Nature had not only painted there in all her hues
But there the sweetness of a thousand scents
Was blended in one fragrance strange and new.
Dante Aligheri, The Divine Comedy, Purgatory VII, 79–81
The National Gallery of Victoria’s exhibition British Watercolours 1760–1900: The Age of Splendour (2011–12) provided the Gallery’s paper conservation studio with the opportunity to undertake non-destructive technical examination of the much-treasured collection of watercolours from William Blake’s Divine Comedy series, 1824–27. The NGV’s holdings originally formed part of a suite of 102 works commissioned in 1824 by the landscape painter John Linnell. The original commission was retained by the Linnell family until 1918, when it was sold at auction in London. At that time the NGV was part of a consortium bid, together with the Tate, British Museum (both in London) and a few other parties, to secure the entire group. A careful process was devised for the division of the works, and after a painstaking selection Robert Ross, London adviser to the Felton Bequests’ Committee, recommended the purchase of thirty-six works which became known as the ‘Melbourne Blakes’.1 Morris Eaves (ed.), The Cambridge Companion to Blake, Cambridge University Press, Cambridge, UK, 2003, p. 34. The Divine Comedy is divided into three parts: Hell, Purgatory and Heaven. The ‘Melbourne Blakes’ include 26 scenes from Hell, 6 from Purgatory and 4 from Paradise.
In the lead-up to British Watercolours each work was examined and photographed under ultraviolet (UV), infrared (IR), transmitted and raking light to help identify pigments, detect underdrawing, note compositional changes and confirm the origin and characteristics of the paper. Stereo microscopy was used to help classify media and determine how they had been prepared and applied. This time spent examining the watercolours provided an insight into Blake’s working methods and uncovered techniques not previously noted. It also highlighted the presence of light-sensitive pigments and the continued need to carefully monitor the display duration of these popular works.
The making of the suite
The series of drawings based on Dante Aligheri’s Divine Comedy commissioned by Linnell, Blake’s last great patron, would later be engraved.2 Alexander Gilchrist (ed.), The Life of William Blake, 2nd edn, The Mayflower Press, Plymouth, and William Brendon & Son Ltd, 1880, p. 351. Linnell provided Blake with a folio volume specifically for the drawings, and the artist ‘set to work full of energy, sketching, while confined to bed by a sprained foot, the first outlines of the whole, or nearly the whole, of this new series’.3 ibid. He worked on the drawings by orienting the folio either portrait or landscape, depending on the needs of the composition.4 Of the works examined, 21 are landscape orientation and 14 are portrait.
Blake was plagued by poor health while working on the series, and his letters to Linnell recount attacks of shivering that rendered him bedridden. Although he described his illness as ‘that Sickness to which there is no name’, it is now thought that Blake’s symptoms were due to biliary cirrhosis caused by sclerosing cholangitis, a condition possibly related to his prolonged exposure to the fumes produced when copper plates are etched in nitric acid.5 Lane Robson & Joseph Viscomi, ‘Blake’s death’, in Blake: An Illustrated Quarterly, no. 30, 1996, pp 36–49. Despite his fluctuating health, Blake’s letters convey a great enthusiasm for the Divine Comedy watercolours, describing them as ‘all I care about’.6 Geoffrey Keynes (ed.), The Letters of William Blake, Rupert Hart-Davis, London, 1968, p. 161. In 1826, while recuperating from a period of ill health and making preparations to stay with Linnell at Hampstead, Blake wrote: ‘I intend to bring with me, besides our necessary change of apparel, Only My Book of Drawings from Dante & one Plate shut up in the Book’.7 ibid. His visit to Linnell’s home took place some months later, and there he found a clump of trees on the skirts of the heath where he took his folio of drawings and worked on the series. Years after Blake’s death, close friends still referred to this area as the ‘Dante wood’.8 Gilchrist, p. 373.
Some months after returning to London, Blake continued to paint the series despite his declining heath, which forced him to work bolstered up in bed.9 ibid., p. 379. Under technical examination a microscopic down feather, which may have come from Blake’s bedding, was found attached to the watercolour media of one of the most highly finished works, The Vestibule of Hell and the Souls mustering to cross the Acheron.10 Wayne Longmore, Ornithology Collection Manager, Museum of Victoria, Melbourne, email to the author, 27 Sep. 2011.
