We have launched a new website and are reviewing this page. Find out more

Now you see it, now you don't: the conservation of a turkey work chair

Gill Owens
Textile Conservator, Conservation Department

Fig 1. Pattern of knots in Turkey work

Fig 1. Pattern of knots in Turkey work (click image for larger version)

Turkey work, or 'sette', or 'carpet work', which was produced in Britain from the second half of the 16th century, was inspired by the knotted woollen carpets imported from Anatolia (Turkey). It was woven with a strong linen or hemp warp and weft. The worsted, long staple, wool knots were added alternately with the weft, or after every two wefts. (Fig.1) Sometimes cross-stitch was added to an area. At first the designs too were inspired by Turkish carpets but quite soon they came to resemble the embroidery of the Stuart period, depicting angular flowers in glowing colours, often on a very dark or black ground. This ground is often degraded or missing, because of the iron mordant used in dyeing. Turkey work was robust and had many uses. It was used for carpets (a carpet was defined as late as 1772 by Chamber's 'Encyclopaedia' as 'a sort of covering to be spread on a table or trunk or a staid [dais], or even a passage or floor') and for wall and window hangings.

By the start of the 17th century upholstered chairs were becoming popular. Turkey work, which was warm and strong, was used to upholster 'farthingale' or 'back chairs' which had padded seats and a low padded back. This article will discuss the conservation of a chair of this type, acquired by the Victoria & Albert Museum in 1894 for £4 from a Yorkshire dealer. The frame was not contemporaneous with the covers (which are dated 1649 and initialled G.S.G.) and so they were transferred to a more suitable one. In 1989 this chair came to be conserved when the Early English Galleries were being rearranged. At this time the back and seat covers had large missing areas and the pattern that should have been on these areas had been painted on the canvas to which the covers had been stuck. The wool braid from which the fringe hung was in reasonable a length running along was missing. The fringe itself was very dirty and broken, with large areas missing.

Fig 2. Chair back cover during conservation showing chequered selvedge

Fig 2. Chair back cover during conservation showing chequered selvedge (click image for larger version)

The chair frame was beech (which was cheaper than oak and commonly used outside London). It was riddled with worm holes and the legs had been cut off just below the junction with the horizontal bars. The left back leg had been repaired with a metal plate covered in leather and there were two large cracks on the back upright. A 13 cm wide piece of oak had been fixed under the original nailer (the piece of wood to which the bottom of the back cover is fixed). This must have been added to accommodate the full length of the back cover when it was moved to a new frame. Unlike most upholstery fabrics, Turkey work covers were woven to a specific shape with a chequered selvedge outlining the shape rather than a cut from a length of fabric. (Fig.2)

The conservation brief for this chair required not only that the frame should be mended and the fringe conserved, but that the covers should be cleaned of as much  glue as possible and supported by stitching techniques. It was also essential that painted areas should be removed so that the extent of the original Turkey work was made clear (Fig.3).

The top layer of glue, which we had assumed to be of animal origin, was confirmed as such by laboratory analysis. The covers were removed from the canvas to which they were glued, using a metal sculptor's spatula. They were very dirty and mechanical cleaning with a small vacuum cleaner, through a layer of protective net, made little difference. There was also a lot of animal glue around the edges and over worn areas.

Fig 3. The chair before conservation. A large area of the pattern at the front of the seat has been painted onto the backing

Fig 3. The chair before conservation. A large area of the pattern at the front of the seat has been painted onto the backing (click image for larger version)

It was decided to test the wool dyes for colour fastness and if possible to wash the covers. All 11 colours were tested using loose fibres from the back of the Turkey work; some areas on the covers themselves were also tested. Cotton wool swabs soaked in the washing solution then in use in the studio (carboxymethyl cellulose and Synperonic N), were placed on top of the wool fibres which had been laid out on drying paper. Everything was covered in an impermeable layer of polythene and left to dry. When everything was dry, the swabs and drying paper were examined for traces of colour and none was found. Washing is a procedure which is always approached with a great deal of caution in textile conservation because it is irreversible and can cause damage or loss of degraded fibres by sudden swelling on wetting and contraction on drying. Glycerol, which is very viscous, is sometimes used to help the process. It enables the degraded fibres to expand slowly and gently, but it is only used on relatively robust textiles and must be carefully washed out. A particular worry when washing chair covers is that they may shrink or distort. Before washing, templates were prepared by drawing round the seat and back cover onto paper. The paper was fixed to softboard and the whole covered with a waterproof layer of polythene or Melinex. After washing, the covers could be straightened out and pinned over these templates before being left to dry.

