10 
Paper
10 DAVEN CHRISTOPHER CHAMBERLAIN
Paper is a sheet material consisting of overlapping vegetable fibres that bond together to form a compact mat. Its origin can be traced to China, where a court official named Cai Lun is said to have invented it c. AD 105. Paper remained exclusive to the Middle East and East Asia until around 1151, when there is evidence of its being made in Spain during the Moorish occupation. Slowly, the craft spread across Europe, from Italy in 1276 to France in 1348, Germany in 1390, and eventually into England by 1495.
Each sheet of paper bears physical signs of the manufacturing process it has undergone. Coupled with analysis of its constituents, these can be used to provide evidence of provenance. This essay presents information on paper as an historical source; although it is chiefly concerned with occidental manufacture from its inception to the present day, its generalities can be applied, with due consideration, to paper from any period or origin.
Paper can be made from all manner of vegetable fibres, in a process that has remained largely the same for two millennia. Early Asiatic papermakers used various specially harvested and prepared plants as their fibre sources. By contrast, the first European papermakers used textile waste, cordage (rope), and other pre-processed material. It was only in the 18th century that occidental papermakers and scientists investigated raw plants for their papermaking potential; full commercialization of suitable materials did not start until at least the 19th century.
Regardless of which source of fibrous raw material is used, the same basic steps are followed to turn it into paper. The first process involves cleaning and purification. For rags and cordage this means cutting them into small pieces, dusting them to remove fine matter, cleaning, and softening. The earliest cleaning method involved steeping in water and allowing them to rot partially, and then washing them. Adding alkali, such as lime, helped to speed the disintegration, though this was banned at various times because of the detrimental effect it could have on fibre quality. More recently, processing involved boiling the shredded textile with alkali. The shade of the pulp produced by these methods depended upon the initial textile colour and water quality; by the 1790s, to improve colour and to remove dyes, chemical bleaching was applied in papermaking.
Processing of raw plant matter follows a similar path, except that the unwanted constituents need to be dissolved rather than washed away. Strong alkali or acid is used as a chemical treatment, sometimes following a retting (or soaking) stage in which biological pre-treatment softens the fibres. Again, bleaching may follow before the pulp is ready for use.
Plant or textile fibres prepared in this way are still not suitable for paper-making; they need mechanical treatment to separate the individual fibres, modify their length, alter their flexibility, and increase their surface area to promote bonding. In the earliest European mills, this process was performed using either kollergangs or stamping mills. Kollergangs were large millstones set on their rims, running in a stone trough; they twisted and crushed batches of fibres without cutting them, in a way similar to milling grain or pressing oilseed. The other process involved pounding the wet fibres with large wooden hammers shod with nails; these ‘stampers’ tended to flatten fibres and break their outer layers, resulting in more fibrillation than kollergangs. Both processing methods resulted in a strong, very long-fibred pulp. The main change in macerating techniques occurred in the late 17th century, when the Hollander beater was introduced. This processed larger batches of fibre faster than either of the previous machines, and by the late 18th century its use was widespread. It could produce in hours what took days in a stamper or koller-gang; however, there was a trade-off. The rapid processing resulted in more stress to the fibres, causing significant breakage and a consequent reduction in their length and strength; there were also more fibres of varying length, including many very short fibres. In the late 19th century, as continuous machine production became dominant, the last fibre-processing method, refining, gained prominence. Beating is a batch process, whereas refining is continuous, with the fibres being fed by pipeline through a series of rotating barred discs or cones; the end result is more homogeneous than beating, with much more uniform fibre lengths.
After mechanical processing, the fibre suspension is diluted to the required consistency, mixed with chemicals, and introduced to the forming stage. In a handmaking operation, the vatman uses a mould to scoop a quantity of the stock, which drains through the porous wire mesh at the bottom to form a sheet. This is then passed to a second worker, the coucher, who transfers the wet fibre mat on to a textile, usually of felt, to support it during the next process, pressing. A stack of alternating wet mats and felts, called a post, is introduced into a mechanical press, which expels much of the surplus water. Pressing may be repeated further times with the paper sheets placed together rather than with felts interleaved, before they pass to the drying stage.
