The assumption is made here that up till, say, the last few decades of the 19th C., the practice of using alum in paper can be confined mainly to European and American papermaking; and further, that for most of this period the quantities used by the paper industry were perhaps small compared to its use in mordanting. However, with the demise of natural dyes, after 1858, in favour of synthetic products (not requiring alum) any further comparison would clearly shift, temporarily, to other uses for alum. But the replacement of alum by hydrated aluminium sulphate, manufactured from bauxite, was already imminent; so, in effect, later comparisons of this kind are no longer relevant.
Whereas in the distant past alum might conceivably have been added during the preparatory stages of papermaking, or to the paper itself, for some special purpose, its principal application has been associated with the sizing process. The latter can be divided, broadly, into papers sized with gelatine mixed with alum; and, more recently, those sized with rosin, acting as a water-repellent coating and precipitated with alum directly onto the cellulose fibre. With hindsight, we can now assign the first method to what may be termed Class I papers, papers with a high degree of permanence; and the latter to other classes of paper, where permanence was not of prime importance.
The main advantage of the second method lay in the fact that rosin could be added during the preparation of the pulp, engine sizing, i.e. before sheet (or web) formation had taken place (Bib.2 Vol.II pp.251-260). It thus eliminated the expensive, but long standing, after-process of sizing with gelatine. Wastage, and hence costs, were reduced very significantly which, in a highly competitive age, was a popular innovation. Unfortunately, the shortcomings of rosin size, in general use after ca.1830, were not appreciated for some time after its introduction. Unlike the gelatine process, which protected the cellulose fibre (see below), the acidity of the alum, used in rosin sizing, attacked the cellulose and led, ultimately, to the papers concerned becoming brittle and discoloured. This became increasingly noticeable when wood pulps formed a normal component of the furnish. The chemical extraction of cellulose from wood produces a more reactive, and thus a much more easily attacked, substance than the highly crystalline cellulose of cotton or linen.
Many paper makers, in an attempt to try and regain the status of their products as well as maintain lower costs, tried combining the two sizing processes for paper made on papermachines. But it was discovered that this belt and braces approach could not be applied successfully to papers made on fast running papermachines. Under these circumstances the gelatine component was often no more than a "lick and a promise". Various methods were devised to find a compromise, but none of them could match the quality of papers sized with gelatine in the traditional way in a size bath. To size a sheet properly with gelatine, the liquid size must be allowed time to soak into the paper and displace entrapped air. After this immersion any excess of gelatine is expelled by pressing; the sized sheet is then allowed to cool gradually until the size is in the gel state, and it is from this state that it is dried. This was, by definition, a slow multi-stage process, and it was not until equipment for temperature and humidity control was devised that it became possible to mechanize and speed up this operation. Even so, compared with other methods for sizing paper at more economic speeds, this was still a slow and expensive procedure.
The foregoing is a digression from the real purpose of this Appendix, which is concerned with the Class I papers produced in the 18th and earlier centuries, and those that survived through all the changes of the 19th C. into the 20th. Rosin sizing was just one among many elements that affected the quality of paper as it progressed through, what the author has termed, "the chemical era" in papermaking. The first change was chemical bleaching introduced ca.1790; followed by a long series of other changes practised before their possible effects on the quality of paper were fully understood chemically; and before adequate means of monitoring them became available. Consequently, with the rapid expansion of output of both handmade and machinemade papers in the 19th C. it is not surprising to find so much paper in need of conservation today, no doubt a significant proportion of this attributable to the acidic action of alum.
But the dangers that might be encountered in the manufacture of the earlier Class I papers were potentially no less damaging to the cellulose than those introduced in "the chemical era". For example, severe degradation of cellulose could result from acids produced in the fourth stage of the retting of linen or, equally, damage caused by the fermenting of rags, still widely practised up to the mid-18th C. (Bib.2 Vol.II pp.193-198). Some of the affected material was no doubt washed out in the papermaking process, but an element of chain scission most probably remained. There were other hazards such as the use of quicklime to "soften" rags for beating in lieu of fermentation, a practice forbidden by law, but nevertheless employed by unscrupulous paper makers. Certain methods of laundering textiles could also lead to severe oxidative damage (Bib.2 Vol.II pp.205, 213 and 253 n.142). These examples should be sufficient to demonstrate that, apart from the effects of subjecting paper to a hostile environment, the traditional handmade papers of the past, made from rags, were potentially far from being immune to factors (other than alum) likely to lead to their degradation.
