How Were Medieval Stained Glass Windows Made?

The glass itself was not painted with colour; the colour was built into the material at the melting stage. The Benedictine monk who wrote under the name Theophilus, working in the early twelfth century, gives the earliest full recipe: washed sand or ground quartz pebbles was mixed with roughly twice its volume of beech wood ash and melted in a furnace. Beech ash supplied not only the potash needed as a flux but also iron and manganese oxide impurities that were themselves important for generating colour. This wood-ash formula produced what specialists call potash-lime-silica or "forest glass," and its widespread adoption across northern Europe around the end of the first millennium marked a deliberate shift away from the soda-based natron glass of the Roman and Mediterranean world, changing not just the recipe but the furnace design and the location of production. Deliberate colourants were added to the molten batch: copper produced red, cobalt gave blue, manganese produced pink or purple, and iron oxide gave various greens or a bright yellow, with the final shade set by the amount of oxide and the firing temperature. Deep pot-metal red was a particular problem because the dark ruby glass was not translucent enough for light to pass through, so glassmakers dipped a bubble of colourless glass into a pot of molten red glass to produce a thinner, streaky red layer over a clear base, known as flashed glass. Once a batch of coloured glass was ready it still had to be formed into flat sheets a glazier could cut. Two blown methods were used side by side. In the muff or cylinder method, a lump of glass on a blowing pipe was inflated into an elongated balloon shape, then split along its length and flattened while still hot, yielding sheets probably no larger than about 25 by 31 centimetres, sometimes identifiable by a rounded muff edge or by straight parallel air bubbles. In the crown method, a blown sphere was transferred to a metal rod, the pontil, and the far side broken open, leaving a hole; the bubble was then reheated and spun so centrifugal force opened it into a flat disc, which tapered from a thick lump at the centre (the bull's eye, where the pontil had been attached) to a thin edge. In England the crown method itself was not practised domestically until the seventeenth century, though imported crown glass from Normandy was already used in medieval windows, while home-produced sheet glass came chiefly from the muff process.

Before any glass was cut, the design had to be agreed and laid out at full size. Medieval sources call the preliminary sketch shown to a patron for approval the vidimus, Latin for "we have seen" — the process of making a stained-glass window began with a design known in medieval times as the vidimus. Working out the exact placement of every piece of glass and every lead line, though, required a full-size master drawing, and for most of the medieval period that drawing was not made on paper at all. Before paper became more readily available around the year 1400, stained-glass painters drew the scaled-up design directly onto a wooden trestle table coated with chalk or whitewash, marking it with letters and symbols for colour and painted detail, and the drawing on the table then served as the guide for cutting the glass, the guide for painting, and finally the workbench on which the window was assembled. Because that master drawing was the work surface itself rather than a separate portable sheet, for much of the medieval period the cartoon was not preserved at full size, and only from the seventeenth century were cartoons marked out on paper so that they could be stored and reused. This is one reason full medieval working drawings for specific surviving windows are rare: the design lived and died with the table it was drawn on. As the Victoria and Albert Museum's own glass specialists put it plainly, the exact working method by which the initial design was translated into that table drawing is not fully recorded — we don't know exactly how the design process of medieval stained-glass windows worked.

With the coloured sheet glass and the table drawing both ready, the glazier selected a sheet of the right colour for each shape and laid it over the corresponding area of the drawing. Medieval glaziers had no diamond or wheel cutter; the cut was started with heat. Sheets were cracked away with a red-hot iron: applying the hot iron to the edge of the sheet started a crack that could be guided more or less in the direction the iron was moved, allowing the glazier to break away a piece of approximately the right shape. This rough-cut piece was then refined to its exact outline with a second tool. A flat, grooved metal tool called a grozing iron was used to cut and shape glass throughout the medieval period, and glass cut with it shows a distinctive nibbled and chamfered edge — the tell-tale mark by which conservators can still identify medieval cutting technique on surviving fragments today, distinguishing it from the smooth edge left by a modern wheel cutter.

