A History of Handmade Paper. The Basic Problem

The problem of identity is ever with us. I am afraid that many people have used the phrase ‘the same watermark’ without any clear idea of what it should mean. A common example which turns up in book descriptions is: “This book has just one watermark throughout”. If this sentence merely means that the marks are similar, that they belong to the same type, the description should say that. If it means that all the paper in the book was made on the same mould, then the writer of the sentence probably does not know what he is saying.

Allan H. Stevenson, Observations on Paper as Evidence (1961)

“I think the problem, to be quite honest with you, is that you’ve never actually known what the question is”.

Computer “Deep Thought” in Douglas Adams, The Hitch Hiker’s Guide to the Galaxy (1979)

The basic problem is all too simple. (I hope we are not going to have too many of these titles banged in just for effect ...).

The basic problem is that paper is a vast, vast phenomenon.

Some scholars have speculated about how many sheets of paper might have been made in the history of civilisation. But any number becomes too impossibly big to comprehend. And who cares anyway? When you start to joust with one or more pairs of watermarks in your own personal bibliographical combat, the scale of the universe becomes a very secondary problem.

From its first introduction – exactly when, where, and by whom we do not know, which is a nice simple historical fact, but some two thousand years and more ago in China is as good a time and a place as any – up to the present day, paper has been the principal vehicle for any and all texts mankind has wanted to communicate and keep. It has had its rivals of course, but has usually seen them off ignominiously. Clay tablets, which are sometimes round egg-shaped objects, beat everything else for resistance, but don’t hold much text and are cumbersome when you move house. Its ancient Egyptian counterpart, papyrus, is flimsy and has to be kept in rolls. Its most serious contender has been parchment (or vellum, if, and only if, we are talking about calfhide), an extremely durable material, but difficult to print on, whose utility and effectiveness is further limited by its scarcity, its cost, and by the fact that it can be difficult to store (if allowed to warp by being placed upright, a medieval manuscript will rip a fancy bespoke modern binding apart in a matter of generations). Celluloid, as in photographic negatives, film reels and microforms, seems to have had its day, while electromagnetic supports ... well, apart from the fact that the technology seems to hoppity-hop along rather too often for comfort (so that instead of being able to read one’s collection of floppies and CDs on the latest generation of computers, you can moan to your offspring – or rather have them explain – about the concept of built-in obsolescence, which includes you), their ability to subsist for more than a handful of years is very much open to doubt. Of course you can pop it into the cloud, but who is going to find it there? The loss of a password can make everything disappear.

So, for ubiquity, convenience, resistance, endurance, strength, cost, texture, pleasingness, delicacy, perfume, and softness, paper has no rival.

But the brutal scale of the phenomenon is also a challenge (one that most scholars have preferred to decline). Except perhaps for buildings (but then again how many of those have not been thoroughly mucked around with in the interim and have reached us in an unadulterated state?), hand-made paper is the artefact from the Middle Ages, Renaissance, and Early Modern period that has come down to us in the greatest quantity (and is much more interesting than potsherds). Apart from the ancillary circumstance that most of it has been written on or printed on (or both), this paper is often in pristine condition and only in a few cases has been messed up by restorers. Of course what has survived is only a minuscule proportion of what was made at the time: differently from other supports, paper is easily recycled through pulping and remaking in a slightly inferior quality, and it has other secondary uses (an indescribable amount over the centuries has satisfied mankind’s seriously serial physiological needs and thus has vanished into cesspits and sewers: soft-toilet paper was being made in China as early as the Fourteenth century, but in the West paper specifically for use in the water-closet was first manufactured and marketed by Joseph Gayetty in 1857; toilet-rolls followed in the 1890s. Did we need to know this? no, but it is interesting).

What has scholarship done therefore to place paper studies on a proper footing? The correct answer is (virtually) nothing. It remains a scholarly S.E.P. (Someone Else’s Problem, but don’t you hate unnecessary acronyms?). Since it involves just about everybody, it is up to everybody (else) to do something about it. And this sort of attitude gets nobody nowhere in no time whatsoever.

If the most obvious place to start is at the beginning, at least as far as Western Europe is concerned, outside the venerable pages of Briquet, it is more than a trifle absurd that no systematic census has been conducted (or is being conducted) of early watermarked paper in order to denominate and describe material produced up to 1300, when Dante went on his little walking tour through the realms of Hell, Purgatory and Heaven. Why is this the case? One reason is that Briquet did the job so superlatively that it is difficult to improve on what he achieved, however incomplete that now seems. Another is that there is too much of the stuff. Yet another is that it is kept for the most part in inconvenient places, such as the archives of numerous small cities, mostly in Italy, also in France, where even specialist curators are rarely interested in paper. And yet another again is that palaeographers are somewhat snobbish and consider paper less interesting and important than parchment, which is absurd, since paper could tell them a great deal, if they were willing to learn (which for the most part they are not).

If we take the next benchmark down the chronological line, or the year 1400 (just to get our literary bearings, the year of Chaucer’s death), the quantity of material surviving, still mainly in Italy, but also elsewhere in Europe, often on Italian paper exported at the time, becomes truly daunting, especially if we add a significant number of dated or datable manuscripts in libraries all over the world, which are not always easy to examine.

As for the Fifteenth century, which ends on Thursday, 31st December 1500 (thus allowing everybody to take a long weekend), well, don’t even think about it, since printed books now come into the equation and they are nothing but trouble.

One essential fact about paper has to be remembered, though it is all too easily and too often forgotten. Just about everything we say about the whole wonderful subject, all the chronological and topical coordinates (that means when and where), rarely derive from the substance itself, but from what has been written or, more latterly and laterally, printed on the same. The consequential hypothesis, all too rarely voiced, is that in terms of the time and the place, the paper was made reasonably close to where it was used. As a general assumption, it is absolutely splendid; as a specific assumption, it bristles with danger.

The picture, any picture, every picture, of the way paper is brought in and used to create a record or a text, whether in a municipal archive in the late Middle Ages, or in a printing shop with the Renaissance in full swing, or in the manuscripts of an Eighteenth-century writer, has to be built up with reference to as full a context as possible. It also has to decide at what level the analysis is going to be conducted. In certain lines of research it is necessary to recognise individual watermarks, or rather the twin individuals characterising the two moulds employed in alternation at the vat and sometimes even different states of these watermarks; in others, especially when we are dealing with printed artefacts extant in multiple, widely dispersed copies, it is enough to identify a particular purchase of paper, where the watermarks set one supply apart from another, without going to all the hassle of establishing twinship; in yet others, especially when dealing with cancels and forms of substitution in printed books, the emphasis falls on other features, such as the distinction between the mould/felt sides of the sheet or the pattern made up by the distribution of the watermarks.

So, as ever, Forty-two may well be the correct answer, but that is no help if you don’t know what the question is.

The Essential Early Chronology, or One Day, Somewhere, Long Ago, in China

If you flick through any sort of standard history of the book over the millennia (probably not a good idea, unless it is a wet Sunday afternoon and you have plenty of time on your hands), looking for information about paper, the same basic dates always pop up. The cute thing is that they are mostly wrong, or at the very best incredibly misleading. So what follows is akin to reconstructing the universe from a piece of fairy-cake.