It is not known when gelatine sizing was introduced into papermaking, but it has been proposed that: (i) it was first used in European papers; (ii) that it replaced sizing with starch, a process inherited from Moorish and Arabian papermaking; and that this change took place from the 13th C. onwards. What is also uncertain is when alum was first added to gelatine to create what became the traditional composition of size. It may have been well known from the outset that alum was an essential ingredient, if gelatine or glue was to be applied to, say, fabrics or, more specifically, to any cellulosic artefact. It is conceivable, however, that to begin with gelatine was used by itself to coat paper, giving the vulnerable waterleaf sheet greater rigidity and protection from damage. If this was so, paper makers and paper users must have very soon realized that, as in the mordanting of dyes, alum was needed to bind the gelatine to the cellulose substrate. It is a fact that gelatine alone is very easily removed from paper, if the latter is wetted, a weakness that may have been exposed in the early stages of printing.
Alum would not have become an established component of gelatine size had it not served a very definite function. It was obvious that it had a mordant-like action in bonding the gelatine to the paper (see Note 17); it increased the jelly strength of the gelatine film; and it possessed some biocidic properties in retarding, if not preventing, a degree of mould growth. The above answers the second question posed at the beginning of this Appendix; some points will be amplified in the section that follows, particularly when we consider the quality of the gelatine used by paper makers.
A variety of products are derived from animal hides. Where the skin tissue is sought, the hides have to be de-haired. In the 20th C. this has been achieved by using either quicklime or sulphurous acid, producing ultimately (see below) a calcium gelatinate (pH 6) on the one hand, and an acid gelatine sulphate (pH ca.3) on the other. It seems certain that, in earlier centuries, the limed hide would have been the source of the paper maker's gelatine. The undamaged parts of the hide would have been saved for leather or parchment and treated accordingly. The waste from the hides was kept for other uses, mainly as a source of glue or gelatine.
Gelatine is a derived protein obtained by treating the principal protein of animal connective tissue, collagen. Two sources of this are used today, fleshings off the inner side of the hide and the trimmings, referred to above. We need only concern ourselves with the latter. Nowadays these are treated in lime pits for several days, to remove hair and to convert the natural protein, collagen, into gelatine. The action of the lime also converts some of the amide groups in the gelatine into carboxyl (acid) groups enabling it to combine with calcium ions at pH 6, usual for a lime hide product. This is only a bare outline of the process, which includes other processes such as the removal of fats etc. In essence the stages of this process were probably followed in earlier periods when preparing gelatine for use in paper mills; but in some instances they may not all have taken place on the site of the paper mill, e.g. the removal of the hair, the liming process, could have been undertaken at a tannery leaving the paper mill to wash out surplus lime (Bib.8 Art.98). The paper mill would have, as a matter of course, extracted, strained off foreign matter and concentrated the gelatine, from an aqueous suspension of the trimmings, in heated vessels. Without modern methods of control the quality of the gelatine produced in past times would have varied considerably. Le Francois de Lalande (Bib.8 Art 97) believed that the size produced from Tanners' cuttings (scrowls) was the strongest, but that it reduced the whiteness of the paper; he maintained that sheepskin trimmings produced a weaker size, but a whiter paper. Finally, he recommended that size extracted from parchment should be used for the best qualities of paper. However, to go a stage further, the Society of Antiquaries instructed the younger Whatman to use only Kid-Leather size for the Antiquarian paper required for Basire's engravings. (For Basire's full specification see Bib.1 p.31: for an interpretation of this and what the stipulations signified see Bib.2 Vol.II pp.259-260).
