We all bring different perspectives to our study and all benefit from the findings of those working in other sectors. I came to the forensic analysis of paper from papermaking. The practical experience of making sheets, by hand, for a wide range of uses stood me in good stead as my work evolved away from production into analysis. In the late 1970s I began to collect paper and to analyze the particular properties of individual papers in order to make better paper myself. The ways individual makers, at different periods in history, had approached similar problems – the continuity of raw material supply, pulp preparation, surface treatments and finishing techniques – in order to achieve particular results in the paper when it was in use, have become, for me, an absorbing study. Over the years the collection has grown and grown and now threatens to overrun all available space. However it has served as the foundation of my forensic study of paper. Given the variety of papers one is called upon to examine, reference material is essential: mine consists of a library of fibers and raw materials, a collection of paper samples, and a library of books and journals.1 The growth of fraud and counterfeiting in so many fields (e.g., banknotes, drawings, books, watercolors, forms and documentation, packaging of pirated goods) is demanding much more of the Forensic Paper Analyst than in the past. In trawling for information relevant to any particular piece of paper, one has to cast one's net widely. Reading the Paper: Context and Interpretation in the Analysis of Paper peter bower _____________ notes 1. Currently my collection includes some 1,000 microscope slides of fibers, 90 different papermaking plant samples, some 200,000 paper samples, several hundred publications, and a gallery of thousands of color transparencies and black and white prints. 2. In dealing with colored paper, stock publications such as Sandoz Chemicals' Acid and Basic Dyestuffs on Paper (Basel, 1951), which contains 233 tipped-in examples of various dyes applied to three different fiber furnishes, has proved indispensable. 3. J.J. Francois de Lalande, L'Art de Faire Le Papier (Paris, 1761), plate II, fig.1. 4. Eric Hebborn, Drawn to Trouble: The Forging of an Artist (London, 1991). Charles Black and Michael Horsnell's Counterfeiter: The Story of a British Master Forger (London, 1989), is another highly selective tale, an amusing but rather shallow "autobiography" of Black who specialized in English currency." autobiography" of Black who specialized in English currency. Examining a seventeenth-century sheet at a "Reading the Paper: Indentification of Paper "Examining a seventeenth-century sheet at a "Reading the Paper: Indentification of Paper" workshop at the Art Gallery of New South Wales, Sydney, one of a series of workshops and workshop at the Art Gallery of New South Wales, Sydney, one of a series of workshops and lectures in Australia and New Zealand in October and November 2004. All photos courtesy lectures in Australia and New Zealand in October and November 2004. All photos courtesy of the author. of the author. 24 • hand papermaking 24 - hand papermaking The literature of the papermaking industry, whether it be recent The literature of the papermaking industry, whether it be recent publications or material published years ago, is indispensable. publications or material published years ago, is indispensable. For example, when determining the probable age of a toned For example, when determining the probable age of a toned business paper, the identification of the coloring agent used can business paper, the identification of the coloring agent used can tell you much about the source of the sheet, which in turn can tell you much about the source of the sheet, which in turn can help in dating it.2 Mills regularly send me samples of their latest help in dating it.2 Mills regularly send me samples of their latest products together with relevant technical data. Any information products together with relevant technical data. Any information on current or recent practice is of course highly sensitive, but Ion current or recent practice is of course highly sensitive, but I have found most mills to be extremely helpful where suspected have found most mills to be extremely helpful where suspected forgery or fraud are concerned. However it should be remembered forgery or fraud are concerned. However it should be remembered that the literature is often not going to answer whatever your specific that the literature is often not going to answer whatever your specific problem is, but it is essential in pointing the way. Every case problem is, but it is essential in pointing the way. Every case is unique and what is described, as best practice, in much of theis unique and what is described, as best practice, in much of the technical literature, is more theoretical than most authors would technical literature, is more theoretical than most authors would care to admit: the mills have developed very specific variations of care to admit: the mills have developed very specific variations of basic techniques, some of which are recognizable under the mos tbasic techniques, some of which are recognizable under the most careful examination. One area of research that sometimes needs to be treated with some caution are publications on papermaking from earlier centuries: in particular encyclopedia entries or learned society treatises that discuss paper and its making. One classic instance of this problem is an engraving of a waterwheel and water filtration beds found in Lalande's L'Art de Faire Le Papier.3 Although the settling ponds (filtration beds) illustrated are an excellent idea and were used by many mills at a later date, there is no archeological evidence that any mill in Western Europe was actually using such a technique for insuring the cleanliness