Colour alone could not render a face, drapery folds, or lettering; that fine linear detail was painted onto the glass surface afterward with a paint made for the purpose. Glass painters used a vitreous or glass paint that had to be fired to adhere to the glass, made of iron or copper oxide, ground glass, gum arabic, and a binder such as water, wine or vinegar, and the paint's colour varied from grey-black to brown and could be diluted to different textures, with a dense ink-like consistency used for trace-lines and a thinner wash for shading and modelling. Painters worked the design in stages, applying the paint with brushes of various animal hair and using sticks, needles, quills or fingers to scratch back the wet paint to add depth and control the light passing through. Once painted, each piece went into a kiln; the pieces were fired at a temperature not high enough to melt the glass itself but high enough to cause the glass paint to vitrify and fuse permanently to its surface. From around the year 1300 a second pigment entered the glazier's kit that changed what a single piece of glass could show. Silver stain, introduced to stained glass in the early 1300s, involved painting silver compounds such as silver nitrate, chloride or sulphate onto the glass and firing it in the kiln, whereupon the yellow stain actually became part of the glass itself rather than sitting fused to the surface like ordinary paint, and it could range from a pale lemon to a dark orange. Because it let a single piece of clear or blue glass carry both its base colour and a yellow highlight — hair, a crown, a halo — without cutting a separate piece and adding another lead line, silver stain is usually credited with giving the whole craft its English name: glass that had literally been "stained."

The last stage joined the painted, fired pieces into a single rigid panel. The pieces of glass were assembled on a flat surface with their edges fitted tightly together using H-profiled strips of lead known as cames, which formed part of the design, and the work was then soldered at every joint to fix it. A finished window was rarely one seamless sheet; it was built up as a series of leaded panels, and those panels needed external support to survive wind loading and their own weight over a large window opening. A shaped iron framework called an armature was inserted into the lancet window opening to provide support for the panels of stained glass, and such armatures were common until the late thirteenth century, after which stone tracery increasingly took over the structural role. The panels were tied to iron bars with wire within the window opening and soldered to hold the whole composition in place. This combination of glass, lead and iron was as much an engineering solution as an artistic one: the same lead lines that a modern viewer reads as outline and shading were structurally load-bearing, holding a wall of coloured glass in place against the weather for centuries.

Gives step-by-step instructions for glassmaking, colouring, cutting and painting written by a practising craftsman; it is the only complete treatise on the visual arts to survive from the High Middle Ages, so it cannot be cross-checked against a rival contemporary manual, only against surviving objects.

Confirms potash-lime-silica composition consistent with wood-ash glassmaking, though the specific claim that beech wood was the dominant fuel/flux source has been questioned by researchers such as Jackson and Smedley, who argue the archaeological picture is more regionally variable than the text alone suggests.

Directly shows the tool marks of hot-iron cutting and grozing and confirms the lead-and-iron assembly method, but cannot by itself date when a given technique first appeared.

Establishes dating and stylistic sequence for individual windows well, but the design and workshop-organisation stage of the process is comparatively under-documented, as conservation specialists themselves note.

Would settle itSystematic regional chemical surveys of ash residues and glass composition tied to identified wood sources across a wider sample of medieval glasshouse sites.

The beliefPeople often assume stained glass windows were painted with colour, the way a picture is painted on canvas.

The recordMost of the colour was built into the glass itself during melting, by adding metal oxides to the molten batch; surface painting was reserved chiefly for linear detail such as faces, drapery folds and lettering, applied with a vitreous paint that was then fired in a kiln.

The beliefA common assumption is that the term 'stained glass' refers to the overall colouring process.

The recordThe name is usually traced specifically to silver stain, the yellow-producing silver compound introduced to the craft in the early 1300s, whose name for one particular technique came to describe the whole art form.

Wikipedia's entry on Theophilus Presbyter and the Britannica entry on De diversis artibus both date the compilation to this window, based on the two oldest surviving manuscripts (Vienna, Codex 2527, and Wolfenbüttel, Cod. Guelf. Gud. Lat. 69).

Cited directly from De diversis artibus in a peer-reviewed materials-science study of medieval glass leaching (ResearchGate/Melcher and Schreiner-related analysis, 2024 publication date on the citing article).

Multiple specialist sources (the Corning Museum of Glass's Ask-a-Glass-Question service citing glass historian Meredith Lillich, and the Boppard Conservation Project run by Glasgow Museums) place its introduction to stained glass around 1300, while noting the underlying lustre technique existed earlier in Egyptian and Islamic glass.

The York Glaziers Trust's illustrated glossary of medieval glazing terms states this dating for the armature's structural role before stone tracery took over.

Figure given by the Norfolk & Norwich Stained Glass conservation resource describing the muff method; treated as an approximate, commonly cited figure rather than a single documented measurement from a primary medieval source.

Stated in the Southwell Minster fact sheet on stained glass production and corroborated by the Wikipedia entry on crown glass (window), which notes the technique passed from Normandy to England in the 1600s.

