From at least 1647 onwards, Rembrandt executed many of his prints on papers of oriental origin.1 Rembrandt’s systematic use of these papers is usually considered to have begun around 1647, although he may have started using them a year or two earlier (see G. Biörklund, ‘Old Paper’, in G. Biörklund & O. Barnard, Rembrandt’s Etchings: True and False – A Summary Catalogue, rev. edn, Stockholm, 1968, p. 172; C. White, Rembrandt as an Etcher: A Study of the Artist at Work, vol. I, London, 1969, p. 16). While his European papers have been the subject of intense scrutiny,2 I am referring to the Rembrandt watermark project begun in 1981 by Nancy Ash and Shelley Fletcher at the National Gallery of Art, Washington DC. This project has resulted in the collection of over 900 watermarks and has given rise to a cataloguing system that groups them by general type and by degree of similarity. The results of the project will soon be published. The Conservation Department at the National Gallery of Victoria has recorded the fifty watermarks found on papers in the Melbourne collection. These watermarks have been catalogued, according to the National Gallery of Art’s system, by Rembrandt scholar Drs Erik Hinterding of the University of Utrecht in the Netherlands. Eight of the watermarks occur in papers carrying etchings by Rembrandt that were included in the recent exhibition Rembrandt: A Genius and His Impact, and are noted in the exhibition catalogue (see I. Zdanowicz, ‘Etchings by Rembrandt’, in A. Blankert, Rembrandt: A Genius and His Impact (exh. cat.), National Gallery of Victoria, Melbourne, 1997, cat. nos 103, 106, 107, 109, 112, 113, 115B, 116). the oriental papers remain somewhat neglected. It is hoped that this article will make some contribution in this area.
Print scholars have traditionally divided oriental papers into three categories: China, Japan and India.3 The earliest reference to Rembrandt’s use of oriental papers appears to be the allusion to ‘Indian’ paper that occurs in a letter from the English traveller Edward Browne (1644–1708), written to his father from Amsterdam in 1668: ‘Here is a strange variety of excellent prints … Here are divers good ones of Rembrandt and some upon Indian paper that look like washing, though scratched after his manner’ (Edward Browne, letter to his father, 1668, cited in E. Hinterding, ‘The History of Rembrandt’s Copperplates, with a Catalogue of Those That Survive’, Simiolus, vol. 22, no. 4, 1993–94, p. 264 n. 62). Browne’s use of the term ‘Indian’ was probably the beginning of many misunderstandings, caused by the vagaries of language. As Mayer Haunton points out, ‘Indian’ in this context may simply be a reference to ‘from the Indies’ or, more particularly, to ‘imported by the East India Company’ – rather than to India itself; a small number of papers used by Rembrandt do, apparently, bear a strong resemblance to India papers, and Rembrandt’s ownership of Mogul miniatures adds circumstantial evidence in support of his also having had access to Indian papers (K. Mayer Haunton, ‘Rembrandt’s Papers’, in A Collection of Etchings by Rembrandt Harmensz. van Rijn (1606–1669), Formed by Joseph R. Ritman (sale cat.), eds N. Bialler, A. T. Eeles, R. Godfrey & K. Mayer Haunton, Artemis & Sotheby’s, London, 1995, n.p.).The seven oriental papers used for prints by Rembrandt in the National Gallery of Victoria collection exhibit characteristics associated with China and Japan papers.
The terms China and Japan, though widely used, are essentially rather subjective, referring to the appearance of a paper, and not necessarily to its place of origin. The following characteristics are associated with Rembrandt’s ‘China’ papers: they are generally thin, and when viewed in transmitted light they reveal the bamboo or reed screen pattern imparted by the papermaking mould on which they were formed; in colour they vary from white and greyish white to ivory. The ‘Japan’ papers, by comparison, are thicker and are often referred to as ‘vellums’. In colour, they range from pale buff, and gold, to ivory. Frequently smoother in appearance than western papers, when viewed in transmitted light they do not generally reveal the screen pattern of the papermaking mould on which they were formed.
The composition and origin of these papers have long been matters of speculation. Recourse to historical references has often confused rather than clarified the issues.4 For example, Biörklund, ‘Old Paper’, p. 172, notes [in] the auction catalogue of 1755, when the important De Burgy collection was sold, the Dutch text mentioned 48 etchings printed “Op Oostindisch papier” which the French translator of the auction catalogue erroneously interpreted as “Papier de la Chine” (see also note 3 above). The most authoritative study of the oriental papers remains that conducted by George Biörklund in 1968.5 Biörklund, ‘Old Paper’. After examining the fibres from a number of these papers, Biörklund concluded that they were made up purely of gampi fibres and were Japanese in origin.6 Gampi is a deciduous plant from the family Thymelaeacea. A few varieties are used for papermaking. Diplomorpha sikokiana Nakai is generally considered the best. Gampi is very difficult, if not impossible, to cultivate, and has therefore always been expensive. Although it grows throughout Southeast Asia, it is most closely associated with papermaking in Japan, where it has been used since at least the eighth century. Biörklund does not discuss China papers in any depth, although he does refer to a ‘nearly white paper with a pale, silvery tone and perfectly smooth surface’ used in etchings after 1654, a paper he also believed to be composed of gampi.7 Biörklund, ‘Old Paper’, p. 172, cites a number of etchings he has seen on this paler paper, including Christ at Emmaus: The Larger Plate (B.87). The National Gallery of Victoria’s impression of this print is on what we will be calling Paper 6, which may well be this same paper. Biörklund in fact rejected the proposition that any of Rembrandt’s papers may have come from China, because of the supposedly low quality of Chinese papers at that time, and because of the political instability in the country.8 Biörklund, ‘Old Paper’, p. 172.
