Hand papermakers are very aware of the physical characteristics of paper. For most people, it is such a common object that they pay almost no attention to it, even though they may use it a great deal. This is especially true with printed paper: there is so much around that most pay it no mind as an object. Those of us who study the history of printing and books spend a major portion of our time looking at physical details in a printed piece that nobody else either notices or cares about. Usually we are searching for evidence that shows how a given book was made each step of the way, especially if it was made hundreds of years ago, since the chances are especially poor that anything other than the book itself--i.e., publisher's or printer's archives--has survived to help explain how the book came to be. Often the only chance we have of discovering what went on in a printing business, or even in the minds of famous authors who chose to proofread their own work, is through the printed products that survive. This makes studying old books a lot like detective work. Our goal as book detectives or descriptive bibliographers, as we call ourselves, is to recreate the scene of the crime as closely as possible. This sort of detailed examination can be applied to newspapers, lottery tickets, blank forms--anything that has been under a printing press of some sort. Close examination of an imprint can tell you when and how it was produced, how it was designed, and what materials went into its manufacture, data that can provide the information needed to reconstruct the nature of the workplace and to identify the people who took part in the manufacturing process. One of the most important and perhaps most obvious pieces to this bibliographical big picture is the paper. The investigation of a printed object requires the appreciation that paper is a lot more than a vehicle for ink. After all, the paper ultimately determines the product--one would not print playing cards on thin onion skin, or newspapers on the best stationary. The size and thickness of a sheet of paper can help determine the kind of press that was used and how it had to have been set up. A paper's origin can indicate the kinds of buying practices a particular printer engaged in. (Maybe the printers you are interested in did not keep their records but their distributors or other firms they did business with might have. These records may help identify the printer under investigation and his practices.) One of the principle clues in researching paper is its watermark. It is always interesting to watch people take a close look at paper, as one of the first things they do is hold it up to the light to look for a watermark. These watermarks are not necessarily just trademarks or a form of advertising. In addition, watermarks in paper produced during the hand press period (up to about 1800) can provide important information regarding its date of manufacture, quality, and place or origin. Bibliographers like Allan Stevenson have looked at watermarks in undated (or questionably dated) printed works in comparison with watermarks in dated books, and drawn conclusions based on their similarities.1 In the 1960s, Stevenson began a study of what was thought to be one of the very first printed books, called the Missale speciale. Through a series of examinations, one of the most important of which involved a study of the watermarks, Stevenson was able to cast serious doubt on the purported age of the book and to suggest a later date of production. Consideration of the watermarks found in the Missale speciale led Stevenson to the discovery that the same watermarks appear in the Compendium theologicae veritatis, which is marked with the date 1473. This identified the Missale speciale as being much younger than originally thought. In addition, Stevenson found the same watermarks in another undated book, the Missale abbreviatum. Stevenson deduced from the watermark evidence that all three were printed in 1473 and, in so doing, affixed a date to two previously undated books. He also suggested a monthly production schedule for the workers who printed them in that year. In 1908, Sir Walter Greg used watermarks to help disprove marked dates on a set of separately-issued Shakespeare plays.2 A series of ten plays in one volume all appeared to have been printed by two separate printers in the years 1600, 1608, and 1619. Greg found it unusual that none of the ten plays was printed on paper which contained a watermark that was not also in at least one (and in some cases up to four) of the other ten. For these plays to have been printed in the years indicated, the watermarks in the paper should have been very different from year to year. Greg pointed out the high unlikelihood of two different printers in two different decades having access to the exact same supply of paper. More likely, suggested Greg, a single printer copied earlier editions and reissued them with false dates and under different names, probably to avoid problems with copyright laws then in force. Greg proposed that all ten of these plays were, in fact, printed in 1619. Finding a watermark, comparing it with others, and then using this information as evidence is an important part of solving a biblio-mystery. In essence, a watermark can turn out to be the missing piece to a whole line of facts or provide the first solid information you have on a book.3 As many hand papermakers know, watermarks are placed in a sheet of paper during its formation in the mould. They are not actually "water" marks at all, but impressions made by a wire design, a "cold brand" which thins the pulp when it is scooped onto the mould during sheet formation. (FIG 1) The marginally thinner density of pulp produced by this contact makes watermarks visible when the sheet is lit from behind.