura (fig.28a), a pale blue-grey to outline the bridge (fig.28b); a broad, deep red brushstroke applied directly underneath the bridge, warming the surrounding sky next to the descending fireworks (fig.28c); a pink-grey layer above the horizon to structure the middle sky, creating the effect of residual sunlight (fig.28d), and finally a pale grey to lighten the water (fig.28e).

These mixed colours were made in all cases from lead white and bone black, with the addition of Prussian blue for the blue-grey; red lake and vermilion for the deep red; red lake, yellow ochre and brown ochre, for the pink-grey; and yellow ochre, brown ochre (and possibly indigo) for the pale grey. The bridge and figures immediately above it were partially left in reserve or wiped back, making the imprimatura, integral to the final image (fig.29).52 Although largely concealed beneath the finished surface, these coloured underlayers reveal the extent to which Whistler anticipated the distribution of light, form and placement of preparatory colours to shape the composition before applying the unifying blue paint on top.

The blue veil

In the seven Thames twilight nocturnes, Whistler’s approach to establishing the principal tonal relationships shows notable consistency (fig.30). After the preparatory layers were well dried, he applied highly fluid blue paint with wide, long-haired brushes in broad, horizontal sweeps across the surface. Where the dragged bristles deposited paint unevenly, the darker imprimatura peeks through (fig.30a–c) or is made visible where Whistler wiped paint back (fig.30d–f), introducing subtle warmth and tonal variation within the predominately cool palette. These applications created a thin, semi-transparent ‘veil’ of blue that unified sky and water into a single continuous atmospheric field. By doing so, Whistler realised in paint his conviction that pictorial harmony arose from the arrangement of colour rather than articulated motifs or details.

Way recalled that Whistler thinned paint until it was ‘almost as thin as water’ – so fluid that drips over the edges of Chelsea confirm the panel was laid flat during the painting process.53 Material analysis of the blue veil identifies a restricted palette dominated by blue pigments, cobalt or Prussian blue, modified with lead white and small amounts of vermilion, used sparingly to introduce warmth within the predominately cool colour. The same pigment combinations appear consistently across the blue twilight nocturnes although their precise distributions and proportions vary between works.54

The fluidity of these applications was achieved at least partly through dilution with turpentine, a practice that, if taken to extremes, carries the risk of poor wetting out or loss of saturation, producing excessively matt surfaces, yet this is not a surface feature of these nocturnes.55 Whistler was able to control the translucency, flow and sheen of his paint by modifying it with his self-developed medium, which he referred to as his ‘sauce’.56 This is better understood as an adaptable medium than as a fixed recipe, adjusted to achieve particular effects of translucency, saturation and finish.57 Research by Georgina Rayner and Saffie Patel, reconstructing Whistler’s paints using Rayner’s analysis of his paint medium in the Harvard nocturnes, suggests Whistler at times added resin to his oil paint to adjust its handling and optical properties.58

The studio trials demonstrated the close relationship between paint consistency, brush pressure and resulting marks. The paint needed sufficient fluidity to travel across the surface in a continuous stroke, retaining enough body to prevent uncontrolled spreading (fig.31). To achieve this the oil paint was thinned with turpentine and a small amount of linseed oil, which produced a fluid, translucent paint that retained sufficient body and cohesion for controlled brushing. Using stiff brushes, comparable to those that survive from Whistler’s studio, it was possible to produce long, ribbon-like bands in a single stroke across the length of the support.59 Their relatively even width and uninterrupted edges depended on maintaining consistent pressure and movement (fig.32). The trials also revealed the limitations of repeated working. Further brushing disturbed the translucent veil, increasing its opacity and compromising the transparency essential for achieving the twilight effect.

The coloured preparations beneath the blue veil are integral to a nocturne’s appearance. The function of the reddish and brown-grey imprimature was to create warmth and depth beneath the cooler blue surface layers, making the blues appear luminous rather than flat and producing the effect of a twilight sky in which the last traces of daylight persist as darkness deepens (fig.33a and b). The dark grey imprimature established a lower tonal register from which lighter passages could emerge, contributing to the impression of encroaching darkness without requiring successive opaque applications (fig.33c and d). Across the five canvas nocturnes, Whistler increasingly made use of the dark grey imprimatura, coloured underlayers or tops of the canvas weave as a visible component of the image by subtracting the upper layers of paint or leaving it partially in reserve for key compositional features and in the foreground (fig.34).60 This interaction between preparation and the veil was central to Whistler’s developing technique, maintaining continuity of tone across the sky and water.