Biörklund also points to the fascinating trading relationship between the Netherlands and Japan during Rembrandt’s lifetime. From 1639 until 1854 the Dutch were the only Europeans permitted to enter and live in Japan. They were, however, restricted to the small, artificial island of Deshima at Nagasaki, receiving annual visits of two to three vessels.9 Deshima measured a mere 200 x 600 feet. Some twenty Dutchmen at a time were in residence on the island. They lived under severe restrictions but the Dutch East India Company, for which they worked, maintained a very profitable business: according to records, profits averaged 81 per cent in the period 1649–62. See A. Hyma, The Dutch in the Far East, Ann Arbor, Michigan, 1942; G. K. Goodman, The Dutch Impact on Japan (1640–1853), Leiden, 1967. Remarkably, following an examination of Dutch East India Company records, Biörklund was able to identify two records that actually named Japanese paper as a cargo: the first ‘an invoice of the ship “De Swaen” dated October 1st 1643 for two casks of Japanese paper’; and the second ‘in a memoire [of 6 November 1644] of merchandise to be delivered by the Japanese “both for the Netherlands and for India”, 3000 sheets of Japanese paper’.10 Biörklund, ‘Old Paper’, p. 173. I am also grateful to Dr Christiaan Jörg of the Groninger Museum, Groningen, for the following information: ‘The amount of goods mentioned in the [Dutch East India] Company records never covers the total export from Decima [Deshima] by the Dutch in a given year, as the merchants were licensed (officially, it was not smuggling) to buy and ship their own private merchandise on return ships from Japan to Batavia/the Netherlands. So the two shipments of 1643/44 can in no way be conclusive’ (Dr Christiaan Jörg, letter to the author, 6 January 1997).
The assumption that most, if not all, of Rembrandt’s oriental papers are composed entirely of gampi fibres and are probably of Japanese origin is often repeated in the literature.11 The only mention of fibre analysis found by the author to date in the literature published since Biörklund reported on his findings in 1968 appears to be that undertaken in the Asiatic Department of the Museum of Fine Arts, Boston, in 1970, but complete details have not been published; the conclusions drawn on the basis of this latter analysis appear to be similar to Biörklund’s (see S. W. Reed, ‘Types of Paper Used by Rembrandt’, in Rembrandt: Experimental Etcher (exh. cat.), Museum of Fine Arts, Boston, 1969, pp. 178–83). China cannot however be ruled out as a source of some of the papers. While all Europeans but the Dutch were excluded from Japan after 1639, the Chinese continued to trade there, having warehouses and a compound not far from Deshima.12 See Goodman, p. 19. It should also be remembered that between 1624 and 1661 the Dutch occupied Taiwan, a strategic location for trade, and for piracy.13 The Dutch imported silk, deerskins and other products to Japan from Taiwan (see Hyma, p. 155).
In view of the little recent attention given to Rembrandt’s oriental papers, and because of the often repeated generalisations associated with them, in early 1997 it was decided that a careful examination of the seven papers represented in the National Gallery of Victoria collection would be timely in the lead-up to the exhibition Rembrandt: A Genius and His Impact, of October that year. This article documents that examination, and its findings.
Papermaking in Japan and China: a brief overview
The Chinese are credited with the invention of paper around the first century AD. The art of papermaking then appears to have spread slowly, reaching Japan only in the early seventh century.14 For an account of the history of papermaking in the Orient, see D. Hunter, Papermaking: The History and Technique of an Ancient Craft, New York, 1974. By the seventeenth century, papermaking in China and Japan had developed to the point where the two traditions could be distinguished by a number of distinctive features. In China, a wide variety of plant material was being used for fibre sources, and papers were more likely to consist of a mixture of different fibres.15 See H. Mullock, ‘Xuan Paper’, Paper Conservator, vol. 19, 1995, p. 26; see also Tsuen-Hsuin Tsien, ‘Raw Materials for Papermaking in China’, Journal of the American Oriental Society, vol. 93, 1973, pp. 510–48. Major sources of papermaking fibre included hemp, jute, flax, ramie, rattan, bamboo, rice, wheat stalks, and the bast fibres of some trees, notably paper mulberry (usually Broussonetia papyrifera) and than (Pteroceltis tatarinowii Maxim.). In contrast, papermaking in Japan was dominated by three bast fibres: kozo, gampi and mitsumata.16 Hemp played an important part in early Japanese papermaking but seems to have been replaced by kozo and gampi by at least the twelfth century (see K. Masuda, ‘Japanese Paper and Hyōgu’, Paper Conservator, vol. 9, 1985, p. 32). Kozo (Broussonetia kazinoki Sieb.) is a variety from the family Moraceae, the family to which Broussonetia papyrifera also belongs. Mitsumata is usually cited as Edgeworthia papyrifera Sieb. et Zucc. (strictly speaking, it should be known as Edgeworthia chrysantha (see Y. Harvey, ‘Edgeworthia chrysantha’, Plantsman, vol. 14, no. 3, December 1992, p. 130)). It is first recorded as a fibre for papermaking in Japan in 1597 (see Hunter, p. 481). For gampi, see note 6 above.