of its water for nearly a century after the original date of this 1698 engraving. The plates found in Lalande were originally engraved by L. Simonneau for a text by Gilles Fileau de Bilettes. On a lighter level, the autobiographies of counterfeiters have occasionally proved useful, but they need to be treated with great caution. Eric Hebborn's recent autobiography, documenting his production of "Old Master" drawings, is perhaps as truthful as the works he produced, adding yet another layer of deception to an already complex web.4 Sometimes no special techniques are needed. In a recent case I was asked to examine the paper ground of an oil sketch by Raphael, over which there was some confusion. The art historians were arguing about the handling of the paint and whether this was by the hand of the master or one of his contemporaries. A brief examination of the paper – simply holding it up to the light – showed both hypotheses to be wrong: the paper was wove (only invented in the 1750s5); and judging by the details of both the relatively coarse profile of the forming wire and the configuration of the support wires underneath, I determined the paper to be English, dating from around 1800.6 Not a Raphael but an interesting copy by a much later hand. Over the past century and a half, an accelerating technological progression in the manufacturing side of the industry has been coupled with an increasing specialization in the products. There is a vast difference between a basic wrapping paper and a quality bookpaper, despite the fact that they will probably both be printed on. In the eighteenth century the main difference between these two types of papers was simply the quality of the raw materials and their preparation. Lower grades of the same rags were used for coarse wrappings, but strong wrappings might well be made from rag grades equal to those used for the finest writings and printings, although less care would have been taken in the quality control of the finished sheets. Waterwheel and filtration beds from an English reprint of Lalande's L'Art de Faire Le Papier [The Art of Making Paper], published by John Hinton in Universal Magazine (London, 1751). In the past, the similarities in raw material and production methods made most papers of a particular type usable for other purposes. Hence it is common to find various wrapping papers being used by artists for drawing and painting – quite simply they liked the colors, tones, and surface characteristics. The increasing specialization of most paper mills has led to much less versatility in terms of what any one paper is capable of, but nevertheless it is still common to find artists, in particular, using papers for purposes quite other from what they were originally designed for: fax papers for ink drawing; business stationery being used for drawing, woodcuts, and even engraving; and legal and security papers being used for calligraphy with a quill or a steel nib. There is such richness of possibility for artists in the choice of paper available: presently there are some 14,000 designed end uses for papers now, compared with the half dozen distinct uses 200 years ago. But this very diversity can actually help rather than hinder the search for origins or dates for a particular paper. Basic types can be recognized and a search can be initiated into specific categories. But the difficulty of such searches is compounded by the level of import and export of both raw materials and finished paper across the world today.7 Scientific analysis of the sample works complements the research into the possible background history of the sheet, its results suggesting further avenues for possible origins for the paper. The combination of the scientific data and the historical and contextual interpretation of that data is crucial. Too many reports are written in such vague terms as to be practically useless. All too often the analysis of paper appears to come down to general interpretations of erroneous facts. In the case of an old master drawing or an early nineteenth-century manuscript, it is infuriating to read this simplistic analysis: because it is made of linen rag, nothing about it contradicts the purported date. Linen rags were in use all over Western Europe and beyond. They were processed in different ways at different times and places, and particularly from the nineteenth century onwards, linen was found in combination with a wide range of other fibers. Much can __________________________ notes 5. The actual date of the invention of wove paper is placed c. 1754–56. James Whatman, the elder, made the first wove paper for Baskerville, who wanted a printing surface that would not distort the serifs of the smaller point sizes of his new typefaces. We will have to wait for the publication, after years of research, of John Balston's forthcoming The Whatmans and the Invention and Development of Wove Paper in the West, for aand Development of Wove Paper in the West, for a more accurate dating. 6. Occasionally one is very lucky in one's research: in this instance the vatman who had been forming the paper had not kept his mould clean and pulp had stuck between the wires of the forming surface and the grid of support wires underneath. This distorts the proper drainage of the mould during use, leaving thinner areas in the sheet in the pattern of the support wires which can be "read" as easily as a watermark. 7. The best international directory of the world's paper trade and one way into this maze, is the annual edition of Birkner. It covers every aspect of papermaking and the paper trade. It is available in book form or on CD-ROM from Birkner & Co., Postfach 54 07 50, D-22507, Hamburg, Germany. 8. R.H. Clapperton. Paper, An Historical Account of its Making by Hand from the Earliest Times down to the Present Day (Oxford, 1934), pp. 1–26. 9. T.J. Collings and W.D. Milner, "An Examination of Early Chinese Paper," Restaurator 4 (Copenhagen, 1979), pp. 129–151. Discussing what one needs to be able to see in a sheet at a "Reading the Paper: Identification of Paper" workshop at the Art Gallery of New South Wales, Sydney, one of a series of workshops and lectures in Australia and New Zealand in October and November 2004. Polarized light image of a blend of linen rag and hemp fibers found in a paper dating from 1804. Polarized light is used to bring out individual details of the fiber structure.details of the fiber structure.