Traditional Chinese sources attribute the invention of paper to Cai Lun (or Ts’ai Lun, or some other spelling), a eunuch of the imperial court (63-121), sometime around 105 A.D. To be more exact, such is the date at which the discovery was formally reported to the court and officially adopted; experimentation had been going on for some time previously, at least a couple of centuries, since archaeologists have made solidly-grounded claims for much earlier dates on the basis of scraps of paper discovered in tombs. The main source for Cai Lun’s life and achievements is a chronicle compiled in the Fifth century by Fan Yeh as the official record of the Han dynasty, at least three hundred years after the event. It is a reasonable guess therefore that a trifle, and more, of historical rewriting took place, so that what was most likely an anonymous invention, developed over time by a number of different people, could be attributed to a figure at the imperial court, allowing them to grab the glory (and this, of course, is just what happened). Even more intriguingly, by the by, Communist, or post-Communist China, through its Ministry for Light Industry, has sought to discredit the archeological discoveries, on the grounds that this would imperil the genius status of Cai Lun (with effects akin to throwing away a boomerang). Such anti-historicism might appear absurd, but of course in Europe there have been plenty of analogous squabbles for issues of municipal or national pride, such as Coster vs. Gutenberg.

Knowledge of the discovery slowly moved westwards, along what would later be known as the Silk Road (the term itself, in German Seidenstrasse, was first coined in 1877 by the geographer Ferdinand von Richthofen, who – what’s in a name ? – was the uncle of the “Red Baron”, Manfred von Richthofen, the First World War flying ace). Here account has to be taken of one of the most extraordinary discoveries in the history of archaeology. The Mogao caves, or grottoes of the Thousand Buddhas, are a complex of 492 temples, mostly dug into the rock, near the Chinese city of Dunhuang. In one of these, now known as the Library cave (n. 17), in June 1900 a Taoist monk called Wang Yuanlu (c. 1849-1931), banging his pipe against the wall of the neighbouring shrine, heard an echo and uncovered a wooden door, behind which was hidden an enormous cache of documents: the current estimate is 1,100 bundles of scrolls and some 15,000 paper books [2]. The latest date recorded in the documents of the collection is 1002 A.D. The most widespread, and even sensible, interpretation is that the library was sealed up to protect it from an external threat and forgotten about for nearly nine centuries. After its discovery, some manuscripts were gifted to local dignitaries by the monks and the news soon reached the ears of the “foreign devils”, as Western travelers in China were flatteringly known. First on the scene was the British archaeologist of Hungarian origin, Aurel Stein (1862-1943), who in 1907, with a mixture of threats, cajolery and bribery, was allowed to take some 9,000 documents, which, not being able to read Chinese, he chose mainly on the basis of their physical condition, and an assortment of art works. In 1909 these were bestowed on the British Museum in London and Stein was rewarded with a knighthood. In 1908 he was followed by the French professor of Sinology at Hanoi, Paul Pelliot (1878-1945), who obtained some 1,500 items, chosen on the basis of their textual quality, which in 1910 made their way to the Bibliothèque Nationale in Paris. A Japanese mission in 1912 obtained a further 400 scrolls for the National Diet Library in Tokyo, a Russian one followed in 1914, while the remnants of the collection were brought to the National Beijing Library in the 1920s. Any count of the documents has to take account of numerous fragments, since the walled up cave seems also to have served as a deposit for sacred waste. The looting or safeguarding of the library (as with the Elgin marbles) is a controversial issue: on the one hand the discovery would probably have been completely dispersed by the monks of the time, or destroyed in the troubled history of the area (in the 1920s the caves were occupied and vandalised by White Russian soldiers); on the other the loss of the collection to foreign libraries all over the world is a grievous one for the history of Chinese culture.

Among the materials procured by Stein is the Buddhist Diamond Sutra, which is one of the oldest, securely dated documents on Oriental paper, which scores double points by also being the oldest dated extant example of printing, made in 868 A.D. (well, to be more exact, the woodblocks with which it is printed have the said date in the colophon, which actually reads “reverently [caused to be] made for universal free distribution by Wang Jie on behalf of his two parents on the 13th of the Fourth moon of the Nineth year of Xiantong [11 May 868]; so, as with all printing done from blocks or from stereotyping, the impression of the document itself might have happened at a later date), and triple points by being the oldest dated printing to contain an illustration. It is formed from seven pieces of paper, printed on one side, and stuck together to make a scroll over five metres long, which can now be unrolled and electronically perused on the site of the Dunhuang project hosted by the British Library (well worth doing) [2].

What the discovery of the Mogao cave library shows is the existence of a vast culture of manuscript and printed documents, which otherwise has been lost without trace, while the other lesson to be learned from these early survivals is that Chinese civilisation did not consider paper primarily and exclusively as a writing material, but recognised its use as multiform, as wrapping paper, toilet paper, and the infinite other uses it has in our own time (One Chinese chronicle, written in the Seventh century A.D., mentions the existence of tea bags, or at least bags to store tea, which might explain the stale taste of some British motorway café cuppas).

The next date that invariably pops up in histories of paper, even quite respectable ones, is 751 A.D., as a consequence of the battle of Talas between the Chinese Tang dynasty and the Arab Abbasid Caliphate. The actual site of the battle is not known, but it was somewhere on the Syr Darya (or Talas) river, which at the time flowed into the Aral sea. Victory allowed the Arabs to return to their base at Samarkand with captured Chinese papermakers, who were induced to reveal the secrets of their trade (best not to think about how the inducing actually happened). The problem is that the whole account is provided by an Arab historian, ’Abd al-Malik al Tha’alibi, writing some three centuries after the event in a work entitled the “Book of Curious and Entertaining Information”, or a medieval chronology. Where did he get his facts from? Of course he does not tell us and it is quite possible that the anecdote was made up to explain the flourishing industry in Samarkand in the author’s own day. There is in fact evidence that paper, and perhaps even papermaking, were known in the area previous to the battle of Talas. For instance, Arab merchants travelling in China communicated with their base in Samarkand through letters written in Sogdian, or the lingua franca of the Silk road, of which some were on Chinese-made paper: a packet of such letters, dated 313 A.D., was discovered in 1907 in a ruined watch-tower near Dunhuang. Similarly, driven out by Arab invaders, the last king of Panjakent retreated to the fortress of Abargar on Mount Mugh, about 130 km East of Samarkand, where he died in 722 A.D.: a cache of 76 documents from the royal archive, of which 22 on imported Chinese paper, were discovered by a Russian expedition in 1933 and published thirty years later.

A further fascinating, albeit slightly later, reference to papermaking technology in the region is a fleeting mention in the Kitab al-Jamahir fi al-jawahir, or “Book Most Comprehensive of Knowledge about Precious Stones”, by the great Medieval Islamic scholar, al-Bīrūnī (973-1048), who, in a discussion of hydraulic mechanics, talks about stones “fixed to axles across running water, as in Samarkand with the pounding of flax for paper”. The fact that he considers the industrial process to be so well known as not to merit further explanation suggests that the sight was a common one.