(FIG 2) Up until the 1840s, all watermarks were made by fastening a design of shaped wires directly to the screen of the paper mould.4 The first appearance of these types of watermarks was in European papers in the thirteenth century. Whether the early watermarks were meant to indicate something about the papers in which they appear or were made merely for the amusement of Europe's early papermakers and their customers has long been a matter of conjecture. By the eighteenth century, watermarks and their associated secondary marks or "countermarks" became a pretty dependable indicator of the region where the paper was made, paper size, or both.(FIG 3) Occasionally, a watermark may also indicate the quality of the paper or even the mill that made it. ------------------------------------------------ In order to begin studying a watermark you have to compare it to other watermarks. Like a fingerprint, if a watermark has not been seen before it can be very difficult to identify. Just as a criminologist pours over the prints of known felons looking for a match to an unknown print, descriptive bibliographers search through watermark catalogs to make the same kind of connection. A fair number of these catalogs are available, especially for earlier papers and for countries with substantial paper industries; these important (but expensive) reference sources can be found in larger research libraries. They are usually indexed in several different ways, e.g., alphabetically by subject or region. Not all watermarks contain a characteristic or easily- identifiable image, however. One of the devices used at the Williamsburg paper mill during the eighteenth century, for example, depicted the coat of arms of the colony of Virginia.(FIG 4) Where do you start? Luckily the word 'Virginia' is visible, but one of the very first verbal reports of this watermark described it as "two men in their pajamas holding a basket of fish." This description is not very helpful; especially since no watermark catalog I know of has a "pajamas" listing. Sometimes imagination and an ability to be innovative are required. Another difficulty is unique to paper found in books (as opposed to unbound manuscripts or blank sheets): a complete watermark may not appear in a leaf because it may be obscured in the gutter between two leaves, or part of it may have been trimmed away at the outer margins. Books are made of large sheets of paper folded several times and the watermark might have been separated into several parts. It is possible to assemble parts of the watermark from different leaves in a gathering from a book to make a single, composite watermark, although the results may only give you a general idea of what the watermark looked like in its entirety. Still, this kind of piecing together is better than nothing. To add to the difficulty, in many cases a watermark will not be listed in any of the catalogs to which you have access. An initial failure here does not mean you have hit a dead end, however, because of other possible resources.5 Because a watermark is formed on a sheet of paper while it is still in its most fragile state on the surface of the mould, significant variations in definition and even shape can result between one sheet and the next. A given mark can sometimes turn up in ten sheets of paper looking like ten different watermarks.6 Try it with fingerprints: a smudge here, a twist there, incomplete contact on this one, only half a print on that one. There is a lot of room for originality. Because papermakers have historically been paid by the sheet, you can find some bizarre derivations of the exact same watermark, depending on the crush of work that day. The paper moulds themselves suffer quite a bit of abuse in this kind of environment. Although the wire design might be damaged, the mould may continue to be used for hours, if not days, before it is repaired. Should the watermarks resulting from a repaired design be considered separate from the pre-repaired watermarks? Are they a new generation?7 Another aspect of how paper is made complicates matters even more. During the hand-press period, papermakers usually worked with a pair of moulds at a time, alternating between these twin moulds in forming sheets of paper, using a single deckle for the two. Two moulds meant one person could cast while the other couched, which meant greater efficiency. Each twin mould contained a watermark design on its surface that was very similar to that on the surface of its mate. So, watermarks are twins, too, and even though they are fraternal and not identical twins, telling one from the other can be extremely difficult. The presence of twin watermarks is one of the reasons why detail is so important in accurately identifying a watermark.(FIG 5) The wires forming the watermark design are attached to the screen of the mould by very thin sewing wire, the resulting tacks forming little clumps at intervals throughout the design. These sewing clumps are unique to individual watermarks; their position will differ even between twins and their locations are often noted as evidence. The position of the watermark design on the surface of the mould will also vary between moulds and their twins. Close examination in a sheet of the relation of a watermark to the adjoining chainlines and wirelines position (in a laid mould) will often help distinguish between a watermark and its twin. Finally there is the problem that papermakers tended to copy each other's watermark designs, sometimes with the intention to deceive, but more usually because a certain watermark came to represent a certain size or quality of paper. The result can be a great many different watermarks with only very minor variations. When it comes to identifying watermarks, a