Journal

‹ Back

The work of tile making II - pattern and finishing

 

Medieval tile makers produced various different types of floor tiles, including regularly shaped — usually square — undecorated tiles, mosaic tiles (tiles cut into different shapes, each of a single colour, which could be laid in interlocking formations to produce larger scale patterns); plain relief tiles — where patterns were formed by impressing a stamp into the tile and leaving the surface in relief; and two-colour tiles — where a design was impressed into the tile and then filled with a different coloured clay to produce an flat tile face with an inlaid pattern. At Lossenham, the priory remains have yielded examples of the first and last of these, and it is the methods for producing the latter which I want to discuss here. The journal entry “Conversation with a medieval floor tile” provides a good example of one such tile.

 

As previously noted, while no documentary records exist of how such tiles were made, examination of surviving tiles suggests, in general terms at least, how the decoration process might have worked. Once the tile had been formed, but while the clay was still not too hard, the design — carved in relief on a wooden block — would have been pressed, or stamped, into the tile to produce an impression. This would then have been filled, proud of the tile face, with a white-firing clay. Once this clay had firmed up somewhat the tile face would need to be scraped back to remove the excess white clay and reveal the pattern inlaid into the tile ¹. However, as with tile forming process, lots of questions remain about the details. Was the tile stamped while still in the mould or later? How much later? Was the white clay applied as a paste, or in a more liquid state (slip)? When was the excess clay removed, and how? How was the a smooth tile face with sharply defined pattern achieved when using clay with grit and stones in it? If the edge bevel was cut was this after the addition of the white clay or before?

 

Whilst these all seem to be quite minor details, the effectiveness of each step is important to the production of good individual tiles, and the way the steps are combined needs to allow efficient production of tiles in quantity. Tile makers were often paid per thousand tiles (typically five to ten shillings) ² and, with a tile kiln perhaps taking between 4000-11000 tiles in a firing ³, there was a strong incentive for efficiency in the tile making process.

 

Following on from some thoughts on the tile-forming process previously discussed, the following summarises some of our experiments so far on the different approaches to pattern forming and finishing.

 

The first consideration is the nature of the stamp used to make the pattern impression. It is generally assumed that stamps were carved from wood - though no examples from the period have ever been found. The main clue to the material used is that in some medieval tile remains, where the white inlay clay has fallen out, the bottom of the impression shows marks which seem likely to be from wood grain. The way these marks align across multiple impressions in the tile, clearly suggests a stamp made from a single piece of wood.

Marks in the impressions on a medieval tile from Lossenham

 

While this does not necessarily mean all stamps were wood — other suggestions include metal (likely lead), fired clay and wax ⁴ — it seems likely that the availability of materials and the skills to carve them, and the durability of finished stamp made wood the favoured choice. The need for durability, when making many hundreds of tiles, would argue for a robust hardwood, and being able to carve finely detailed relief for a fine grained timber. It is often stated that beech would have been the preferred wood ⁵, being dense and hard, but fine grained, although oak might have been more readily available. We have made wooden tile stamps with both oak and beech, and the coarser grain of the former does make the fine detail relief on some tiles hard to carve — though in neither case is the process helped by our limited expertise in wood carving!

 

We have also found that while some designs can be carved entirely with standard woodworking gouges (albeit of a small gauge) for others with more finely detailed designs it is useful, sometimes necessary, to have finer tools. This has necessitated making some finer gouges ourselves ⁶.

 

Two early wood block stamps, both from oak

 

It is not clear who drew the original tile designs used by medieval tile makers, and whether the blocks would have been carved by the tile makers themselves, or by experienced wood carvers. Even if the latter, the finished stamps represented many hours of work ⁷ and, if paying another craftsman to make them, presumably a significant cost. In some cases particular marks, or flaws, in the impressions left on the tile allow particular stamps to be identified, and evidence from some of these suggests stamps were kept in use for long periods, sometimes at different locations, and mended as they became damaged ⁸. Having carved a number of blocks, it is not hard to understand why this would be the case.

 

When it comes to applying the stamp, in principle the tile could be stamped at any point until the clay is too hard for the stamp to make a sufficient impression. However, between the moulding and leather hard stage of drying (after which the clay is too hard to stamp) the tile undergoes some shrinkage (in the case of our clays, 3-4 mm is typical), so ideally the stamp should be made to the size of the tile at the stage when the pattern is applied ⁹.

