How Were Medieval Shield Planks Joined?

Reading the Evidence: A Seam Is Not a Glue Recipe

Five kinds of statement are kept apart throughout this account.

There is also a basic structural distinction. Boards placed beside one another form an edge-joined core. Wooden layers bonded face-to-face form a laminated core; when their grain directions cross, the construction is cross-laminated. Parchment, leather or textile over wood belongs to the shield's composite structure, but does not turn a single wooden layer into cross-laminated timber.

Finally, glue at a plank seam, glue beneath a facing, a binder in a painted ground, and a later conservation adhesive have different functions. Identifying one does not identify the others.

Before the Medieval Shields: What Dura-Europos Can and Cannot Prove

Dura-Europos is therefore a technological comparison from many centuries earlier, not the missing chemical analysis of Seedorf or Szczecin.

The 11th Century: Lenzen and an Early Laminated Core

The fragmentary shield from Lenzen is the earliest medieval construction in this comparison. Its archaeological context is assigned to the second half of the 11th century: contextual dating, not a scientific date obtained directly from the shield wood.

This matters because lamination already appears at the beginning of the medieval sequence. The 14th-century Bartolomei shield cannot therefore be presented as the point at which medieval makers first discovered crossed wooden layers.

The Late 12th Century: Two Szczecin Shields, Square Edges and an Unresolved Bond

The two Szczecin shields are especially valuable because they preserve thin alder boards and join-related observations. They are separate objects: the painted shield, catalogued as MNS/A/22250/1 and dated by the museum to 1176–1200, and the much more fragmentary shield conventionally called "unpainted," dated more broadly to the late 12th century. Their measurements, coverings and fittings must not be merged.

Together, the Szczecin objects establish that square-edged boards were a real medieval shield-making option. They do not establish that glue was unnecessary, identify a particular glue, or justify universal claims about the absence of mechanical joinery.

Around 1200: Seedorf

The most useful statement is consequently straightforward: Seedorf's makers deliberately prepared the mating surfaces to increase the gluing area. Seedorf proves an important construction choice, not a universal medieval joint or glue formula.

Older literature also places the Oña shield D 59 at the Real Armería around 1190–1200. That date belongs to its shape, not to the surviving object, which the Real Armería's own technical study dates much later; it is discussed at the end of this article.

The 13th Century: Plank Assemblies Within Layered Shields

The broader Marburg group prevents Konrad from carrying the whole 13th-century argument. For the Vögte von Keseberg shield, conventionally dated around 1250, Kohlmorgen describes six linden boards about 12 mm thick, with one parchment sheet on each face. For W. von Liederbach, conventionally placed in 1250–1300, the description gives six visible linden boards around 8 mm thick. The Counts of Arnsberg shield, assigned the same broad period, has five visible linden boards around 10 mm thick. Current institutional dating for Liederbach and Arnsberg is broader or later, including 14th-century labels.

Their coverings are relevant without being a substitute for that missing evidence. Parchment sheets, thin parchment-like skin, and documented turnovers show that the wooden assembly was incorporated into a layered body. Where a facing spans several boards, it can mechanically bridge their seams; how much load it carried in each original shield remains an engineering question, not something quantified by the surviving descriptions.

Around 1300: Variety Within the Surviving German and Swiss Shields

The Marburg lion shield that Kohlmorgen identifies as Heinrich I of Hesse's, an identification I follow, is described as five or six linden boards around 15 mm thick. Its 1292–1308 bracket depends on that attribution, and a conflicting published identification as Heinrich "der Junker" remains unresolved. Parchment and modelled textile decoration provide evidence for a layered object, while plank-seam geometry and adhesive remain unknown.

The Nordeck zu Rabenau shield, conventionally late 13th century or around 1300, has a three-piece linden core described as glued. Its sources differ in their wording of board orientation, so a definitive plank diagram would exceed the evidence. Kohlmorgen also records a transverse strip partly let into the upper rear of the core, an important reminder that core reinforcement and edge-to-edge joinery are related but separate questions. It should neither disappear from the account nor be generalized into a universal transverse batten; its medieval originality and exact structural role require further examination.