little knowledge can indeed be a dangerous thing. Entropy, economy, and counterfeiting all have to be considered. ------------------------------------------- The very nature of watermarks is rather elusive. They can not even be seen unless the paper is viewed the right way: this is very much in character. Describing fishmongers in their pajamas to somebody else or trying to remember what the cross you saw looked like when you are staring at a catalog of ninety watermarks of crosses, some of which are twins, is difficult. The ability to compare a watermark against all the other watermarks you have run across is a prerequisite for making any real headway in its identification. To do this, you need to be able to reproduce a watermark cleanly and correctly. In most instances, the watermark you are curious about will not be something you can take with you to study or reproduce. Therefore, to take full advantage of the watermarks you encounter, you have to capture the images you see using an accurate, portable method. There are several different ways of obtaining a good copy of a watermark. Some are quite simple while several procedures, developed in recent years, are more technically-demanding. One of the oldest methods involves reproducing the watermark with nothing more than a pencil and another sheet of paper, either by tracing it or by taking a rubbing of the watermarked paper against a hard surface. Tracing the watermark, even on a light table, will not be totally accurate. This has historically been a very popular way to capture a watermark, but there are a number of factors limiting the success of the method. It is easy to miss a detail or to simply get a part wrong. Tracings also traditionally do not include the mould wires, only the wire which makes the watermark itself. Without the location of the wires it is almost impossible to separately identify twin watermarks. As when distinguishing a watermark of a cross from ninety others, in order to differentiate between many similar watermarks, you need a more faithful reproduction method. Rubbings of watermarks can be made by using a soft pencil on a blank piece of paper pressed against the sheet containing the watermark, lying on a smooth and hard surface. Any uneven surface pattern will make itself evident when you start rubbing the top piece of paper with the side of the pencil. While people have been doing this for years on book bindings to capture the decorations tooled in leather, the technique has only recently been tried with watermarks. A perfectly sized image is revealed that can have a lot of detail given the proper technique. The representation consists of only the watermark; it will not be obscured by any printing or writing that happens to be on the original sheet. However, there is a risk of damaging the original sheet if the pencil goes clear through the necessarily thin rubbing paper and into the original sheet below it. This possibility makes those who care for old books understandably nervous and reluctant to let bibliographical interlopers perfect their technique on a priceless collection. There is also a certain level of extreme detail missing from a rubbing, a level that may nevertheless be necessary for telling the difference between one watermark and another, almost identical one. In recent years, the most widely-accepted approach to accurately recording watermarks in paper has been to use a technique developed in Russia during the 1950s called "beta-radiography." The process involves a plastic sheet impregnated with a carbon 14 isotope (a "beta source") and x-ray film. The idea in this approach is to sandwich the paper containing the watermark you wish to capture between the beta source and the sheet of x-ray film. The film emulsion on the x-ray film is exposed, typically for several hours, to the carbon 14 isotope as it passes through the paper. The process is not unlike shining a flashlight through the paper: you get a good image of the watermark. But, since electrons are used instead of visible light, the exposure is based solely on the thickness of the paper. Like a rubbing, the resulting image on the x-ray film, called a "beta-radiogram," shows only the watermark, without any printing or handwriting obscuring potentially important details. This exact identification of the watermark with a beta-radiogram is much easier even than looking at the original sheet, where the printing has a tendency to obscure the watermark's finer details. (FIG 6) Unfortunately, getting a beta source is both difficult and expensive. A single ten-centimeter-square sheet is presently priced at over $3500 (although it will last about 5200 years). The x-ray film is also costly, at roughly $5.00 an exposure. In addition, the Nuclear Regulatory Commission demands a license fee for having the carbon-14 isotope. This agency also requires periodic inspections, some training fees may be incurred to insure compliance with federal regulations, and other state and federal agencies demand advance notice every time the beta source is moved. Accordingly, few independent bibliographers are able to use beta-radiography and its use is limited almost exclusively to institutions. Luckily, beta-radiography is not the only modern technique for reproducing watermarks. A more feasible alternative is to use an enlarger to expose photographic paper through the original paper, making a contact print of the watermark (like beta-radiography, but using visible spectrum light radiation, as opposed to carbon 14). It is also possible to simply photograph the original sheet of paper while it is backlit. Both methods produce an image of a watermark in high detail that can by studied and compared. Unlike beta-radiography, however, both photographic techniques will also capture any printing or