Hand papermaking from G. A. Böckler’s Theatrum Machinarum Novum (Nuremberg, 1661). With water-powered hammers, the linen rags are pulped. The vatman stands at the vat with a mould; the coucher presses the post; the drying sheets hang on ropes above, ready to be sized, calendared, gathered into reams, and packaged. Courtesy of Alan Crocker
The still-damp sheets are then separated from the pile in ‘spurs’ of four or five, which are dried as a group, as this produces a flatter sheet than if they are dried singly, when excessive curl and cockling can result. The commonest drying method involves draping the spur over ropes in lofts, where air flow regulated by louvres allows them to dry over a period of days. Other methods include pegging the spurs by the corners and hanging them in a loft, and ‘sail drying’, where the spurs are laid horizontally upon canvas. Draping the sheets over ropes results in a ‘back’, and pegging produces localized compression at the corners; both of these methods leave tell-tale signs in sheets.
If sized paper was required, it was made traditionally in a separate process: the dried waterleaf (unsized) sheet was passed through a solution of size, such as animal glue or starch, before being pressed and re-dried. The sheets were then left in stacks to mature, during which time the dried-in stresses relaxed and flattened. By contrast, some modern hand-mills have employed reactive sizes, that are mixed with the fibres prior to forming; in such cases the sheets are fully sized after drying and do not pass through a separate sizing process.
The machine manufacturing process is very similar to that outlined above. The macerated fibre is diluted and mixed with required additives, including sizing chemicals, after which it passes to a moving mesh on which it drains, then through a press section, where it is supported by felts. On early machines the pressed wet web was wound on to a drum, then slit by hand, and the wet sheets were taken to a drying loft; in 1820, however, steam-heated drying cylinders were invented, which allowed the web to be dried prior to reel-up. During this process, the web is pressed against the heated cylinders by stretched textiles, which help maintain contact between paper and hot surface, and exert restraint that helps counteract lateral shrinkage.
Finally, it should be noted that some sheets are made from a multitude of layers. They can be formed from plies brought together in the wet state and glued or pressed to form a single entity, or from previously dried sheets glued together to form a board. Multi-ply sheets made on hand-moulds are usually of board weights; those made on paper machines could be of board or paper weights. A recent invention in machine manufacture allows the various furnish fractions to be separated, then recombined at the forming stage to create a multi-layer product in which the properties of the outer plies are tailored to the needs of the final paper, while the inner section contains cheaper components. An example of this is a printing paper whose outer layers consist of virgin chemical wood fibre, and the inner layer of cheaper recycled fibre.
Although this description of the process of fibre preparation and sheet formation, as practised over almost 900 years, is highly simplified, it should be obvious that practices have developed so much during this period that clues will reside in each sheet to aid its identification. The following sections describe various ways that such clues may be read.
The heterogeneous nature of paper can be observed by holding a sheet to a light source and looking at its cloudy appearance; this is known as formation. It is a measure of how evenly the various furnish components are distributed throughout the sheet. Distinction depends mainly upon the fibres, their length, and how well they were macerated and individually separated; it also indicates the degree of their dilution and agitation prior to forming. The presence of undispersed fibre bundles, filler agglomerates, dirt, and other contaminants will also be disclosed by looking through the sheet, as will any fault caused during forming, pressing, or sizing operations—air bubbles and pinholes, which disturb formation, and bloom of sizing, which gives a milky sheen to the sheet.
The condition of the mould is also apparent at this point. For example, Bower has noted a case where a sheet showed a preponderance of fibre in the centre compared to the edges; he attributed this to poor attachment between the mesh and frame of the mould, resulting in sagging of the forming surface.
If one sheet is being compared with another, the degree of similarity between them may also be ascertained. In general, if they have been machine-made, the inter-sheet variability will be much less than for hand-made paper. Furthermore, in machine-made papers, especially those made by modern machines, the presence of dirt should be minimal, because efficient cleaners are used before the forming stage; however, a notable exception is with sheets made from recycled fibre, which may have dirt as a contaminant even with current methods.
Looking through the sheet will also reveal the impression of any relief surface upon which the wet fibre mat was laid prior to drying, or that was pressed into the damp or dry web or sheet after forming; these include: watermarks; wire or mesh impressions; laid (wire) and chain lines; and press, felt, rope, and emboss marks. This last group are formed by compaction of the sheet rather than by lateral movement of mass.