Subtraction, iteration and reworking

Following the application of the blue veil, Whistler developed his compositions through successive additions, adjustments and selective removal of paint, returning to them over several sessions. The Pennells recalled that ‘If the tones were right, he took them off his palette and kept them until the next day, in saucers, or gallipots, under water, so that he might carry on his work in the same way.’61 Storing prepared colours beneath water limits the formation of a dried surface skin, allowing the paint mixtures to be preserved and reused between painting sessions.

Over the translucent veil, Whistler articulated skylines in lighter blue paint and by selectively exposing the darker preparations beneath. Whistler worked from broader continuous strokes in the foreground towards smaller, more varied strokes in the sky, contributing to the impression of spatial depth. Boats and reflections were reinforced with thin touches of grey or blue-grey, while darker blues gave definition to the water alone. This supports the Pennells’ account of Whistler preserving his prepared colours, enabling him to use them as a tonal base for successive layers. Applied wet-in-wet within each session, these later passages are generally smoother and more opaque than the underlying veil, as seen in Chelsea (fig.35). Using a variety of smaller, stiff, brushes, Whistler could create waves or suggest spatial depth in the sky with relatively few strokes (fig.36).

Technical examination of the paintings considered alongside the studio trials suggest that Whistler used small brushes to add touches as well as subtract the still workable paint to carve out details within the horizon, physically ‘pushing back’ the blue veil to allow the silhouetted skyline to emerge from the darker preparation beneath (fig.37). Broader passages may have been wiped back with a rag dampened with mainly turpentine. The expression ‘scraping down’, sometimes used to describe Whistler’s technique, implies a more forceful removal than physical evidence suggests for these nocturnes; ‘rubbing’ or ‘wiping’ is more representative of the surface effect where the weave is visible by subtraction of the paint.

Across the Thames twilight nocturnes, Whistler developed a pictorial device described by the conservator Joyce Hill Stoner as ‘weavism’, exploiting the visible canvas texture as a part of the finished surface.62 In the nocturnes the canvas contributes to the appearance of the image in different ways; it is revealed through localised rubbing back of the upper paint layers, exposing the coloured preparation or raw canvas, or remains visible beneath the sufficiently thin preparation and paint layers for the overall weave pattern to show through. Some areas were left in reserve, while others were wiped back without disturbing the imprimatura. In the studio trials the most convincing results for creating the effect of weavism occurred when touch-dry paint was rubbed with a cloth dampened with turpentine: if the paint was too fresh, it smeared, but when fully dried, it resisted removal (fig.38). Whistler used weavism to create a unifying effect across his nocturnes while preserving subtle tonal variation, the network of canvas texture reflecting light at multiple angles and contributing to the soft, diffuse atmospheric quality fundamental to his twilight aesthetic.

The interaction between thin surface paint, coloured underlayers and canvas texture becomes progressively more pronounced across the Thames twilight nocturnes. In the earlier nocturnes, paint was removed locally to expose the imprimatura, coloured preparation or weave of the canvas (fig.39a and b). Increasingly, subtraction served not merely to adjust the composition but as a deliberate means of modulating tone, opacity and to delineate motifs. This effect is particularly apparent in Old Battersea Bridge, where several thin layers of different preparation colours exist beneath the abraded blue surface (fig.28). In places, small nubs of the underlying colours emerge as a result of Whistler having wiped back to the canvas weave, introducing subtle tonal variation without disrupting the overall atmospheric unity.

In the later nocturnes, weavism becomes increasingly prominent and subtraction was controlled so that it stopped when it reached the imprimatura without exposing the raw canvas. This is evident in the three twilight Battersea Reach nocturnes, where the paint layers are so thin that the entire weave is able to contribute to the surface appearance (fig.39c–e). From the earlier panels to the canvas nocturnes Whistler’s brushwork decreases in prominence, as the interaction between the thin paint, coloured imprimatura and support texture assumes a greater role in establishing tonal variation and atmospheric depth.

Considered alongside nineteenth-century painting practices, Whistler’s methods reveal both affinities with, and differences from, those of his British and French contemporaries. The Pre-Raphaelite William Holman Hunt (1827–1910) advocated the use of brilliant white grounds and pure pigments to maximise luminosity and colour intensity,63 while some French Impressionists preferred pale grey or cream imprimature, establishing broad relationships of colour and form through an initial ebauche or painted lay-in, as seen in Cezanne’s Auvers, Panoramic View 1873–5 (Art Institute of Chicago).64 These early applications could remain visible in the finished painting, contributing to its tonal and chromatic effects. Whistler exploited the underlying colour but achieved a more subdued luminosity through darker imprimature, locally varied coloured underlayers, translucent blue paint and selective removal. Rather than serving as a foundation, these preparations were integral to the finished image, modifying the appearance of overlaying colours or emerging through selective exposure.