Blake was still working on the series when he died at his home on 12 August 1827. As reported by Alexander Gilchrist, and reflected in the Melbourne Blakes, the watercolours are at various stages of completion and therefore present the artist’s conceptions in all their stages – from the bare outlines of the initial drawings to highly finished watercolour paintings. This variation in finish provides tremendous insight into Blake’s technique.
The key to the Blake watercolours entering the NGV collection was the appointment of Oscar Wilde’s close friend Robert Ross as London adviser to the NGV’s Felton Bequests’ Committee in March 1917.11 Irena Zdanowicz, ‘Introduction: the Melbourne Blakes – their acquisition and critical fortunes in Australia’, in Martin Butlin & Ted Gott, William Blake in the Collection of the National Gallery of Victoria, Robert Raynor Publications in Prints and Drawings, no. 3, National Gallery of Victoria, Melbourne, 1989, p. 10. Ross was deeply interested in Blake’s work, and shortly after his appointment as adviser contacted the Felton Trustees about the forthcoming sale at Christies in London of the Linnell collection, which included Blake’s illustrations to Dante’s The Divine Comedy.12 ibid., p. 11. The watercolours, having been separated for exhibition at the Royal Academy of Arts, London, in winter 1893, were no longer bound together in a single folio book.13 Exhibition of works by The Old Masters and by Deceased Masters of the British School, Royal Academy, London, and Clowes and Sons Limited, London, 1893. As no British institution could afford to buy the entire collection, a consortium was established to ensure it stayed within the empire. In this proposed arrangement, Melbourne was well-placed because of the scale of the Felton Bequest, the professionalism and knowledge of Ross and the willingness of the NGV Trustees to place their trust in him.14 Zdanowicz, p. 11.
Although the series was purchased in March 1918, the watercolours did not reach Melbourne until 1920; primarily because Ross had concerns about them being damaged or lost on their journey, and the risk of enemy activity on the seas. In addition, he believed:
The only trustworthy packers in London is the French firm of Chenue, who act for the British, Victoria & Albert Museum. All their French hands are at the War … The art of scientific packing is unknown to the English.15 Robert Ross, letter to E. Armstrong, Esq., dated 17 Sep. 1918, Felton papers, National Gallery of Victoria, Melbourne.
Upon arrival at the NGV, the series was framed and displayed in the Buvelot Gallery in August 1920. At the time, the works received fairly hostile reviews in the local papers. The Argus critic wrote:
Many of his pictures, considered artistically, are grotesque in the extreme, and in some cases repulsive in treatment … no justification can surely be shown for the purchase of so many artistically inferior pictures, which will no doubt before long find their way to the cellars.16 The Argus, 11 Aug. 1920, p. 6.
It was twenty years until the watercolours received more favourable reviews. In 1940 The Argus critic wrote: ‘Few Felton Bequest acquisitions have reflected more credit on the purchasing committee than the William Blake drawings … praise of them could hardly be too lavish’.17 John Harcourt, The Argus, 12 Nov. 1940, p. 6.
Methodology of examination
In preparation for British Watercolours 1760–1900 the entire series was unframed for re-photographing. The watercolours had not been out of their frames since preparations for the 1989 exhibition William Blake in the Collection of the National Gallery of Victoria, so this provided an opportunity to examine the whole collection and arrange for additional photography (including IR, UV, raking and transmitted light) to be undertaken.
Working within a tight timeframe, all works were examined under a stereo microscope, with a selection of photomicrographs taken that represented the various materials and techniques present in each work. Over the following months research into the art practice of Blake was carried out and the various photographic images examined and compared with findings from past studies.