Wire marks are the imprint of the mesh through which the water drained as the sheet was formed; they are further enhanced during couching, when the wet sheet is pressed against a textile material and the wire is pushed into the wet fibre mat. The shape of the mark will not be an exact representation of the original wire, because the sheet shrinks as it dries. This is especially significant for machine-made paper, where lateral shrinkage is high—especially at the web edges—typically leading to the formation of a diamond-shape for the wire mark. Even two hand-made sheets from the same batch can also appear dissimilar, however, if they experience different degrees of shrinkage during drying.
The gross nature of the forming surface will also be visible during drying. In particular, the presence of laid and chain lines will be apparent; wove moulds only became available after 1756. Shadows around the laid lines give information on the barring pattern under the forming mesh, including the presence of tranchefiles, which were added at the edges of deckles to improve drainage. Papers made on Fourdrinier machines have laid lines imparted by a dandy roll pressing on to the felt side of the sheet; on cylinder machines and in hand-made products, these lines come directly from the forming wire and are a feature of the wire side of the sheet. A patent for a dandy roll to impart laid lines was granted in 1825; before that date, all paper from Fourdrinier machines was wove.
Felt marks come from the textile on which the wet sheet is couched. They are enhanced during pressing, when both sides take on some of the surface characteristics of the felts. Sometimes this is exacerbated by using heavily textured fabrics for the final pressing, which impart specific characteristics to the sheet surface. For example, some sheets are given a ‘linen’ finish by pressing against a coarse-textured fabric with a square weave. Alternatively, some machine-made papers are textured on-machine by having various materials attached to the press fabrics; this is called felt-marking. Use of a textured sleeve on a press roll produces a press-mark.
Prominent wire and felt marks are typically not found on very modern machine-made papers, as recent industrial developments have resulted in machine clothing that minimizes marking of the finished sheet. Rope marks come from the cord over which wet sheets were draped during drying. Occasionally, hairs from the rope can also be found embedded in this area. Emboss marks are imparted to a plain wove material when dry, by passing the sheet through a calender (a machine for smoothing or glazing paper) containing a textured roll.
In hand-made paper, watermarks from adjacent sheets may transfer when they are pressed without intervening felts. This is more likely with bulky sheets, and is generally less problematic with lightweight papers. Hand-made sheets more commonly suffer from occasional random imperfections, such as the slur of laid lines, caused by the coucher slipping during the act of couching, or spots of water from the vatman’s hands flicking on to the newly formed sheet.
Marks made while the sheet was still very wet, such as laid and chain lines and watermarks, are permanent. Although some relaxation may occur during humidification or re-wetting, such as happens during wet conservation treatment, they remain largely unchanged because their formation involved physical displacement of fibre. By contrast, marks made when the sheet was drier, such as emboss marks, can be removed almost completely by wetting or humidification. One way of differentiating a ‘true’ watermark from those formed in the press section, or by printing or embossing techniques, is by imaging, using beta-radiography rather than optical techniques; only a ‘true’ watermark is visible to radiographic processes.
This discussion concerns the distribution of materials within the plane of the sheet. There also tends to be stratification of material through the thickness of the sheet, however, though this is less obvious during simple visual inspection. Microscopic examination shows quite conclusively that fine materials tend to be concentrated more at the top and in the middle of the sheet, with far less at the surface that contacted the mesh of the forming stage. This is most notable on machine-made paper—especially that made on fast, modern machines where high suction is used to remove water rapidly during sheet formation. The construction of multi-ply sheets is also only revealed by microscopic investigation.
Watermarks are one of the main means of identification for assessing paper provenance. Technology follows a simple chronology whereby early examples consisted of bent pieces of wire, tied to the mould face by thinner bits of wire that leave sewing dots. When watermarking was first started on Fourdrinier machines (c.1826), the same procedure was used, with bent wire being tied to the dandy roll cover in a similar manner. Embossed mould covers were invented c.1848; these allowed the development of more complex watermark designs, with a wider optical density range, resulting in complex pictures that are known as light-and-shade or shadow marks. Eventually, this technology was transferred to dandy roll covers. By 1870, soldering was introduced as a means of attaching the bent wire that formed the simple binary watermark image; soldering often replaced sewing because it was quicker. Finally, in more recent years electrotyping was developed as a means of mass-producing the raised designs for binary watermarks; these images could be attached either by sewing or soldering.
