Archaeometallurgy

Guidelines for Best Practice

Summary
This guidance document provides an introduction to the ways that the
archaeological evidence for metalworking is studied. Archaeometallurgical
evidence can include whole landscapes, buildings, features, artefacts and
waste materials (eg slag and crucibles). Archaeometallurgy includes fieldwork
investigations (survey and excavation) and the subsequent study of these data
as well as any artefacts and residues recovered. Scientific approaches provide
insights into the techniques used to produce different metals and how these
were fabricated into artefacts.

The previous edition of this document was compiled in 2001 by Justine Bayley,
David Dungworth and Sarah Paynter with the assistance of the Historical
Metallurgy Society’s Archaeology Committee, with contributions by Peter Crew,
Vanessa Fell, Brian Gilmour, Gerry McDonnell, Cath Mortimer, Peter Northover,
David Starley and Tim Young. This edition was revised in 2015 by David Dungworth.

This guidance is published in association with Historic Scotland, CADW, the

Environment and Heritage Service and the Historical Metallurgy Society.

It is one is of three Historic England publications on guidance concerning

materials science and industrial processes:

■ Science for Historic Industries. Guidelines for the investigation of 17th- to

19th-centuries industries;

■ Archaeological Evidence for Glassworking. Guidelines for best practice; and

■ Guidance for Archaeological and Historic Pottery Production Sites.

First published by English Heritage September 2001.

This edition published by Historic England April 2015.

All images © Historic England unless otherwise stated.

www.HistoricEngland.org.uk/advice/

Cover images
Left: 16th-century blacksmithing (Agricola) and
Right: Modern experimental blacksmithing.
[photo © David Starley]
Contents
Introduction. . ................................. 1 4 Archaeometallurgical processes
and finds: copper and its alloys... 40
1 What to expect. . ...................... 2
4.1 Background.................................................40
1.1 Bronze age. . ...................................................2 4.2 Smelting and alloying................................41
1.2 Iron age. . ........................................................3 4.3 Casting.........................................................43
1.3 Roman.. ..........................................................4 4.4 Wrought metalworking. . .............................48
1.4 Early medieval .. .............................................5
1.5 Medieval........................................................5
1.6 Post-medieval...............................................6 5 Archaeometallurgical processes
and finds: lead.........................49

2 Standards and good practice 5.1 Background.................................................49

for archaeometallurgy. . ........... 7 5.2 Smelting......................................................49
5.3 Lead working. . .............................................52
2.1 Project planning and the formulation of
research designs...........................................8
2.2 Fieldwork: survey . . .......................................9 6 Archaeometallurgical processes
2.3 Fieldwork: excavation. . .................................9 and finds: other metals........... 53
2.3.1 Structures and context...............................10
2.3.2 Finds and sampling....................................11 6.1 Silver and gold............................................53
2.3.3 Dating..........................................................12 6.1.1 Refining . . ......................................................54
2.3.4 Site archive. . ................................................13 6.1.2 Parting.........................................................55
2.4 Assessment of potential for analysis........13 6.2 Tin................................................................56
2.5 Analysis and report preparation. . ..............13 6.2.1 Mining..........................................................56
2.5.1 Quantification.............................................14 6.2.2 Smelting......................................................56
2.5.2 Scientific techniques..................................15 6.2.3 Alloying........................................................57
2.6 Dissemination.............................................15 6.3 Zinc..............................................................57
6.4 Other metals . . ..............................................58

3 Archaeometallurgical processes
and finds: iron and its alloys.....16 7 Non-metallurgical residues
and materials....................... 59
3.1 Background.................................................16
3.2 Smelting......................................................18 7.1 Geological materials. . .................................59
3.2.1 The bloomery process. . ..............................18 7.2 Debris from non-metallurgical industries...59
3.2.2 The blast furnace........................................24 7.3 Conflagration residues...............................61
3.2.3 Refining pig iron . . ........................................29 7.4 Heat-magnetised residues.........................61
3.3 Making steel................................................32 7.5 Plastics and tarmac....................................61
3.4 Smithing......................................................34
3.5 Casting.........................................................39
8 Scientific techniques applied
to metalworking.......................... 62

8.1 X-radiography. . ............................................62

8.2 Geophysics..................................................62
8.3 Archaeomagnetic dating............................63
8.4 Microscopic examination...........................63
8.5 Chemical analysis.......................................65
8.6 Isotope analysis..........................................66
8.7 X-ray diffraction.. .........................................66

9 Where to get advice. . ............. 67

9.1 Historic England.. ........................................67

9.2 The Historical Metallurgy Society. . ............67
9.3 Regional Research Frameworks. . ...............67
9.4 Published archaeometallurgical research...68
9.5 Historic England reports............................68
9.6 ADS Grey literature.....................................68

10 Glossary............................... 69

11 Bibliography. . ....................... 71
Introduction
Archaeometallurgy is the study of metalworking structures, tools,
waste products and finished metal artefacts, from the Bronze Age
to the recent past. It can be used in the field and during the post-
fieldwork phases to identify and interpret metalworking structures
and waste products, such as slags, crucibles and moulds. The nature
of technologies used in the past (as well as their social and economic
impact) can be reconstructed from the archaeological evidence.
Scientific techniques can provide information about the manufacture
and consumption of a range of metals.