What is certain is that a knowledge of paper and papermaking methods was spreading through the Arab world from the Eighth century onwards. Although the importance of Islam as a filter between China and the West should never be underestimated, the only real innovation of the Arab world was the substitution of rags for the mulberry bark and other bast fibres employed in the Far East. The first great centre was Baghdad, where – according to encyclopaedist Yaqut, writing a mere five centuries later, and thus not entirely reliable – during the reign of the munificent and unforgettable caliph, Haroun al-Rashid, a papermaking factory was established in 794-795 (again this date has become canonical, but the documentary basis is slender). Afterwards factories and shops, taking advantage of the plentiful water supply afforded by the Tigris, seem to have proliferated, while the availability of this new, relatively cheaper, writing material led to a sort of cultural, or at least literary, explosion (to the extent that things got written down). The oldest dated manuscript written on Arab paper currently extant was produced in 848 A.D. and was discovered comparatively recently in the Regional Library at Alexandria (Egypt). It is followed by a codex in Leiden University Library, dated 866 A.D. (Cod. Or. 298), while a Greek manuscript, now in the Vatican Library, ms. Vat. Gr. 2200, was probably copied in Damascus sometime around the year 800 A.D. and is plausibly the oldest known document in a Western script on Arab paper.

By the Tenth century papermaking in Egypt had ousted the traditional papyrus industry, which wholly disappeared and had to be reintroduced in the Nineteenth century. Obtaining rags became a major business, with some nasty stories about linen wrappings being recycled from mummies (as long as Hollywood doesn’t get hold of this, I don’t mind what happens). Almost all this material has disappeared or has survived by pure chance. A few exceptional discoveries, however, have been made. In 1881-82 excavations in central Egypt, around the towns anciently known as Arsinoe or Crocodilopolis (nowadays Faiyum, or El-Faiyūm) and Heracleopolis Magna (an abandoned site some 15 km West of the modern city of Beni Suef), uncovered thousands of bits of fabric and 10,000 written documents on various supports, which were obtained by the antique dealer and carpet trader, Theodore Graf (1840-1903). He duly sold the written materials a couple of years later to the Archduke Rainer of Austria, who in 1899 donated the collection to the Imperial Library in Vienna. Although the prime interest was for the older documents on the more traditional support, papyrus, a significant number of items were on paper and stimulated the ground-breaking studies of the library’s director, Joseph von Karabacek [3]. These were added to over time, so that today's Papyrussammlung in Vienna holds some 16,000 examples of Medieval Arab paper. Another equally extraordinary discovery happened around the same time in Cairo. In Jewish synagogues any document containing the name of God, or more simply written in Hebrew, the language of God, could not be destroyed or thrown away, so they were accumulated in a storeroom known as the Geniza. Generally they were disposed of in a ritual burial, but sometimes this did not take place, meaning that the pile simply grew through the centuries. The best known instance is the cache of approximately 280,000 items discovered in the Nineteenth century at the Ben Ezra Synagogue in Old Cairo, covering a period mostly between 1002 and 1266. Again discovery meant dispersal, so that the largest single nucleus is nowadays in Cambridge University Library, comprising some 193,000 fragments [3]. The Geniza documents are on a variety of supports, including parchment and paper, but perhaps due to their fragmentary status and the immense difficulties in dating have received little attention as physical documents.

Chronicles and other documents show that by about the year 1000 papermaking as an industry had spread along North Africa and reached Arab Spain. Famously the Arab historian El-Edrisi in about 1150 praised the paper made in Xativa, mid-way between Valencia and Alicante (rather annoyingly a lot of histories of the book and the ilk talk about papermaking being first introduced in Spain at the said date, but, from what El-Edrisi says, production is well established).

Now here comes the rub. Dated manuscripts written on Arab paper are relatively few and on account of their antiquity very jealously guarded by libraries. As well as the wear and tear of time, one reason for the extreme rarity of Spanish Arab manuscripts is the reconquista in 1492, during which ensued a cultural genocide in which books and other testimonies of the scale and depth of Islamic civilisation were systematically torched. Paradoxically one of the largest collections to have survived is in the Vatican Library.

A Digression about the Forme: Floating or Dipping

Here it is necessary to digress somewhat, in order to explain that in primitive, and by “primitive” I mean “very primitive”, papermaking there were two fundamental techniques and technologies.

The oldest was the floating mould, in which the container was placed on the surface of the water and the pulp was ladled or poured into it and smoothed out. A floating mould obviously cannot employ a metal mesh, and has to be made of a light wood, such as bamboo, since otherwise it would sink. Floating moulds were not supported underneath by rods, since making the sheet did not involve the sharp lift that is typical of the dip mould. On the other hand the openings could not be too large, because otherwise fibres would be lost into the water. So the surface of the mould was formed by a rough cloth, such as calico, drawn as tightly as possible. When the sheet was made, the cloth bellied slightly, but not enough to create an uneven sheet, after which it was lifted out and the mould was put to dry in the sun, which on a hot day can be a matter of minutes. One feature, therefore, that identifies such a mould is being able to see the imprint left by the fabric on the mould side of the sheet of paper. Floating moulds were, and still are, used in Oriental papermaking, but they have the disadvantage that, if the papermaker is going to maintain a steady daily output, a large number of individual moulds is called for, though of course this does not represent a significant expense. On the other hand the floating mould has several advantages: it is low-cost in terms of materials and technology, since it can be used in the open air, on the edge of a stream or a pool (though more permanent structures obviously progressed to a purpose-built vat), and it requires small amounts of pulp, which is a consideration when the fibres have to be hand-beaten. It is also possible to employ this technique to make very large sheets of paper, by constructing an appropriate mould and eventually using several people to lift it out of the water. In his trips to Korea and China in 1933, to Siam in 1935, and to India in 1937-38, Dard Hunter found floating moulds still being employed and documented them photographically, as well as bringing back examples for his paper museum, now in Atlanta [33]. I have a sneaking suspicion, on the other hand, that their existence was prolonged, or renewed, by hippy culture in India in the 1970s, since the unsophisticated technology and the lack of skill required allowed footloose westerners to improvise laboratories.