Watermark imagery and design are also important. Before the 19th century, watermarks were simple binary line-marks that Dard Hunter divided into four main categories: simple images, such as crosses or circles; human forms and works, such as heads and hands, keys and pottery; animals, including mythological beasts; and flora and images from nature. Eventually, some items from the second category, depicting human creativity, became synonymous with various paper sizes: beakers and pots (pot), and the fool’s cap (foolscap). By the era of machine-manufactured watermarks, only these images were commonly in use. Some decades later, the trademark was devised: manufacturers, stationers, and latterly customers, introduced named grades of paper to the marketplace, the most famous being Conqueror from Wiggins Teape. Today, trademarks and geometric designs are undoubtedly the commonest form of watermarks from machine-makers; by comparison, hand-makers produce all manner of images, the main requirement being that the design is not too intricate to reproduce and that the image can be fixed to the mesh cover without danger of removal or fouling during normal working practice.
Many catalogues have been produced in which the general imagery of watermarks can be studied. However, it must be remembered that many of these catalogues (such as Charles Moïse Briquet’s or Edward Heawood’s) were developed from rather limited sources, and that altogether they contain only a small portion of the whole canon of marks produced since the earliest Western examples, from Fabriano, Italy, in c.1282. Furthermore, most watermark publications contain simple tracings that supply insufficient detail for full identification.
Countermarks, often found associated with watermarks, give added information such as dates, mill, and maker identification. Their uses have varied through history, but they are probably the main method by which papers from certain periods are assigned to specific mills. However, the best any watermark or countermark can actually do is identify the original mould or dandy cover used to make a specific sheet, since both are portable and can be used at sites other than their origin. Cases in point include the removal of dandy rolls used to mark postage stamp paper, from Chafford Mill to Roughway Mill in 1878, and the various mills from which Whatman watermarks originate, especially prior to the opening of Springfield Mill in 1807.
Dates in countermarks are another common source of problems, since the apparently simple information they convey may not always be correct. For example, a French ordinance of 1741 required all paper makers in the country to include a date in the countermark; many manufacturers complied by using the date ‘1742’ for a number of years, since it was not stated in the original directive that the date should change every year. A similar order, this time containing the stipulation that dates should change annually, was issued in England, in 1794.
This should mean the date in English paper from this period is more reliable; Kelliher, however, has highlighted an intriguing publication printed in 1806 on paper dated 1807. In addition, Hunter described the case of Joseph Willcox, an American hand-maker known to have continued using a mould dated 1810 many decades later. In sum, then, the date in a countermark must be treated with as much caution as any other mark in a sheet.
The formal comparison of watermarks and countermarks can be performed on two levels. Superficial comparison requires the imagery and countermark information in the sheet under examination to agree with either a catalogue picture or other reference sheet of paper. For catalogue images, exact agreement is rarely possible, because most published images are low-quality reproductions. Detailed comparison requires the exact size, orientation, position relative to chain and laid lines and sheet edges, and points of attachment, to be assessed and recorded. These permit greater certainty as to the similarity or differences between two sheets. They also allow the identification of twins, which are pairs of moulds used in tandem to make a single post of paper. Information regarding points of wear and signs of repair to the mould should also be sought. During the working life of a watermark design, parts are subject to abrasion and wear, with sharp edges becoming rounded; Allan Stevenson analysed the chronology of paper used in the Missale Speciale by looking at such deterioration. Parts may occasionally need to be removed and replaced; for example, it was common to change the last digit in a date when this became necessary.
Lastly, the watermark’s clarity should be noted. The thickness of wire used for producing the image, and its cross-sectional shape, have a bearing on this, as do the condition both of the mould and of the furnish. Good watermark quality requires a short-fibred, well-beaten stock, which is able to conform closely to the watermark design during couching. Insufficient beating—and hence a preponderance of long fibres—significantly affects clarity.
The sheet surface carries a huge amount of information about its manufacture and processing. Initial investigation can be performed by holding the sheet up to a diffuse light source and viewing it by glancing illumination.