There were also differences associated with the moulds upon which the Japanese and Chinese made their sheets, and some interesting variations in techniques of sheet formation.
The oriental papermaking mould consisted of a flexible screen, made of bamboo splints, grass or reeds tied together with silk or horsehair to form a horizontal ‘laid’ pattern. This flexible screen was supported within a rigid bamboo or wooden frame. In Japan, the screen and frame are referred to as the su and geta, respectively. By the seventeenth century the Japanese su was likely to be made of bamboo splints, with pampas grass and other grasses being more often used prior to 1300.17 See T. Barrett, Japanese Papermaking: Traditions, Tools, and Techniques, New York, 1983, p. 82. Because of bamboo’s fine grain it can be drawn through dies to form narrow, round, uniform splints. The distance between ‘chain’ lines in a su (that is, the silk or horsehair threads connecting the splints) traditionally varied between 2.5 cm and around 4 cm. By contrast with the su, the Chinese mould traditionally lacked uniformity, with its bamboo, grass or reed splints being of various diameters and shapes, and its chain lines closer together (1.5–2 cm).
In addition to making ‘laid’ papers, Japanese papermakers would also occasionally produce ‘wove’ papers by attaching to the su a finely woven piece of silk, referred to as the sha. Papers made with a sha do not usually reveal chain and laid lines when viewed in transmitted light, although sometimes chain lines can be apparent (see discussion of Papers 4 and 5 below).
In addition to the idiosyncrasies associated with the papermaking moulds used in China and Japan, there also developed in both countries significant differences in techniques of sheet formation. Basically, a sheet of paper is formed by plunging the mould into a vat of water and fibres, lifting out the mould, and shaking it to and fro to distribute the fibres evenly before all the water drains away. This is essentially the papermaking technique introduced to Japan from China in the seventh century. This technique came to be known in Japan as tame-zuki, which has been loosely translated as ‘the fill-and-hold way to make paper’.18 ibid., p. 51. Because water drains rapidly, any fibrous clumps (knots) or impurities that find their way into the vat during this process are quickly incorporated into the sheet.
By the ninth century, however, the Japanese appear to have modified this process by adding neri, a viscous plant material, to the vat.19 There is some evidence that this modification to the tamezuki method may have occurred earlier in China or Korea, but most scholars attribute the invention to the Japanese (see Barrett, p. 70). Neri traditionally came from either the root of Abelmoschus manihot Medikus (known as tororo-aoi in Japanese) or the inner bark of Hydrangea paniculata Sieb. (noriutsugi in Japanese) (see Barrett, pp. 56, 287). The role of this viscous agent was to disperse the fibres evenly, both in the vat and during agitation, and also to slow the drainage of the water through the su. The addition of this agent led to the development of a very dynamic sheet-forming method known as nagashi-zuki, or ‘the flowing or sloshing way to make paper’.20 See Barrett, p. 51. According to this method, a sheet of paper is formed by dipping the mould into the vat, lifting it out and tilting it back and forth, allowing the slurry to flow across the mould; because drainage is slowed, any excess slurry, and with it any impurities or fibrous clumps, can be tossed back into the vat before it settles. This process is repeated several times, slowly building up a sheet of paper on the su. This complex and demanding action gives the papermaker immense control, resulting in thin and uniform sheets of great strength and beauty.
The nagashi-zuki method is ideal for making thin, lightweight papers, rarely heavier than 80 gsm.21 ibid., p. 70. When it comes to making thicker, heavier sheets, however, the viscous additive actually appears to work against the papermaker, slowing drainage to such an extent that the fibre accumulation eventually becomes uneven, resulting in an uneven sheet. Thicker, heavyweight papers are therefore usually made using the tame-zuki method or by laminating together thin sheets formed by the nagashi-zuki technique. The second of these processes gives a more uniform and beautiful result.
The preceding discussion is necessarily brief and is not intended to be definitive. It is important to be aware of the pitfalls that are always inherent in attempts to simplify the distinctions between the Chinese and Japanese traditions of papermaking.
Aim and procedure
The aim of this study has been to gather objective information on the oriental papers used by Rembrandt for printmaking, of which there are seven examples in the collection of the National Gallery of Victoria. The research was also extended to a posthumous impression of a Rembrandt print, reworked by William Baillie, and to a print after Rembrandt, by James Bretherton, on oriental paper. The investigation of all nine papers included careful visual examination, ß-radiography22 The ß-radiography process involves placing the paper between a poly[Carbon 14] methylmethacrylate plate and a sheet of Kodak Biomax MR film. A 1:1 black and white photographic contact print can then be made. ß-radiographs are useful because they reveal details about the paper structure, including fibre distribution and quality, and any characteristics imparted by the mould upon which the sheet was formed. Thin areas appear lighter on a ß-radiography. and fibre analysis.
