Non-ferrous Metalworking
 

Introduction

Non-ferrous metalworking was an important activity at the site particularly during Phases 8, 9 and 10 (mid 14th to mid 17th century). Four major types of debris were identified as relating to copper alloy working: crucible, mould, metal waste (dross, spillage, offcuts, fragments) and fired clay. Detailed analysis confirmed that copper alloy vessels, probably cauldrons, were cast, using thick-walled crucibles and moulds made in sections.

Analysis methods

This analysis relates to nearly 40kg of material associated with non-ferrous metalworking processes. A small proportion may have been incorporated with the ferrous slag which was not examined in the course of analysis (see ironworking evidence). The material was examined visually and weighed to the nearest 2g. In some cases, it was not feasible to separate each different material type, for example, where copper alloy waste was attached to fired clay. Non-destructive surface X-ray fluorescence (XRF) analysis was carried out in a few cases to identify the major chemical elements present.

Results

Crucibles

Only a very small number of crucible fragments (less than 1kg) were recovered at the site (see table). All the crucible fragments originated from Phase 9 contexts, with five examples coming from 1910, the backfill of a large pit (Phase 9.3) on the eastern side of the site. Several other pits in Building Y in Phases 9.5 and 9.8 also contained crucibles, as did levelling deposits in Building Y and to the east of it in Phase 9.9. Within the southern room of Building U, a metalworking occupation deposit (2310, Phase 9.6) also contained a crucible fragment, which suggests that this building was used for copper alloy casting.

Two types of crucible, one thick-walled and straight-sided and the other thin-walled, were identified. The thick-walled forms (SF01092, SF01098, SF01671 and SF01672) were typical of those used to cast large copper alloy artefacts such as cauldrons and vessels in the late medieval period; they are similar to those analysed from excavations at St Andrewgate and Bedern (AY 10/03, AY 17/15 and AY 10/07). The walls were about 12 to 13mm thick, straight-sided and in a highly reduced well-made fabric. These crucibles probably would have had flat bases originally, but no bases survived at this site. It is difficult to estimate the diameter of the complete crucibles, but they would probably not have exceeded 150mm. Several of the fragments show evidence of having failed at high temperatures, with vitrification and copper alloy waste flowing over interior and exterior walls and across broken surfaces. XRF analysis of two of the samples (SF01098 and SF01671) confirmed the presence of copper, lead and zinc, the major elements in brasses. Analysis of another (SF01672) showed detectable levels of copper, lead, tin and antimony. These elements are found in some late medieval bronzes which have significant levels of antimony. Such examples were discovered amongst scrap at Bedern (AY 17/15, 2709–10) and in vessels elsewhere in England (Blades unpublished; Brownsword and Pitt 1981).

The thin-walled forms (SF01685 and SF01686) may also have been used for melting copper alloys, but on a smaller scale. However, one of the examples (SF01686) has a quartz lining suggesting it was not used for simple copper alloy melting but another process, possibly for the cupellation of precious metals. Traces of a copper-rich material are visible on the vitrification of the exterior, but XRF analysis was inconclusive.

A heavily vitrified lump (SF01638) may originally have been from a crucible or possibly another type of fired clay.

Moulds

In contrast with the scarcity of crucibles, about 25.4kg of mould material was recovered from the site, nearly all of which originated from moulds used to make large vessels. These pieces showed the classic reduced internal surfaces where the clay was in contact with the hot metal. Typically the mould fragments relating to the wall of the vessel were about 20 to 25mm thick, but other areas of the moulds were thicker, particularly where two pieces of mould were luted together, for example, at the rim, foot or handle areas of the vessel. The size of mould fragments varied but, in many cases, the pieces were very small, suggesting the mould was thoroughly destroyed when the cast object was retrieved. Moulds were constructed with inner and outer sections separated by the casting void. Where larger mould pieces existed at the site, both core (inner) and cope (outer) fragments were identified. Fragments from distinctive areas such as those relating to the rim of the vessels, from handles and from leg areas were also discovered. Some pieces from near the rim show two raised lines running just below the rim (e.g. SF01442); these are quite common and may have been decorative or due to the process of manufacture. The cope was clearly created in several sections (SF01431), a feature which has also been seen elsewhere (e.g. at Prudhoe Castle, J. Bayley pers comm). Thin layers of paler clay slip are visible on some of the inner surfaces.