The first features to be noted are whether the individual fibres can be seen, and how glossy the surface looks. If no fibres are visible, the sheet has been coated; this can be confirmed by rubbing the surface with a fingernail or coin, so that the affected area becomes polished and more glossy. If no coating is present, the fibres and filler materials should be visible as shiny filamentous or particulate constituents of the surface. For paper made on a fast machine, especially of Fourdrinier design, the fibres will tend to align in the direction of travel on the forming wire; this allows the grain direction to be assessed. Slower machines, such as cylinder moulds, show less obvious orientation, and hand-made paper shows little or none.
For uncoated papers, the degree of gloss shows whether the surfaces have been calendered, or for older papers, subject to friction- or plate-glazing. Generally this is performed in a paper mill; however, John Baskerville developed a secret process that involved hot-pressing a sheet after printing, which also resulted in a glazed appearance. A gloss finish can also be imparted to embossed paper; however, embossing leaves a gloss effect only in the troughs of the pattern, whereas the other methods only polish the peaks, while the deep parts of any surface pattern remain matt. Finally, if only one surface is gloss, and the other is both matt and rough, the sheet has been machine-glazed—formed on a paper-machine and pressed against a large highly polished cylinder to smooth one surface preferentially.
If a coating has been applied, several other factors can also be deduced by using low-angle illumination. The degree of gloss is a measure of calendering; highly glossed papers have been super-calendered, while matt papers tend to be calendered lightly. Secondly, the relative thickness of coating can be deduced by whether any fibre shapes are visible. If none can be seen, the sheet has probably had two or three applications of coating; where they are partially visible a single application is more likely; if they are fully visible, but are indistinct, it is likely a thin coating—applied by a size-press or other applicator—which is termed a ‘lick-coating’, is present. Finally, the method used to apply the coat may be discernible in some circumstances: a series of well-defined evenly spaced lines, running mainly in the machine direction, indicates a wire-wound bar applicator (introduced in the mid-1980s); defects such as thin, straight scratches, again running parallel to the sheet grain, suggest a blade applicator. Other methods can be deduced by microscopic examination.
Careful observation of the surface by glancing illumination from diffuse light will often show the wire side, on which a repeating pattern of geometric marks should be visible; turning the paper whilst viewing it helps to highlight their presence. The side to which the watermark was applied should also be apparent by this method; on hand-made and cylinder-mould-made paper, this will also be the wire side, while on paper from a Fourdrinier machine it will be the felt side.
Visual observation reveals much information about the surface, but detailed measurement, especially of repeating marks, is only possible using image capture and mathematical analysis; the basic methodology for this is described by I’Anson.
First, the type of paper should be noted, because grades were introduced to the market at different times: for example very thin India paper was introduced into Europe c.1750; coated paper was produced and patented in 1827, as was transparent paper.
The weight of a paper can be gauged by hand, as can other aesthetics, often grouped together as ‘handle’. These are judged by gripping the sheet firmly, shaking it, and listening to the timbre of the rattle. Assessing paper by this method requires some experience, but it can give information on the degree of beating, possible furnish composition, and extent of surface sizing, all of which contribute to sheet bonding and rigidity. The process used for cleaning fibres may also have a bearing on handle: Bower attributes the softness of 18th-century French papers to their makers’ use of alkali, rather than soap, to clean rags before maceration.
Colour should be examined. In most early cases this was a compromise between the collection of rags used, their treatment during pulping, and water quality. Indeed, water quality may vary with the season: heavy rains churn up sediment, making it impossible at times to produce white paper. The colour of early handmade sheets was never very uniform; even colouration only became possible after synthetic dyestuffs were introduced, along with full chemical bleaching.
Determining the size of the original sheet is difficult, but it can be estimated by carefully measuring the existing trimmed sheet. First, a book’s gathering that corresponds to a single sheet needs to be identified; the orientation of chain lines in bound, laid paper helps determine this. In a sheet folded in half, the chain lines run parallel to the fold and book spine, producing a folio. A second fold, perpendicular to the first, will produce chain lines running perpendicular to the spine, and thus a quarto. A third fold restores the original orientation, producing an octavo, and so on. Once a single sheet has been identified, any watermark and countermark can be noted, and their relative positions in the sheet assigned. This should allow the sheet’s original dimensions to be estimated from a knowledge of common commercial sheet sizes, taking into account the trimming that is necessary during the book-binding process, and the fact that printing papers were generally slightly larger than their equivalent writing-paper sizes, in addition to being more softly sized.