Transmitted light photography enabled the chain and laid lines, watermarks and countermarks to be observed. The paper is antique laid, with half the sheets examined bearing a fleur-de-lis watermark, and half bearing the W. Elgar 1796 countermark (both belonging to the W. Elgar paper mill).18 William Elgar operated Chafford Mill on the Medway, outside the village of Fordcombe, near Tonbridge, between c. 1785 and 1802. The mill specialised in the production of fine writing paper, well suited to drawing and watercolour work, used by both Blake and J. M. W. Turner (Peter Bower, ‘The vivid surface: Blake’s use of paper and board’, in Joyce Townsend [ed.], William Blake: The Painter at Work, Tate Publishing, London, 2003, p. 59). The paper, which was originally ‘Imperial’ size (56 x 75 cm), was folded in half and gathered into sections for sewing, resulting in a group of pages containing the watermark followed by a group containing the countermark. Images of the watermarks and countermarks were sent to Peter Bower, a paper analyst based in London, and he determined there were two variants present, as one would expect from a pair of moulds.19 Peter Bower, email to the author, 12 July 2012. The raking light photographs indicated approximately two-thirds of the works are painted on the wire side, and one-third on the felt side. Unfortunately, in some cases the paper surface is severely flattened (possibly from past treatment) and it is not possible to be absolutely certain which side the work is executed on.
A number of the supports contain distinctive creasing and tiny losses, possibly due to the pulling of sewing thread at the original locations of the sewing stations. Five works have binding evidence along the left edge, indicating the work is from the right-hand page of the folio book. Dante adoring Christ is the only work examined that exhibits binding evidence along its right edge, indicating it was a left-hand page.
Blake’s initial drawings were made using a combination of graphite-type pencil and black friable media variously identified as black chalk and charcoal in previous studies. When sketching-out most of the examined works, Blake’s first compositional marks were made using shimmering graphite lines, with small enhancements added using the friable media. Examination of the IR photographs indicates few digressions from these initial marks, and those that were found were minor in nature. Blake attached great importance to these first lines, and is recorded as having said, ‘First thoughts are best in art, second thoughts in other matters’.20 Gilchrist, p. 390.
Although the series was left at varying stages of completion, all of the Melbourne works have tonal washes and watercolour applied. According to past studies, Blake’s palette included carbon black (used in both ink and watercolour form), Prussian blue and indigo, vermillion and red lake, gamboge and chrome yellow and brown, yellow and red ochres.21 Gilchrist, p. 72; Anne Maheux, ‘An analysis of the watercolour technique and materials of William Blake’, Blake: An Illustrated Quarterly, no. 17, 1984, p. 127; Noa Cahaner McManus & Joyce Townsend, ‘Watercolour methods, and materials use in context’, in Townsend, pp. 61–79, p. 68. He extended this basic palette by mixing indigo or Prussian blue with gamboge to make green, gamboge with vermillion to make orange, and indigo or Prussian blue with red lake to make purple.
When viewed under magnification, sizeable chunks of pigment were noted in the red and blue media of several works. The appearance of solid masses and clumping of pigment is possibly the result of incomplete grinding which failed to evenly distribute the pigment particles within the binder. Blake is known to have ground his own colours from powdered pigment on a statuary marble slab.22 Raymond Lister, Infernal Methods: A Study of William Blake’s Art Techniques, G. Bell and Sons Ltd, London, 1975, p. 39. According to Raymond Lister, Blake mixed powdered pigments with diluted carpenter’s glue to form water soluble paints. Although commercially prepared watercolour cakes would have been available to Blake, he may have found these early prototypes too hard, or he may just have preferred not to change his methods of working and continued to grind his own paints.23 ibid. Technical examination undertaken at the Tate revealed Blake used a range of gums, including Tragacanth and Arabic, which could have been mixed with glue or used alone as a binder for his watercolours.24 Cahaner McManus & Townsend, p. 74.
UV photography of the works was undertaken by the NGV’s Photographic Services department, and in addition the department photographed two colour charts of English watercolours dating from the nineteenth century that included most of the pigments known to have been used by Blake. Examination of the UV photographs indicated the presence of red lake (possibly cochineal carmine or Kermes carmine sourced from scale insects) which fluoresced a bright pink in twenty-five works.25 Helmut Schweppe & John Winter, ‘Madder and alizarin’, in Elizabeth West FitzHugh (ed.), Artists’ Pigments: A Handbook of Their History and Characteristics, vol. 3, National Gallery of Art, Washington DC and Oxford University Press, New York, 1997, pp. 109–42. Because Dante’s robe is always depicted as red throughout Blake’s series, the most extensive use of red lake was found in Dante running from the Three Beasts, in which Dante plays a prominent role. When viewed under UV light, gamboge (which appeared olive green) was noted in thirty of the Melbourne works.26 The gamboge on the two colour charts also appeared olive green under UV light.