Jackson and Smedley's chapter 'Theophilus and the use of beech ash as a glassmaking alkali' reviews documentary and analytical evidence together and proposes a revised, more regionally nuanced history rather than treating the treatise's beech-ash preference as settled fact.

Theophilus Presbyter (trans. Hawthorne, J. G. and Smith, C. S.), On Divers Arts: The Foremost Medieval Treatise on Painting, Glassmaking, and Metalwork (c. 1120)primary source

The original medieval craftsman's manual behind almost every technical claim in the modern literature on how stained glass was made.

Theophilus Presbyter (ed. and trans. Dodwell, C. R.), De Diversis Artibus (The Various Arts) (1961)scholarly

The critical Latin edition with facing English translation and scholarly apparatus, considered the definitive text of the treatise.

Hayward, Jane, English and French Medieval Stained Glass in the Collection of the Metropolitan Museum of Art (2003)scholarly

A museum catalogue study grounding the treatise's techniques against a large body of surviving medieval panels, cited by the Metropolitan Museum's own essay on the subject.

Brown, Sarah and O'Connor, David, Glass-Painters (1991)accessible

A concise, illustrated introduction to the medieval glass-painting trade and its workshop organisation, aimed at a general reader.

Theophilus's De diversis artibus was compiled by an early twelfth-century Benedictine monk, probably between 1100 and 1140.

Wikipedia's entry on Theophilus Presbyter and the Britannica entry on De diversis artibus both date the compilation to this window, based on the two oldest surviving manuscripts (Vienna, Codex 2527, and Wolfenbüttel, Cod. Guelf. Gud. Lat. 69).

Theophilus's glass recipe calls for roughly one part washed sand to two parts beech wood ash.

Cited directly from De diversis artibus in a peer-reviewed materials-science study of medieval glass leaching (ResearchGate/Melcher and Schreiner-related analysis, 2024 publication date on the citing article).

Silver stain was introduced into stained-glass practice in the early 1300s.

Multiple specialist sources (the Corning Museum of Glass's Ask-a-Glass-Question service citing glass historian Meredith Lillich, and the Boppard Conservation Project run by Glasgow Museums) place its introduction to stained glass around 1300, while noting the underlying lustre technique existed earlier in Egyptian and Islamic glass.

Medieval iron window armatures were common only until the late thirteenth century.

The York Glaziers Trust's illustrated glossary of medieval glazing terms states this dating for the armature's structural role before stone tracery took over.

Muff-glass sheets were probably no larger than about 25 by 31 centimetres.

Figure given by the Norfolk & Norwich Stained Glass conservation resource describing the muff method; treated as an approximate, commonly cited figure rather than a single documented measurement from a primary medieval source.

Crown glass was not produced domestically in England until the seventeenth century, though imported Norman crown glass was used earlier.

Stated in the Southwell Minster fact sheet on stained glass production and corroborated by the Wikipedia entry on crown glass (window), which notes the technique passed from Normandy to England in the 1600s.

Whether beech ash was the universal or merely a regionally dominant alkali source for medieval glass is a live question among archaeological scientists.

Jackson and Smedley's chapter 'Theophilus and the use of beech ash as a glassmaking alkali' reviews documentary and analytical evidence together and proposes a revised, more regionally nuanced history rather than treating the treatise's beech-ash preference as settled fact.

Was beech wood ash genuinely the dominant, near-universal source of alkali for medieval northern European glassmaking, as Theophilus's treatise implies, or has this become an overstated orthodoxy read into the archaeological record?

The traditional view, following Theophilus's own preference and widely cited in archaeological literature, holds that beech ash was established as the first-choice alkali source across medieval northern Europe. vs Researchers including Caroline Jackson and James Smedley, reviewing documentary, environmental and analytical evidence together, argue the picture is more regionally variable and that the beech-ash-as-standard narrative needs revision rather than simple acceptance of the text.

How exactly did a medieval workshop translate an approved vidimus sketch into the full-size cartoon on the whitewashed table, and who performed that step?

The design stage between patron approval and physical cutting is the least documented part of the process, and conservators at the Victoria and Albert Museum state plainly that this transition is not well understood.

How many people typically worked in a medieval glazing workshop, and how were tasks such as glassmaking, cutting, painting and leading divided among them?

Without workshop account books or guild records surviving in quantity for most medieval glazing centres, the social organisation of production behind large cathedral programmes remains largely inferred from the finished windows themselves.