The analysis of oriental paper fibres is a very specialised and often difficult task. It has been pointed out by various researchers that some fibres, particularly gampi and mitsumata, belong to the same family and are very similar,23 For the similarities between gampi and mitsumata fibres, see G. Biörklund, ‘Paper Analysis’, in Biörklund & Barnard, p. 175; S. Beauman Murphy & S. Rempel, ‘A Study of the Quality of Japanese Papers Used in Conservation’, in The Book and Paper Group Annual, vol. 4, ed. J. Abt, USA, 1985, p. 68. and the situation is also complicated by the fact that, ethically, only very small samples may be removed from a work of art. Katsuhiko Masuda, Director of the Department of Restoration Technique at the Tokyo National Research Institute of Cultural Properties, was approached, and volunteered to organise the analysis on the Gallery’s behalf. The analysis was carried out in June 1997 by Akinori Okawa, Senior Researcher at the Kochi Prefectural Paper Technology Centre. Kochi is one of the three oldest papermaking prefectures in Japan.
Sampling of the papers in the Gallery’s collection involved the removal of very small (literally millimetres in length) representative fibres from the verso of each.24 This process was documented in detail (National Gallery of Victoria (Conservation Department) files). These samples were then sent to Japan for analysis.25 In carrying out this analysis, Akinori Okawa used water to disperse the fibres on a glass plate, dyed them with C-stain solution, and then examined them microscopically.
Findings
The papers used by Rembrandt for the seven prints on oriental paper in the collection of the National Gallery of Victoria fall into three distinct categories. The primary differentiating feature is the botanical source of the paper fibres, although ß-radiography has also revealed clear differences in the ways in which the sheets were formed. Table 1 identifies the fibre type found in each paper. Further details of these seven papers, and of the two other papers examined (Papers A and B), are provided in the Appendix at the conclusion of this article, accompanied by reproductions of the ß-radiographs made in the course of the examination.
Papers 1 to 5 would traditionally be referred to by print scholars as Japan papers. The small samples taken from each of these five papers indicate that they are probably composed entirely of gampi. Samples taken from Papers 1 and 2 also contain rice powder, added as a filler.26 Akinori Okawa points out that examples of gampi papers with rice powder fillers date back to the Heian period (9th–11th centuries) (letters to the author, 1997 (translated by Katsuhiko Masuda)). All five papers are similar in colour, ranging from a golden yellow to buff, are quite thick and range in basis weight from 106 to 200 gsm.
A careful visual examination of certain physical features, namely damaged areas on the versos, and the characteristics displayed in the ß-radiographs, indicate that all of these papers are multi-layered.27 This has been suggested previously (see Biörklund, ‘Old Paper’, p. 171).
In general, the fibre quality of all five papers seems quite high (that is, the fibres are of uniform length and have few knots or shives). Fibre distribution is also relatively regular. All this points to the likelihood that the two or more sheets making up these gampi papers were formed individually by the nagashi-zuki technique, and then couched and pressed together when wet, drying together to form a laminate. Gampi is an ideal fibre for this process because gampi sheets have a natural tendency to adhere to one another when wet.28 See Barrett, p. 67. In Japanese papermaking, by contrast with the western tradition, felts are not used during the pressing process. The newly formed wet sheets are piled directly on top of each other, lightly pressed, then separated and brushed onto boards to dry in the sun. The tendency of gampi papers to stick together after pressing causes problems for papermakers.
The ß-radiographs of Papers 4 and 5 reveal some interesting similarities. Broad, vertical and light linear bands are apparent, and the spacing of these bands is consistent with that of the chain lines of a Japanese su. This banding is further indicative of papers made using a sha on top of rather raised chain lines (that is, chain lines formed with thick and rough thread rather than with fine silk). A less regular, more diffuse, and horizontal wave-like pattern is also apparent in the ß-radiographs of Papers 4 and 5. These horizontal ‘variegations’ in the paper are due to the accumulation of slight irregularities in the thicknesses of the individual sheets making up the laminate. A similar effect can be seen in the ß-radiograph of Paper 1, only in this case the orientation of the print means that the variegations are vertical.
The ß-radiograph of Paper 1 reveals a laminate made up of sheets formed on a sha over a very fine su. Here there is no sign of chain lines, so it would appear that the su had low-profile chain lines made with fine thread such as silk.29 I am very grateful to Timothy Barrett for pointing out to me the significance of the broad and diffuse chain lines seen in the ß-radiographs of Papers 4 and 5. I am also indebted to him for providing me with a unique sheet of handmade Japanese paper, which, when ß-radiographed, exhibited many of the characteristics seen in the ß-radiographs of Papers 1, 4 and 5. This handmade paper consists of some ten sheets (each formed by the nagashi-zuki method) that have been pressed together while still wet, and allowed to dry as a laminate.
Papers 2 and 3 are relatively small and have sustained areas of damage that make it difficult to discern the types of patterns apparent in the ß-radiographs of Papers 1, 4 and 5. Slight damage on the versos of Papers 2 and 3, however, indicates that they are laminates, and the ß-radiographs of both papers further exhibit the wavelike variegations associated with laminates consisting of sheets made using the nagashi-zuki method.