A perfectly uniform surface finish, devoid of texture, was neither Whistler’s aim nor is it a characteristic of the finished nocturnes. Examination using transmitted light reveals the thin paint (fig.40), and with raking light, reveals the grain of the panels and the weave of the canvases, together with variations in the relief of the preparatory layers and brushwork, traces of paint subtraction and evidence of earlier designs (figs.41 and 44a). In Nocturne in Blue and Silver (Harvard Art Museums), an earlier skyline executed during an initial stage of painting was abandoned before the paint dried. Short vertical incisions, possibly made with the end of a brush, cut through the grey paint to expose more of the dark imprimatura beneath (fig.42). Whistler then turned the panel upside down and developed the present composition over the earlier design.

A comparable reuse of the support is evident in Cremorne Lights. Infrared traces of the earlier design persist in faint relief. Infrared reflectography (IRR) and X-radiography reveal an earlier figurative composition, including the outlines of a woman possibly in a Japanese inspired robe, a second less distinct figure and a fan motif (fig.43). Much of the composition was substantially rubbed down before the nocturne was painted, although traces of the earlier design persist in faint relief (fig.44a). The canvas was extended on the right edge by 25 millimetres at its widest point (fig.44b); this alteration appears to predate the earlier figure composition in terms of layer structure. In Old Battersea Bridge the canvas was considerably narrowed after the initial blue veil had been applied, with the scene developed within the newly reduced format: at least 10–12 centimetres of excess canvas remains folded around the back of the stretcher, partially painted (fig.29b). These examples demonstrate the range of Whistler’s revisions: from local incision and the removal of the paint, to reuse, inversion and physical reformatting of the picture plane.

Together, evidence from X-radiography, IRR, transmitted light and raking light examination reveals the complexity of these interventions. Some passages suggest that Whistler adjusted or removed paint before it had fully dried, reconsidering local relationships of colour and form without abandoning the composition. Other areas preserve evidence of successive applications and partial exposure of earlier coloured preparation layers. Whistler may have made multiple attempts at the image, wiping away unsuccessful passages until the desired balance emerged. The apparent economy of the nocturnes should not, therefore, be mistaken for an absence of preparation or revision. The illustrator George du Maurier (1834–1896), who had known Whistler since their student days in Paris, recalled that ‘no difficulty discourages him’ and he was ‘fastidious with his work’.65 This reinforces the idea of disciplined, systematic practice concealed beneath surfaces often mistaken for spontaneity.66 Greaves also recalled that ‘Mr Whistler made us paint those nocturnes; he would have us paint them’, suggesting that aspects of his method were sufficiently established to be repeatable and dependent upon controlled preparations rather than individual inspiration.67 Neither spontaneous nor a linear procedure, the paintings reveal an adaptable process in which addition, subtraction and the physical properties of the support were used to refine the relationship between the image and its atmospheric effect.

Finishing touches

The final stages involved the addition and adjustment of small, carefully placed brushstrokes to add fine details: pinpoints of light along the distant shore, the defining touches to boats and their reflections, fireworks and, lastly, the artist’s butterfly signature. These touches are generally more opaque, slightly impasted, and glossier than the surrounding paint, since the paint was not thinned so much. Technical examination indicates that Whistler sometimes removed and repositioned these accents by wiping them away while the paint remained wet, rather than scraping off dried paint. Their placement could then be adjusted to achieve the desired optical balance yet retain spontaneity.

Across the nocturnes, the lights were most consistently applied in small touches of creamy, relatively thick paint, charged with pure pigments to achieve vibrancy (fig.45). In some instances, the paint was allowed to dry before thin, rich glazes were applied above (fig.45c) and for the reflections, paint is applied deftly, first vertically, then with a precise horizontal stroke while still wet, to merge it with the water (fig.45d) Yellow ochre and occasionally vermilion mixed with lead white are the likely pigments used for creating such accents.68 These small additions establish points of emphasis without interrupting the continuity of the twilight atmosphere.

