By Lori Covington

Fig. 1 Terracotta Mother & Child relief, in glazed, raku-fired earthenware

What is it? The broken object on my workbench; my first "live" project, was a modern, terracotta relief sculpture, Mother and Child, probably made in the 1970s or '80s by the Head of Art at St. Paul's School, in London. The owner was a student of the potter and cherishes this piece, holding as it does memories of happy days at school. It is heavily potted and covered in semi-translucent green and white glazes. Raku-style firing created pockets of black encrustation in the crevices and highly textured portions of the sculpture, such as the hair and heads of the joined figures. The Damage The sculpture was recently broken when a plaster cornice fell from the ceiling. Both heads were broken off at the necks and along the woman's hair, where it met her shoulder. Her feet, at the end of the piece, were broken off, and the wing-like cape was shattered, with around 25 fragments having come from an area of around 2 cm2. Made to rest flat against the wall, the heaviness of the sculpture places the thinner portions of it at some risk. On the workbench it needed to be carefully supported while being conserved.

Fig. 2 Unglazed Back of Mother & Child

How It's Made The clay flowerpots you can find for 69 pence at the garden shop are made of "terracotta"-earthenware. Earthenware is fired in the kiln at lower temperatures (around 1100°C) than stoneware, bone china or porcelain, which are fired at around 1250-1400°C. High firing creates clay bodies with more melted, vitrified (glassy) material, which makes them less permeable, harder. Earthenware is a softer, thicker material, which tends to absorb liquids. Because this piece is earthenware, its body is porous, and that influences the way it needs to be treated. In this terracotta piece, it's important to make sure water doesn't get in, as it may pick up and dissolve loose dirt, creating stains. Intact, glazed areas on the front would be less likely to absorb liquid: the glaze, a glassy, hard coating, keeps moisture out. But the back of the piece is unglazed and porous, and any break in the glaze made the piece vulnerable to liquids. So, although the sculpture was dusty, it was important to keep water away from any unglazed part of the piece! The Conservation Cleaning Conservation began with an overall brush gentle cleaning, followed by swabbing unbroken glazed areas with deionized water. At the start of a conservation treatment, gentle, thorough brushing removes dust: even a speck of dust can ruin bonding, getting in the way and ruining the fit between pieces. Planning-the "Dry Run" Next, I spent some quality time with the sculpture, putting the broken pieces back in place with bits of Magic Tape, seeing which parts were missing and would need to be replaced by filler later on. The three bigger breaks-the heads and feet-were simple enough, but I could see the mother's head was going to present a challenge. Nothing's Perfect, Darn It! Clay can undergo an amount of tension when being worked into shape and, on breaking, the fired object alignment can be impaired and we call that "sprung". When I placed the head onto the neck area as perfectly as possible, there were small discrepancies of alignment nearby. The myriad tiny bits broken from the cape were too small to fix with tape in the "dry run"- the practice trial without adhesive. Some came from the top side, edged in green, black or white glaze, the reddish terracotta body showing through the translucent colouring. Other pieces came from the centre of the body, unglazed earthenware chips too small to realistically fit back together. A few had broken all the way through the cape. Although I could piece some together in the dry run, many pieces had broken in layers-for example, a top bit of glaze, with a piece of the mid-body sandwiched onto it, and a third piece resting below and extending just past the others. The pieces were uneven, and different layers had to built separately before coalescing into a whole. I had to accept that some of the work would remain a mystery until certain pieces were bonded in place. Only then, it might be possible to add others. Stick Like Glue When bonding (gluing) ceramics, you have to consider the ceramic body type before choosing an adhesive. Different types of ceramic require different types of adhesive. An epoxy perfect for bonding a high-fired, highly-vitrified (glassy and hard) and thus non-absorbent piece of porcelain, for example, might be totally wrong for a low-fired, porous piece. A mobile adhesive, for example, could migrate into an earthenware body, stain it, become impossible to remove and, adding insult to injury, fail to bond the object anyway! Ideally, the adhesive is strong enough to hold the piece together, but can also be taken apart one day if the need should arise. In the case of a low-fired ceramic such as this one, a super-strong epoxy resin would be overkill: so I decided to use an adhesive called Paraloid B-72. Paraloid B-72, like many adhesives, cures by evaporation of a solvent. To make it, solid granules of the resin are first dissolved in a solvent, in this case acetone, to whatever consistency is needed. Once it's applied, the bulk of the acetone evolves away in a few minutes, leaving only the adhesive film behind. Paraloid B-72 is popular with ceramics and glass conservators because it doesn't discolour over time, as many epoxy resins do, and its viscosity can be tailored to fit the job; thinner for lighter needs, thicker for heavier ones. Another major benefit in using Paraloid B-72 is that it's "reversible": it will hold beautifully, but if you need to take the bond apart, it can be done with applications of acetone. On the downside, Paraloid B-72's glass transition temperature (Tg) is only 40°C, which prevents its use in circumstances where objects may be exposed to heat. Originally, I had hoped to use 50% w/v Paraloid B-72 in acetone, but it proved much too thick. Bubbles were forming in the adhesive even as I was attempting to use it! They couldn't be pressed out: they hardened instantly, pushing small, fragile bonds out of alignment. After consulting with a much more experienced conservator, I switched to cellulose nitrate, which bonded quickly and well too but without the problems of bubbling. For the larger pieces, cellulose nitrate was perfectly acceptable. Doing, Undoing, Redoing However; the cape's many tiny pieces made tolerances between pieces too small for the thickness of even a minute dab of cellulose nitrate. In those cases I used a thinned version of Paraloid B-72-around 30% w/v. Until the last two pieces, it seemed to be working. But the next-to-last piece was "out"-nothing would make it fit, and it was easy to see the last piece wouldn't fit at all. Putting the piece under the microscope showed each small piece was just slightly out-a bubble of Paraloid B-72 here, a too-thick layer of adhesive there-under microscopy, the "good fit" conjured up the Cheddar Gorge! Mindful of the conservation mandate to do less, not more, I hesitated to start over. The pieces were tiny but robust: I decided to try again. Removing one piece at a time, dissolving the bonds with careful applications of acetone (luckily, cellulose nitrate is also reversible with acetone) I hoped to find one wrong bond that would let me stop, but they were all wrong-ever so slightly. Finally, I removed 16 pieces, cleaning adhesive off each. While the piece was "resting" from my latest interventions, I did some reading and experimenting, using a variation on Paraloid B-72's recipe. I had to get rid of those bubbles! The popular way of making Paraloid B-72 involves adding fumed silica, as a bulking agent, or thickener. (If you've ever blended icing sugar into butter, you've "bulked" the icing with sugar!) Paraloid B-72 was made without the fumed silica and the few bubbles it did make escaped, rather than hardening in the adhesive. What good luck! In a few days, the cape was nicely re-bonded, even down to those last stubborn pieces. Colouring in the (Break) Lines After bonding, missing areas were filled with a plaster-based filler (Polyfilla®), and break lines filled with Fynebond ® epoxy. Fynebond, bulked with fumed silica and tinted with pigments, mimicked the missing glaze layer on the front, and the unglazed back of the sculpture. The break lines were small, so I worked with a cocktail stick and sometimes the tip of a scalpel. Texture was smoothed with a brush, slightly dampened with water. To match the glaze, I polished the dried, hardened resin with very fine sandpaper and micromesh. Seeing Mother & Child in one piece is satisfying, especially as, before conservation, I had only seen the relief in pieces. After the challenge of putting a piece to rights, it's a good feeling to see the piece as the artist made it.

Fig. 3 Mother & Child, Reunited