The other sort of mould is the dip version, in which the sieve is plunged into a vat containing the fibres diluted in water and lifted out again. In terms of the general method, a vat specifically for this purpose had to be constructed with masonry and filled with a large quantity of water and pulp, some of which necessarily went to waste. On a dip mould, an unsupported cloth surface would belly and create a very uneven sheet of paper; if on the other hand it were held up by rods and the equivalent of wires, the pulp would have difficulty in draining. The Chinese therefore constructed the mould in a completely different fashion, from an early date, certainly by the second century A.D., by laying thin strips of rounded bamboo side-by-side and tying them together with threads of flax, silk or animal hair, leaving the equivalent of the mark of a chainline on the surface. As regards the subsequent problem of removing the sheet from the mould, there were two potential solutions, depending on whether the sheet was dried on the surface or immediately couched. All moulds, whatever the technique, have to have a deckle (the word comes from Dutch or German, and means “cover”), or a wooden frame surrounding the sieve, fixed or mobile, that holds the fibres in when the frame is lifted out of the water. In the first instance, as with the floating mould, the structure was fairly lightweight and had a fixed deckle. The obvious disadvantage was that once again it required the papermaker to have a large number of moulds, which were used in sequence, after which the sheets were exposed to the sun to dry and peeled off, possibly while still damp. Nevertheless, such moulds, especially to make large sheets of paper, were and are used extensively in traditional Chinese papermaking. In the second instance, the mould was constructed with two separate side deckle sticks, which were slipped off after making the sheet, and the sieve with its still fresh layer of pulp facing down was couched onto a flat surface. In some instances this was a board with a piece of cloth on it: the mould was rolled up to separate the sheet, which was immediately taken and brushed onto a heated wall in order to smooth out any irregularities and to dry it; otherwise, it was couched onto a post of previously made sheets, and afterwards dried on a heated wall or in the sun. As the earliest account of papermaking in China, published in 1637 [5], explains, a resin, often a vegetable gum from the hibiscus plant, was added to the fibres in the vat, ensuring that there was no need to interleave the sheets in the post with some sort of cloth. The same resin also sufficed as sizing, which Oriental paper required to be much less stiff and rigid than its Western counterpart, since in this part of the world calligraphy is written with a small brush on only one side of the sheet.

What did the Arabs learn from the Chinese? Unfortunately, whereas Chinese papermaking continued unchanged up to the Twentieth century and still continues in some areas, so that ancient pieces of paper can be compared to observed procedures, in the latter Middle Ages the techniques and tools of “Arab” manufacture were displaced by Italian and French products, including in the Muslim world, and thus totally disappeared. What the Arabs certainly learned from the Chinese papermakers, supposedly captured at the battle of Talas, was the floating mould, and it is the knowledge of this procedure that spread westwards and reached the Mediterranean. The famous Umdat al-Kuttab, which is the oldest extant description of the papermaking process in Arabic, written by Mu’izz ibn Badis sometime in the Eleventh century, unquestionably describes a floating mould. Did the Arabs reach the next stage, i.e. the dip-mould? Almost certainly, yes, although the evidence is necessarily deduced from the paper itself. Rather than showing a fabric imprint, where visible, sheets of Arab paper usually show sequences of thin, closely-set lines: twenty of them take up a space varying between 20-30 mm; in comparable Western paper, in which the lines are certainly metal wire, twenty occupy anything between 34 to 52 mm. The most likely explanation is that in Arab paper these lines were formed by flaxen or hempen thread boiled in oil or pitch to give it rigidity, explaining both the fineness of the wires and their density. Also evident in Arab paper, at right angles to the lines forming the surface of the mould, are series of knots, or chain-lines, possibly formed from flax or horse hair, serving their well-known purpose of preventing the lines of the sieve from shifting and opening under the weight of the water. Whereas in Western paper, especially in Italian Renaissance paper, the chain-lines are regularly spaced, in Arab paper the distances are variable, often with groupings of two, three or four chain-lines followed by a wider gap, and are at closer intervals, generally 15-30 mm. The construction of the surface of the mould in this fashion of course favours the hypothesis that the paper was made on a dip-mould, while the fact that sometimes the impression of a wall or a board is visible suggests that the sheets were couched while freshly made, probably with the same rolling mat procedure employed in Chinese papermaking. On the other hand, the fineness of the lines and the tight grouping of the chainlines make it improbable that the Arabs were the first to introduce a metal mesh, as has sometimes been claimed.

A final characteristic of Medieval Arab paper, which should be mentioned since it can puzzle those who meet it for the first time (as happened to me on my first encounter with the phenomenon), is that sheets were often glued together, in both cases with the uneven mould side inwards, in order to form a single, stiffer sheet. The intent was most likely to give a more rigid, “parchment” feel to the paper, but it was probably also a consequence of the fact that the sheet was not pressed while it was still wet and so it was impossible to flatten out all the mould-side irregularities.

Paper Reaches the West

At some point paper takes a huge technological stride forward.

When and where? Almost certainly in Italy, sometime around the middle of the Thirteenth century.

In the Arab world paper had become the principal communication medium, supplanting papyrus and other supports, though parchment kept its role for more prestigious and expensive documents, such as copies of the Qur’ān. Once the industry was well established and was producing significant amounts of material, more perhaps than local markets could absorb, inevitably it was exported.

One proof of this early circulation of paper is to be found in the word ream. The word comes into English from French reyme, which in turn derives from the Spanish resma or the Italian risma (in Medieval Italian sometimes lisima); behind these shared terms lurks the Arab word rizmah, meaning bundle or bale, often of cotton cloth, which obviously shows how Western users first became acquainted with paper, arriving in packages from ships trading along the Mediterranean coast.

The oldest known surviving piece of paper to have been written in Europe is a document in the State Archive at Palermo in Sicily, with texts in Arabic and Greek, which is dated 1109 [4]. Likewise, the State Archive at Genoa has a paper register in which the first entries were made in 1154, although the paper therein, which reuses the remnants of an Arabic scroll, was certainly imported from outside Italy. In 1231 Holy Roman Emperor Frederick II, reiterating earlier decrees of 1145 and 1220, issued legislation, ordering that all official documents written on paper be recopied on parchment. (Scholars have traditionally attacked this seeming boycott as Medieval obscurantism; in reality, given that Arab paper has a vegetable-based sizing, which attracts microbes and insects, and is more fragile than its Western successor, Frederick, or whoever was in charge of information technology at his court, was actually being quite sensible, while the decree does not imply that the emperor was averse to other, less prestigious, uses of paper.) For a legislator actually to prohibit something, there has to be enough of it around to create a nuisance (like binge-drinking, or is that not forbidden?); so we have to deduce that, although very little of it might have survived, by the beginning of the Thirteenth century conspicuous amounts of paper were in circulation.

Where was this paper made? Somewhere in the Arab world and imported? or was some of it being made in Italy itself? The latter is a beguiling hypothesis. And there is a historical pointer in the distribution of the traditional papermaking centres in Italy, which include cities such as Genoa and Amalfi. Now these localities were among what was known as the maritime republics, of which there were four altogether (the others were Pisa and Venice, which for reasons deriving from the local terrain, did not develop paper industries) and which specialised in sea-trade, especially with the Arab world. The records are sketchy, but it is a persuasive guess that Arab papermaking techniques were brought across the Mediterranean and established in or near these cities. Just to get the dates given above into some sort of perspective, the British Library holds what is believed to be the earliest classical text copied on paper, a copy of Aratus, probably written in the first half of the Thirteenth century in Southern Italy or in Sicily, in Arundel ms. 268, ff. 75-103 (viewable on the library’s website). Likewise the earliest known dated manuscript on “Western” paper (i.e. unwatermarked, but seemingly produced with the procedures described here below) is a commentary by Ioannes Zonaras on the Octoechos, a Greek orthodox service book, written somewhere in an unspecific Eastern Mediterranean in 1252, again held by the British Library, Add. Mss. 27359 (similarly viewable on the website).