The first nocturne, Chelsea, includes the most legible human figure in the series, representing a waterman. Even here, the waterman is only faintly defined, constructed through thin, fluid strokes resolved with a few opaque accents. In subsequent nocturnes, the presence of figures diminishes, while boats and their occupants are reduced to faint silhouettes rendered by a few strokes of the brush and appearing absorbed into the surrounding veil (fig.46). As Whistler made greater use of translucent paint, exposed preparations and the texture of the support, these motifs became less descriptive and more closely integrated into the overall tonal arrangement. As the prominence of weavism increases in the later works, figures and vessels come to act only as compositional markers and their increasingly abbreviated forms allow colour, tone and surface texture to carry a greater pictorial effect. The final accents belong to the same carefully controlled process as the underlying veil; rather than describing the scene in detail, they establish just enough visual information to make the atmospheric relationships legible, contributing to the series’ broader movement towards abstraction and optical equilibrium.

Conclusion

The seven blue twilight Thames nocturnes reveal a consistent but adaptable painting process rather than a spontaneous one, in which apparent simplicity emerged from sustained observation, careful preparation, controlled applications and at times extensive reworking. Technical analysis demonstrates the importance of Whistler’s selection and preparation of the supports, the colour of the preparatory layers, his restricted palette and the handling of his exceptionally thin paint films. These nocturnes should therefore be understood not as fleeting impressions or poetic evasions but carefully constructed interpretations of twilight in a rigorously organised pictorial method, one in which water and sky were deliberately constructed as a single atmospheric field.

Technical analysis reveals Whistler as a technically strategic painter, attentive to the optical and handling properties of his materials. Throughout the twilight Thames nocturnes every material choice was deliberate and consequential, demonstrated by his selection of supports; the precise tones of the imprimatura and localised underpainting; his economical pigment choices; the sequencing of paint application through thin layering; and a strategic use of medium to determine flow, translucency and adhesion. His attention to drying intervals and understanding of the behaviour of paint likely developed through his earlier experiments of the 1860s, some of which exhibited mechanical failure. The experimental nature of the nocturnes lies precisely in this close negotiation between materials behaviour and pictorial intention, and relates directly to Whistler’s subsequent statements about finish.

In his ‘Ten O’clock’ lecture, Whistler declared that ‘The work of a master reeks not of the sweat of the brow, suggests no effort, and is finished from its beginning’.69 For Whistler, visible effort threatened to break art’s spell, drawing attention to means rather than ends. His ideal of apparent effortlessness did not necessarily imply rapid execution. Rather the labour of painting was absorbed into the finished image, leaving the viewer attentive to the relationships of colour, tone and form rather than the processes through which they had been achieved. His method was neither bravura nor alla prima but could involve both fluid, rapid applications and deliberate stages of adjustment, removal, repetition and reworking. The material evidence indicates a process involving brief periods of active painting separated by drying intervals, further observation and preparatory labour, including tasks undertaken by assistants. The apparent contradiction between this extended process and rapid execution can be resolved by redefining ‘finish’ for the twilight nocturnes not as material completion but as optical equilibrium. Whistler’s notion that paint should resemble ‘breath on the surface of a pane of glass’ provides further insight into his ambition for surface finish.70 The visibly worked surface of Chelsea, with its sweeping brushstrokes pressed hard onto the panel might initially seem at odds with this ideal. Yet when the twilight series is viewed chronologically, it reveals a growing integration of brushwork, translucent colour and support texture, and an evolution towards an increasingly diffuse and optically unified surface becomes evident.

The greater contribution of the preparatory layer and the canvas weave marks a significant development in Whistlers handling of the painted surface. From the more visibly brushed early panels, through the softened tonal relationships of Cremorne Lights to Old Battersea Bridge, and the three versions of Battersea Reach, the coloured preparations and texture of the support playing a more prominent role in the final image. In some passages subtraction of paint exposes the canvas weave and coloured preparations, elsewhere, the weave remains perceptible beneath the exceptionally thin paint, the imprimatura undisturbed by any wiping back or areas left in reserve (fig.29). As weavism becomes more pervasive and the visual prominence of the brushwork reduces, these effects modify the perception of colour and light, allowing the entire woven support to contribute to the softened appearance of the later paintings as dusk falls.

This technical evidence describes a practice in which mood and method are inseparable. Twilight provided the ideal subject for Whistler to achieve his aesthetic intentions because of its transitional character, neither clear nor obscure, static nor dynamic. This is mirrored in the passage of time that unfolds through the twilight nocturnes, from the relatively legible riverscape at the start of dusk in Chelsea almost to its end and the increasingly abbreviated forms of the later Battersea Reach (Isabella Stewart Gardner). The treatment of preparation, translucency, brushwork and support texture reveals Whistler’s sustained exploration of how the effects of diminishing light might be expressed through paint. His achievement lay in not concealing the physical character of materials, but in bringing their differing properties into such delicate accord that the labour of making gave way to the poetic evocation of twilight.