A single unstratified clay fragment (SF01466) may relate to the casting of a smaller copper alloy object. The reduced appearance and fabric type resemble those of other moulds, but it is not possible to say what object was being cast on the basis of such a small fragment.

Many of the fragments were deposited in sizeable groups of more than 1kg; in one case nearly 6kg (SF00815) came from a levelling deposit (1610) in Phase 10. Overall, the largest quantities came from Phases 9.6 and 10 (totals of 6.8 and 7.4kg, respectively), but other phases, such as 8, 9.3, 9.8 and 11 also yielded significant amounts. These quantities do not take into account the weight of the mould debris which was crushed and used as hard-core within floor surfaces (see the discussion section).

Casting of copper alloy vessels using moulds appears to have commenced in the vicinity of the excavation area in Phase 5 (12th to early 13th century). Mould fragments were used as levelling material in construction work on Buildings R and S in Phases 7 and 8, but it could not be ascertained if either building was used for this craft activity. In Phases 9.1 and 9.2, Building U may have been the focus, with moulds being dumped to the west of it in the yard area. Moulds were also used during construction work on Buildings V and W. Floors, a hearth (1859) and dumps associated with metalworking in Building W confirm its use for this craft activity in Phases 9.3 and 9.4. In phase 9.6 non-ferrous metalworking seems to have taken place in the southern room of Building U and in Building Y, though moulds were still used for levelling in Buildings V and W. Buildings U and Y continue to be the focus of non-ferrous metalworking in Phase 9.8 and into Phase 9.9, with the east end of Building V also being used. Later phases reveal the redeposition of mould fragments within levelling and backfill deposits.

Mould weights by phase
Phase Weight (g)
554
7258
82280
9.2119
9.34144
9.4534
9.528
9.66770
9.82583
9.982
107435
111039
1271
1326

Copper alloy fragments

Most of the copper alloy debris at the site is classified as ‘waste’. Some of this waste can be further identified as having come from particular processes; for instance there are many small dribbles and spills, indicating small-scale loss of metal during casting. Other relatively solid pieces of copper alloy seem to be pieces of flash, trimmed from the completed artefact after casting. Dross is the lighter, usually more corroded, copper alloy material that originates from scooping the slaggy material from the top of the heated crucible load. It is not possible to suggest the origin for many of the pieces of waste material, often because the fragments are very small.

About 8.5kg of copper alloy waste was recovered, the majority of pieces weighing less than 100g. The very small size of these fragments suggests that the metalworkers ensured that waste metal was recycled (remelted) whenever possible. One notably large deposit (comprising 1kg of large lumps of metallic material) came from a Phase 9.5 pit backfill (2322) at the western end of Building Y. Overall, the largest quantities of copper alloy debris came from Phases 9.5, 9.6 and 9.8. The distribution of copper alloy waste, both chronologically and spatially, correlates with that of the mould fragments (see above). From Phases 9.3 to 9.6 copper alloy waste was associated with floors, dumps and hearths in Building W. Deposits (but no structures such as hearths or furnaces) within Building Y produced copper alloy waste in Phases 9.4 to 9.9, Copper alloy waste was found in Building W in Phase 9.6, in Building V in Phase 9.8, and in Building U from Phase 9.6 onwards.

Copper alloy fragments by phase
Phase Weight (g)
38
6.3430
74
846
9.2230
9.3755
9.4759
9.51161
9.62996
9.81251
9.9147
10168
11128
13427

Seven of the copper alloy waste fragments are sheet-like: one seems to be a small piece of vessel rim (SF01127) and another is a thin strip wound round a piece of stone or mortar (SF00786). Several of the sheet-like pieces of waste are roughly square and about 2mm thick; these could be chopped-up items prepared for remelting. Others are more irregularly shaped and may represent offcuts of trimmings from sheet which was being formed into shapes. Apart from SF01528, from Phase 6.3, these offcuts and strips all derive from Phase 9 levels (SF01202, SF01319, SF01410 and SF01411). Miscellaneous sheet fragments (including SF01559, SF00878, SF00697) may have been destined for recycling. A wire fragment (SF00556) appears to have been drawn, and may result from pin-making.