 

We have experimented with stamping when the tile is still in the mould and, after removal from the mould, up to the soft leather stage (tile is firm, but edge can still just be deformed with the pad of a finger). A usable impression can be made at all stages, but when leather hard the stamp needs applying with considerable force — usually 2-3 hard blows with a wooden mallet. The disadvantage to later application is that the stamp can shift slightly and leave slightly displaced impressions, and with long term use it seems much more likely to result in damage to the stamp block. The advantage is that the clay is not sticky at this stage, so there is no issue with the stamp adhering to the clay and damaging the impression when being removed.

 

Applying the stamp when the clay is softer, either while still in the mould or shortly after demoulding, has the advantage that the stamp can be impressed with less force, even just by pressing it by hand, and the evenness of the impression is more easily controlled, but the disadvantage that sometimes the adhesion of the clay to the stamp can make clean removal of the stamp difficult. It is possible that the stamp surface was dusted with wood ash before each application, to reduce the adhesion, though we have not found this is always effective. What does seem to help is ensuring the edges of the relief carved pattern are smooth and slightly angled inward towards the top. When the clay is at this softer stage applying pressure to the stamp on the tile can result in some sideways spread of the clay in the upper part of the tile. Stamping the tile while still in the mould avoids this problem.

 

Tile stamped in the mould

 

Once the tile has been formed and stamped, then another question of timing is when to apply the white clay slip to fill the design impression. In theory this could be done at any stage from just after stamping right through to the tile being quite dry, but one important issue is matching the shrinkage of the tile body and the slip. By altering the time at which the white slip is applied, and its consistency (water content), in principle it would be possible to alter the relative shrinkage of the tile and slip. A medieval tile maker would no doubt have made any such adjustment by trial and error to accommodate the different clays available. However, in our experiments, at least so far, we have found that application of slip can be be successful at almost any stage of the process, the timing perhaps being less critical than the types of clay being used.

 

Whether the slip was applied as a liquid (about the consistency of cream), or as paste (about the consistency of thick yoghurt) is not known for sure. For the most part we have used the former, and found that in liquid form the slip fills the indentation without needing any working in, and it is possible to apply the slip quite economically so that relatively little needs to be scraped off, and potentially wasted. Thicker slip has to be applied with some sort of blade and worked into the pattern to ensure the indentations are properly filled, but this is perfectly possible ¹⁰. We have found the application of liquid slip, using a slip trailer constructed from cows horn and quill ¹¹, is perhaps slightly quicker ¹², although both processes are quite time-consuming.

 

Applying white clay slip to a tile using the cowhorn and quill trailer

 

Applying white clay as a paste

 

Once the slip has firmed up it is possible to carefully scrape off the excess slip to produce a flat tile face, with the pattern clearly defined.

 

Scraping the surface of a leather hard tile

 

A tile after scraping back to reveal the pattern

 

 

This all sounds very straightforward, and indeed in most writing on medieval tile making it gets no more explanation than that. In practice, however, there is a bit more to it. The state of dryness of the slip is important. In our experience it needs to be left to the point where it is firm (hard to dent with the pad of a finger), but still somewhat damp. At this point scraping the surface with a straight metal edge causes the slip to crumble off in fine shavings, without pulling out the clay below it. If the slip is left to get drier still — even to the point of the whole tile being bone dry — then scraping is still possible, but the clay comes off as a granular powder, and the scraping process is harder work. If the medieval tile makers worked this way, then the only tool required would be a straight-edged scraper — a metal edge being most effective and long lasting, though if scraping tiles while still damp a hard wood might also have worked. One point to note is that, with care, one can separate the initial scrapings, which are entirely composed of white clay, from the mixed scrapings that result from the later stages of the process. Recycling the pure white clay scrapings would have made the best use of the rarer white clay.

 

One significant issue with the finishing process which rarely gets discussed is that when using clay that has had little processing, and may have grit and other larger impurities in, the scraping process almost inevitably results in pieces of grit being dragged across the tile surface by the scraper. This can create noticeable scoring of the tile surface and, more importantly, can drag clay from the tile body into the areas of white clay, creating streaks of red or brown in the pattern.

 

A tile showing coloured score marks in the pattern

 

This is rarely observed on medieval tiles, and indeed is not a significant problem when working with manufactured and processed clays which are grit free. Although the problem can be minimised if sufficient care and time is taken in the scraping process, this obviously adds to the time taken to make a tile, and it is interesting to consider whether there are other solutions. One approach we have tried is to rub the tile surface on a wooden board (or piece of leather) flooded with water to abrade away the slip and produce a flat tile face. It is a tricky process, and has to be done swiftly to avoid overwetting the tile, but it seems to have potential with such clays ¹³.