In 1888, in his famous article on the watermarks of Genoa, Briquet published the text of a document of 1235, in which Walter the Englishman (Gualterius Englesius) agrees to work with an Italian colleague in order to make paper, promising to keep the method secret, i.e. “nec alicui persone docere sive monstrare dictum misterium” [6e. Liguria]. Although fears about Medieval industrial espionage often feature in such contracts, the insistence on the misterium and the involvement of a lawyer to draw up the document suggest that the procedures were still relatively little known. It is reasonable to suppose that these early factories made paper with much the same procedures and tools as in Africa and in Spain. At a later stage, as knowledge of the revolutionary techniques being introduced elsewhere in Italy spread outwards, these centres adapted to Western methods, discarding the previous ones. But the procedures were sufficiently similar to make the change-over a seamless one.

If up to now, most of what has been said has been guesswork, what follows is pure unadulterated blind-man’s buff. So let’s play!

At some point in the Thirteenth century, somewhere in Italy, someone, or more likely several different people, took the centuries-old method of papermaking, more or less as it had come from China, without profound innovations in its passage through the Arab world, and transformed it. The changes that they made have remained more or less standard, wherever paper is still made at the vat, up to the present day. These innovations, which transformed paper into a major commercial product, have had enormous, far-reaching consequences for the history of records and culture (but we don’t want to go into that).

Where did all this happen? The evidence is fragmentary, but a great deal points to the small Italian town of Fabriano.


(Well, yes, Where? is a pretty good question. Actually a little Question and Answer session might be the best way of getting through the next bit).

Q. Where?

A. It is a town, not a very exciting town (if the truth be told), in central Italy, in the region known as the Marche, 325 metres above sea-level, population in 2016 a bit less than 32,000 inhabitants.

Q. Why Fabriano?

A. That is a very good question, you know. It is not asked often enough. To be honest, it is hardly ever asked at all. Which is strange, since it is the most important single question one could ask.

Q. So why Fabriano?

A. I detect a touch of impatience on your part, but I am glad you asked that. Of course the local scholars, beginning with the brothers Aurelio and Augusto Zonghi over a century ago, have searched at great length for documents that might explain the origins and rise of papermaking in Fabriano. And found absolutely nothing. They were moreover scrupulous in avoiding formulating just the question you ask and thus did not provide any answers whatsoever to what is a very intriguing question.

Q. Therefore why Fabriano?

A. I see. You would like me to answer the question. I had not realised that.

Of course what I am going to say now is a guess, but of an informed, intelligent variety. A bit as happened in Germany three centuries later, some time around 1450, when an individual we like to think of as Gutenberg did not really invent printing, but sort of cobbled together bits of know-how belonging to several different fields, so the huge leap forward in papermaking in the middle of the Thirteenth century draws on expertise in three other domains.

Q. So, can you get to the point, please?

A. Most certainly. Fabriano, as the name itself implies, was a metal-working centre, albeit with iron and blacksmithing, rather than the softer, more malleable metals such as copper and bronze. The skills involved in metal-working, especially in shaping the wire, was certainly very necessary when we come to talk about mould-construction.

Furthermore, there are hints that, when papermaking was introduced into Italy, the extant parchment industry somehow got in on the act. With its rolling hills and pleasant meadows, which accommodated large numbers of animals, in the Middle Ages livestock was an important feature of Fabriano’s economy. Animal size for paper is obtained from the collagen present in the skin, connective tissue, and bones of animals.

Here a little bit of basic (very basic) chemistry might be in order; or just jump this paragraph. Collagen is defined as the main structural protein in the intercellular space in the various connective tissues in animal bodies (also human, but generally we do not use them to size paper). The term derives from the Greek κόλλα (glue) and its suffix –γέν (producing), referring to the centuries-old practice of boiling down animal hide and sinews to produce glue. The boiling process hydrolises, or breaks down, the collagen, transforming it into gelatin that, much diluted, becomes the size. The various techniques involved, including the purification of the gelatin by passing it through cloth filters, were certainly known to Medieval parchment producers, while the final phase in papermaking, calendaring or polishing the surface with a dense, rounded stone, such as basalt or marble, was familiar to Oriental and Arab papermakers, but can again be related to the parchment shops.

Furthermore still, Italy’s Medieval wool industry, which covered Tuscany, Umbria and the Marche, reached as far down as Fabriano and had perfected the hydraulic stamping mill as a way of “fulling” the woolen cloth. For a long time in Italy wool and paper mills – both of which require large amounts of clean, running water – occupied much the same premises and used similar equipment. Likewise the couching (pronounced kooching, from the French coucher, since like other papermaking and printing terms it enters English from Belgian or Dutch French) process, where sheets of paper are transferred from the mould onto felts, was thought up by someone who knew about wool and knew that it would not adhere to the linen or hemp fibres in the paper. Woolen mills also made an ample use of screw presses and this characteristic switched easily into the nascent paper industry.

Q. Can you steer clear of “Furthermore” and just say what the innovations were?

A. Apologies! A little tick I have. But let us announce them properly. All six of them.

Q. If you really must …

Six Inventions that Lasted Six Centuries

A. If these several strands of know-how are woven into a single strand, somewhere in or near Fabriano, sometime around the middle of the Thirteenth century, there was an inventor, probably more than one, perhaps belonging to the same family, whose name remains unknown, but who, as much as Gutenberg and possibly more, has changed the history of the world. As more recent happenings have shown – for instance, the “chair triangle” around Manzano in the Friuli, or Luxottica spectacles in a valley near Trent – Italians are very good at small-scale, family-centered, concentrations of knowledge and innovation that have far-reaching consequences.

It is only a guess, but something like this happened.

First, the bamboo or reed or thread sieve of the Oriental and Arab mould was substituted with a metal mesh, formed by aligning wires of a copper or bronze alloy parallel to the long side of the mould. These are held in position by two, or sometimes three, strands of a more finely drawn metal wire, wound around small wooden bobbins, as in lace manufacture, which are knotted or plaited around the wires of the mesh (Lalande compares the process to basket-making). This sort of plait leaves the trace known as the chainline, running at right-angles to the wirelines, on the surface of the sheet. The chain-lines were in turn supported, although it was not strictly necessary to make them coincide, by triangular wooden struts or ribs, set at regular intervals.

Q. That is number one. And next?

A. Well, yes. Second, the couching process, or the removal of the sheet from the mould, was made quicker and more effective.

When in the West papermakers first experimented with a metal construction for the sieve, the new material necessarily made the mould heavy, rigid, and inflexible. A different technique therefore had to be devised to get the sheet off the mould, while an additional problem was posed by the fact that the laid surface, formed by parallel brass or copper wires, was much less regular than its Oriental counterpart and left a deep indentation in the surface of the sheet that had to be smoothed out. The solution involved, first, changing the nature of the deckle, which became a single, removable frame, and, second, alternating the sheets with pieces of woolen cloth or felt, somewhat larger than the paper. In other words, seconds after the removal of the deckle, the mould is turned upside down and the sheet is couched with a rolling, semi-circular pressure onto the felt (best to see it rather than have me describe it). It is better if the woolen surface is slightly yielding, so a pile of felts gives the best result. In most Western papermaking the sheet is added to the top layer of the “post”, i.e. the pile of interleaved felts, usually 250, that forms a unit of work, and another felt is laid on top of it before continuing with the next sheet. At Fabriano, however, the sheet is couched onto the topmost felt of the pile, which is immediately taken by its corners and lifted across to the post. The pressure exerted in couching required the mould to be extremely robust, in other words metal and hard, seasoned wood; anything else would simply disintegrate in a short space of time.