The Welfenberg/Helfenberg shield has reported diagonal grain in a linden core around 9 mm thick. Its board count and seam map remain unknown; diagonal wood fibres do not, on their own, establish diagonal planks or cross-lamination. The Stromberg shield, conventionally around 1300, is described as five linden boards around 12 mm thick, thinning toward the edges. The Steinau/Steinrück shield, conventionally around 1300–1320, has a linden multi-board core of approximately the same reported thickness, but no secure board count. Their exact join profiles and adhesives remain unknown, and present curvature in both should not be read as a record of manufacture: deformation and warping can alter a shield after construction.

The eagle shield MV 172 at Sion, around 1300, long attributed to the lords of Raron and now thought more likely to belong to Antoine von Weingarten, was described by Laking as built of wood, leather and canvas. The museum notes that its reverse is lined with leather, so the leather may belong to the back rather than the face. No secure wood species, board count, seam profile or complete layer order has been published. It belongs in the comparison as evidence for composite shield construction; it is not evidence that all Sion shields shared the Bartolomei core discussed below.

Around 1320: Karl von Trier, Glued Boards and Wooden Tenons

The shield associated with Karl von Trier, Grand Master of the Teutonic Order from 1311 to 1324, is not one of the Marburg shields. It came from Burg Reifenstein in South Tyrol and is now in the Tiroler Landesmuseen in Innsbruck, and it is described in the heraldic literature as a processional or display shield. Its date of around 1320 rests on its painted heraldry; around 1900 Freiherr von Rahden argued that the board was an older shield painted later, and Bernhard Engel rejected that. Kohlmorgen gives four fir boards approximately 15 mm thick. Engel gives the same thickness but calls the wood Fichtenholz, the usual German word for spruce, not fir (Tanne). The two descriptions disagree on the species, neither rests on a published wood analysis, and I leave the question open.

An Einsteckzapfen is a separate wooden piece let into both boards across the seam. Kohlmorgen's word for the older Schenk shield, gedübelt, means dowelled. The two words may describe similar fittings or different ones; neither source gives a shape or size, so the two shields should not be treated as one method. What they share is that each is reported with wooden fittings in addition to glue.

The Mid-14th Century: Continuity and Cross-Lamination

The Muschenheim shield, conventionally around 1350, is described as three linden boards around 8 mm thick, with one parchment sheet on each face. Its unusual strap attachments, loops and toggles of thick hide, are relevant to carrying the shield, but do not identify the plank joints. The younger Schenk von Schweinsberg shield, also conventionally around 1350, has a reported linden core around 15 mm thick and parchment on both faces; its board arrangement and joinery remain unknown, and the older shield's dowels must not be transferred to this separate object.

Two further Marburg originals, the unidentified lion shield, Iv-Nr 16, and the shield traditionally labelled von Dernbach, Iv-Nr 9, have especially limited construction evidence. No published description I have found establishes their original core architecture or seams, and later copies and their conservation records cannot answer those questions. Including these objects is not a reason to invent technical detail; it makes clear why a collection of researched survivals does not equal a collection of fully exposed and analysed joint systems.

Bartolomei securely expands the range of medieval construction. Lenzen rules out treating it as the invention of that approach; neither object establishes a continuous regional tradition between the two.

The Later 14th and Early 15th Centuries: Layered Surfaces, Hidden Joints

The Canterbury shield associated with Edward of Woodstock belongs to his funerary achievements of 1376, an association that does not prove manufacture in that exact year. Laking describes it as made of light wood joined down the centre; later accounts call the wood poplar, but I have not found a published identification. Laking's front-layer sequence is unusually elaborate: white canvas, gesso, paper and leather over the wood. Modern scanning and conservation have taken place, but no complete seam analysis has been published. The central wooden join's detailed engineering and original adhesive remain unknown.

The large Sion pavois, MV 1, is stylistically assigned to the first half of the 15th century. Museum evidence records poplar and a Chapter-painted textile over an earlier textile layer; its precise board arrangement, seams and adhesive are not established. MV 1 must remain distinct from both the eagle shield MV 172 and cross-laminated Bartolomei MV 4: its layered coverings do not prove a layered wooden core. It shows a familiar boundary: a surviving composite object can preserve substantial surface and carrying evidence while still concealing how its wood was joined.