It can be seen then that several different qualities of gelatine might have been used in the paper mill. The final colour, for example, could have ranged from a pale yellow to a brown, conditional on the state of the raw material, which might have been fresh trimmings or "pieces" in a quasi-putrified state. That undesirably dark or cloudy size caused problems for paper makers is confirmed by the practice of adding "white vitriol" (zinc sulphate) or, worse, "copperas" (ferrous sulphate) to bleach or clarify it. The use of both materials was frowned on, partly because they affected writing inks and partly because the ferrous sulphate (white) oxidized with time to brown ferric hydroxide, discolouring the paper.
Needless to say, the manufacturers of top Class I papers would have ensured that the materials they used were of the best quality, so avoiding blemishes that might have spoiled their products. The elder Whatman, for example, inherited a tanning business and continued to run it until he was thirty-eight. Thus he would have been in a position to know what scrowls to accept or reject. The same regard for the quality of their materials was undoubtedly true for many other paper makers of those times, witnessed by the survival of their papers in good condition to this day.
In the preceding sections some of the potential agencies that might lead to the deterioration of rag papers, a hostile environment apart, have been outlined. It is claimed that manufacturers of Class I papers would have been aware of the effects of these dangers based on practical experience and taken what steps they could to keep clear of trouble spots. Much would also have depended on the standards maintained in their finishing departments. But, as pointed out earlier, papers of this quality have survived successfully for centuries, in spite of the fact that, in general, they may have a pH significantly lower than neutral. Acidity in paper has been condemned and rightly so in certain circumstances; and alum has been included among the agents responsible for this condition (Plenderleith & Werner Bib.61 pp.53ff.). How then can we reconcile these two conflicting positions?
A suggested explanation for this anomaly is outlined below. This is based on the author's experience of 20th C. practice applied to the production of Whatman writing and drawing papers (manufacture regrettably discontinued in 1962). Consequently, it may not receive universal acceptance; but it could lead to a better understanding of the slightly "acid" condition found in the sized paper. The solution proposed is certainly backed up by evidence from the past.
One of the factors to be considered before accepting the delivery of gelatine to the paper mill was its response to the addition of "alum" to a gelatine solution. It will be recalled that the paper makers in Le Francois de Lalande's time noticed that alum in the size gave the paper more "body" (plus ferme) and "rattle" (plus pétillant); and, in addition, it could raise the viscosity and gel strength of the size in very hot weather. These features were as important in the 20th C. as they were in the 18th C. There were, of course, many other considerations to take into account, such as the colour, clarity, mineral content etc. But the reaction to alum was, perhaps, the most important. The limed hide class of gelatine gave a very positive response to the addition of alum, whereas the acid type did not and was usually rejected on these grounds, possibly for the wrong reasons. So for our purpose here, interpretation of this reaction will be confined to the limed hide class of gelatine.
On adding the alum to the limed hide gelatine sol, there is an immediate and dramatic rise in the viscosity, so much so that in this state it could not be used to size paper. With time, the viscosity falls to an almost constant and acceptable level. In practice the size, after making up, was left to stand overnight before use. If one examines this behaviour more closely it can be interpreted as follows. As alum is added, so the pH of the gelatine sol is lowered, the alum being converted into calcium sulphate and aluminium hydroxide precipitated (maximum insolubility pH 4-9) within the gelatine sol. The precipitate, normally gelatinous itself, is not overtly visible; but, if ferric or chromic sulphate (or their corresponding alum form) is added in place of aluminium, brown or green streaks can be seen forming in the sol. The reaction may be expressed as:-
Al2(SO4)3 + 3Ca(OH)2 = 3CaSO4 + 2Al(OH)3
Alum is added to a point just below the isoionic point (pH 5) of the lime hide gelatine, in practice to a pH of ca. 4. It was under these conditions that Whatman papers were sized. When dried and finished for despatch the sized sheets of paper had a surface pH of 4.7-5.0, probably raised there by calcium carbonate, derived from hard water dried into the waterleaf paper (a typical 20th C. size contained ~ 0.55% hydrated aluminium sulphate and ~ 6.7% air-dry gelatine by weight).