Historically, paper has permeated almost every aspect of life in Japan. The population of Japan is estimated to have been between fifteen and twenty million people at the time of the arrival of the Dutch in the early seventeenth century.30 See Hyma, p. 140. It has been noted by one writer that during the Edo period (1615–1868) paper was second only to rice as a source of tax revenue for the Japanese government.31 See S. Hughes, Washi: The World of Japanese Paper, Tokyo, 1978, p. 45. Paper has been used in Japan for functional purposes, such as clothing and shelter (for example, umbrellas and wall screens), as well as for cerebral and aesthetic pursuits such as calligraphy. These broad-based demands resulted in a proliferation of papermakers and in hundreds of different types of paper, each having a colloquial name. These names often described a paper’s colour, place of origin, uses and any number of other details. Many of the names have survived, although the modern Japanese papers to which they are ascribed frequently bear little real resemblance to the earlier papers.32 ibid., p. 171; A. Dwan, ‘A Method for Examining and Classifying Japanese Papers Used by Artists in the Late Nineteenth Century: The Prints of James Abbott McNeill Whistler’, in Conservation Research, Studies in the History of Art, vol. 41, Washington DC, 1993, p. 108.
Papers 1 to 5 show characteristics consistent with those of a type of Japanese paper known as torinoko. Torinoko can be translated as ‘bird’s egg’ and refers to the brownish colour of the paper. Examples of torinoko, which is traditionally made from gampi, date back as far as the Nara period (710–94).33 See Hughes, p. 191. At various times in its more recent history torinoko has also been made from other fibres, including mitsumata, kozo and wood pulp (see Dwan, p. 113). The paper is first mentioned in Japanese literature in the fourteenth century, and by the middle of the fifteenth century appears to have become extremely popular among the upper class. Despite being very expensive, it seems to have enjoyed even greater popularity in the seventeenth century. Interestingly, in the Jesuit Japanese-Portuguese dictionary Vocabulario da lingoa de Iapam com adeclaraçao em portugues, published in Nagasaki in 1603, ‘the names of paper made of gampi, i.e. torinoko and its derivatives, outnumber all other kinds of paper’.34 B. Jugaku, Papermaking by Hand in Japan, Tokyo, 1959, p. 16. Only thirty-six years separate the publication of this dictionary and the confinement of the Dutch to the island of Deshima, at Nagasaki.
A bamboo and mitsumata paper
Paper 6 would traditionally be referred to by print scholars as a China paper. It is a relatively lightweight (53 gsm), pale, thin paper. It is also more absorbent than the gampi papers, probably because of its bamboo fibre component, and the printing ink has penetrated through to the verso. The chain and laid lines imparted by the su are clearly visible. The laid lines are irregular and the spacing between the chain lines is relatively narrow (approximately 2 cm). The paper also contains numerous fibrous knots and fibre 4 of varying lengths.
Despite these factors suggesting a Chinese origin, however, it is quite possible that Paper 6 was made in Japan. The Japanese have for centuries made a type of paper called gasenshi, or ‘imitation Chinese calligraphy paper’.35 See Dwan, p. 111. The ß-radiograph of Paper 6 is almost identical to a ß-radiograph of a gasenshi paper reproduced by Dwan (p. 122 (The Duet, no. 2)). Gasenshi is an ideal paper for ink painting (sumi-e) and for calligraphy (shoclō). One of the most famous of the Chinese calligraphy papers it probably seeks to imitate is Xuan paper, which was first produced in Anhui province (central China) during the Tang dynasty (618–906). Xuan papers have always enjoyed a tremendous reputation among calligraphers, and continue to do so today. The highest grade of this paper is made from the bast fibre of the than tree (Pteroceltis tatarinowii Maxim.), while lower grades (often used by artists for practice purposes) are made from various ratios of than and rice straw.36 See Mullock, p. 28. Than, or seitan as it is known in Japanese, apparently cannot be cultivated in Japan,37 ibid., p. 26. and it is probably for this reason that gasenshi, the Japanese version of Xuan paper, is traditionally made from mitsumata and bamboo fibres.
Although mitsumata is native to Korea, China and Japan, it is almost exclusively associated with papermaking in Japan. It seems most probable, then, that Paper 6 is a Japanese gasenshi paper, though the possibility that it is Chinese in origin cannot be ruled out.
A bamboo paper
Paper 7 is a wonderfully large sheet with broad, uninked margins around the etched image and would traditionally be referred to by print scholars as a Japan paper. Superficially, it is visually similar in many respects – including colour and thickness – to the gampi papers. Close examination reveals it to be made up of two similar sheets that have been formed separately and then pasted to each other, rather than formed and dried together.38 The recto sheet extends 0.5 mm beyond the verso sheet, at the bottom edge. The adhesive exposed as a consequence on the back of the recto sheet was found to be starch-based.
The ß-radiograph of the bottom left corner reveals a darker rectangular area at the top right; this is the inked image area, the lighter surrounding areas being the uninked margins. There is a vague horizontal linear pattern, reminiscent of laid lines. The long diagonal lines, most apparent in the upper half of the ß-radiography, are due to the starch paste applied by brush to adhere the two sheets together. Like the mitsumata and bamboo paper (Paper 6), this paper contains fibres of varying lengths, and knots (the small black dots seen in the ß-radiography).
Although bamboo has a history of usage in Japanese papermaking, its role appears to be that of a filler rather than a fibre per se (it is added to bast fibres to produce a paper with certain qualities, such as gasenshi). The use of bamboo as a fibre on its own is not a practice commonly associated with Japanese papermaking. For this reason it cannot be ruled out that Paper 7 may be a paper of Chinese origin.
