The primary method used to produce metal figurines was lost-wax casting, and two types of moulds, investment and piece, allowed slightly different techniques. Investment moulds utilised a wax model of the object to be cast. Clay was then formed around this model and allowed to dry. The mould was fired, during which process the wax melted and was poured out and the molten metal was then poured into the finished clay mould. In order to remove the casting the mould had to be broken and so was used only once. This process produced a solid object. Piece moulds, however, produced a hollow casting. This technique utilised a clay model which was then coated in wax, followed by clay around the whole. The inner core was held in place by chaplets, small metal bars placed between the core and the mould. Like investment moulds the piece mould was fired, the wax poured out and molten metal poured in. However, while investment moulds had to be broken to remove the casting and so could only be used once, piece moulds allowed the reuse of the internal pattern if it was removed from within the hollow casting (Bayley 1988, 202; Hemingway 2000, 39-40). Although a number of figurines are very similar, suggesting that they may have been made in the same workshop, there is almost no evidence for the reuse of moulds to produce figurines apart from two nearly identical hollow cast heads of Diana as Luna from Penllyn, Vale of Glamorgan (1173 and 1174). In a study of Latin, Etruscan and Roman figurines, Galestin (2002) was able to find evidence for the use of moulds in this manner only for a series of 6th-century BC kouroi from a sanctuary at Lavinium.
Other types of objects were being produced using investment moulds in Iron Age Britain and large numbers have been found at a few sites such as Weelsby Avenue, Grimsby, Lincolnshire (Foster 1995), Gussage All Saints, Dorset (Spratling 1979; Foster 1980) and Silchester, Hampshire (Northover and Palk 2000, 406-13). The use of piece moulds is often associated with Roman metalworking and the mass production of objects, such as the large group for making spoons and enamelled vessels found at Castleford, W. Yorkshire (Bayley and Budd 1998). Meanwhile, examples of both mould forms have been found on a site spanning the late Iron Age and Roman periods at Prestatyn, Denbighshire (Blockley and Day 1989). At this site, while investment moulds were used to produce typically Iron Age objects such as horse gear, piece moulds were used to make a mixture of Iron Age and Roman objects, indicating that metalworkers would adapt their techniques to suit the type of objects being produced (Blockley and Day 1989, 187). Foster (1980, 37) notes that the horse gear being produced at Gussage All Saints comprised high-quality items. Perhaps there was a desire to produce unique pieces that would be better suited to the use of investment moulds, while the mass production allowed by the use of piece moulds was seen as beneficial for simpler, less prestigious items such as button and loop fasteners. The majority of figurines were cast using investment moulds, and it seems likely that the complex nature of the pieces was better served by making a wax model of the object to be cast.
As well as being cast whole, sometimes figurines were cast in pieces and assembled. This is indicated by Arms 457 from St Albans, Hertfordshire, and 1029 from Colchester, Essex. Both have a dowel at the shoulder for attachment to the figurine. Harpocrates 726 from St Albans has traces of solder within the arm sockets, suggesting that the arms were cast separately (Worrell 2005, 460), while Fragment 1137, also from St Albans, is drapery from a left shoulder with a hole for the insertion of an arm. In addition, it is quite common for the hands of a figurine to be empty, although they are curled to hold an object. This is usually an attribute such as a spear, club or shield that must have been cast separately and then placed with the figure.
Unfortunately, the direct evidence for the production of metal figurines in Britain is sparse. A single lost-wax investment Mould 762 has been found at Gestingthorpe, Essex. The fragments show the stomach and thigh of a male as well as an ivy-leaf wreath, which indicates that this was probably a Bacchus, or possibly a character associated with him. Other mould fragments, crucibles and fragments of bronze from Building 2 on the site, as well as the metal finds themselves, indicate that bronze working was taking place there in the 3rd and 4th centuries (Draper 1985, 11).
Other indications of bronze working are seen on unfinished pieces such as the small head from Wroxeter, Shropshire, which is still attached to the casting excess from the sprue (Bushe-Fox 1914, 11 and fig. 5 no. 12). Isis 734 from Thornborough, Buckinghamshire, and two crude lead figures of Fortuna 1166 and 1167 from Springhead, Kent, have casting flanges on their sides, and this lack of finish suggests that they were made locally (Green 1983, 139; Schuster 2011, 270-71). Casting flaws might also indicate a local production for pieces such as Horse 793 from Frocester, Gloucestershire, which has a flaw on the left side of its rump (Foster 2000, 57 no. 374), Lar 517, possibly from Kingsholm, Gloucestershire, with a flaw at the rear of its lower right leg, and Mercury 692 from Great Walsingham, Norfolk, with a flawed bulbous left calf and several holes in the left arm. While these flaws are all relatively minor, the rider on Horse and rider 974 from Lincolnshire has no left arm. Thus it would appear that even figures with fairly serious flaws were still considered good enough to sell, possibly due to the effort that went into the production of such a prestige item, or its function as a religious object.
A final form of evidence supporting production of figurines in Britain comes from inscriptions on the bases of two pieces. Mars 31 from Foss Dyke, Lincolnshire, bears the inscription 'To the god Mars and the Deities of the Emperors the Colasuni, Bruccius and Caratius, presented this at their own expense at a cost of 100 sesterces; Celatius the coppersmith fashioned it and gave a pound of bronze made at a cost of 3 denarii' (Collingwood and Wright 1965, 274). The base of Equestrian 646 from Martlesham, Suffolk, is inscribed on the side 'To Mars Corotiacus Simplicia for herself willingly and deservedly set up this offering' and underneath 'Glaucus made it' (Collingwood and Wright 1965, 213). These inscriptions indicate a level of literacy and wealth needed to commission such a work. The names listed show a variety of origins for the people involved. Bruccius and Caratius are thought to be Celtic names and Colasunius a Romanised gentilicium, or family name, while Simplicia was a popular name among the Romanised population (Birley 1979, 131). The name Celatius may be derived from the Latin word 'celatus' meaning concealed, while Glaucus is a Greek name, possibly indicating an immigrant bronze-smith (Birley 1979). Toynbee (1964, 66) suggests that the Roman name Celatius might indicate an immigrant artist, while Birley (1979, 131) believes it to be a Celtic name, and so could belong to a local or Gaulish bronze-smith. Although perhaps a chance occurrence, it is interesting that all of the names associated with the Foss Dyke Mars, a fairly standard provincial piece depicting a popular deity, appear to be Celtic, while those associated with the more unusual equestrian figurine might indicate a slightly more cosmopolitan patron and artist.
In addition one might consider the sum paid to create the Foss Dyke Mars. It is likely that the figurine was made in the 2nd or 3rd century and Rathbone (2009, 311 and 316) has estimated that the gross salary of an infantryman rose from c. 1,200 sesterces per annum in the late 1st century to c. 3,600 in the 3rd century, wages for rural labourers were slightly higher and for craftsmen perhaps twice this. Assuming that the Colasunni were men of some standing, the sum of 100 sesterces between the two of them, although representing a reasonable investment, was not a huge sum to pay.
© Internet Archaeology/Author(s)
University of York legal statements | Terms and Conditions
| File last updated: Tue Mar 20 2012