Blake preferred to apply aqueous media with a camel hair brush, and in most of the works examined various washes have been applied on top of one another.27 Maheux, p. 125. In several cases the works are essentially tonal drawings; Ephialtes and two other Titans and Lucifer, for example, are comprised predominately of black wash with minimal red lake washes applied after the black layer had dried. To paint figures in most of the works, Blake started with a tonal drawing to delineate musculature and to help model form, before adding varying degrees of colour. In many instances, two layers of colour were applied, one on top of the other, to form secondary colours. This layered colour mixing technique was noted with greater prevalence than the use of secondary colours, mixed on a palette, was. In most of the works examined Blake allowed the separate layers of watercolour to dry prior to adding subsequent colours so they would remain clearly distinct. In some works he applied subsequent layers of colour using a dry brush to avoid bleeding.
Blake frequently laid different colours next to one another, resulting in a dynamic, vibrant effect. This technique is seen in Dante running from the Three Beasts where gamboge, red lake, vermillion and Prussian blue have been placed in close proximity to depict the radiant sunrise over an expanse of water.
While examining the works under the stereo microscope, it was occasionally possible to gain a sense of the vigour with which Blake had wielded the camel hair brush. Pilling of the surface fibres and localised abrasions could be seen where he had worked vigorously with a loaded brush. His brushwork ranges from broad sweeps of colour to tiny strokes delicately applied with the tip of a fine brush. There are a small number of examples within the most highly finished works that include stippling, an aesthetic Blake would have been familiar with through his previous use of stipple engraving. In Dante adoring Christ, for instance, the face of Christ has been delicately rendered in great detail using tiny brushstrokes. From 1800 Blake worked as a miniature painter, when he would have perfected stippling with brush and aqueous media. The technique was increasingly used by contemporary watercolourists, such as J. M. W. Turner, in the first decades of the nineteenth century.28 Lister, p. 49.
Several works in the collection contain areas of sponging, a technique not noted in previous studies of the series. Sponging is used to soften effects, blend various pigments or to create highlights by removing pigment to reveal the glow of the underlying paper support. In cases where a compositional change is desired, sponging can be used to intentionally erase elements the artist is unhappy with. Having been given a copy of Trattato della Pittura (1821) by John Linnell, Blake was familiar with the teachings of Cennino Cennini, who suggested wetting out an unsatisfactory area of the composition with a large brush steeped in water, then rubbing over it to remove surface fibres so the area could be reworked.29 ibid., p. 43.
During Blake’s time it was common for a small piece of bread rolled into a ball, or a sponge, to be used to rub and draw the pigment away from the paper’s surface. Within the collection there are examples where the technique has obviously been used to create highlights – such as some of those in the lightning bolts in Capaneus the Blasphemer – while in other works such as The Six-Footed Serpent attacking Agnello dei Brunelleschi pigment removal is more likely to be erasure Blake planned to later repaint, but died before he could do so.
Another technique not noted in previous studies, but observed in one of the most highly finished of the Melbourne works, is scratching out. In rendering the raindrop-like forms descending from the clouds above the head of Antaeus in Antaeus setting down Dante, Blake has used a penknife or similar tool to remove the pigment sitting on the paper surface and some of the underlying paper fibres to create pristine white lines that act as highlights.
Quill and ink
In most of the works it appears that once the watercolour washes were laid down and had been manipulated to Blake’s satisfaction, the composition was outlined using black ink lines applied with a quill. Blake had strong opinions about the importance of line: ‘The great and golden rule of art, as well as of life, is this: That the more distinct, sharp, and wiry the bounding line, the more perfect the work of art’.30 Gilchrist, p. 525. According to research undertaken at the Tate, Blake used ready-made ‘India ink’ (a carbon black ink) for the application of lines.31 Cahaner McManus & Townsend, p. 73. The ink was made by combining soot with a water-based medium such as gum or glue. In the nineteenth century shellac dissolved in borax was added to render the ink insoluble in water, and this property would have allowed Blake to use the ink over or under passages of watercolour without the risk of the ink running.32 ibid. Red/brown coloured ink was used in addition to black ink in the work Minos, but this is the only example of a different type of ink found during this study.