A posthumous impression of a Rembrandt etching on an imitation oriental paper
Paper A, which bears an impression of the Rembrandt etching Cornelis Claesz Anslo, Mennonite Preacher, 1641 (B.271, 2nd state), reworked by the Irish printmaker and dealer Captain William Baillie (1723–1810), illustrates the degree to which oriental papers could be imitated in Europe by the late eighteenth century. Hinterding notes that Baillie refers to such ‘oriental’ papers as ‘Indian paper’ or as ‘the finest Japan paper’, in his own catalogue, The Works of Captain William Burnie, of 1792.39 Hinterding, p. 272. The ß-radiograph of Paper A shows that it convincingly replicates the features of a Chinese paper, but as Biörklund revealed in 1968 the imitation oriental papers do not contain any oriental fibres at all.40 Biörklund, ‘Paper Analysis’, p. 175. Fibre analysis of Paper A shows it to be made of cotton.
A further point of interest: An eighteenth-century copy after Rembrandt, on a true oriental paper
The National Gallery of Victoria has in its collection a copy after Rembrandt’s print The Rat Catcher, 1632 (B.121). Made in the late eighteenth century by the English artist James Bretherton (active 1770–1781), this copy is printed on authentic Xuan paper (Paper B), probably 100 per cent seitan.
Conclusion
Within the context of the National Gallery of Victoria’s relatively small collection of Rembrandt etchings, it has been possible objectively to describe and group seven oriental papers in terms of fibre type and other information obtained by ß-radiography and careful visual examination. Although the majority of the papers reflect previous findings in terms of the dominance of gampi, there is such a startling diversity (particularly in the case of the gampi laminates) that any generalisations about Rembrandt’s oriental printing papers should be made with great care. It could be said, however, that in this small group all but Paper 7 exhibit characteristics closely associated with Japanese papermaking traditions and techniques. Clearly, more research is required. The examination of many more prints by Rembrandt on oriental papers, and of similar oriental papers of the same era, is essential before an understanding of this interesting aspect of Rembrandt’s oeuvre can be reached. It is hoped that this modest research will encourage others.
Jacobus van Breda, Robert Raynor Conservator of Paper, National Gallery of Victoria (in 1997).
Appendix
Areas shown in the ß-radiographs are indicated in the reproductions of the etchings. All ß-radiographs are reproduced at actual size. Triangular markings accompanying ß-radiographs indicate position of chain lines.
Paper 1
Comments
Gampi laminate with uncooked rice powder filler. High fibre quality; even distribution.
Approximate basis weight 168 gsm.
Paper 2
Comments
Gampi laminate uncooked rice powder filler. High fibre quality; fairly even distribution. Approximate basis weight 106 gsm. Note that some printed lines from the image are visible in the ß-radiograph.
Paper 3
Comments
Gampi laminate. High fibre quality; fairly even distribution. Approximate basis weight 123 gsm.
Paper 4
Comments
Gampi laminate. High fibre quality; even distribution. Approximate basis weight 200 gsm. Note diffuse vertical chain lines.
Paper 5
Comments
Gampi laminate. High fibre quality; fairly even distribution. Approximate basis weight 110 gsm. Note the diffuse vertical chain lines.
Paper 6
Comments
Bamboo and mitsumata. High to medium fibre quality, with small knots; fairly even distribution. Approximate basis weight 53 gsm. Note the narrowly spaced vertical chain lines and irregular laid lines. Note also that some printed lines from the image are visible in the ß-radiograph.
Paper 7
Comments
Bamboo (two sheets pasted together). Medium to low fibre quality, with some very long fibres and with small knots. Approximate basis weight 133 gsm.
Paper A
Comments
Cotton. Medium to low fibre quality, with fibres of various lengths and with small knots; fairly even distribution. Approximate basis weight 35 gsm.
Paper B
Comments
Pteroceltis tatarinowii Maxim. (known as than (China) or seitan (Japan)). Medium to low fibre quality, with fibres of various lengths and with small knots; fairly even distribution. Approximate basis weight 28 gsm.
Acknowledgements
Although I gratefully acknowledge the generous assistance of the following individuals, I am responsible for the conclusions reached in this article, and for any errors. I am deeply indebted to: Akinori Okawa, Senior Researcher at the Kochi Prefectural Paper Technology Centre, for undertaking the fibre analysis of the oriental papers; Katsuhiko Masuda, Director of the Department of Restoration Technique at the Tokyo National Research Institute of Cultural Properties, for organising the fibre analysis and for translating correspondence from Mr Okawa; and Timothy Barrett, of the University of Iowa Center for the Book, for examining the ß-radiographs and reviewing my comments on the gampi papers. I am also very grateful to the following people for their assistance: Jill Thurlow, Royal Botanical Gardens, Melbourne; Professor William Coaldrake, University of Melbourne; Dr Geoff Irvine and John Ward, CSIRO, Melbourne; Neil Holland, Scientific Document Services; Rosie Freemantle, Ian Potter Conservation Centre, University of Melbourne; Johann Alcock, State Library of Victoria; Ruth Shervington, Art Foundation Development Conservator of Paper, National Gallery of Victoria; Rodney Manning, Helen Skuse and Gary Sommerfeld, National Gallery of Victoria.
Finally, my sincere thanks to Irena Zdanowicz, Senior Curator of Prints and Drawings at the National Gallery of Victoria, for her encouragement and advice; and to Lyndsay Knowles, Senior Conservator of Paper, for her constant enthusiasm, unfailing support and ever-patient ear.