Two cake-shaped objects appear to be lead-rich copper alloy (SF00626 and SF00901). One (SF00626) has a slice out of it, suggesting that it was used as a source of metal for casting.

A tiny droplet of lead alloy (SF00785) may have resulted from melting lead for recycling. Other lead alloy fragments appear to represent items destined for recycling (e.g. SF01291, SF01228, SF01533, SF01362) or are offcuts trimmed from sheet (SF01150, SF01153, SF01510, SF01229). Fragments of lead alloy roofing sheet (see Everyday Life Part 1) were also thought to have been intended for recycling.

Fired clay

Fired clay could have originated from many industrial or domestic processes and it is often difficult to identify the source of fired clay fragments, unless there are signs of copper alloy or ferrous deposits. This was noted in some instances while in others the clay was obviously fired at a high temperature, sufficient to cause vitrification, and this may also suggest a connection with metalworking, deriving from hearth or furnace structures. There are also pieces of tile with traces of copper alloy waste on them (pieces from contexts 1910 and 2394) which were obviously in close contact with the copper alloy working. A total of 5.6kg of fired clay was recovered from the site, mostly from Phases 8, 9.3 and 9.6, but this does not include material from within fired clay structures such as hearths.

Miscellaneous fired clay, by phase
Phase Weight (g)
3186
4168
5274
6.1138
6.2270
6.382
748
81085
9.260
9.31054
9.4175
9.61258
9.8144
9.966
10601

Slag

Although copper alloy melting processes sometimes produce a small amount of slag-like material, that which was examined was either associated with ironworking, or could be identified as heavily vitrified fired clay.

Other material types

Fuel, in the form of coal and charcoal, was also recovered from the excavations, but only a small sample of this was collected and the results discussed below should be treated with caution.

The hand-collected charcoal fragments were generally associated with dumps and backfill deposits rather than metalworking activity although this might be a reflection of sampling.

Coal was first recovered in pit backfills from Phase 8. One of these pits was associated with the dumping of iron smithing waste and it may be that coal was used as a fuel for this industry from this phase onwards. In Phases 9.2, 9.3 and 9.6, coal was associated with ironworking in Buiding W and with metalworking in Building U (Phases 9.6 and 9.8). The retrieval of this fossil fuel from metalworking dumps in Phase 9.9 may suggest the use of coal as a fuel for this industry (possibly within Building U).

Total weights of material types recorded
Material Type Weight (g)
Copper alloy waste8467
Crucible809
Fired clay4775 (includes daub)
Mould25428
Other263
Total:  39742

Conclusions

The copper alloy metalworking material found at Walmgate is typical of a medium-sized casting workshop. A single casting process has been identified, making large cauldron-like vessels. The debris relating to casting was concentrated in Phases 8, 9 and 10. Buildings U (Phases 9.1 and perhaps 9.2, and 9.6 to 9.8), V (Phase 9.8), W (Phases 9.3 to 9.6) and Y (Phases 9.4 to 9.9) appear to have been used for this craft activity. Buildings U, V and W all produced structures (hearths and furnace bases) which were associated with non-ferrous metalworking, while Building Y may have been used for storage or finishing. The relationship between the evidence for non-ferrous metalworking and the structures within these phases is discussed further in the discussion section.



Click to enlarge
Crucible fragment SF01098

Click to enlarge
Crucible SF01098

Click to enlarge
Mould fragment SF01442

Click to enlarge
Cope fragment SF01431

Click to enlarge
Cope fragment SF01431

Click to enlarge
Copper strip wound round a piece of stone and mortar (SF00786)

Click to enlarge
Copper alloy off-cut (SF01528)

Click to enlarge
Wire fragment (SF00556)

Click to enlarge
Plain tile with copper alloy concretion from context 1910
© Copyright York Archaeological Trust 2003