This pH range has been found from experience to produce the most satisfactory sized papers. On the basis of the above, the acidity of the alum has been neutralized and we have in its place aluminium hydroxide as a suspensoid or lyophobic colloid which, initially, is difficult to disperse and might, therefore, account for the large increase in the viscosity of the gelatine sol and its slow decline. The vital feature of the process is the conversion of the alum into the corresponding hydroxide. It is thought that in this state the hydroxyl groups of the cellulose and the gelatine are attracted into the hydration shell of the complex aluminium ion and that this acts as a cross-link between the cellulose substrate and the gelatine, somewhat similar to its behaviour when acting as a mordant. Provided then that the paper is sized under these conditions, and that alum has not been added in excess, there should not be any residual acidity in the paper to attack the cellulose. In any case one would expect the gelatine to have a buffering influence at its iso-electric point. It is hoped that this explanation illustrates that the system, described above, is a stable one and vindicates the age-old practice of using alum in gelatine size with certain obvious provisos.
In the foregoing pages, an attempt has been made to answer the questions posed at the beginning of this Appendix. Taking the last question first, "alum" today, customarily but incorrectly, refers to hydrated aluminium sulphate. In truth it should signify one of the double salts; for paper made in the past, potash alum. Regarding the second question, what is the function of alum in papermaking, the distinction has been made between the function of alum in gelatine size and its role, as a sizing agent, in precipitating rosin onto the unprotected cellulose fibre. In the first instance, assuming the absence of a hostile environment, alum, precipitated as aluminium hydroxide, acts as a safe bonding agent in marrying gelatine to cellulose; in the latter, one has a situation where the cellulose fibre is exposed to attack by acid. Finally, does alum in paper deserve the evil reputation often accorded it? As indicated above, this depends on the kind of paper under discussion. The answer, in the case of Class I papers, is in the negative. The author has handled and inspected papers made in the 16th, 17th, 18th and 19th Cs. and found them to be in good condition in the 20th C. Perhaps we should now let Byron have the last word:-
"...to what straits old Time reduces
Frail man, when paper - even a rag like this -
Survives himself, his tomb, and all that's his."
 The younger James Whatman's Ledger (1780-1787) shows that he bought 55 cwt. alum per annum (in two qualities) compared to the North Yorkshire output for 1860 of 4000 tons (Scotland 5000 tons and Spence 5,500 tons. Almond). Whatman's paper mill was one of the largest paper mills in England at that time (1780s).
 According to Labarre (Bib.22 pp.66-67) both the Chinese and Arabs dyed their papers at times. The Arabs had three methods of applying dyes (J.Karabacek) two of which were wet methods. It is conceivable that in these instances alum may have been used as a mordant, although alum is not mentioned in the description of the process in the Umdet el-kuttab.
 For more detail see under Internal Sizing (Bib.2 Vol.II p.252). Rosin is a gum residue from pine trees, consisting mainly of abietic acid (C20H30O2), previously water-solubilized with alkali. Alum can be added before or after the addition of rosin to the pulp.
 Until the introduction of steam power and the mechanization of process equipment in the 19th C., paper was sized in a metal tub, the contents maintained at a moderate temperature (ca.50-60C.) by a charcoal brazier underneath and replenished from a nearby heated receptacle used for preparing the size. Sheets of paper, held by hand, were immersed in the tub until saturated and then removed to a press where excess size was expressed. The sized sheet was allowed to cool and then taken to the drying loft. The younger Whatman took this process a stage further by sizing a whole wad of paper contained in a wooden box, which was lowered by pulley into a size bath and left to absorb the size for half an hour. Later, in the 19th C., handmade paper was sized in a long size bath called a traveller, the paper being inserted between endless felts at one end which carried it through to squeeze rolls at the other. The sheets were deposited on another set of travelling felts, exposed to warm air, allowing the paper to cool gradually and the size to gel.