The woolen felts absorbed a proportion of the moisture of the sheet, but at the same time prevented the wet fibres from sticking together (Anyone in the textile trade will tell you that it is difficult to mix vegetable and animal fibres, so the couching process was thought up by someone who knew this fact). They also made it possible to press the post while the sheets were still densely hydrated. Pressing not only removed the excess liquid, allowing the easy separation of the sheets from the felts, but also flattened out the indentation left by the wire-lines on the mould side of the sheet. In early Medieval paper, in which the wires and thus the gaps between the wires were particularly thick, this process was extremely important. Without pressing in fact, it would have been impossible to write on more than one side of the sheet, with the additional disadvantage that sheets of paper would have easily broken along the chainlines when folded.

Like the chicken and the egg, what came first, the rigid mould with its metal sieve or the couching onto a post of woolen felts? The best answer is that the two processes were intimately connected and probably evolved simultaneously as a way of overcoming the limits posed by the flexible mould.

Q. I count two innovations so far, and this is taking a long time. And the others?

A. Certain things cannot be expounded in haste. So, third, as has already been said, instead of beating the rags with a foot-actioned treadle or a crushing wheel activated by a donkey or mule, Fabriano modified the fulling-mill, or gualchiera, widespread in the Medieval textile industry harnessed to a water-wheel. Fulling actually involves two phases. The first, known as “scouring”, was required to remove the dirt, oil, and impurities from the sheep’s wool: in Roman times the cleaning agent was urine, which was trampled by slaves in appositely constructed vats known as fullonicae (a pleasant task!), whereas by the Middle Ages urine was substituted with fuller’s earth, a soft clay derived from prehistoric deposits of volcanic ash; the second, known as “thickening”, consisted in compacting or felting the fibres to give them strength and resistance.

In arguing that the application of the stamping mill to beat rags for paper first happened in Italy, account has to be taken that hydraulic mills were known and widespread in Islamic and in Christian Spanish culture. In particular, sweeping claims have been made by Spanish scholars for the existence of hydraulically-powered stamping mills to produce pulp, both in Islamic Spain and in the subsequent Christian industry, which have however been denied by more recent (and less partisan) research, since in most cases the mills concerned were fulling mills for the textile industry. The earliest certain reference to a water-powered papermill in the Spanish Kingdom of Aragon occurs in 1282 and involves a dispute between the crown and the community of Muslim papermakers at Xàtiva, the main centre of the industry. Faced with luddite-style protests from the papermakers, who wish to continue with their traditional handbeating methods, the king exempts them from service in the new mill. The obvious implication is that the beating technology is a novelty being imposed top downwards. Of course, by that date Italians had mastered the procedures involved in water-powered beating and were beginning to experiment with watermarks.

Progress in technology, especially Medieval technology, where secrecy abounded and there was no concept of free sharing of knowledge, was rarely linear or straightforward, so it is quite possible that experiments were made at various dates in the mechanical beating of rags, without success or a permanent solution. What has perhaps not been sufficiently understood is how, when the technological leap-frog took place, all the various innovations interacted. In Oriental papermaking, the rate of production was necessarily slow, since thin sheets of paper were fabricated in a process that either required the sheet to stand on the mould to dry, or to be couched by taking the mould apart. Likewise, the subsequent drying on boards or walls required quite a lot of effort and space. Smaller amounts of pulp were consumed therefore and were easily supplied by hand- or foot-treadle beating. It ought also to be remembered that the raw pulp has to be used quickly, especially in a hot climate, otherwise it will ferment and go bad, so a hydraulic stamping mill might have been excessive to requirement. The introduction in the West of twin moulds and couching on woolen felts meant, instead, that a sheet, often of thicker paper, could be made every twenty seconds, while the post, usually containing 250 wet, freshly-formed sheets, took up less room, and subsequently the drying of the sheets happened on the meadow outside or by taking them up to the attic. So much larger quantities of pulp were called for and this impulse saw the adaption of the fulling mill traditional in the textile industry.

The mechanization of the process allowed the mill to beat larger quantities of rags in relatively shorter periods of time, thus obtaining a smoother mix of fibres, as well as to wash the rags, if necessary, by adding soap or cleansing substances, such as ashes, in the initial stages and filtering water through to export the dirt. In this phase of experimentation, not necessarily immediately, papermakers discovered that a controlled rotting, or retting, process (as in the compost heap at the bottom of the garden), before beating, made it easier and quicker to reduce the rags into their constituent fibres. Using rags had a further advantage, as well as being cheap (as anyone who goes round the charity shops knows). Years of rubbing against human flesh, and subsequent washing (not too often in the Middle Ages), and wearing again, broke the fibres down, making them more suitable for paper. In fact papermakers rarely employed new material, for example cuttings from tailoring shops, since it resisted the retting process and was only usable for rough paper.

Q. We are only halfway through the list. Can you speed things up a bit?

A. I’ll try. Fourth, not long after the introduction of the rigid mould and the immediate couching onto a felt, as in a Fordian organization of labour, where the aim is to produce more objects for less cost, it was discovered that the most efficient procedure was to employ a two-man team, one acting as the vatman and one as the coucher. The consequence of this discovery was the introduction into the process of a second or twin mould, perfectly uniform in terms of size and shape to the first, since otherwise either the vatman or the coucher would have been inactive for 50% of the time. While some sort of detachable mould surface, as has been said, was a characteristic of Chinese and later Arab papermaking, here the deckle had to fit exactly onto both the moulds in the same fashion, but also be easy and rapid to remove. The craftsmanship involved saw the rise of the specialist mould-maker, who in due course also took on the task of shaping and attaching the twin watermarks. The process led to a differentiation of the moulds, which over time involved a placing of the watermarks alternately in the right- or left-hand half, as member of a pair of twins.

Q. At least that was quicker. And next?

A. Fifth, and penultimately, as a substitute for the lichens or vegetables – usually rice, sometimes wheat – employed in sizing oriental or Arab paper, animal sizing was introduced, in which the sheets of paper received an infusion of dilute collagen solution. Again, the innovation shows the debt of early papermaking to the parchment industry, since the best sizes derived from scraps of skin left over when the membranes were cut into rectangles. This sizing, essentially glue, not only bettered the impermeability of the paper, but also made it much less prone to microbe or insect attacks. Perhaps even more importantly (most scholars who talk about the early history of papermaking rarely grasp this point), animal sizing, when it dried made the surface hard, more like parchment, and thus it was easier to write with a goose-quill pen, in which the ink is made to flow by exercising a light pressure on the point. As noted above, much the same effect was obtained in Arab paper by glueing two sheets together.