Finally, the edges of the sheet give information on sheet formation and finishing. Rough borders can be a sign of deckle edges, as from a hand-mould or cylinder-mould machine; equally, however, they could be due to having been torn, in an attempt to simulate the raw edges of a mould. A deckle edge from a hand-mould is generally not straight, and shows an uneven gradation of fibre at the very edge; cylinder-mould machines produce much straighter edges, with a wide fringe of thin fibre at the edge; torn edges are generally straightest and show a preponderance of fibrous filaments along the entire edge. Cut edges can be due to guillotining, die-cutting, or slitting and chopping of the web on a sheeting machine. Guillotined or die-cut edges tend to have a pronounced slur caused by the blade pushing through a mass of sheets; this can be felt by gripping a sheet between thumb and forefinger and pulling it so that the edge is drawn though the digits. Guillotine blades also tend to slice at an angle, which can leave a diagonal scratch on the book edge; die-cutters only move perpendicular to the sheet. Slitting produces a finish known as ‘mill’ edges.
Sheet edges (from left to right): hand-made; machine-made cylinder mould; hand-torn; guillotined or cut edge. Courtesy of Daven Chamberlain
Analysis of the furnish used for a particular paper requires both specialist knowledge and equipment, but is often necessary to verify what has been deduced from visual observation.
Fibre analysis is the most obvious point at which to start. For occidental papers made prior to the 1750s this should yield only linen or hemp, perhaps with some non-vegetable fibre included, such as wool. Cotton only started to make headway thanks to the introduction of mechanized cotton spinning in the last quarter of the 18th century. By 1800, straw was used; wood in various forms was introduced by 1845, but did not become popular until much later; esparto was used from 1857, with bagasse (refuse from sugar-making) added c.1884. Pulping methods for wood can sometimes be assigned by means of stain tests: soda wood came first; mechanical groundwood next, from 1851; sulphite from 1872; kraft from 1884, with various semi-mechanical methods starting in the early 20th century. The exact date for the introduction of any fibre is subject to revision, as early trials certainly antedate the years given, which are when their use was commercialized.
Sizing chemistry is probably of next greatest importance. Early papers used gelatine or starch, which was applied as a post-manufacturing surface treatment; smell will often allow gelatine to be detected, and iodine will indicate the presence of starch. The introduction of machine manufacture coincided with a new method, internal sizing, which used rosin precipitated on the fibre surfaces by alum prior to sheet formation. For machine-made papers, however, coating with starch is also a common post-forming treatment, accomplished at a size-press; in addition, some specialist machine grades produced into the second half of the 20th century were ‘tub-sized’ by running the waterleaf web through a vat of gelatine solution prior to drying and reel-up. Synthetic internal sizes, such as Alkyl Ketene Dimer (AKD) and Alkyl Succinic Anhydride (ASA), were introduced in the 1950s and 1980s, respectively; these are very popular today in most modern mills.
Inorganic pigments were added to paper from the early 19th century, although they were at first used mainly as cheap extenders to the fibre, and as such were viewed with suspicion by many papermakers and customers. Indeed, concern about their detrimental effect on paper strength was so great that in some countries they were banned. In England, for example, before c.1800 no pigments could be added, since they were considered adulterants. However, the benefits of adding some amount of pigment were eventually recognized: pigments increase opacity and improve ink transfer during printing; they also increase ink hold-out, maintaining it at the surface and decreasing its penetration into the sheet during printing, resulting in a better-printed image. The earliest (c.1820) common filler was barium sulphate; gypsum was added soon after, while clay, which had been tested before barium sulphate, was used from 1870. Titanium dioxide, calcium carbonate, and various zinc compounds all came into use during the early 20th century.
The final class of major additives worth assessing are colourants and optical whiteners. Early papers used the intrinsic fibre colour, or natural dyestuffs, such as ultramarine or indigo, or coloured earths like ochre. Indeed, James Whatman II is credited with the innovation of adding blue dyes to ‘whiten’ paper, when in 1765 he started to use indigo; however, continental makers from the 16th century had already developed this practice, adding smalts (an oxide of cobalt), indigo, or even small amounts of blue rag to the furnish for the same reason. Synthetic dyestuffs became available after W. H. Perkin’s pioneering work—aniline dyes were used from around 1870, with synthetic pigments added by 1901. Optical brightening agents, which fluoresce under ultraviolet light and give paper a bluish tinge, were introduced around the 1950s, and are particularly common in modern papers, where they are used to enhance whiteness. Indeed, fluorescence due to use of such reagents was a major factor in Julius Grant’s exposure of the Hitler Diaries as fraudulent; these chemicals were not introduced to the paper industry until after the Führer’s death.
