2007: Early Medieval Handbell Reconstruction

The handbell was one of the iconic attributes of the priesthood in the Early Christian church of the celtic areas of Brittany, Cornwall, Ireland, Scotland and Wales in the Early Medieval period. The earliest handbells (Bourke Type 1) are made of wrought iron, forged and riveted into shape, with a bronze coating applied by brazing to produce a ringing sound. Related techniques were used until modern time to produce smaller animal bells. The project aims to reconstruct the techniques used in the fabrication of such handbells. This work has arisen partly from the desire of Amgueddfa Cymru - National Museum Wales (NMW) to produce a working replica of St Ceneu's Bell for their new archaeology gallery "Origins: in search of early Wales", and partly from research by GeoArch into metallurgical remains from a monastic site at Clonfad, Co. Westmeath, Ireland (excavated by Valerie J Keeley Ltd, for the National Roads Authority, as part of the N52 realignment), which include vitrified clay fragments interpreted as debris from the brazing of handbells. The work is being jointly funded by NMW and the NRA.

Site of experiments: Church of St. Teilo, St Fagans museum

The brazing experiments are being conducted within the churchyard of the church of St Teilo, recently reconstructed at the National History Museum, St Fagans, Cardiff

Two mild steel trial bells

Two trial bells forged from thin mild steel as part of the development process. As well as being used to develop the forging technique, these bells will be used as trials for the brazing process

The experimental work is in two parts. In the first the NMW blacksmith, Andrew Murphy, is working on the problems of forging the wrought iron bells. Andrew is working in the St Fagans Smithy. Although the basic mode of construction of wrought iron handbells has been described by Bourke (1980, 1986), the practical details are far from clear, particularly the join of the body of the bell to the handle.

The first trial forgings will be undertaken with thin mild steel, with the techniques developed being later applied to wrought iron. The original bells appear to be forged (with the exception of the handle/suspension loop) from a single sheet of iron, which would in turn have been forged from a single bloom. Many of the surviving bells are very large and it appears that the sheet from which they were forged must have originally weighed over 6kg in some cases. Forging such a sheet from bloom is an enormously demanding task and requires very good quality blooms. The production of handbells therefore represents the pinnacle of iron production and working in the Early Medieval period in the Celtic areas. In our experiments we will be using rolled puddled wrought iron for the final bells, since our own blooms are not large enough.

The second part of the experimental work is the reconstruction of the brazing technique. This is being undertaken by Tim Young (GeoArch) and Thérèse Kearns (UCL) in a temporary forge within the churchyard of St Teilo's church. This site is reminiscent of the location of the Clonfad iron-working site within the concentric ditches of the monastic enclosure there.

The initial experiments into brazing are being conducted within a hearth similar to those we have worked with before, to avoid introducing too many new variables. Hopefully some experiments will be made using an Irish-style ceramic tuyère later in the programme. Our larger double action blacksmiths' bellows are being used, blowing a hearth formed of bricks with a thin clay lining and a supportive external earth bank.

The hearth is being blown manually for the brazing experiments

Two test pieces in their clay coatings drying in front of the hearth as a crucible of bronze is being melted

It has long been known that wrought iron handbells were coated in copper alloy. Antiquarian accounts describe the coating as having been applied by dipping the bell. More recently it has been recognized that such coatings were applied by brazing. Brazed coats are also now known from other early medieval artefacts assembled from multiple iron components, with Viking padlocks being the best known.

The brazing shroud fragments from Clonfad suggest that the brazing metal was allowed to melt in a reducing atmosphere inside the clay coating and to be drawn by capillary action across the face of the artefact. The clay fragments have yielded no evidence for the size or form of the brazing metal fragments, but in some cases at least a cloth wrap between the clay and the iron may have held the brazing metal in place. In recent cowbells, fabricated using a similar technique, the brazing metal is brass (Laurence 1991), but the use of a (leaded?) bronze is likely for the early medieval ecclesiastical bells.

Update 1/12/07

coil-built clay coating

Following work on the small test pieces, the two mild steel trial bells have been prepared for brazing. The clay coats must now be allowed to dry before firing.

The bells were wrapped in thin cloth before the clay was placed around them (left). The coated bells were then inverted and the bronze placed close to the base of the bell (right). Finally the base of the bell was closed (apart from a small hole) with further coiled clay.

bronze placed next to base of upturned bell

Update 5/12/07

The clay coats on the steel bells are now dry. The small bell (coated with clay tempered with fine sand) is to be fired first. This bell is wrapped in canvas and has about 500g of bronze included. The firing started well, reaching the point of emission of zinc vapours after about an hour. However, a segment of the coat broke away, exposing the bell, after about 1:30h. The firing was stopped at this point. The bell was quenched to avoid excessive oxidation of the exposed metal. The bell coated fairly well, particularly on the side illustrated.

The bell as retrieved from the package (right) and after cleaning (far right). The crack in the bell may have been caused by the failure of the remaining clay coat as the bell was removed from hearth with tongs.

small steel bell as it emerged from clay package small steel bell after cleaning

Update 6/12/07

Following the successful experiment with the small bell, the large bell was to be fired. This differs in two major respects (other than size). Firstly the clay coat has been tempered with much coarser sand (commercially purchased sharp sand, burnt in a crucible to break down the shell fragments, then gently ground and washed), and secondly the cloth wrap is a much thinner muslin.