be deduced or interpreted from a thorough examination of what is present. The nature of the "whiteness" of paper can tell one much. White papers, for most of papermaking history, were rarely as white as we see today, particularly since the advent of optical brightening agents in the 1950s. Tones of white, in conjunction with other evidence, are very useful in showing one where to begin to look for the origin of a particular sheet. Grayness or yellowness are definite indications of different areas where papers might have been made. The soaps used to wash the raw materials, before they were rags, varied from country to country and these soap residues had distinct effects on the processes of fermentation used to begin the breakdown of rags until the latter part of the eighteenth century. Have we one type of linen or more than one type present in the furnish? Has the color of the paper been "corrected" by the use of blueing agents, such as blue linen fiber, smalts, or dyes such as indigo, ultramarine, or more rarely, Prussian blue? Were the fibers beaten by stamper or Hollander? Stamper beaten rag is often recognizable. Is there any trace of bleaching of the fibers, i.e., the presence of shives and specks, not so amenable to the chlorine bleaching powders that came to be so prevalent in the nineteenth century? Are there other "rag" fibers present, such as cotton or hemp? What, if any, sizing agent has been used? Are any loadings present, such as china clay or pigments? How has the surface finish been achieved: plate glazing or or calendering? Microscopes and testing can tell one much about specifics, but unless the particular details are placed in their correct context in terms of actual historical practice – whether that be from 1500, 1700, or yesterday – and are actually understood, all one has is raw information rather than knowledge. Over the years different analytical techniques, better equipment, and deeper experience has led to the re-evaluation of some earlier researchers' results. Recent analysis of some of the earliest surviving Chinese papers, deposited at the British Museum in 1909 by Sir Aurel Stein, shows that hemp, derived from waste cloth and rope, was one of the most important fibers used in early Chinese papermaking. Clapperton discussed the use of hemp extensively in early Chinese papermaking, but his actual fiber analysis, done in the 1930s, of 15 different papers from the Stein collections, showed no hemp present. Clapperton's interpretation was ramie and paper mulberry.8 In 1970 Collings and Milner were given permission to remove small samples of paper from points adjacent to those from which Clapperton had removed his samples. Using different techniques Using different techniques and more modern optical systems, their studies were based on comparisons with the fundamental identifying characteristics of individual fibers derived from botanical specimens. Their conclusions were startlingly different.9 While blends of ramie and paper mulberry were present in 14 of the samples, 10 of them also contained often large amounts of hemp, including five samples which Collings and Milner estimated as containing more than 90 percent hemp.10 Few watermark researchers seem aware that there are a whole series of factors, determined by the actual making of the sheet, that produce subtle and, sometimes, not so subtle changes in a watermark image. Evidence can also be gathered by looking at the wire profile of the sheet. Paper made on the same mould with the same watermark can exhibit enormous differences due to varying drying times, whether or not the sheets are dried in spurs, the number of sheets to the spur, the thickness and density of the sheet, couching faults, fiber choice, and beating details. Raking light image of a writing paper drawn on with graphite. Raking light image of a wrapping paper drawn on with very soft graphite. These factors all contribute to the finished size, proportion, and clarity of the mark. The dried sheet size can shrink from the original formed sheet size by as much as three quarters of an inch in any direction during drying, with a resulting difference in the watermark. If the sheet shrinks proportionally more in one direction than the other, then the watermark's actual proportions will change as well as its size. Published in 1987, Phillip Pulsiano's bibliography of published books and articles on watermarks documents well over 100,000 watermarks. Although incomplete, his listing is the single most useful bibliography of watermarking so far.11 Given the dramatic increase in watermark research in recent years, both in North America and Europe, someone could do us all an inestimable service by producing a supplement, bringing it up-to-date.12 Some of the problems encountered with watermarks can be illustrated by discussing a particular and very common watermark, 1794 / J WHATMAN. In my research into the use of paper by J.M.W. Turner R.A. (1775–1851), I became aware that many art historians had often assumed that all the different examples of the papers watermarked found amongst the 20,000 works on paper in the Turner Bequest, are the same.13 Papers were being made for many purposes and each different use had its own sheet size and weight, even if the papers were given the same name. In Turner's case, 