While Oriental and Arab methods seem to have preferred sizing with a brush (when and if they did size), Fabriano probably introduced dip-sizing, in which a handful of sheets at a time are briefly immersed in a vat. One probable cause was that animal size required the fluid to be warm, otherwise the collagen would solidify and form a jelly on the surface. The solution was therefore to heat the liquid in a copper cauldron over a small charcoal brazier. Dip-sizing also meant that the afterwards the sheets could be pressed once again, both to remove the excess fluid and to distribute the size evenly through the pile of sheets.

Q. Good. And to finish?

A. Sixth, last, but by no means least, by stitching a piece of wire bent into a distinctive shape to the surface of the mould, which duly left its indentation in the surface, papermakers found a way of marking the sheet with a sign of its provenance or quality or anything else one might want to say. Obviously, but obvious things are not always obvious, watermarks were made feasible only by the introduction of a rigid metal sieve, since on a flexible mould the rolling involved in couching would soon have broken or bent the watermark wire.

Dates, Mistakes, and Further Progress

Q. What a tiresome list! Can we date any of these innovations?

A. Apart from watermarks, which appeared in the mid to late 1280s (not quite as early as the “1282” claimed by Briquet) [18], only very approximately. The numerous links with the wool industry mean that, even if the word “gualchiera” appears in a document, it does not necessarily refer to papermaking. And matters are not helped by other muddles.

Q. Explain yourself.

A. Quite a lot of general histories of the book (especially those that take everything in secondary sources at face value, can’t read any language outside of English, and copy large chunks of information off the Internet) do indicate the introduction of the paper industry in Fabriano, or in Italy in general, or even in Europe in general, as 1275 or 1276, but herein lies an amusing little tale of scholarly ineptitude [6e. Marches]. The story is a somewhat complicated one and begins with the publication of the first edition, by the great Estense librarian, Girolamo Tiraboschi (1731-94), of his gigantic, multi-volume Storia della letteratura italiana, first published in nine tomes, comprising 13 volumes, from 1772 to 1782. (The real novelty is the invention of “letteratura italiana” in the title: the literature existed previously, but not the concept; on the other hand Ugo Foscolo’s venomous suggestion that the work be re-entitled Archivio ordinato e ragionato di materiali, cronologie, documenti e disquisizioni per servire alla storia letteraria d’Italia is not so very far from the mark!). In tome V (1775), the author touches briefly on the history of papermaking in the Middle Ages, and makes a fair mess of things, culminating in a vague claim that the use of linen rags was first introduced at Treviso around the middle of the Fourteenth century. Fabriano pride was touched to the quick and Tiraboschi received a lengthy communication from a local erudite and aristocrat, Luigi Mostarda (1723-1801), which made its way into his second, even more lengthy, edition, again nine tomes, this time in 16 volumes, published from 1787 to 1794. To give credit where it is due, Mostarda’s note, which Tiraboschi included in the most uncritical fashion imaginable, has a lot of pertinent and helpful information in it, but it also included, fatally, reference to a deed dated 1275, but which correctly had to be 1276 (thus explaining the oscillations in the many mentions by subsequent scholarship), that in his opinion contained a term describing a papermaking factory, i.e. cartere or carterem.

Q. This all sounds implausible.

A. It is pure unvarnished truth. As might all too easily have been guessed, but wasn’t, the text of the original documents actually read carcere or carcerem, i.e. normally a prison, but here designating the cell of a Benedictine nun (to read “c” instead of “t”, and vice versa, in Medieval handwriting is a standard slip of the pen taught in any basic course on Latin palaeography). The other fact that makes the interpretation implausible is the ownership. Of course, a document such as the Diario di Ripoli two centuries later does tell us that a Dominican nun, called suor Marietta, in Florence set type to print the Morgante, but here we are constrained to believe that Benedictine nuns owned or ran a papermaking factory! But believed it was, for a very long time.

Q. That sounds sexist! but I get the point. So?

A. To the credit of Fabriano’s home-grown erudition, in 1930 the mistake was identified by local scholar, Romualdo Sassi, who did take the trouble to go and read the original (and thus gets lots of brownie points). But of course, once the virus has got into the academic bloodstream, it is almost impossible to get it out (it still appears in the Wikipedia entry and there is no point in removing it, since some well-intentioned person would just put it back in). A few years ago I was reading through a draft for the synthesis of the history of paper for the 2010 Oxford Companion to the Book. The article was carefully documented and well-informed, but up popped the date 1276 for the introduction of paper into Italy. I drew attention to the erroneousness of the same, but with no success, apart from the addition of “ca.” (in other words a transition from “harmless” to “mostly harmless”). Likewise, the much vaunted, and emphatically promoted, new volume The Paper Trail. An Unexpected History of the World’s Greatest Invention by Alex Munro (2014): well, even the blurb on the Penguin Book website tells us how “Paper finally reached Europe in 1276 and was indispensable to the scholars and translators who manufactured the Renaissance and Reformation from their desks” (one would like to think that the printers got a look in somewhere, but no matter).

Q. Something of a blooper, I must admit. So, when did paper really come to the West?

A. To my mind, the introduction of papermaking with Arab techniques in Italy has to be pushed back as early as the 1220s, certainly no later than the 1230s, as is confirmed by the 1235 Genoa document, first published by Briquet in 1887 [6e. Liguria]. The subsequent metamorphosis through contact with the wool industry probably happened between 1240 and 1250, so that by the second half of the century the new procedures were beginning to turn out a significant quantity of material. Since this paper started as an inferior and cheaper substitute for parchment, at least until the sizing problem was worked out, it is understandable that very little of it has remained. Elementary common sense suggests that, if Fabriano’s papermakers had overcome most of the technical obstacles by the 1280s and were playing around with the frills, such as watermarks, then the real developments must have taken place at an earlier stage, probably a much earlier stage. In recent times attention has been drawn to documents in the archive of the Matelica, a town some twenty km south of Fabriano, which in 1264, and again in 1268, mention purchases of paper from an unnamed locality, but almost certainly Fabriano. We don’t know what this paper looked like, but since Fabriano seems to have been supplying the stuff on a fair scale, it is reasonable to assume that the industry was up and booming.

Q. So why has all this not been explained before?

A. Without accruing too much merit to myself, researchers are always bad at seeing what is not there, as in this instance, where there is a curious, even amusing, black hole in the scholarship. Almost all inquiry into the history of paper, quite legitimately, has developed out of specific areas of interest, which have led scholars to highlight and privilege some chronologies and geographies with respect to others. The most enlightening work has been done by bibliographers of the printed book, such as Alan Stevenson and Paul Needham (just to give two of the names that pop up most often in this piece): now, of course, since printing only appears in the middle of the Fifteenth century, such scholars have had no reason to go back any further in time. Codicologists have manifested an attention for earlier paper (as is shown by the valuable contributions of Jean Irigoin), although most of their interest focuses on parchment and of course, since paper began as a brownish-coloured, poor quality substitute, some attitudes appear mere palaeographical snobbery. Most manuscripts are also deracinated from their context and time of making: even if a colophon tells us when and where it was written, rarely is it conserved in the same place, and equally rarely is it possible to relate the paper to other documents in the same collection. Filigranologists, to give a fancy name to people interested in watermarks, spread their net more widely, but of course their starting point is c. 1282, with the earliest instances of dated marks, and given the vast quantities of watermarked paper still unstudied, they have little reason to adventure into earlier periods (unwatermarked paper is as taciturn as a headless corpse). As a result there is a gap of some sixty years that has never really been looked at, by anyone! Although there is an abundance of material in Italy’s city and state archives, which provide sequences of paper over long chronological periods. Now archivists, in my bitter experience, rarely if ever know anything about paper, except sometimes for conservation purposes, and archive ‘cataloguing’ (for want of a better word) is extremely poor on the whole at telling a user anything about the physical support of documents. The only solution is a hands-on one, in which one goes in prima persona to the archive and looks at the documents there (and, having done just that several times, it is also a good idea to enlist the support of a competent Medieval historian).

