Edo G. Loeber: Experiences of an Inexperienced Filigranologist

1. One does not need to be a Rembrandt to make a watermark tracing. The only things required are good eyes, a steady hand to hold a well-sharpened pencil, and – above all – patience and love for the work.
2. This love will grow the more one penetrates into the world of imagination and artistry that this wonderland of watermarks represents.
3. Watermarks are as a rule copied in pencil and this copy is again copied in ink. However meticulously the work may be done, deviations inevitably occur. This is impossible to avoid, even when working with a large magnifying glass resting upon the chest.
4. Copying watermarks with a pencil has the advantage that one studies them more closely and thereby learns how the wire profile was constructed. Moreover, watermarks can be compared by putting one tracing over another.
5. The pencil has to be well sharpened, but preferably not of too hard a quality since this may leave a mark in the paper. One should try to follow the watermark line in the middle. For this a strong magnifying glass and good light are, however, required.
6. Drawings of watermarks should not show sharp turns since a wire profile never has sharp bends. This is also the reason why one should draw the pencil or pen around a corner before lifting it from the paper and putting it down again.
7. Normal tracing paper is very sensitive to humidity and heat. When working on a light-box with electric bulbs one should be prepared for changes in the size of the paper. Fluorescent tubes are far better since they do not develop a great heat even when used for hours on end.
8. Even when well glazed, tracing paper shows unevennesses that may impede the pen. Synthetic foil has a more even surface, but is more expensive. In most cases tracing paper has a smoother and a rougher side.
9. Really complicated watermarks should be copied on transparent plastic foil (e.g. Kodatrace) and not on tracing paper. The foil has a more even translucency and is more stable and transparent than tracing paper.
10. As a rule the results are more promising when one can continue the copying over a number of hours or even days. One acquires more skill, even though the accuracy may on the other hand suffer through flagging attention.
11. When a pen-and-ink copy is required it will be a great help to put the pencil copy on a background of graph paper. This helps to get the chain lines in position and the lettering more regular.
12. For tracings in ink a Redis-pen is recommended, or Rotring pens of 0.3 to 0.4 millimetre; in very complicated cases one should use a finer pen of 0.2 millimetre.
13. When copying watermarks in ink one may be inclined to make one‘s drawings more elegant than the original, especially by making curved lines more streamlined. This should, of course, be avoided.
14. When a watermark has to be done in ink from a pencil copy it is convenient to have a photocopy of the original sheet by its side. This is very useful for verifying whether anything has been overlooked or faultily copied.
15. After having worked on the light-box for several hours one seems to see watermarks that are not there at all. This is another case where a photocopy may be of great help in checking the tracing one is doing.
16. For an aesthetic appreciation of a watermark or the look-through of a paper a pencil or a pen-drawing are of no use: the lines are much too sharp and the contrast between light and dark too great. Photography, beta-radiography, electronic radiography, or Ozalids give a more faithful picture of the watermark and of the surrounding paper‘s structure. A reasonable result can also be obtained by making a rubbing from the wire side of the paper.
17. A photocopied watermark on e.g. Océ/Ozalid paper is obtained through direct contact. Since during the process this copying paper is but superficially moistened to a very low degree, the size of the reproduction is practically the same as the original.
18. Not a single line in a watermark – not even a laid or a chain line – is absolutely straight and no circle is completely round, except in imitation watermarks produced by plate-pressing into ready-made paper.
19. Hardly any chain line is parallel to the next. Moreover the distance between them varies, so that the fields differ in width. In addition, the ‘watermark field‘ may be wider or narrower, e.g. when divided by a specially added chain wire.
20. Chain lines should be indicated at top and bottom of the pencil tracing and again somewhere near the watermark. When they are more or less straight they may be drawn along a ruler; when bent, crooked, or even interrupted (broken), they have to be drawn by hand.
21. Even though it is impossible to indicate the exact gauge of the chain wire, it is recommended to make a note of whether the chain line is thick or thin since this may be typical for a certain paper.
22. The ordinary chain line consists of a series of somewhat obliquely orientated small oval dots, brought about by the crossings of the chain wires (twists) between the pairs of laid wires, and the crossings over each wire in the pair. Indications of various chain-wire constructions were found in very old Italian papers, as also in more recent Spanish cigarette papers made upon laid facings bound by a single chain wire instead of by pairs of twists.
23. When carefully copying the laid lines one may find that a greater spacing occurs every tenth or eleventh line at presumably the place where the binding wire crosses over a chain wire and a rib, attaching the one to the other. // p. 79//
24. Measurements on 20 laid limes are not sufficiently precise for comparing identical papers. Here one will have to follow Th. Gerardy‘s method of counting the laid limes contained over 100 millimetres, preferably in the centre of the sheet.
