Firing the wood kiln IV (tiles)
Rather late in the year, for a variety of reasons, but a few days of dry weather with a bit of sun and modest breeze, gave the opportunity to fire the wood kiln with some of the 250 or so decorated floor tiles made this year. This presented some new challenges.
Firstly, the kiln was really made for pots, not tiles. Tile kilns of the period tended to be rectangular and wider/longer than they were high; both of which make packing the kiln with tiles much more practical. In addition, our kiln has quite wide spaces between the arches that form the floor of the firing chamber, and these were too wide to allow tiles reliably to span the gaps. We couldn’t do much about the shape of the kiln, but it was possible to build a new, more suitable, firing chamber floor, using smaller bars laid and levelled in a bed of cob, on top of the original arches. This was done a couple of weeks before the firing using well dried, but unfired, cob bars made earlier in the year.
The new firing chamber floor, after the firing
The next challenge was to get the hang of stacking tiles in the kiln, using the methods which seems likely to have been used by medieval tile makers. The problem is that the glazed faces of the tiles cannot touch, otherwise after the glaze melted and re-solidified the tiles would end up fused together and, because there are no kiln shelves the tiles cannot simply be laid flat on the floor of the firing chamber as this would preclude using anything beyond a single layer. So, the tiles have to stand on their edges, and be able to be stacked more than one layer high. This is complicated by the fact that they have bevelled edges so do not stand up on edge by themselves. An interesting challenge! Fortunately we did not have to work this method out for ourselves thanks to some clever work by Oliver Kent and David Dawson1, working with tiles from Cleeve Abbey, who used the marks that the edge of one tile leaves on the edge of another when they are stacked to work out what the stacking arrangement might have been. Using the system they devised, on the new kiln chamber floor, along with a little bit of improvisation around the edges where our oval kiln did not lend itself to a regular stacking pattern, it proved possible to get between 50 and 60 tiles in a single layer, and to stack a second layer on top of the first. Tiles were glazed at the kiln, in batches of 10-15, and then immediately loaded into the kiln to minimise handling.
The first layer of tiles (with the start of the second - top left)
The second layer of tiles
A third layer would almost certainly have been possible, but we did not want to commit all the tiles we had to a single firing. Instead, we put a layer of cooking pots, upside down, on top of the second layer of tiles. We then topped off the load with pieces of already fired broken pottery, to provide heat retention to the top of the pack.
The kiln packed with two layers of tiles, one layer of pots, and topped with fired broken pottery.
As the preceding couple of weeks had seen quite a lot of rain, small fires were burned for a few hours in both fire boxes once the kiln had been packed to help ensure both kiln and contents were dry. As with previous firings, the top level of the kiln was raised by the addition of three rounds of bricks, with clay cob filling the corner joints, to increase the draw through the kiln.
The kiln was lit at 6 am the following morning, on a day that proved ideal for firing: dry and mild with a gentle breeze.
The kiln shortly after lighting
We followed much the same firing pattern as in the later previous firings: 3 hours or so at around 100°C, then a steady (or sometimes not so steady!) rise at about 80-90°C per hour to around 900-1000°C, then relying on the cones to guide us to our target temperature of about 1020°C.
10:30 am, clean burning fires, about 300°C
5:00 pm - orange colour in the heart of the kiln - about 850°C
About there: first three pyrometric cones down, and guard cone starting to bend - 7:30 pm
In practice, we are finding that the kiln has a fairly predictable dynamic, with moderate stoking and a gradual shift from larger, slower burning, wood to smaller pieces, it seems to follow much the same pattern each time. Some of our wood was a bit damp from recent rain and, once it was dark - by about 4:30 pm - it became more difficult to judge the colour of the smoke, which can be helpful in judging when over-stoking results in periods of reduction but, otherwise, there were no obvious problems, and no unnerving sounds of things collapsing inside the kiln! Once the 1012°C cone had fully dropped and the guard cone (1030°C) was starting to bend, we stopped stoking and, having let the fires settle a little, the fire box entrances were roughly bricked up and the top of the kiln covered by about 8 pm.
Although the kiln had cooled to about 150°C by the next morning, because of the limited daylight available later in the afternoon, and the dry weather promising to hold, we left opening it until the following day.
Closed up kiln the morning after firing
Unpacking revealed a very encouraging, and interesting set of results. All the pots were well fired, though one was not as round as it had been when it went in! Most importantly, the tile stack had held up well, with nothing shifting or collapsing. In addition, as intended with the stacking method, there were no places where tiles had stuck (even in minor ways) to each other, or to the kiln.
Tile stack after removal of the capping layer, and the cooking pots.
One tile had a stone in, which had fractured, breaking off a piece of the tile face, and there were a handful others with what might be regarded as faults, though in most cases they would probably still have been usable. One example is that a few tiles had slight warping. It seems likely that this was a result of uneven temperature exposure on the front and back of the tile, as it is not something that has occurred in the electric kiln, using tiles made in exactly the same way.
Overall, there was an encouraging success rate. And, of course, at some point the minor flaws and variation in the tiles, become ‘character’ capturing the variability of the processes the tiles have been through, and adding to the visual interest rather than detracting from it.
There are also, of course, interesting points in terms of understanding the sources of some of the variation, and how this relates to what is sometimes seen in medieval tiles. The most obvious of these is the colour variation, with tile body colours ranging from orange/pink to more or less black, sometimes within a single tile. It is the tile body colour that varies, not the glaze or the white slip. All the tiles in the firing were made in the same way, using the same clay, and glazed with the same glaze.
The explanation for this variation is almost certainly that we must have had, late in the firing, a period in which there was a reducing atmosphere in much of the kiln. The clearest indicator of this is that the pots, which under oxidation fire to a bright terracotta colour, have largely come out black.
Cooking pot, showing darkening caused by reduction
The tiles are clearly also responding with the same darkening of colour in the clay body depending on how they have been exposed to oxidising or reducing conditions. There is no clear relationship between the colour change and the part of the kiln a tile was in, but individual tiles which happened to be standing in such a way as to be protected from the direct flame path do seem to have avoided the darkening effect.
The colour variation certainly adds character and visual interest to the tiles, and it corresponds to effects we have seen in some of the tile fragments from the priory excavations at Lossenham. It seems likely that such variation also occurred in firings of medieval tiles - though no doubt those tile makers were better at knowing and managing their kilns than we are!
Medieval tile from Lossenham Priory showing the dark body colour characteristic of reduction
The same tile design from the current firing
Having got such variation, it seems possible that, when laying a floor, and having hundreds of tiles to work with, the tiles might have been grouped or arranged in such a way that the colour variation was itself used to create pattern, or lighter and darker areas, across the floor as a whole. There was almost certainly more to the paviors’ art than just following a plan.
We don’t have enough tiles for a floor (yet!), but putting together the various patterns from this firing of the wood kiln, we begin to get a sense of the visual effect that these sorts of tiles might have had, even over modest areas, compared to the beaten earth or stone-flagged floors that preceded them.
There is much to learn from our attempts so far to recreate medieval inlaid floor tiles using the techniques and tools of the period, including questions and insights raised by this latest firing. However, autumn is upon us, and winter approaching, and this stage of the pottery/tile project is drawing to a close - for the moment at least.
Notes
- Kent, O., Dawson, D., 1998. The Packing of Medieval Floor-Tile Kilns. Medieval Archaeology 42, 45–53. https://doi.org/10.1080/00766097.1998.11735616
- Many thanks to Lesley Warren, Rebecca Brown and Rosy Cousins for help with the firing and unloading