 

The final stage of the making process is glazing the tile. It is well established that potters and tile makers of this period, used simple lead-based glazes ¹⁴, which produce a transparent glossy glaze, sometimes tinted yellow with iron or green with copper, and flux well at the relatively low temperatures used for tile making. We use a glaze made from a lead frit (lead bisilicate) mixed with a small proportion of china clay. Lead bisilicate alone has a tendency to craze (develop very fine cracks in the glaze as it cools after firing), and the china clay alters the shrinkage of the glaze reducing that tendency. I don’t know whether medieval tile makers would have worried about crazing in their tiles — probably not. I have seen medieval tiles with crazing in the glaze; it would not be visible in a floor unless you got down on hands and knees and looked very closely, and it doesn’t materially affect the functionality of the tile.

 

Application of lead glaze to a dry tile

 

The experimental tiles we have made so far have all been fired in an electric kiln (it is useful to be able to replicate firings, and fire small batches), but the real test will be to fire a batch in the wood kiln. So far we have completed about ninety experimental tiles, using different combinations of clays and the variations in process outlined here and the in the previous journal entry (we calculate there are at least 288 different permutations of the possible processes — but we are not doing them all!).

 

In addition to the processes already mentioned, we have also experimented a little with two other ideas that have been suggested: ‘printing’ wet slip onto tiles ¹⁵, and the ‘stamp through slip’ method ¹⁶. With the former, which seemed rather implausible to start with, we have had no success at all. The latter does show some promise, but is sensitive to the stage at which the stamping is done and needs some more experimental work to really test its potential.

 

A selection of the experimental tiles

 

From all these experiments, we need to settle on a suitable clay, and a set of methods which seem to work most reliably to make a tile and, importantly, which fit together to allow an efficient production process. When we’ve made the first thousand we might begin to have a glimpse of what the medieval tile making was all about!

 

 

Notes

1. Eames, E., 1980. Medieval lead-glazed earthenware tiles in the Department of Medieval and later Antiquities, British Museum; British Museum Publications, British Museum, London; Eames, E., 1985. English medieval tiles. British Museum Publications, London; Van Lemmen, H., 2000. Medieval tiles. Shire Publications.

2. Lewis, J.M., 1999. The medieval tiles of Wales. National Museums and Galleries of Wales.

3. Estimate of 5000 from Green (Green, M. 2003. Medieval Penn floor tiles. Miles Green.); Drury and Pratt estimate around 11,000 tiles for the Danbury tile kiln (Drury, P.J., Pratt, G.D., 1975. A late 13th and early 14th century tile factory at Danbury, Essex. Medieval Archaeology 19, 92–164.)

4. Eames, E., 1980. op. cit.; Barbier, M., 1999. La conception des carreaux du moyen âge et de la Renaissance, in: Mayer, J. (Ed.), Pavement: Carreaux de Sol En Champagne Au Moyen Âge et à La Renaissance. Centre de Recherches sur les Monuments Historique, Paris.

5. Beech was well established in Britain by the medieval period (https://www.forestryengland.uk/article/beech); although naturally occurring in the south-eastern and central southerly regions, it was probably planted elsewhere.

6. We are grateful to John Winterbottom who makes wood stamps used by Diana Hall, for suggestions and advice.

7. In an experimental example, Barbier (1999, op. cit.) estimates about 15h.

8. Eames, E., 1980. op. cit.

9. I have not discussed here the idea that the tile might be made oversized, then stamped and the excess trimmed off using the stamp as a template — although we have made a few tiles this way and it is a feasible method. However the trimming adds another step to the process and produces clay waste that needs recycling.

10. Diana Hall uses this approach very successfully.

11. I am not sure what medieval tilers would have used for this purpose (assuming this was how they applied slip), but the cow’s horn / quill trailer (inspired by an example from the Company of Artisans) works very well, and uses materials they would have had available.

12. For a moderately complex tile design it takes me about two and a half minutes to apply the liquid slip.

13. After starting to experiment with this approach we came across a similar experiment in Barbier, M., 1999. op. cit.

14. Eames, E. 1980, op. cit.; McCarthy, M.R., Brooks, C.M., 1988. Medieval pottery in Britain AD900-1600. Leicester University Press; Van Lemmen, H., 2000, op. cit..

15. Haberly, L., 1937. Mediaeval English pavingtiles. Blackwell, Oxford.

16. This consists of putting a white slip layer on a plain tile, then stamping the design when the slip and tile are leather hard to drive a layer of slip into the tile (Eames, E., 1980. op. cit..).