25. Even when the number of laid wires per 20 millimetres is the same, the laid design in the paper may turn out to be quite different due to variations in wire type, wire gauge, and wire spacing. Ozalid, beta-radiographic, or electronic radiographic copies will reveal this quite accurately; copies in ink or pencil do not.
26. Laid lines and watermarks in Fourdrinier paper appear about 1840-50; they always occur on the felt side. Paper made on a cylinder mould machine has its laid limes and watermark on the wire side, and therefore resembles handmade paper.
27. A laid mould cover made on a loom is much more uniform than one made by hand.
28. It would be useful to start a collection of photocopies of laid papers typical for certain mills or mouldmakers. This might be important in view of Prof. J. Irigoin‘s studies (Irigoin, ‘La datation‘) on early unwatermarked papers which he found to be datable by a combination of the width of their fields and the density of the laid lines.
29. In many watermarks one can determine in which way the mouldmaker guided the profile wire, i.e. where he began and where he ended and how the wires were crossed over. In copying watermarks it is exceedingly useful to pay attention to these factors.
30. In places where the profile wire crosses over, the watermark will show a small oval light spot, with its longest axis pointing in the direction of the uppermost wire.
31. Sometimes it is rather difficult to find out where a profile wire starts and where it ends. Sewing dots may prove to be a help: as a rule the mouldmaker fastened the ends of his profile wire by sewing several stitches tightly together, thereby securing this vulnerable spot of the wire profile more firmly.
32. A more complicated tracing can, unfortunately, not be relied upon in all its details. The most experienced filigranologist‘s eye can err and even old hands in the trade are liable to it. Moreover, watermarks from one and the same mould may show certain deviations owing to deterioration of the wire profile.
33. Watermarks in thin paper as a rule appear as very sharp limes. In bower quality paper, on the other hand, their outline is usually rather vague and a pencil or ink copy may therefore turn out to be somewhat inaccurate.
34. Many eighteenth and nineteenth-century papers are very cloudy and dirty in look-through. One gets the impression that the papermakers did not bother about the visibility of the watermark and often used moulds with a damaged facing and wire profile. In these cases it is very difficult to make an accurate tracing of the laid and the watermark and we may therefore expect serious errors.
35. The mouldmaker as a rule tried to avoid the use of short bits of wire in his wire profiles since those made in a single piece proved to be Ute soundest as each wire-end is liable to loosen and cause a hole in the sheet. Letters are therefore as a rube also made up of one single wire.
36. A watermark that consists of two parts can never originate from a wire profile made of a single bit of wire and vice versa. Genuine variants can thereby be distinguished from spurious ones.
37. In some cases a wire profile that obviously could have been made out of a single bit of wire was made of two because, (a) the wire broke during the work or (b) the wire proved to be too short and a supplementary bit had to be added.
38. It is often difficult to distinguish the watermark proper from a loosened end of sewing wire or from a laid wire that got raised because the wire profile had been attached to it too tightly.
39. Watermarks in paper from deteriorating moulds often inform us about the build-up and construction of the wire profile. They are extremely useful for studying the technique of the wire profile.
40. It was not unusual for the mouldmaker to use various gauges of wire in one and the same wire profile. This is confirmed by a small booklet, now owned by the Stichting Erven Honig(h), Zaandijk (NL), written and drawn about 1790-95 by one of the ancestors of the Honig family, an expert papermaker.
41. Occasionally the mouldmaker bundled (two or three) thin wires together instead of using one single profile wire. In these cases it is quite impossible to tell by the watermark how the wire profile was constructed.
42. The facing (and backing) of a mould often went to pieces through corrosion or as a result of rough handling when it was cleaned with a brush; most often, however, the underlying cause was metal fatigue, due to stress during dipping and couching.
43. Among the equipment of various mouldmakers a simple wire-draw was found upon which – if required – they could alter the gauge of the copper wire they had bought (Lessebo and Tumba (S); Eikers Bruk (N); Ambert (F); Fabriano (I), etc.).
44. Some mouldmakers had an outspoken preference for fixing the wire profile symmetrically over a chain wire. Others chose to put their wire profile between the chain wires – or over a certain number of chain wires when the profile was a large one. Still others do not seem to have bothered about the chain wires at all. They possibly used old moulds (bought at an auction) and replaced the wire profiles.
45. Apart from the palaeographical side, watermark lettering and its wire guiding are of great interest. The shapes were derived from written or engraved lettering. Technically, however, the bending of the wire bits exercised a great influence on the shape of these letters.