Q. Can you give some examples?

A. One series of documents I have personally handled, and which convince me that an Italian papermaking industry was established, and possibly even thriving, as early as the 1220s, are the acts of the city of San Gimignano in Tuscany, held mostly in Florence’s State Archive, though some volumes have remained at the city’s Biblioteca Comunale. This sequence of 494 registers are the object of an ongoing transcription and study by historian Oretta Muzzi, but also provide a unique example of a paper supply that shows a remarkable evolution in the space of relatively few years. The series begins in 1228, and the sheets of paper display irregularly distanced (oscillating between 38 and 47 mm), wide-set, chain-lines, while the surface has a sort of mushy feel, like soft toilet-paper. Within some twenty years, however, the surface has become harder and much more resistant, probably due to the introduction of animal sizing. It ought to be possible, especially using digital scanning techniques, to decide whether these sheets are produced on a common set of moulds, perhaps a single pair, which would point to a vatman and a coucher working together; or on a large set of similar moulds, which would suggest a continuation of the Arab method. These are however single sheets, not a pair of sheets glued together, as is the norm in Arab manuscripts of the time, and they also show the strange zig-zag striations known from Spanish-made paper of the same period, which have remained largely unexplained. Most importantly the study of a long sequence of documents, used and conserved in the same place, should make it possible to see significant innovations when they first appear. Another famous series of documents is the Liber plegiorum, or the oldest sequence of documents written on paper in the Venice State Archive, dating from 1223.

What scholarship is crying out for, therefore, is an exhaustive census of Thirteenth-century paper in archives and manuscripts, employing non-destructive methods such as Fourier Transform-Infrared Spectroscopy, which will tell us not only what the paper is made from, but also whether the sizing is vegetable or animal, whether there is evidence for moulds with a metal sieve, whether twin moulds are being used, and other interesting things.

Sorry, have I been going on too long?

Q. A bit too long. What happened next?

A. Very little for nearly five centuries. As will also be the case with printing, a remarkable technological step forward, achieved in a mere handful of years, transforms itself into a stable, wealthy manufacturing industry, with little need for innovation. Culture had to catch up with the consequences of having a new, much cheaper support for texts and took a fair amount of time to adapt.

During this period Fabriano continued to dominate the Italian industry, though other Italian states, and even other European countries, gradually lured away people with the necessary know-how and so other centres were set up. Papermaking required an ample supply of clean, running water, as well as a nearby market, both to collect the rags and to sell the finished product. Venice, in particular, made determined efforts to develop its own industry, which grew up principally in the valley of the cartiere above the town of Toscolano, on a promontory of Lake Garda [6e. Lombardia]. Lesser centres appeared near Voltri, to the West of Genoa (or rather, as at Amalfi, the Arab papermaking methods gave way to the new industrial process), and in Tuscany at Colle Val d’Elsa and Pescia. Elsewhere in Europe, important factories were set up in Basle [6k] and in the Auvergne in France [6c], but up to the Eighteenth century Italy remained dominant and exported on a huge scale.

Q. Were there no novelties at all?

A. There were plenty of small changes, mainly of the improving variety or to do with watermarks, with which we shall deal with later.

A very important improvement occurs towards the end of the Fifteenth century, when the wires forming the mesh of the mould become much finer and are more closely set, with obvious advantages for the quality of the paper, which was not only more even, but also much thinner. It is probable that this development derived from a not entirely recent discovery in metallurgy, in which wire was drawn instead of being hammered. The wires plaited together to form the chain-lines also become much thinner, to the point that it is no longer possible to decipher their number or how they are being tied.

As the industry expanded Northwards, in the late Fifteenth century, mould-makers started to introduce tranchefiles to reinforce the narrow ends of the moulds [13]. A tranchefile is a thin wire placed on the underside of the mould between the last rib and the short edge, usually at a distance of 18 mm from the former and some 10 mm from the latter, but of course the binder’s plough means that this last measurement is often uncertain. A chainline is plaited along it, but there is no supporting rib underneath. (Images of how and where they were placed are provided in the illustrations to Lalande in 1761 and to the Encyclopédie in 1765.) The presence of tranchefiles in a book can be helpful in establishing a format or the imposition of a printing forme, since they tell us where the short edges of the sheet happened to be.

My own experience suggests that tranchefiles appeared some time in the first half of the Fifteenth century, somewhere in the area shaped by South-East France, Western Switzerland, and North-West Italy. Work on the Gutenberg Bible, which can be dated to 1454-55, has drawn attention to the presence of tranchefiles in sheets of Royal paper and has argued that the provenance of the paper was one or more mills at Celle, near Turin in Piedmont. On the other hand tranchefiles are conspicuously absent from the moulds of major Italian papermaking centres, such as Toscolano and Fabriano.

Apart from watermark practice, there were no significant changes in papermaking processes in the Sixteenth and Seventeenth centuries. The Eighteenth century brought however two major innovations. The first was the Hollander beater, originally invented in the Netherlands in the Seventeenth century to work with a windmill, and in the Eighteenth adapted for hydraulic power. There are detailed early accounts of how the machine – basically a giant-sized Moulinex – worked in Lalande (1761) and in the entry on papermaking in the Encyclopédie (1765) [5]. The Hollander was much quicker – according to Lalande it could reduce a load of rags to pulp in eight to ten hours instead of twenty-four to thirty – and it had a larger capacity, but traditional papermakers claimed that it chopped the fibres too short and often left knots of material. In Italy therefore it never entirely replaced the traditional stamping-mill, which remained cheaper to construct and run, something the small family firms characteristic of the peninsular industry preferred.

The second, extremely important, change was the introduction of wove, instead of laid, paper (actually a reintroduction, since a wove surface, usually a thin piece of cloth, was characteristic of Oriental papermaking) [15]. Laid is the term for traditional paper, in which the fibres deposit themselves directly on the wire and chainlines, which leave a visible mark on the sheet. In wove, as the name implies, a thin mesh is placed on the surface of the mould and the fibres deposit thereon, so that signs of the chainlines and wirelines disappear. It is a metamorphosis whose inception has an exact date, since wove was famously employed for the first time in Baskerville’s Birmingham edition of Virgil in 1757, for which the paper was made by James Whatman at the Turkey Mill in Maidstone. Wove was also the essential technological step forward for the next stage in the process, the invention of the mechanical papermaking machine at the beginning of the Nineteenth century; but that is another story.

Q. Thank Heaven that is over. Is there much more?

A. We are not even half-way.