The presence of particular pigments such as gamboge, red lake and indigo, discovered in the Melbourne Blakes has resulted in a greater awareness of the light sensitivity of the works. This knowledge will inform decisions about their future display, to ensure they remain as rich and vibrant as they are at this time. Display durations for each work will continue to be carefully monitored, and light levels during display periods will be maintained in accordance with the NGV’s lighting policy for ultra-sensitive works. In addition, blackout roller blinds will continue to be drawn over the works, to prevent exposure to incidental light, while the watercolours are in storage.
Louise Wilson, Conservator of Paper, National Gallery of Victoria
1 Morris Eaves (ed.), The Cambridge Companion to Blake, Cambridge University Press, Cambridge, UK, 2003, p. 34. The Divine Comedy is divided into three parts: Hell, Purgatory and Heaven. The ‘Melbourne Blakes’ include 26 scenes from Hell, 6 from Purgatory and 4 from Paradise.
2 Alexander Gilchrist (ed.), The Life of William Blake, 2nd edn, The Mayflower Press, Plymouth, and William Brendon & Son Ltd, 1880, p. 351.
4 Of the works examined, 21 are landscape orientation and 14 are portrait.
5 Lane Robson & Joseph Viscomi, ‘Blake’s death’, in Blake: An Illustrated Quarterly, no. 30, 1996, pp 36–49.
6 Geoffrey Keynes (ed.), The Letters of William Blake, Rupert Hart-Davis, London, 1968, p. 161.
8 Gilchrist, p. 373.
9 ibid., p. 379.
10 Wayne Longmore, Ornithology Collection Manager, Museum of Victoria, Melbourne, email to the author, 27 Sep. 2011.
11 Irena Zdanowicz, ‘Introduction: the Melbourne Blakes – their acquisition and critical fortunes in Australia’, in Martin Butlin & Ted Gott, William Blake in the Collection of the National Gallery of Victoria, Robert Raynor Publications in Prints and Drawings, no. 3, National Gallery of Victoria, Melbourne, 1989, p. 10.
12 ibid., p. 11.
13 Exhibition of works by The Old Masters and by Deceased Masters of the British School, Royal Academy, London, and Clowes and Sons Limited, London, 1893.
14 Zdanowicz, p. 11.
15 Robert Ross, letter to E. Armstrong, Esq., dated 17 Sep.1918, Felton papers, National Gallery of Victoria, Melbourne.
16 The Argus, 11 Aug. 1920, p. 6.
17 John Harcourt, The Argus, 12 Nov. 1940, p. 6.
18 William Elgar operated Chafford Mill on the Medway, outside the village of Fordcombe, near Tonbridge, between c. 1785 and 1802. The mill specialised in the production of fine writing paper, well suited to drawing and watercolour work, used by both Blake and J. M. W. Turner (Peter Bower, ‘The vivid surface: Blake’s use of paper and board’, in Joyce Townsend [ed.], William Blake: The Painter at Work, Tate Publishing, London, 2003, p. 59).
19 Peter Bower, email to the author, 12 July 2012.
20 Gilchrist, p. 390.
21 Gilchrist, p. 72; Anne Maheux, ‘An analysis of the watercolour technique and materials of William Blake’, Blake: An Illustrated Quarterly, no. 17, 1984, p. 127; Noa Cahaner McManus & Joyce Townsend, ‘Watercolour methods, and materials use in context’, in Townsend, pp. 61–79, p. 68.
22 Raymond Lister, Infernal Methods: A Study of William Blake’s Art Techniques, G. Bell and Sons Ltd, London, 1975, p. 39.
24 Cahaner McManus & Townsend, p. 74.
25 Helmut Schweppe & John Winter, ‘Madder and alizarin’, in Elizabeth West FitzHugh (ed.), Artists’ Pigments: A Handbook of Their History and Characteristics, vol. 3, National Gallery of Art, Washington DC and Oxford University Press, New York, 1997, pp. 109–42.
26 The gamboge on the two colour charts also appeared olive green under UV light.
27 Maheux, p. 125.
28 Lister, p. 49.
29 ibid., p. 43.
30 Gilchrist, p. 525.
31 Cahaner McManus & Townsend, p. 73.