Also during the 19th C. equipment was devised to try and bridge the gap between the manual methods described above and on-machine methods using steam-heated drying cylinders. For instance, spar drum dryers were developed, by means of which "a web of paper was passed through a size bath and dried over a long series of 'skeleton' drums inside of which are a number of fans rapidly revolving; sometimes there are forty or fifty of these drums in succession, the whole confined in a chamber heated by steam. ....The advantage of drying the paper in this manner over so many of these drums is, that it turns out much harder and stronger, than if dried more rapidly over heated cylinders". (Herring, Richard Paper and Paper Making, Ancient and Modern, Longman, Brown, Green and Longmans 1855 p.73). As mentioned in the text, gelatine size should be dried from the gel state; drying from the sol state, as it would have been if dried over heated cylinders, produces an entirely different, weaker film.
 The term "hostile environment", referred to in the text, embraces a variety of conditions including climatic extremes; moisture from condensation or immersion; exposure to acidic or sulphurous or saline vapours; ultra-violet radiation; dimensional changes etc. (For details see Plenderleith & Werner Bib.61).
 Chinese and Arab papers were often sized both in the pulp and in the sheet stages. The former (it has been claimed by Clapperton) using fish or vegetable glues; the latter, wheat starch as a filler (Labarre, E.J. Bib.22, citing Karabacek).
 Singer, Charles (Bib.57 p.109) The Arte dei medici e speziali, besides controlling drugs, directed certain alum-using trades, notably those of saddlers and paper and parchment makers", in the 13/14th Cs. The use of alum in paper is also implied (ibid.p.132) in a 14th C. reference to Nuremberg as a papermaking centre.
 As mentioned in the text, it is not known exactly what prompted the addition of alum to gelatine in papermaking size. The idea may have come from some entirely different usage. On the other hand, in Europe, the use of paper for writing was preceded by materials such as parchment and papyrus. It was some time, and then only gradually, before paper replaced parchment (Bib.2 Vol.I p.1; p.81+Note 24). Parchment from early times had been treated with alum to harden and protect it, a form of tanning known as "tawing". (It is worth noting that, unlike its role in bonding gelatine to paper, the alum in this instance is water-soluble; chromium salts were needed to produce effective tanning). Although for the wrong reason, early paper makers may have thought that what alum did for parchment it might equally do for gelatine, a derivative of parchment.
Alum was added "in due proportion" to gelatine size as a matter of course in the 17th C. (Bib.2 Vol.I p.40), but there is no reference here to its function. Le Francois de Lalande (Bib.8, 1761, Art.103) is much more specific not only with respect to the quantities used, but he adds "Ce sel styptique & astringent sert à faire tenir la colle sur le papier, comme dans la teinture il rend les couleurs plus adhérentes à l'étoffe; le papier en est plus ferme, & comme disent les Ouvriers plus pétillant". He mentions another practice, well known in more recent times, "Si l'on craint les grandes chaleurs, on augmente quelquefois la dose de l'alun...".
An interesting sidelight on the modern view that the use of alum in paper is harmful, some paper makers, it seems, have tried sizing paper with gelatine only. The author was given a very presentable sketchbook, in which the paper resembled the description given in Note 13 above, namely, "harder and stronger" than the limp drawing papers usually on offer today. There was nothing to say that the paper in this sketchbook had been sized with gelatine only. But, whereas the surface was reasonable for drawing on, it was very disconcerting to find that when one applied a wash to it, everything moved with the brush !
 Le Francois de Lalande (Bib.8. Art 98-102 and Plate XII)) describes the extraction of gelatine at the paper mill from trimmings of limed hides brought from a tannery. After dusting out the lime, the material was simmered, not boiled, in cauldrons lined with straw to prevent the scrowls from sticking to the sides of the vessel. A perforated frame was later pressed down onto the bouillon to separate the liquor from the straw. The liquor was then ladled out and filtered through a cloth into another vessel. This removed the last traces of the trimmings, which were tested from time to time to find out if they were still yielding glue, fingers being immersed in the extract to see if they stuck together or not. This description, much abbreviated, was published in 1761 and it is interesting to note that Joshua Gilpin recorded that much the same process was still practised in some British paper mills at the end of the 18th C.
