Watermarks in antique paper: identification and dating guide

The watermarks of old paper are the secret signature of bygone mills. A mark traced with a brass wire that, when viewed by transparency, reveals who made the sheet and often to within a few years. Reading a watermark correctly can sometimes add two centuries of precision to a bibliographic record. Here is the method we apply to our pieces, without a laboratory, using only a lamp and a little patience.

What is a watermark and how it is formed in the sheet

A watermark is a mark impressed into the wet pulp as the sheet is formed. The papermaker sews a brass wire, bent into the shape of a design or letters, onto the mould: a sieve made of tightly stretched parallel wires (the laid lines) held by more widely spaced cross wires (the chain lines). As the pulp settles on the sieve, it is slightly thinner where it lies over the wire. When viewed by transparency, this area appears lighter. That is a watermark.

The technique originated in Fabriano, in the Italian Marches, at the end of the 13th century. Paper historians date the earliest known watermarks to the years 1280–1290, in the form of simple motifs (crosses, letters, monograms). Italian mills were then seeking to distinguish their papers from those of the Arab world, which never bore watermarks. The process reached France in the 14th century and spread throughout Europe. From the royal decree of September 1739, French mills were required to watermark the name of their province (Auvergne, Limousin, Angoumois) and, gradually, the name of the mill and a date. One can literally read “AUVERGNE 1742” within the sheet.

Examining a watermark without damaging the book

First rule, non-negotiable: never wet an old leaf, and never fold it to see better. Keep the book flat and work with the light. Four techniques are used.

1. Transparency through a window: hold the gathering open and flat against a north-facing window, out of direct sunlight. This suffices for most clear watermarks and costs nothing.
2. Light table or cold LED lamp: slip a white sheet behind the page and illuminate from below at low intensity. Essential for thick, stained or lined papers.
3. Photography in raking light: book flat, lamp tilted at 15–20 degrees grazing the surface. The laid lines and watermark appear in reverse relief, without interference from the printed ink. The method of choice when heavy printing obscures the mark when viewed by transparency.

Beta radiography: in a heritage library, a radiosensitive sheet is placed under the paper with a low-energy beta source. The result is a perfect facsimile of the watermark and laid lines. Inaccessible to the collector, but useful to know when studying a facsimile published by the BnF or the IRHT.

In the bookseller's practice, raking light photography with a simple architect's lamp yields remarkable results. This is the method we use for the detailed entries in our catalogue of rare books.

Major iconographic families

Motifs are not distributed randomly. Each region, each mill, and sometimes each master papermaker has a repertoire that evolves slowly. Knowing these families helps pre-orient dating even before consulting a reference work.

Family	Dominant period	Production areas	Clues
Cross, Gothic letters	late 13th – c. 1400	Fabriano, northern Italy	Simple outlines, thick wires, wide laid lines
Bell, anchor, balance	14th – 16th century	Italy, southern France	Motifs often topped with a star or cross
Jug, vase, fleur-de-lys	c. 1500 – 1700	Champagne, Comtat Venaissin, Auvergne	Master papermaker’s initials on the jug
Grape, bunch	17th – 18th century	Auvergne, Angoumois, Limousin	Gave its name to the ‘raisin’ format
Coat of arms, heraldic shields	16th – 18th century	France, Netherlands, Germany	Often those of towns or royal privileges
Horn, lion, mill name	1739 – 19th century	Entire France, after royal decree	Province name in full, sometimes year

The three reference works every bibliophile should know

Three works still structure research today, and all are available online. This is one of the rare areas of bibliophily where amateurs have access to the same tools as a heritage librarian.

Briquet, Les Filigranes (Genève, 1907) remains the standard reference work for the medieval period and the Renaissance, up to around 1600. Charles-Moïse Briquet spent decades scouring European archives and published a dictionary of over 16,000 facsimiles, arranged alphabetically by motif. The digitised edition is available on Internet Archive; it is navigated by watermark number rather than page.

Heawood, Watermarks, mainly of the 17th and 18th Centuries (Hilversum, 1950) takes over for the following two centuries: approximately 4,000 watermarks reproduced at actual size. Original print run limited, but widely distributed in libraries and accessible in digitised form.

Piccard, Wasserzeichenkartei (Hauptstaatsarchiv Stuttgart) is the world’s largest collection, freely accessible online at piccard-online.de. Several tens of thousands of records covering crowns, ox heads, towers, letters, scales, anchors, horns, and keys. Gerhard Piccard demonstrated that an exact match between a dated watermark and one to be dated often narrows the range to a few years.

Since the late 2000s, the portal Bernstein – The Memory of Paper has aggregated these reference works and several dozen other European databases into a single search interface. For a bookseller, it has become an everyday tool.

Watermark dating: real precision and limitations

Let us be honest about what can be expected. A watermark well identified in Briquet or Piccard generally provides a production window of five to twenty years for the design itself. However, a sheet may lie in storage before being printed or written on. The cautious rule in material bibliography holds that paper is used within ten years of manufacture, sometimes longer for high-quality stocks or bindings.

That said, the watermark remains the strongest argument in several very concrete situations. To distinguish an original edition from a later reprint, the watermark often settles the matter where typography is ambiguous. To date an anonymous manuscript, it becomes the only objective material evidence. To spot a facsimile or a forgery, a watermark dated after the alleged publication date kills the case instantly. This is why we systematically check the watermarks of major pieces before purchase.

From watermark to mill: tracing back to the workshop

Before 1739, it is the initials and accessory motifs that betray the workshop. A pot surmounted by a fleur-de-lys with the letters I HONIG points to the Honig family of Zaandam, major Dutch exporters; a bunch of grapes with the word “FIN” usually comes from Auvergne, around Ambert or Thiers. After 1739, the royal decree makes identification easier: the province—and soon the mill—can be read directly on the sheet.

This traceability has a direct impact on the value in bibliophily. A text printed on Dutch wove paper with a watermark can easily cost twice as much as a print on ordinary laid paper, and a large-sized paper with a prestigious watermark (Whatman, Auvergne) often forms the central point of the argument when appraising an old book.

Cross-referenced reading: watermark, laid lines, format

A watermark is never read in isolation. It must be cross-referenced with the laid lines and the sheet’s format to validate dating. Laid lines spaced more than 1.5 mm apart indicate paper likely produced before the mid-18th century; thereafter, improvements in moulds tightened the mesh. The presence or absence of wove paper (developed by James Whatman in the mid-18th century) is another strong marker, to be corroborated with our article on the difference between laid and wove paper.

The format names found in booksellers’ descriptions—raisin, jésus, cloche, couronne, grand aigle—originate directly from the watermarks that identified them. Knowing these marks allows one to read this technical vocabulary effortlessly and better understand the history of old papers within the book trade.

Is a watermark sufficient to precisely date a book?

Rarely on its own. It provides a production timeframe for the sheet within a few years, but the sheet may be used several years later. It is always cross-referenced with typography, binding, and, when available, provenance marks.

Is expensive equipment needed to observe watermarks at home?

No. An LED architect’s lamp, a camera (even a recent smartphone), a north-facing window, and a little patience are sufficient for 90% of cases. The rest depends on the eye and free online references.