Notes
1 Rembrandt’s systematic use of these papers is usually considered to have begun around 1647, although he may have started using them a year or two earlier (see G. Biörklund, ‘Old Paper’, in G. Biörklund & O. Barnard, Rembrandt’s Etchings: True and False – A Summary Catalogue, rev. edn, Stockholm, 1968, p. 172; C. White, Rembrandt as an Etcher: A Study of the Artist at Work, vol. I, London, 1969, p. 16).
2 I am referring to the Rembrandt watermark project begun in 1981 by Nancy Ash and Shelley Fletcher at the National Gallery of Art, Washington DC. This project has resulted in the collection of over 900 watermarks and has given rise to a cataloguing system that groups them by general type and by degree of similarity. The results of the project will soon be published. The Conservation Department at the National Gallery of Victoria has recorded the fifty watermarks found on papers in the Melbourne collection. These watermarks have been catalogued, according to the National Gallery of Art’s system, by Rembrandt scholar Drs Erik Hinterding of the University of Utrecht in the Netherlands. Eight of the watermarks occur in papers carrying etchings by Rembrandt that were included in the recent exhibition Rembrandt: A Genius and His Impact, and are noted in the exhibition catalogue (see I. Zdanowicz, ‘Etchings by Rembrandt’, in A. Blankert, Rembrandt: A Genius and His Impact (exh. cat.), National Gallery of Victoria, Melbourne, 1997, cat. nos 103, 106, 107, 109, 112, 113, 115B, 116).
3 The earliest reference to Rembrandt’s use of oriental papers appears to be the allusion to ‘Indian’ paper that occurs in a letter from the English traveller Edward Browne (1644–1708), written to his father from Amsterdam in 1668: ‘Here is a strange variety of excellent prints … Here are divers good ones of Rembrandt and some upon Indian paper that look like washing, though scratched after his manner’ (Edward Browne, letter to his father, 1668, cited in E. Hinterding, ‘The History of Rembrandt’s Copperplates, with a Catalogue of Those That Survive’, Simiolus, vol. 22, no. 4, 1993–94, p. 264 n. 62). Browne’s use of the term ‘Indian’ was probably the beginning of many misunderstandings, caused by the vagaries of language. As Mayer Haunton points out, ‘Indian’ in this context may simply be a reference to ‘from the Indies’ or, more particularly, to ‘imported by the East India Company’ – rather than to India itself; a small number of papers used by Rembrandt do, apparently, bear a strong resemblance to India papers, and Rembrandt’s ownership of Mogul miniatures adds circumstantial evidence in support of his also having had access to Indian papers (K. Mayer Haunton, ‘Rembrandt’s Papers’, in A Collection of Etchings by Rembrandt Harmensz. van Rijn (1606–1669), Formed by Joseph R. Ritman (sale cat.), eds N. Bialler, A. T. Eeles, R. Godfrey & K. Mayer Haunton, Artemis & Sotheby’s, London, 1995, n.p.).
4 For example, Biörklund, ‘Old Paper’, p. 172, notes [in] the auction catalogue of 1755, when the important De Burgy collection was sold, the Dutch text mentioned 48 etchings printed “Op Oostindisch papier” which the French translator of the auction catalogue erroneously interpreted as “Papier de la Chine” (see also note 3 above).
5 Biörklund, ‘Old Paper’.
6 Gampi is a deciduous plant from the family Thymelaeacea. A few varieties are used for papermaking. Diplomorpha sikokiana Nakai is generally considered the best. Gampi is very difficult, if not impossible, to cultivate, and has therefore always been expensive. Although it grows throughout Southeast Asia, it is most closely associated with papermaking in Japan, where it has been used since at least the eighth century.
7 Biörklund, ‘Old Paper’, p. 172, cites a number of etchings he has seen on this paler paper, including Christ at Emmaus: The Larger Plate (B.87). The National Gallery of Victoria’s impression of this print is on what we will be calling Paper 6, which may well be this same paper.
8 Biörklund, ‘Old Paper’, p. 172.
9 Deshima measured a mere 200 x 600 feet. Some twenty Dutchmen at a time were in residence on the island. They lived under severe restrictions but the Dutch East India Company, for which they worked, maintained a very profitable business: according to records, profits averaged 81 per cent in the period 1649–62. See A. Hyma, The Dutch in the Far East, Ann Arbor, Michigan, 1942; G. K. Goodman, The Dutch Impact on Japan (1640–1853), Leiden, 1967.
10 Biörklund, ‘Old Paper’, p. 173. I am also grateful to Dr Christiaan Jörg of the Groninger Museum, Groningen, for the following information: ‘The amount of goods mentioned in the [Dutch East India] Company records never covers the total export from Decima [Deshima] by the Dutch in a given year, as the merchants were licensed (officially, it was not smuggling) to buy and ship their own private merchandise on return ships from Japan to Batavia/the Netherlands. So the two shipments of 1643/44 can in no way be conclusive’ (Dr Christiaan Jörg, letter to the author, 6 January 1997).
11 The only mention of fibre analysis found by the author to date in the literature published since Biörklund reported on his findings in 1968 appears to be that undertaken in the Asiatic Department of the Museum of Fine Arts, Boston, in 1970, but complete details have not been published; the conclusions drawn on the basis of this latter analysis appear to be similar to Biörklund’s (see S. W. Reed, ‘Types of Paper Used by Rembrandt’, in Rembrandt: Experimental Etcher (exh. cat.), Museum of Fine Arts, Boston, 1969, pp. 178–83).