46. We still have not found a satisfactory explanation as to why some mouldmakers put their lettering ‘readable‘ and // p. 80 // others ‘in mirror fashion‘ on their moulds. It may be in connection with the readability of the watermark in the paper, but nothing can be said about it with certainty.
47. One explanation of why a wire profile should be put upon the mould cover mirror-fashion is, however, given in the case of the first postal stamps in the Netherlands (1852): the plate printing had to be done on the wire side of the paper and the text of the border watermark should be readable from the printed side of the paper.
48. Sometimes the wire profile would slide along the laid wires, especially when the mould was brush-cleaned. Another cause may have been the repeated pressure during the couching. Allan H. Stevenson already advanced the theory that watermarks move from right to left, wire side up, a statement based on his studies of a great number of watermarks. It has, however, not yet been possible to find a plausible reason or explanation for the phenomenon (Stevenson, Missale, pp. 248-52). When a watermark appears higher or lower, this indicates that another mould was used or that the wire profile was resewn upon the old mould.
49. Apart from twin marks from twin moulds there are also twin watermarks from two-sheet moulds. In the latter case one will find two pairs of twins (or sets of quadruplets) since the papermaker works with two two-sheet moulds alternately.
50. In early wove paper the watermark was as a rule placed along the lower border of the sheet. The watermark would thus interfere with the printing as little as possible. In writing or note paper the reason may possibly have lain in that the newly introduced steel pen would not catch when it passed over the watermark.
51. In order to indicate how a watermark was copied the following lettering is used (see Pls 124-5): ‘L‘ for ‘left-hand‘ and ‘R‘ for ‘right-hand‘ half-sheet; ‘C‘ for ‘in fold‘; ‘A‘ to indicate that the wire side lay downward and ‘Z‘ upturned during the copying on the light-box; ‘a‘ and ‘i‘ are added when the watermark was traced from the outside or the inside of the folded sheet. Whether one follows the above-mentioned method or chooses another (individually more convenient) one, it is recommended always to adopt a system of tracings and indication of details that will help others to make use of it.
52. In order to determine whether a (wire side) watermark is ‘Z‘ (turned towards) or ‘A‘ (averted, turned away), like. Th. Gerardy proposes, one should look at the laid and chain lines in a raking light. When a watermark has been copied with a hard pencil we may, however, be misled by the grooves left by the pencil.
53. Shadow zones as a rule appear on both sides of the chain lines, though they may also occur on one side only or even in between the chain limes. In the latter case we may discover a very faint, light line dividing the shadow zone lengthwise, the rib mark.
54. Is it possible that ‘shadow zones‘ (in laid as well as in wove paper) only appear when moulds with a single facing have been used? Moulds with a backing seem to have originated after 1750 in either England, France or Holland; yet there are early Italian papers without shadow zones. The makers of these papers possibly tried out some form of double facing or perhaps just fixed a thick wire zig-zag under it in order to obtain an easier de-watering of the mould.
55. A British papermaker told mc that moulds with a double facing work quicker than moulds with a single one, due to easier de-watering. Over and over again one bits upon the problem of the de-watering of the mould. Who amongst the hydraulic engineers will once and for all tackle the problem of the de-watering of the mould?
56. Most early cylinder-mould machine papers show less tearing-strength with than across the grain. This fact helps to distinguish them from hand-made paper which never shows this great difference in tear. Apart from this, every paper expert can tell the difference from the deckle-edge, particularly at the corners.
57. We should always keep in mind that watermark reproductions in books (whether letterpress or offset) may show considerable deviations in size when compared with the original, owing to the photographic reproduction methods used when the blocks are made or the plates prepared.
58. When trying to date a document we are in fact looking for a certain paper mould and compare the imprints of a mould in paper. The moulds themselves have long since disappeared.
59. If the opportunity of making a rubbing of a mould cover should present itself, do make one. Use a strong, thin, manifold paper or wrapping (not over 30g/m²), a soft carpenter‘s lead-pencil, or a stick of brass rubbing war (obtainable from Philips & Page Ltd., London). Rub lightly so as to avoid damage and complete your copy by indicating details about wire guiding, wire ends, and laid design.
60. A final piece of good advice: after having made your tracing of a watermark against the light or on a light-box it is absolutely necessary to re-examine the correctness of the drawing. Put it beside the original and compare the two meticulously for a second time.

Source: Loeber, Edo G.: Paper mould and mouldmaker / E. G. Loeber. - Amsterdam : Paper Publ. Soc., 1982. - pp. 78-80 (Appendix VIII); Originally published in German language: Erfahrungen eines unerfahrenen Filigranologen. - In: IPH-Information. - Hannover N.F. 10 (1976) 4. - pp. 94-99.

-- FriederSchmidt - 22 Nov 2006