Late Medieval: The Oña Shields D 59 and D 60

Two shields from the monastery of San Salvador de Oña, now in the Real Armería in Madrid as D 59 and D 60, have early shapes and were long dated to around 1200. Álvaro Soler del Campo's 1996 technical study, made while both were under restoration, describes their construction and dates their surviving surfaces much later. They close this comparison because their date is uncertain, not because they are the latest securely dated shields.

The earlier date of around 1190–1200 for D 59 belongs to its shape, and the "cedar-like" wood comes from older catalogue descriptions, not from analysis. The unequal boards are real construction evidence whichever date is right, but they cannot stand for a shield made around 1200.

Chronological Comparison of the Surviving Shields

Dates below keep the qualifications given in the text. "Reported" identifies published descriptive evidence, not new laboratory verification. Thicknesses are approximate values from the sources, not standardized measurements. "Unknown" means not established in the published evidence, not necessarily absent from the object.

Dura-Europos, an ancient comparison from many centuries earlier, is left out of the table.

What the Glue Evidence Actually Establishes

Collagen-derived glues, often discussed under hide glue, bone glue or glutin, also belong in the reconstruction comparison and cannot simply be dismissed because casein offers an appealing explanation. A set-shield from Kaufbeuren, probably of the mid-15th century and now in the Bavarian National Museum, is especially instructive. Two glue flakes sampled from its skin layers were proteinaceous and matched glutin reference spectra better than casein, although Martin Siennicki cautions that FTIR can scarcely tell the two apart. The study could not determine the original glue of the shield body, and choosing casein for the replica, which had to be bent with heat and moisture, remained a reconstruction decision. The distinction between sampled facing glue and unknown structural seam glue is exactly the distinction needed for the earlier shields.

The Facing Belongs to the Structure

The chronological record repeatedly returns to parchment, leather and textile. These materials were more than convenient surfaces for heraldry: when bonded over an assembled core, they could connect areas of wood on opposite sides of a seam and contribute to the behaviour of the whole shield. That is a mechanical interpretation supported by the layered construction, and by medieval testimony connecting hide covering with strength. It does not reveal the exact load carried by a particular facing, prove that a facing compensated for an unglued seam, or establish identical coverings on every shield.

The differences matter. Seedorf and many Marburg descriptions support parchment systems. The eagle shield MV 172 has a less precisely mapped leather-and-canvas description and a leather-lined reverse. Bartolomei preserves separate textile campaigns. The Black Prince's shield has an unusually elaborate historical layer sequence. On the painted Szczecin shield, the paint samples lie directly on the wood, and the evidence does not authorize a universal full-face skin reconstruction. Material words also require care: parchment, rawhide, white-tanned leather and leather are not technically interchangeable simply because older catalogues sometimes use them loosely.

A shield should be understood as an interacting wooden and organic assembly. Its facing can be structurally important while the adhesive at its plank edges remains unknown.

Conclusion: Several Documented Solutions, No Universal Medieval Joint

The surviving shields support a more varied answer than "the planks were simply glued together." Square-edged plank assemblies are documented at Szczecin, where the original joining medium remains unresolved. Seedorf directly supports glued boards and deliberately enlarged bonding surfaces, without yielding a measured joint angle or glue chemistry. Two 14th-century shields are described with wooden fittings as well as glue: dowelled glue edges on the older Schenk shield, in Kohlmorgen's summary, and crosswise wooden tenons in the Karl von Trier shield, in Engel's description of around 1900. Both are older or secondary descriptions rather than modern examinations, and the two fittings need not be the same. Lenzen and Bartolomei demonstrate laminated wooden construction at widely separated points in the medieval record.

Many other shields preserve board counts, timber descriptions and substantial coverings while withholding the seam itself. That is not a reason to disregard them; it shows both the breadth of surviving construction and the limits of generalization.

Medieval shield makers joined wood in more than one way and incorporated it into composite bodies. The most authoritative account is the one that explains those differences while leaving the missing evidence visible.