12 See Goodman, p. 19.
13 The Dutch imported silk, deerskins and other products to Japan from Taiwan (see Hyma, p. 155).
14 For an account of the history of papermaking in the Orient, see D. Hunter, Papermaking: The History and Technique of an Ancient Craft, New York, 1974.
15 See H. Mullock, ‘Xuan Paper’, Paper Conservator, vol. 19, 1995, p. 26; see also Tsuen-Hsuin Tsien, ‘Raw Materials for Papermaking in China’, Journal of the American Oriental Society, vol. 93, 1973, pp. 510–48.
16 Hemp played an important part in early Japanese papermaking but seems to have been replaced by kozo and gampi by at least the twelfth century (see K. Masuda, ‘Japanese Paper and Hyōgu’, Paper Conservator, vol. 9, 1985, p. 32). Kozo (Broussonetia kazinoki Sieb.) is a variety from the family Moraceae, the family to which Broussonetia papyrifera also belongs. Mitsumata is usually cited as Edgeworthia papyrifera Sieb. et Zucc. (strictly speaking, it should be known as Edgeworthia chrysantha (see Y. Harvey, ‘Edgeworthia chrysantha’, Plantsman, vol. 14, no. 3, December 1992, p. 130)). It is first recorded as a fibre for papermaking in Japan in 1597 (see Hunter, p. 481). For gampi, see note 6 above.
17 See T. Barrett, Japanese Papermaking: Traditions, Tools, and Techniques, New York, 1983, p. 82.
18 ibid., p. 51.
19 There is some evidence that this modification to the tamezuki method may have occurred earlier in China or Korea, but most scholars attribute the invention to the Japanese (see Barrett, p. 70). Neri traditionally came from either the root of Abelmoschus manihot Medikus (known as tororo-aoi in Japanese) or the inner bark of Hydrangea paniculata Sieb. (noriutsugi in Japanese) (see Barrett, pp. 56, 287).
20 See Barrett, p. 51.
21 ibid., p. 70.
22 The ß-radiography process involves placing the paper between a poly[Carbon 14] methylmethacrylate plate and a sheet of Kodak Biomax MR film. A 1:1 black and white photographic contact print can then be made. ß-radiographs are useful because they reveal details about the paper structure, including fibre distribution and quality, and any characteristics imparted by the mould upon which the sheet was formed. Thin areas appear lighter on a ß-radiography.
23 For the similarities between gampi and mitsumata fibres, see G. Biörklund, ‘Paper Analysis’, in Biörklund & Barnard, p. 175; S. Beauman Murphy & S. Rempel, ‘A Study of the Quality of Japanese Papers Used in Conservation’, in The Book and Paper Group Annual, vol. 4, ed. J. Abt, USA, 1985, p. 68.
24 This process was documented in detail (National Gallery of Victoria (Conservation Department) files).
25 In carrying out this analysis, Akinori Okawa used water to disperse the fibres on a glass plate, dyed them with C-stain solution, and then examined them microscopically.
26 Akinori Okawa points out that examples of gampi papers with rice powder fillers date back to the Heian period (9th–11th centuries) (letters to the author, 1997 (translated by Katsuhiko Masuda)).
27 This has been suggested previously (see Biörklund, ‘Old Paper’, p. 171).
28 See Barrett, p. 67. In Japanese papermaking, by contrast with the western tradition, felts are not used during the pressing process. The newly formed wet sheets are piled directly on top of each other, lightly pressed, then separated and brushed onto boards to dry in the sun. The tendency of gampi papers to stick together after pressing causes problems for papermakers.
29 I am very grateful to Timothy Barrett for pointing out to me the significance of the broad and diffuse chain lines seen in the ß-radiographs of Papers 4 and 5. I am also indebted to him for providing me with a unique sheet of handmade Japanese paper, which, when ß-radiographed, exhibited many of the characteristics seen in the ß-radiographs of Papers 1, 4 and 5. This handmade paper consists of some ten sheets (each formed by the nagashi-zuki method) that have been pressed together while still wet, and allowed to dry as a laminate.
30 See Hyma, p. 140.
31 See S. Hughes, Washi: The World of Japanese Paper, Tokyo, 1978, p. 45.
32 ibid., p. 171; A. Dwan, ‘A Method for Examining and Classifying Japanese Papers Used by Artists in the Late Nineteenth Century: The Prints of James Abbott McNeill Whistler’, in Conservation Research, Studies in the History of Art, vol. 41, Washington DC, 1993, p. 108.
33 See Hughes, p. 191. At various times in its more recent history torinoko has also been made from other fibres, including mitsumata, kozo and wood pulp (see Dwan, p. 113).
34 B. Jugaku, Papermaking by Hand in Japan, Tokyo, 1959, p. 16.
35 See Dwan, p. 111. The ß-radiograph of Paper 6 is almost identical to a ß-radiograph of a gasenshi paper reproduced by Dwan (p. 122 (The Duet, no. 2)).
36 See Mullock, p. 28.
37 ibid., p. 26.
38 The recto sheet extends 0.5 mm beyond the verso sheet, at the bottom edge. The adhesive exposed as a consequence on the back of the recto sheet was found to be starch-based.
39 Hinterding, p. 272.
40 Biörklund, ‘Paper Analysis’, p. 175.