The covers were soaked in a weak solution of glycerol in deionised water (approximately 1%) and then in plain deionised water. The animal glue swelled up during this time and could easily be removed. Under the animal glue, in some degraded areas, was a rubbery adhesive which laboratory reports showed to be Copydex. The application of industrial methylated spirits and propan-2-ol enabled this to be largely removed. Net was laid over weak areas and the back and front of both covers were washed using the solution we had tested, applied with a soft squirrel paintbrush. When the rinsing had been completed, they were pinned, pile side up, to the prepared templates. The covers were not very thick and the studio was warm, so they dried quite quickly.

Fig 4. Original nailer with addition removed

Fig 4. Original nailer with addition removed (click image for larger version)

Choosing a support fabric was difficult. Dyed linen scrim looked very suitable, but it was considered too weak to use alone. By sewing it to dyed linen holland we had a backing which was satisfactory in terms of strength and appearance. This double backing was then attached to each of the covers with horizontal and vertical lines of stitching, incorporating a very small amount of fabric to allow for the amount taken up by repair stitching.

Object and backing were pinned to a padded frame. Starting at the centre, and working out, the covers were sewn to the backing, stitching round motifs, and couching down holes and weak areas. Polyester thread was used. All the edges were closely couched down too. Couching is a stitching technique much used in conservation now and in the past. It is a generally accepted way of supporting a degraded original textile on a prepared, sympathetic new backing material.

Next came the conservation of the wool braid and fringe. These were obviously too fragile for washing. The seat length was cleaned in white spirit, one of the milder solvent cleaning fluids, with a minimum amount of agitation. Even so, the fibre loss, seen with the naked eye in the white spirit bath, was unacceptably high, so it was decided to leave the back of the fringe uncleaned. Working all the time next to a humidifier (to help alleviate the brittleness of the fibres), the braid was couched to a dyed cotton tape, using a finer polyester thread than the one used on the Turkey work. The tape was extended so that it would cover the back of the seat. Where the woollen threads from the fringe had broken, dyed tapestry wool threads were taken from the new tape, through the fringe knots, and back. A diagram was made, using the fringe on the back as a model to clarify the pattern of colour and knots, which was very difficult to read on the seat fringe.

Fig 5. Chair after conservation, Museum no. 428-1896

Fig 5. Chair after conservation, Museum no. 428-1896 (click image for larger version)

While the upholstery covers were being treated in Textile Conservation the chair frame had been receiving attention in the Furniture Section. The woodworm damaged areas had been consolidated with repeated injections of Araldite 778 which had also been used to repair breaks. The additional nailer had been removed (as had the metal repair) (Fig. 4). The nailer had been replaced by a piece of balsa wood, which would be easy to nail into, preventing damaging vibrations when the covers were re-attached. A beech infill had been let in where the metal repair had been before, and all repairs had been coloured to match the frame. Finally, frame and Turkey work covers were reunited. The seat cover was stretched over the seat and excess backing fabric cut off and turned under. The studs were replaced through braid and backing fabric. The back cover was similarly held in place along the bottom rail by studs through the conserved braid and backing fabric. Then the other three edges were covered by the edges of the linen holland back lining, which was turned in and fixed down with the studs. (Fig.5)

In the light of a current interest in ethics this seemed an interesting case history, highlighting as it does a dramatic change of attitude within this Museum towards the replacement of information. Would we still transfer covers from one frame to another, necessitating a back lining which such a chair would probably not have had? Would we perhaps just mount the covers on a board? Perhaps the examination of past case histories will help us in the formulation of our code of conservation practice.