 Le Francois de Lalande (Bib.8 Art.99) extols the use of certain fish glues for sizing, but adds that they are expensive.
 It is not clear exactly what the term "Pieces" refers to. Thomas Balston (Bib.1 p.59) states that the younger Whatman made a clear distinction in his Ledger between Scrowls and Pieces, buying about 12 tons of the former against 8 tons of the latter, the Pieces coming from rabbits, calves, bullocks and oxen at about three-quarters of the price of Scrowls.
 Gravell, Thomas L. (Bib.19 Letter 6:Nov:1980) found that a number of 17th/18th C. books "have sections so brown that the ink is very hard to see, while the section alongside (with the same watermark) will be in good condition". He believed, judging from the thickness of the books (500-600 pages), that different batches of paper must have been used to make them up. Size, in his view, may have been the cause of this. Not having seen the degree of discolouration, one can only guess at the answer. It seems unlikely that the paper maker would have sent out paper in this state, or that the printers would have accepted it. A possible solution is that the paper maker had treated dark or cloudy size with copperas, possibly in excess, and that with time it had oxidized to the ferric (brown) state.
 The "alum" referred to in this context was hydrated aluminium sulphate (see p.197). The latter served as well in papermaking as the former; 1 g. of potash alum 0.7 g. hydrated aluminium sulphate.
 The usual composition of the size was 6.7% by weight of air-dry gelatine in water containing approx. 8% by weight hydrated aluminium sulphate based on the air-dry weight of the gelatine (0.55% of the total solution). It is difficult to compare this directly with the composition of the size described by Le Francois de Lalande (Bib.8 Art.103). It would appear that only 5% of potash alum was added, based on the air-dry weight of gelatine; but we do not know how an 18th C. gelatine would compare with a 20th C. one, nor the two alums with each other.
Le Francois de Lalande states that "Alun de Rome" was preferred to "l'alun de roche". The Alun de Rome referred to here must surely have come from Tolfa. Singer (Bib.59 p.58) claims that "Roman alum" referred to alum from Rum, the arabic name for Byzantium. One would imagine that by the 18th C. this interpretation had lapsed. Desmarest (Bib.27 pp.517-518; 555-556) concentrates primarily on the sizing process and the apparatus in use at the time, indicating a size pick-up of 7% (compared to a 1789 Whatman paper of 8%; and 20th C. Whatman 8-12%). He confirms Le Francois de Lalande's quantities of alum added; the situation is complicated, however, as to whether the rags used for the paper had been fermented or not. Likewise he recommends the use of alun de Rome.
 Alum in excess of this would depress the pH even further. Theoretically below pH 4.0 the aluminium hydroxide should redissolve rapidly. In practice, however, the amino groups in the gelatine might complicate this situation by reacting to form a gelatine sulphate. Excess acidity of this order would appear to be an unlikely event unless it was accidental. A tongue test might have been used to check this (?).
 I am indebted to Dr.Derek Priest (Dept.Paper Science, UMIST, letter 23:Feb:1997) for suggesting a hypothetical bonding relationship in the cellulose/alumina/gelatine complex.
Singer, Charles (Bib.57 p.xviii) "The chemical nature of 'mordanting' or fixing of dyestuffs in fabrics has become better understood in the twentieth century. It is now known that the necessary factor in the process is a triple combination of the substance of the fibre with the insoluble alumina from the alum and the colour base of the dyestuffs".
 A letter to the author from Dr.Vincent Daniels (Dept. of Conservation, British Museum. 21:Oct:96) cites two examples of materials containing alum, one of which showed accelerated natural degradation (a Chinese hanging scroll sized with an alum/starch paste); and the other no evidence of accelerated degradation (a Japanese scroll mounter sized with alum in deer skin glue).
 Inf. Letter 6:Nov:97, Dr David McKitterick.
This essay has four sections, spread over three pages: