Friday 11 December 2020

Meddling with madder - part 1

I have been quite intrigued by the red coloured medieval tablemen that are found in the different museum collections around the world (for instance the Metropolitan Museum of Art, the Victoria and Albert Museum, the Museum für Angewandte Kunst Köln, among others). How were these pieces coloured and is this easy to reproduce? The how was easily answered, as the monk Theophilus provides a recipe in his book 'On divers arts' which was written around 1122 - a similar date as most of the red game pieces, which were manufactured in workshops in Cologne, Germany.

A game piece with Samson slaying the Philistines with the jawbone of an ass. Madder coloured walrus ivory. 6.3 by 1.3 cm. Made in Germany around 1140-1150. Photo Metropolitan Museum of Art, Cloisters, New York, NY, USA.
Theophilus writes in Chapter 94 of his book:

'There is also a plant called madder, whose root is long, thin, and reddish. After it is dug up, it is dried in the sun and pounded in a mortar with a ball. Then lye is poured over it and it is cooked in a raw pot. When it has boiled well, if the bone of an elephant or a fish or a stag is put in it, it will become red. from these bones or horns knops can also be made on the lathe for the staves of bishops and smaller knops for various useful objects. When you have turned these with sharp tools, smooth them with shavegrass. Collect the shavings on a linen cloth and, still turning the lathe, rub them vigourously on the knops which will then become completely shining. You can also polish horn-handles, huntsmen's horns, and [horn] windows in lanterns with sifted ashes on a woolen cloth. But do not forget to smear them finally with walnut oil.'

Of course the bone of an elephant is ivory, the bone of a fish relate to normal animal bone, and the bone of a stag is antler. I am unsure if Theophilus also means that (cow, goat) horns can be stained, as these have a quite different structure more similar to nails and hairs. However, wool for medieval clothing is commonly stained with madder. There is evidence that madder was already cultivated for this purpose in the Netherlands in the 12th century; in later centuries the Dutch madder became famous for its quality.

A similar uncoloured game piece with Hercules throwing Diomedes to his man-eating horses. Elephant ivory. Made in Cologne, Germany around 1150. Diameter 7 by 2 cm. Photo Metropolitan Museum of Art, Cloisters, New York, NY, USA.

As I was making a set of alquerque game pieces from antler, I wanted to make one set coloured red. The recipe of Theophilus was not very specific. In fact it looked more like a medieval cooking recipe without amounts. Searching on internet did show many recipes for colouring with madder, but these all concerned wool or cotton, and all are different as regards to amounts of the ingredients. 
A mordant - alum - is needed to fix the red colour to wool. The wool is first treated with the mordant, and then to the colouring solution. For cotton and hemp, both non-animal (plant) fabrics, an extra pretreatment is necessary before the alum. On the other hand Theophilus, as well as the early 18th century scientists do not mention alum at all for bone colouring. 

There are also other steps in the colouring process that need to be considered: temperature influences the colour; a temperature above 82 degrees Celcius turns the red colour into brown (for wool). Hard water (basic pH) increases the intensity of the red colour. The time in the colour solution and the strength of the colour solution influence the outcome as well: the longer, the darker, and more pigments in the solution also make the end result a darker red.

So what is the best way to colour antler madder red? I decided to do some tests to find out.

The first test

Madder powder as well as madder extract was bought from paint mill 'De Kat' in the Netherlands. Madder extract has already the red pigments (alizarin and purpurin) from the root extracted, and is sold as a plaque of dried crystals, which needs to be dissolved again. The madder powder is finely ground madder root, from which the madder pigments still need to be extracted before colouring the antler (or wool). I also purchased some alum from 'De Kat'.

As a basis I used 'the Maiwa guide to natural dyes - what they are and how to use them'. This free pdf guide (provided by the Maiwa company which also sells these pigments) provides a very clear description on colouring for each dye and as well as for mordanting. 
Left: The two basic madder preparations from powder and extract in the 60 degrees Celcius water bath. Right: the setup in the kitchen with the meat thermometer at 60 degrees.

The amount of dyestuff needed is based on a percentage of the weight of fibre (WOF), in my case the weight of the antler pieces. For madder powder this is 35-100% WOF, for madder extract this is 3-8% WOF. The antler pieces should be just fully immersed in the solution. So the actual amount of water does not matter, the amount of pigment that is available to the fibre does. I used 50% WOF for the powder and 4% WOF of extract.

Left: Pretreated antler pieces: with water in the glass, or with alum in the plastic box. Right: Four smaller test pots in the 60 degrees Celcius water bath.
Half of the antler test pieces were pretreated with alum, 15% WOF (as for wool). The other half was only immersed in water. During mordanting, the solution with the antler pieces was kept around 60 degrees Celcius for one hour. The pots with the solution were heated au-bain-marie (a waterbath) in a pan on a low flame and the temperature was kept in check using a digital meat-thermometer. If the temperature started to rise, a bit of cold water was added to the pan. After one hour, the pots were left to cool.

I made 400 ml of each madder solution. The pots with the madder solution were placed at 60 degrees Celcius for one hour in the waterbath, as described above. During the hour the colour of the powdered solution darkened, the extract solution was quickly dissolved and had a similar dark red colour.
Then the pretreated antler pieces were divided over smaller pots and the colouring solutions added. To the pots with antler pieces without alum an extra 2 gram of household sodium was added to each madder solution. The sodium carbonate was used to raise the pH of the solution to see if this influences the intensity of the colour. The test consisted of five pots:
  • Antler mordanted with alum, madder powder solution, with sodium added during cool-down
  • Antler mordanted with alum, madder powder solution, no sodium
  • Antler mordanted with alum, madder extract solution, with sodium
  • Antler with water, madder powder solution, no sodium
  • Antler with water, madder extract solution, with sodium

The pots were kept around 60 degrees Celcius in a waterbath for 2 hours, after which they were left to cool down for 24 hours. Then the antler pieces were rinsed with cold water and dried.

Colour after 2 h 60 C degrees bath. The darker red coloured pieces on the right are without alum  pretreatment and with sodium.

The colour after 24 h cool down with the same oder of test pieces. The piece on the far left (powder + S + alum) was treated with sodium during the cool-down period and gained in colour strength.

As it turned out, mordanting with alum did have no positive effect on the colouring process for antler. Sodium, on the other hand, did. All antler pieces in a solution containing sodium were dark red, the one with sodium added later medium red, while those without were only lightly red coloured. Both the madder extract and the madder powder did colour equally well. Madder powder is cheaper than madder extract, but madder extract is a clearer solution and does not have the risk of the powder grains to become affixed to the antler. So the extract was used for further experimentation (in the next blogpost).

A side note

Science also had an early interest in colouring bones red. The first scientific study on the effects of madder on bones is from the early 18 century by M. Du Hamel du Monceau and published in the Philosophical Transactions of the Royal Society of London. An observation from a surgeon eating a pig with red bones, led to experiments where pure ground madder was (forcibly) fed to chickens, which died after a few days. Indeed, the chicken bones had become all red, but not the feathers.
Nowadays the synthetic red pigment (alizarin) that is also found in madder is still used to study the growth of bones and their calcium deposits.

Diaphonisated Smooth newt (Triturus vulgaris). Short-term xylene treatment. Cartilage area (white arrows), ossified structures (black arrows). Bar = 5 mm.
Smooth newt (Triturus vulgaris) with bone coloured with alizarin. Bar = 5 mm.


  • G.C.H. Derksen (2001) Red, redder, madder. Analysis and isolation of anthraquinones from madder roots (Rubia tinctorum). PhD thesis, Wageningen University, Wageningen the Netherlands.
  • M. Du Hamel du Monceau (1739) Observations and experiments with madder-root, which has the faculty of tinging the bones of living animals of a red colour, by M. Du Hamel du Monceau, F.R.S. & c. Communicated in a letter to Sir Hans Sloane, Bart. Pr. R.S. Translated from the French by T.S. M.D.F.R.S. The Royal Society of London Philosophical Transactions Vol. 41, pp. 390-406.
  • Antje Kluge-Pinsker  (1991) Schach und Trictrac. Zeugnisse mittelalterlicher Spielfreunde in salischer Zeit. Jan Thorbecke, Germany.
  • Vivian B. Mann (1977) Romanesque ivory tablemen. PhD thesis, New York University, New York, USA.
  • Theophilus (1122) On diverse arts. Translated by J.G. Hawthorne and C.S. Smith (1979) Dover Publishing, Garden City, NY, USA.
  • The Maiwa guide to natural dyes - what they are and how to use them. 

Sunday 29 November 2020

A curious medieval chess/backgammon board


A game of "strip"-backgammon in manuscript. Welscher Gast by Thomasin von Zerklaere. Around 1256, Bayern, Germany. Codex Pal germ 389. Folio 11v . University Library, Heidelberg, Germany.

Actually this post is not about the unusual variant of the medieval backgammon or tric-trac game where the players strip during play as shown by these two illuminations. So far no rules have been published on this variant, although more illustrations of this type of play exist (a Roman glass and other copies of  the 'Welscher Gast'. Presumably, one could lose a piece of garment if a game piece is taken by the opponent.


A game of "strip"-backgammon in manuscript. Welscher Gast by Thomasin von Zerklaere,
Cod. Memb. I 120, folio 12v. Around 1340. Forschungsbibliothek, Gotha, Germany

  The 15th century chess and backgammon board in the Museo de Leon, Leon, Spain.  

I encountered an image of a 15th century chess and backgammon board from the Museo de Leon in Leon, Spain in a book that was curiously constructed. The board looks to be made of wood, with an extra rim on the sides. One side contains the backgammon board with 12 inlaid triangles on each side. The rim is smooth. The other side contains a 8 by 8 squares chess board, but here two of the rims have twelve half-round holes. They are decorative and do not have a function on the chess board. Hoever the half-round holes are reminiscent of the Spanish backgammon boards in Le Juegos of Alphonso X 'the wise' of 1283. The twelve half-round holes correspond to the twelve triangles of the backgammon board, and the backgammon game pieces will snugly fit in the holes. I think it is likely that the person who made the game board made a mistake during its construction and put the rim for the backgammon board on the wrong side.

The game of seis, dos, y as from the Libro de los juegos of Alphonso X the Wise. 1283. Folio 75 verso.
Note the board with the half-round holes to hold the game pieces.

Sunday 11 October 2020

The backside story

Normally, when you visit a castle or a museum and look at the furniture, you will only see the front and sides. Only rarely you will see a glimpse of the backside of a chest or an armoire, or equally rare, the inside of the storage compartments of the said furniture pieces. I was therefore happy to discover that some of the medieval furniture on sale at the Prunier Auction of 11 October 2020 in France also did show the backsides. Aside from these auction pieces, I will also show some medieval furniture where I was able to examine the backsides myself.   


The front and back of a 15th century high-backed bench-chest. The back is crudely finished. Interestingly, the vertical stile on the back of the chest is placed asymmetrically to prevent compromising the strength of the horizontal stile. The suggested small horizontal stiles on frontside of the panels of the backrest are in fact part of the panels themselves. height 143.5 cm, length 120 cm, depth 48 cm. Sold for 2100 Euro.

The front and inside of a French 15th century armoire with linenfold panels. The inside also allows us to see the construction of the back of the armoire, which has a similar frame construction as the doors and sides (but undecorated panels, as the sides). The armoire had three shelves, of which one is remaining. You can see two holes where the horizontal supports for the missing shelves were placed. Also a slight discolouration on the backside shows where the shelves used to be. height 191 cm, length 128.5 cm , depth 52 cm. Sold for 6500 Euro.

While the front of this early 16th century oak south German marriage chest is highly decorated, the back is not. Contrary to the frame construction on the front it consist of several slats of wood fixed to the sides with dovetails. Also the hinges are fixed on the back. height 99 cm, length 174 cm, depth 56 cm. Sold at 2000 Euro.

 This late 15th century oak chest is similarly to the previous one, with a highly decorated front and sides, and an undecorated back. The backside consist of one single plank. The hinges are simple, consisting of hooked metal rods. height 78.5 cm, length 154 cm, depth 59 cm. Sold for 1100 Euro.
One of a set of two chests dated around 1500. At the back several repairs have been made, probably including the small low horizontal plank. Likely the backside earlier consisted of one wooden plank, fixed with dovetails to the sides. The construction of the underside, together with the decorative rail at the bottom is typical, suggesting that it is somewhere nailed. Height 68 cm, length 140 cm, depth 59 cm. Both sold for 800 Euro each.
Armoire R.B.K. 1954-7 in Castle Muiderslot, Muiden, the Netherlands dates from the 15th century. The chest is made of two parts that are added on top of one another. During a study, the top was removed, making hidden details visible.
Left: the top part of the armoire. Right: The lower part of the armoire is still in use. A historical spoon was found in it.
Not all restorations have been done according to the latest views. 
Here a wooden block was screwed on the inside to hold the door.
The rail holding the top in place can easily be seen now on the bottom part of the armoire. 
The rail is fixed to the bottom with wooden pins (easily seen on the left photo). 

The top panels are fixed onto the upper part of the dressoir with wooden pins.

At one end the panels are chamfered so they fit into the groove of the next one.

A beautiful armoire from Chateau Bois Orcan, Noyal-sur-Vilaine, France, with four doors and two drawers.
Left: The triangular front panel is nicely carved, but when I looked at the back and beneath the armoire (Right) some interesting thing could be seen: For instance the side triangular panel has some carving on the inside. Apparently the intended carving went wrong and was never completed. Instead, they re-used it by reversing the panel (black arrow). The back panels of the armoire are roughly sawn off and layered with a half lap joint (blue arrow). At the edge stile the groove continues into the legs (red arrow) and the back panel is secured in the groove with a wooden pin (green arrow).

Friday 28 August 2020

Vatican courier chess, part 2 - the board

Just some note before this post really starts. Blogger did change its layout lately, to make it more compatible with mobile phones. This had the side-effect of making blogposts much harder for me, and not being able to present the contents in an agreeable way. I even began thinking to quit the blog. Blogger mentioned that there was a revert option to the old style, but this was nowhere to be found. At the last day when you could revert I decided to check again and the button was there. I am happy to continue this blog. By the way, this blog is best viewed on a computer; you are able to see the details on the photos much better than on a mobile phone. Blogger should understand that not all blogs are written for phones.

The chess board (on the other side) also being a box for the chess pieces.

With the chess pieces complete, it is time to give attention to the making of the board. I decided to make the board into a box, where the chess pieces could safely be stored in. I also prefer light game boards (reason most of our boards are made from poplar), and especially for a relatively heavy chess set, a light box is preferable. So again poplar.

 One half of the board composed of smaller poplar strips. 
The pencil lines show the board squares which align with the glue lines.  

I had some poplar left drying in the workshop. Poplar that was not dry enough when I used it to make a hnfatafl board (the round glass game pieces did not stay put but rolled everywhere on the warped board). I did expect the wood to be dry now, after two extra years in the shop, but to avoid trouble I cut it up into smaller pieces and alternately glued them back together. These smaller pieces had the same with as the squares on the board would be, thus the glue lines would become unnoticeable when the lines were carved.

The half hidden dovetails on the side of the board and the dovetails on the sides of the box.

 The dovetails (test-)fitted together for the box.

Testing if the height of the box is ok for the chess pieces.

Two half game board pieces were made for 6 x 8 squares each. One square approximately 4 cm2, slightly larger than the chess pieces. On three sides an additional edge was added. The sides of the box were connected to the gameboard by half-hidden dovetails. This would also provide extra strength in case of warping. Inside the box, four compartments were made: two small ones for the kings and queens, and two large ones for the remainder of the pieces. The compartments were divided by a thin strip of poplar that was locked inside a groove in the backside of the game board and the sides of the box. The grooves for the squares of the chess board were carved with a carving knife using a metal strip for guidance of the lines.

Left: Glueing the boxparts. Middle and Right: the complete box with and without the chess pieces.

The adapted hinges for the board.

Heating the hinge with a gas torch (left). Note the metal wire that is attached to the hinge 
which was used to quickly drop the red-hot hinge in the linseed oil (right).

Space was chiselled out for the hinges. Two commercially available semi-medieval hinges were bought, but they proved to be too large for the side of the box. These hinges were adapted to a smaller and correct medieval type using cutters, files and a belt sander. Afterwards, the hinges were blacked by heating it with a gas torch and quickly quenching it in linseed oil.

The hare glue (grains) first had to absorb water before they were heated in a glue pot.

The chess board was intended to be luxurious as well, thus half of the squares were gilded, and the other half painted lapis lazuli (I used synthetic lapis lazuli, as the real pigment is far more expensive than the 23.75/24 kt leaf gold used for gilding). An example of such a blue and gold chess board can be found in an illumination in manuscript MS G 24 in the Morgan Library in New York.

Phesonas and Cassiel playing chess on a blue and gold chequered board. Manuscript of Jacques de Longuyon‘s poem 'Les Voeux du Paon', around 1350. Morgan Library, New York, NY, USA, manuscript MS G 24. Folio 025v.

First, the chess box was twice completely covered in hare glue. Then the board and the sides were thrice painted with gesso. For the squares to be gilded this was followed by 3 layers of red bole. Afterwards the bole was polished with steel wool. A gold glue of 3 hours was used to glue the gold leaf to the board. Coincidentally, one gold leaf did just cover a square. Each square was at least gilded twice to provide a stable gold layer.

The board being coated on all sides with hare glue, then several layers of gesso were applied to the sides and top.

  The redbole applied to half of the squares.

Left: Polishing the red bole surface with steel wool. Right: the polished board.

First gold layer added.

Applying gold leaf to the chess board is relatively easy as shown in this video. 
The gold leaf is slightly larger than the square. A second gold layer is being added here.


But then I ran out of gold leaf, and the waiting time for the Siamese gold leaf to arrive was a few months long. It became very warm and dry weather, and despite all the precautions the wood started to warp. So much that the hinges did not fit any more and I had to add some extra wood to the sides to create a flat area again at the hinges side. Furthermore, because of the extra millimetres of wood added, the space for the hinges also had to move a little. The warping on the other side (with all the gilded squares) still exist, but for playing chess this does not matter.

 This photo clearly shows the extra millimetres needed to create a flat surface again.

Left photo: First, the gesso had to be removed from the edges of the box. Then two strips were glued onto the short sides of the box and planed until they were even with the middle of the long side (see also the photo above and below). This was checked with a steel ruler. Right photo: Next, two strips were glued on the long sides. A ruler was used to mark the parts that needed to be planed away to create a flat surface.

The new long side strip being glued. A piece of cloth was used to protect the gold layer on the underside during clamping.

Left: the new flat surface was checked again with a steel ruler. Right: The new flat surface for the hinges of the box. Around 4 mm was added to each side.

Left: Also the space for the hinge needed to move a few millimetres. The remaining space was filled with pieces of wood precisely bended. Temporary nails were used to keep the strips in place during glueing. Right: The wood strips were bend with the heat of a soldering iron clamped to a workbench.

When the box warping was fixed, I still was awaiting the gold leaf. To speed up the process I then decided that I would paint the box first and then apply the second gold layer. I added the black (bone black) and red (Venetian red) colour for the sides, and the lapis lazuli (synthetic ultramarine blue) for the squares. However the drying time of oil-paint can be long, depending on the pigment-oil mixture and the temperature. Now I had the gold leaf, but the paint was still wet and I was unable to finish the gilding. (Actually, I had to order a new set of gold leaf, as the Siamese was of the wrong size. This third package, 24 kt gold, arrived luckily within a week!)

Left: First the sides of the box were painted black and red. When the inside rim was painted, the box was turned and placed on a book - resting on the poplar dividers, so the remaining side and top could be painted. Right: After the back and red, the squares were painted in synthetic lapis lazuli.

At one moment I thought the paint was dry enough to do the gilding. I miscalculated that gold leaf itself can be quite sticky and brushing away the superfluous gold caused it to adhere to the lapis lazuli blue squares. Actually, many lapis lazuli stones have natural gold flakes in them, however lapis lazuli paint does not and should be a bright full ultramarine colour. I had to scratch the gold flakes away with a scalpel from the lapis squares (as well as from the red and black sides). This left all kinds of scratches on the surface, so I had to repaint them.

The second layer of gold being added. The 24 kt gold leaf is much more shiny that the 23.75 kt. Gold leaf does not always stick well to the glue. Sometimes edges are loose and the gold is brushed away. Then a third or even a fourth layer of gold leaf was necessary. These squares needing an extra layer are pointed here with small wooden markers.

Left: Brushing away the excess gold caused the gold flakes to firmly adhere onto the lapis lazuli paint. Right: The gold flakes scraped away, making a second lapis lazuli layer necessary.

Painting the second lapis lazuli layer. You can see the difference between the two board parts: 
the bottom is painted, the top still has the scratched surface.

To avoid the two sides of the board from sticking to each other the lid of small plastic box was used.

Drying the repainted sides of the box, using books and the plastic box lid to stabilize.

Finishing up

Now it was time to fix the hinges with nails as well as add some nails for the 'locking' mechanism. A steel bench block was used to keep the nails against the hinges, while flattening the nails. The inside of the box still was plain poplar wood covered in two layers of hare glue. To protect the chess pieces, I glued a layer of thin red felt on the inside, also with hare glue. A spatula was used to push the felt into the corners of the box.
Left: The nails were bent at the point, and then hammered flat into the wood. Right: a steel bench block - with a piece of wood to set the bench block at the correct height - was used to keep the nails against the hinge while hammering.

The finished hinge.

Left:  Measuring and cutting the felt. Right: Hare glue was added one step at a time 
and the felt was pressed into the corners using a spatula.

The remaining felt pieces were finally cut off with a scizor.

With the box ready, the only the rope and tassel needed to be added for the closing mechanism. The chess board shown in the Vatican Cessolis book used for the chess pieces also has a rope added. Thus a closing mechanism based on a rope and tassel seemed appropriate. The 'rope' was made by de Gelderse Roos in black and red embroidery silk using a miniature rope making machine. There are easy guides on YouTube on how to make a tassel, and I also made use of them.

 The board in the Cessolis manuscript (folio 95r), the rope and tassel marked by the green oval.

The tassel and rope in the red and black colour.

The closed box with the rope alternating between the nails on the box sides.

The complete game

Left: The side of the box showing the hinges. Right: The inside of the box.

The finished courier chess set and board, with the set-up (actually the queen has to stand on the third and sixth row behind the pawn, see the post on playing courier chess).

In the inventory of 1447 of Adolphus, Duke of Guelders  a chess board is mentioned:

Item dat schaeckbret [chess board] mit schaek [chess pieces] ende wortafelspiel [backgammon] as half golt ende silver [half in gold and silver].

Evidently, this is a dual purpose board for both chess and backgammon. This is a quite common combination, also often found with medieval boardgame boxes. Later in the inventory the purchase of a bag for this chess set is mentioned (to protect the board and or chess pieces). The courier chess set is ideally suited for a dual board game purpose as the number of squares on the long side of the board is 12, the same number needed for backgammon. The only extra's needed are a set of dice and three more pawns of each colour (i.e. 15 of each colour). Backgammon then is played only using the pawns as playing pieces.

Two dual purpose boardgame boxes for normal chess and backgammon that were at the auction of 14 July 2020 of Prunier Auction, France. Lot 16: A 15th century certosina-decorated Embriachi box from Venice, Italy. Height 6.3 cm, length 15.7, width 19 cm. Lot 17: Late 15th century Spanish game box. The chess board is on the other side. Height: 4.7 cm, length 25.4 cm, width 17 cm. The lots were sold at 4500 and 3800 Euro (ex auction costs), respectively. Photo copyright Prunier Auction.

The set-up of the courier chess board for backgammon. The folding line of the box neatly divides the board into 4 quarters.

The number of game pieces in courier chess also make it possible to use them for other board games. For instance alquerque needs 12 pawns of each colour, nine men morris needs 9 of each colour, and fox and geese (13 + 2 pawns). The type of game pieces also make it suitable for normal (8x8) and decimal (10x10) chess. It is then only necessary to provide other game boards for these games. 

A three men morris board - needing 3 + 3 pawns for play. 
This small game board was used as a test piece for the courier chess board.