23 different papers have been identified in the Bequest, some of which are found in three different finishes and all in different weights. It should be noted that in many instances only parts of these sheets can be found in the Bequest. I found that there was often considerable variation in size between different batches made on the same moulds. Changes in the furnish used, degree of beating, weight of sheet, and seasonal differences in relative humidity and temperature could all affect drying times and the shrinkage in the sheet as it dried, leading to variations in the size of the finished paper. The basic watermark followed standard positions in the sheet for laid and wove, but the scale of the lettering varied, depending on the size of the sheet: bigger sheets generally had bigger watermarks. The marks were handmade and thus vary slightly from one to another. Turkey Mill used double moulds for the smaller sheet sizes, giving two subtly different marks on each mould in the pair. Some of the Post, Large Post, and Foolscap sheets were watermarked twice in each sheet, giving four marks on each mould and eight to the pair. It should also be remembered that the the 1794 / J WHATMAN watermark, in common with many other British watermarks during this period, was actually in use for many years. After the introduction, in April 1794, of a statutory requirement to include the year in the watermark in papers to be used for particular purposes,the year in the watermark in papers to be used for particular purposes, most English paper was dated in the watermark.14 Many papermakers, however, at least at first, irregularly changed the dates on their moulds, usually when a particular mould needed replacing.replacing. Although the Act's regulations pertain specifically to printed paper, some of the moulds used, say, for Demy Printing could also be used for Demy Drawing. These sheets are often described as having different sizes. As I mentioned earlier, variance in weigh and fiber beating can lead to different amounts of shrinkage in the drying sheets, even when made on the same mould. Pairs of moulds had different degrees of wear and tear depending depending upon the amount of paper that a mill needed to make of each particular size. Consequently, those moulds which were three examples of different 1794 / j whatman watermarks found in watercolors and drawings by J.M.W. Turner R.A. (1775–1851). Note the subtle differences in the letterforms and spacing, for example, the number "7" and the way the "W" is formed. Detail of Large Post (16 - x 19 inches) wove writing paper. used lightly did not need replacing until much later than those moulds in more regular use. In the case of the moulds used by Balston and the Hollingworths at Turkey Mill which carried the J WHATMAN name, the next known date is 1795 in some laid writing papers, but 1797 in some wove papers. Other Whatman wove papers, however, do not appear with a new date until 1801 or, in the case of Double Elephant, until 1804. The next date usually found in Whatman wove papers is 1808.15 The Act was repealed in 1811, but by that date most English makers were in the habit of changing the dates on their moulds quite regularly.16 No investigation takes place in isolation: each object has its own complex history, produced in a certain way, at a specific time, by a particular individual. Much of that evidence is either directly visible or can be identified through vigorous analysis. The investigation of any questioned or previously unseen document or work of art on paper can be hard, painstaking, time-consuming work, sometimes tedious and repetitive but often absorbing. Some projects are cursed with a sort of chaos where communication breaks down – objects are misplaced or simply suddenly not available, cannot be found in the stacks, or permissions not easily forthcoming. Other projects swim along, enjoyable and even relaxing. Perhaps surprisingly I have often found that serendipity and synchronicity (non-causal effects through chance, proximity, or time) have often been essential parts of such enterprises. There have been numerous times when I have been working on a particular piece of research and have come across something of crucial import, maybe not to the job at hand but to a completely different project. It is critical to retain an open mind, be surprised by what you may find, and enjoy the process. I find if I am having fun I do much better work. __________________________ notes 10. Collings and Milner, op. cit., tables 1 and 2, pp. 130–131. The five 90%+ hemp samples were: S 116, early fifth century, 90% hemp, 10% ramie; S 88, early fifth century, 100% hemp; S 312, AD 637, 95% hemp, 5% ramie; S 912, AD 803, 95% hemp, 5% ramie; S 86, AD 991, 90% hemp, 10% ramie. My own examination of Collings and Milner's microscope slides suggests the same fibers and in the same proportions. 11. Phillip Pulsiano, "A Checklist of Books and Articles Containing Reproductions of Watermarks" in Stephen Spector, Essays in Paper Analysis (Washington, 1987), pp. 115–153. 12. For example, just one publication, the British Association of Paper Historians' journal, The Quarterly, has illustrated and documented nearly 200 watermarks in their series "British Watermarks and other articles. 13. Peter Bower, Turner's Papers – A Study of the Manufacture, Selection and Use of his Drawing Papers 1787–1820 (London: Tate Gallery, 1990). Peter Bower, Turner's Later Papers – A Study of the Manufacture, Selection and Use of his Drawing Papers 1820–1851 (London: Tate Gallery, 1999). 14. Act of Parliament (1794) 34 Geo. III c.20. 15. The author would very much appreciate any information on any other dates in Whatman papers from the period of 1794 through 1811, particularly in wove papers. 16. Act of Parliament (1811) 51 Geo. III c.95. Detail of Imperial (22 x 30 inches) wove watercolor paper.