Compared to these four classes, other additives are minor. However, detailed chemical analysis can also reveal traces of bleach residues, formation aids, biocides, fungicides, adhesives, and anti-foam agents, to name but a few chemicals that might be added, intentionally or otherwise, to the furnish. These minor additives, alongside detailed analysis of trace elements found as contaminants in pigments, can be used to provide a paper’s chemical ‘fingerprint’, which may be unique to a locality, mill, or even batch, and which is useful for comparing with other samples of unknown origin.
Determining the provenance of paper is a complex problem that can be approached on a number of different levels, all with varying degrees of certainty.
If there is no sheet for comparison, assessment involves comparing water-mark images with those contained in catalogues or trade listings. Inspection of the surface, looking for coatings or telltale signs of various production processes, should also be undertaken. By these means it may be possible to pinpoint a probable period of manufacture. Chemical and fibre analysis can then be used to identify whether the constituents match those anticipated for the date and origin suggested by visual inspection.
Comparison with a sheet of known origin is far more successful, and is capable of yielding a more exact identification. In this case, visual or instrumental comparison of the two sheet surfaces, markings, structures, and chemistries can be undertaken. In particular, for watermarked paper, an exact comparison of the watermark shape, placement, attachment points, and general quality can be made. However, in many ways this is actually a more difficult analysis to undertake than when no comparison sheet exists, because there is the temptation to look for exact replication of all aspects of the two sheets. Yet, as has been suggested, papermaking is not a wholly controllable process. For sheets made on the same machine or mould at different times, variations in furnish or process conditions can alter appearance significantly. For sheets made from the same batch, differences can also exist for legitimate reasons. Even more importantly, storage and treatment conditions can have a significant bearing on how a sheet looks: two identical sheets, subject to different handling, can appear quite dissimilar.
To conclude, determining provenance is not an exact science. A great deal can be done to analyse paper as bibliographical evidence, but the sheet is only one part of the story and should never be used alone. Evidence from other sources such as printing, typography, and binding should be included, before experience comes into play. Experience, intuition, and judgement must all be used in the final analysis of how the paper helps to determine the provenance of a work of which it is but one part.
J. Balston, The Whatmans and Wove (Velin) Paper (1998)
S. Barcham-Green, ‘An Illusive Image: Some Thoughts about Watermarking Handmade Papers’, TQ 62 (2007), 1–9
P. Bower, Turner’s Papers (1990)
—— Turner’s Later Papers (1999)
—— ‘Watermark Catalogues and Related Texts: A Personal Recommendation’, TQ 56 (2005), 42–4
B. L. Browning, Analysis of Paper, 2e (1977)
N. Harris, Analytical Bibliography, www.ihl.enssib.fr/siteihl.php?page=55&aflng=en, consulted June 2007
D. Hunter, Papermaking: The History and Technique of an Ancient Craft, 2e (1978)
S. I’Anson, ‘Identification of Periodic Marks in Paper and Board by Image Analysis Using Two-Dimensional Fast Fourier Transforms, Part 1: The Basics’, Tappi Journal, 78.3 (1995), 113–19
—— ‘Part 2: Forming and Press Section Marks’, Tappi Journal, 78.7 (1995), 97–106
H. Kelliher, ‘Early Dated Watermarks in English Papers: A Cautionary Note’, in Essays in Paper Analysis, ed. S. Spector (1987)
B. J. McMullin, ‘Machine-Made Paper, Seam Marks, and Bibliographical Analysis’, Library, 7/9 (2008), 62–88
D. W. Mosser et al., eds., Puzzles in Paper (2000)
S. Spector, ed., Essays in Paper Analysis (1987)
A. H. Stevenson, The Problem of the Missale Speciale (1967)
S. Tanner and D. Chamberlain, ‘Chafford Mill: A Short History’, TQ 57 (2006), 37–43