writing on the paper which may hide some of the watermark's important details.8 Using the photographic paper method also requires the use of a dark room facility, making the technique essentially non-portable. In addition, there is grave concern on the part of most curators about allowing old paper into a chemical-rich environment like a darkroom where so much can go wrong. Straight photography, while relatively simple in application, means a delay while the film is processed. If you have only limited or even one-time-only access to the watermark, this process can be problematic. In essence, these two photographic methods leave a lot to be desired. In 1970, a new technique for accurately recording watermarks was developed by Thomas Gravell of Wilmington, Delaware, using ultra-violet light and a photosensitive paper made by Du Pont de Nemours and Company, called DYLUX 503. This application is technically considered a photographic process and is based on the same principle as a contact print: DYLUX is exposed thorough the paper to render an image that is exactly the same size as the original. The DYLUX process is portable and completely safe to the paper, and uses equipment which is readily available and so inexpensive that even casual scholars can try it. Another benefit to the method is the speed with which a good exposure can be made, allowing you to make sure on site that you have captured the watermark with a high-quality image; usually thirty minutes is all that is required. The end result will be a white on blue image of the watermark and mould wires, exactly the same size as the original and detailed enough to use in comparisons to other watermarks. All in all, the DYLUX method seems to fill the bill for those who deal with large numbers of watermarks in many different repositories. It is preferable over beta-radiography from the standpoint of affordability and logistics and, unlike a contact print using standard photographic paper, DYLUX is portable and requires no chemicals for developing. Straight photography of a backlit sheet of paper, in addition to the delay caused by processing, can easily compromise the original size of the watermark in reproduction, while DYLUX preserves the one-to- one ratio. Even though DYLUX does not have these drawbacks, however, it is not the panacea of watermark recordation. As with any photographic process, DYLUX records everything that is in and on the paper. With a watermark that is particularly obscured by surface text, DYLUX may not produce an image that is any better than photography. In these cases, a combination of the DYLUX method and taking a rubbing might be in order. The occasional use of both of these modest techniques to produce usable evidence is the travelling bibliographer's way of taking the best from both worlds. Except in extreme situations, DYLUX can deliver the detail and preserve the scale of watermarks in a way that makes it ideal for field work. ----------------------------------------------------- Through the collection and comparison of watermarks from many places, bibliographers begin to recognize patterns that lead them on to more important data. Once the identity of the paper is confirmed, all kinds of doors to further research are opened. Sleuthing out a watermark for what it is or what it implies is almost like learning to read. It is a skill that can make books a lot more meaningful.9   Endnotes 1. See Stevenson's The Problem of the Missale speciale (Bibliographical Society, London: 1967) for a very interesting discussion of watermarks and their implications. 2. See "On Certain False Dates in Shakespearian Quartos," The Library, (published by the Bibliographical Society, London) 2d ser., IX (1908) pp 113-131, 381-409. 3. It is generally accepted that only about half of printing paper made during the hand-press period was watermarked at all. It is theoretically possible to record the exact pattern in the paper left by the wires and chains of laid moulds, but this is subtle stuff, like working with the vagaries of actual fingerprints. Several plans to record this sort of evidence have been devised, but the logistics are formidable. 4. In about 1845 another method of watermarking papers was invented by William Henry Smith, resulting in light and shade watermarks. These are commonly used in currency and other security papers, and are less frequently found in books. After 1826, some machine-made paper had a watermark applied to its surface after it was formed, by the use of a dandy roll. 5. My suggestion is to join a bibliographical society. Alternatively, you may be able to find someone locally who is already a member, and they may be able to steer you to an individual or institution adept at linking watermarks with known information of a specific mill. One of the greatest things about the bibliographical community is the willingness of a colleague to put down his/her coffee and answer the, "Have you ever seen one of these?" query. Finding these people is not hard; a quick trip to a library and a glance at any of the published papers of one of the bibliographical societies will usually yield a handful of names. 6. Stevenson cataloged watermarks in his own records as he discovered them. He once listed the same watermark on five different occasions five different ways before realizing all the entries described a single large and complicated mark: a lean lion rampant on a twisted escutcheon. 7. Stevenson considers a repaired watermark as being a separate "state" from the original. Gathering enough images of the different states of a watermark can provide an opportunity to build a chronology. From this record one can infer a date of production for papers that are not dated, relevant to the state of the mark. Collecting the states of a series of watermarks was an important part of Stevenson's breakthrough with the Missale speciale. 