The firing went well, with the clay proving exceptionally stable. The package was fired for 3:15h, from building of the fire around the package. The package was rotated several times to allow for even heating. Zinc vapours emerged from the hole for a period of over an hour. The package was eventually removed to a cooler part of the fire, then removed and allowed to cool for an hour, before being quenched in water and cracked open.

The bronze proved to have largely melted inside the bell, passing through to the outer surface and the handle through the small holes around the handle. There was a very thin coating over most of the bell, but almost all the bronze was in this puddle around the top of the bell (base of the package).

the clay package in the hearth

The clay package was very stable in the hearth (above) and showed some good vitrification (upper right). The bronze proved to have flowed to the base of the package, accumulating inside and outside the bell and around the handle in a dense puddle (right).

package after firing

bell after coating

Update 7/12/07

Andrew completed the first wrought iron bell. The process has not been without problems. The folding of the thick wrought iron proved difficult and Andrew now wonders if the sequence of folding is incorrect. The rolled wrought iron also showed a great tendency to crack on the folds; the original bloomery iron would have performed better.

The completed bell was wrapped in canvas and coated in the clay tempered with coarse sand.

components of the bell

The bell comprises two components - the bell body, forged from a sheet, and the handle (above). The bell is partially forged to shape (top right), before the handle is attached (centre right). The bell can then be closed and riveted. The bell is shown (lower right) after cleaning and ready for the application of the clay coating.

forging the bell

the first wrought iron bell

forging the bell

Update 10/12/07

The first wrought iron bell was coated with clay on Friday, and has been allowed to air dry over the weekend. The bell was dried in a low temperature oven for a couple of hours this morning, before moving it to site and continuing the drying first in front of the hearth and then in the hearth. The initial coating and drying was done with the bell right-way-up, but it is to be fired in an inverted position (as with the previous tests). This experiment differs from the previous ones in that the large size of the bell means that positioning the bronze and coating the base of the bell is though to be more easily done in the hearth. This was done with the package in its position for heating, resting on a bed of ash and supported by the surrounding charcoal. Approximately 1kg of bronze was placed in the top of the package and about 150g inside the bell.

The large size of the package also meant that it could not easily be turned in the hearth to ensure even heating as had been done with the smaller packages. It became apparent after about 1:30h of heating that this would mean one side of the package remaining rather cold. The electric blower was moved round to the open end of the hearth and blast conducted through a section of scaffold tube as a tuyère. The manual bellows were used on the normal side of the hearth.

This heating arrangement was operated for about 2.30 hrs until all the remaining charcoal had been added (approximately 110kg used today!). By this time the interior of the bell was glowing nicely, the exterior of the package appeared to be starting to vitrify and zinc fumes were rising from the open end of the package. It was decided that the process had probably gone far enough and we wanted to avoid the complete loss of the bronze to the base of the package as in the last experiment. The remaining charcoal was raked clear of the package which was allowed to start cooling.

As the package cooled, the friable top became loose. This was removed to avoid the clay falling into the bell - but this revealed that the bronze in the top of the package had not melted at all. This was enormously disappointing, given the extreme heat of the hearth (about 0.5m was lost from the scaffold tube tuyère during the afternoon!). For the next experiment some method of better using the heat in the hearth will have to be employed. Perhaps a short while longer and the outcome might have been very different...

clay coating drying in the hearth

sections of bronze rod arranged in the package

the clay package cooling in the hearth

the first wrought iron bell

forging the bell

The main clay coating on the bell received its final drying in the hearth (top left), before the coating was finished by the placement of the bronze (middle left) and the partial closure of the open top of the package (lower left). The package was too large to be turned in the hearth, so part way through the firing the hearth was adapted to use both the manual bellows and the electric blower (upper right). At the end of the afternoon the package was seen to be stable and very hot (lower right), but unfortunately not hot enough at the top to have quite melted the bronze.

Update 12/12/07

Andrew removed the bell from the clay yesterday. The clay coating had fired to become very solid, but did appear to be little thicker than the Clonfad material. Only a single bronze rod (one placed internally) had melted, depositing most of its material in irregular blebs on the handle. The other internal rod and all the ones near the base were intact, although two at hte base had fused together. A crack on the side of the clay coat had allowed gas to penetrate to the bell and half of one of the flaps at the top of the side had burned away.

The impression of the fabric was strong on the shroud, with some marked folds near the base. Such folded (creased) cloth impressions are to be found on some of the Clonfad material, but not that from the main deposit.

clay coating after firing, showing carbonised cloth

The main clay coating had fired well (above), although part of one flap had burned away. Bronze had only been melted from a single rod and had ended up on the handle (right).

the first wrought iron bell with failed brazing

Crucible heating in the hearth


BOURKE, C. 1980. Early Irish hand-bells. Journal of the Royal Society of Antiquaries of Ireland, 110, 52-66.
BOURKE, C. 1986. Early Irish Bells. Seanghas Dhroim Mor, Journal of the Dromore Historical Society, 4, 27-38.
LAURENCE, P. 1991. Cloches, grelots et sonnailles. Élaboration et representation du sonore. Terrain, 16. (