8. Printed or written text is sometimes desirable when recording a watermark as it proves that a particular watermark did in fact come from a particular paper. Beta-radiograms are sometimes subject to questions of authenticity because they do not record any of the landmarks which identify the sheet of paper from which they are made. 9. The author would like to thank Thomas Gravell, Travis Charbeneau, John Bidwell, and especially Terry Belanger for their help with this article. 1� � &  , / / 0 0 0 NORMAL.STY 1 @ � / / 0 �How to Make DYLUX Exposures DYLUX 503-1 paper. This can be purchased from printing supply houses either in rolls or pre-cut sheets. (You can usually locate these in the phone book; if not, call a local printer and ask where they get their materials.) DYLUX was designed for use in the printing industry as a proofing medium for transparencies that would later be used to make offset plates. It is a photographic paper coated with a chemical that changes color when exposed to different wave lengths of light. Make sure the paper has only been treated on one side and that it is the "blue" kind, which yields blue images. The DYLUX paper will be your biggest investment, as it is only sold in large amounts. I recently paid about $190 for 200 sheets, 24" by 20" in size, or about twelve cents for a 5" by 8" exposure. Light sources. Two kinds of light are necessary to make a DYLUX print: an ultraviolet light source, and a standard, visible-light source. For the ultra-violet source, you need long wave black light, the safe kind of UV used for interior lighting effects that will not cause eye or skin damage. (The UV-B or UV-A short wave black light used in tanning salons is dangerous and will not work in this application.) Your local novelty shop may have these, or have a hardware store order one for you. You will also need the fixture that will power the light. For the visible light source, almost anything will work, but you should stick with fluorescent lights, if you can, because of their low operating temperature. Their use will keep you from running the risk of damaging the original sheet of paper whose watermark you are trying to capture. A fluorescent desk lamp (with a bank of two tubes) is probably the best, most easily obtained visible-light source for making DYLUX prints. The more adjustable the light source is the better. "Cool White" tubes work well, but get them in the highest wattage the accompanying fixture will handle. A small piece of plate glass. It should be plain window glass and about 5" by 8". You can get it at framing stores or specialty glass businesses. It should have beveled edges to reduce the chance of damaging you or the paper. Once you have gathered these materials, you are ready to make exposures. Using DYLUX by the book requires that you expose the paper with the UV light and then develop the image with visible light. This is where you should deviate significantly and critically from the Du Pont instructions. Paper will not reflect UV light and, in fact, absorbs it. Because of this and the destructive nature of UV light on paper, you must never expose an important paper to the black light. The solution is to reverse the prescribed method: expose with a visible light source, and develop with a UV light source. I always keep my unexposed DYLUX in a black plastic bag until I am ready to use it, but moving it around in a lit room until you get everything positioned is not a problem. Normal levels of incidental or ambient lighting will not affect the DYLUX, except for direct sunlight, which should be avoided. The process of making the DYLUX print is simple. Begin by cutting a sheet of DYLUX about one inch bigger on all sides than the watermark you want to capture. Lay the cut sheet on a smooth surface with the coated (yellow) side up. Put the original paper containing the watermark on top, and then put the plate glass down on top of them both to be sure everything lays flat (see illustration). When you have everything positioned, turn on the visible light source and leave it in place (about three inches from the glass) for about ten minutes. Once the exposure has taken place, the next step is to develop the image. Remove the glass plate and retrieve the piece of exposed DYLUX. Move it away from the original paper and wave the UV light source back and forth over the DYLUX sheet until the blue dye begins to react with the light. As with photography, be prepared for disappointment with your first exposure. How long the visible light source takes to properly activate the dye in the DYLUX will depend on the thickness of the original paper and the wattage of the tubes. Start with about ten minutes and, if it comes out too blue, add five more minutes to the exposure. The longer the exposure time, the whiter the DYLUX will become. The two 15-watt tubes in the desk lamp I use, held about 5" from the paper, will produce a nice image in about twenty minutes when the watermark to be reproduced is in 22-pound paper. Once you get a little practice with your own lights (I suggest you pick a set and stay with them), you will be able to tell how long the initial exposure will take just by feeling the original paper. Images made with DYLUX are stable (if stored in darkness) for ten years or longer. Thomas Gravell, who discovered this technique, generally has his exposures converted to slides for storage and comparison. DYLUX paper can also be photographed with a Wratten no. 25 (red) filter and printed on high contrast paper, as the DYLUX illustrations in the accompanying article were. In either case, an indication of scale should be included in the frame of the exposure, to document the original size of the watermark.

‍