Archaeometallurgical investigations can provide them to produce project briefs, project designs,
evidence for both the nature and scale of mining, assessments and reports.
smelting, refining and metalworking trades,
and aid understanding of other structural and The Guidelines are divided into a number
artefactual evidence. They can be crucial in of sections. First is a summary of the sort
understanding the economy of a site, the nature of metallurgical finds to expect on sites of
of the occupation, the technological capabilities all dates (p 2). This is followed by a section
of its occupants and their cultural affinities. In entitled ‘Standards and good practice for
order that such evidence is used to its fullest, it archaeometallurgy’, outlining its relationship
is essential that archaeometallurgy is considered with other aspects of archaeological projects
from their outset and at all subsequent stages (p 7). Then come the fully illustrated sections
of an archaeological project. Technological and describing archaeometallurgical processes and
functional aspects of metalworking should be finds: for iron (p 16), copper and its alloys (p 40),
understood within wider socio-economic contexts lead (p 49), silver and gold (p 53), tin (p 56) and
which can include symbolic and/or magical zinc (p 58). A shorter section on non-metallurgical
qualities (Giles 2007). high temperature processes illustrates finds that
are often confused with metalworking debris
These Guidelines aim to strengthen the retrieval (p 59). A section is provided introducing some
of information about all aspects of metalworking of the scientific techniques commonly used in
from archaeological investigations. They are archaeometallurgy (p 62). Finally, sources of
written mainly for curators and contractors additional information (p 67) and a glossary of
within archaeology in the UK and will help common metallurgical terms is provided (p 69).

< < Contents 1

1 What to Expect
It is useful to know what sort of archaeometallurgical evidence to
expect from a particular site. This depends on a number of factors,
such as the location of the site, its date and the nature of the
occupation. For example, archaeological evidence for mining tin
will usually only be observed in areas where tin ores are found, iron
working evidence is unusual before the beginning of the Iron Age, and
precious metalworking is more likely to be concentrated at high status
and/or urban sites.
The following chronological summary of the archaeometallurgical
record for the UK indicates the types of evidence that are likely to
be found.

1.1 Bronze Age similar evidence has been recovered during

archaeological excavation of Bronze Age mining
Copper alloy and gold artefacts of this period sites (Lewis 1990). Early working made use of
show that these metals were worked. Some stone tools or fire to weaken the rock (Craddock
evidence exists for copper mining (Timberlake 1995, 31–7) and this can be distinguished from
and Marshall 2013), while other evidence later working where iron tools or explosives were
demonstrates working, mostly casting, of copper used. The palaeoenvironmental record confirms
alloys. There is almost no direct evidence for how that copper (and lead) mining took place (Mighall
other metals used during the Bronze Age were et al 2009)
obtained. It is generally accepted that the tin
ores in south-west England were exploited from Little is known about how ores were transformed
the Bronze Age onwards but there is little direct into metals in Bronze Age Britain. No smelting
evidence for this (Penhallurick 1997). furnaces have been identified (Craddock 1990;
1994), although some slag has recently been
Evidence for mining can only be expected in found on the Great Orme in North Wales (Jones
regions where ores are found. In England, copper 1999; Williams 2013). Some useful ideas, however,
ores are known in Cornwall, Devon, Shropshire, have been gained from recent experimental
Staffordshire, Cheshire (Timberlake and Prag archaeological work (Timberlake 2007).
2005), North Yorkshire and Cumbria, and other
sources are known in mid and north Wales The earliest copper-based metals (fairly pure
(Timberlake 2003) and Scotland. Old workings copper, sometimes containing small amounts
and hammer stones (Pickin 1990) have been of arsenic) appear in the third quarter of the
discovered during more recent mining and third millennium BC, ie the late Neolithic or

< < Contents 2

 1.2 Iron Age

Iron Age settlement sites generally provide more

evidence for metalworking, and for a wider range
of metals, than Bronze Age sites.

Iron ores, unlike copper ores, are found in many

areas and iron mining and smelting could be
carried out on a small scale almost anywhere in
Britain. No Iron Age iron mines are known, but bog
ores and other surface outcrops were probably
exploited. Several sites have yielded furnaces and
large quantities of iron smelting slag, for example
Brooklands, Surrey (Hanworth and Tomlin 1977),
Welham Bridge, Yorkshire (Halkon and Millett
1999) and Bryn y Castell and Crawcwellt, Gwynedd
(Crew 1986; 1998b).

Evidence for iron smithing is much more

widespread, as at Dragonby, Lincolnshire (May
1996) and Scalloway, Shetland (Sharples 1999).
Figure 1 Iron smithing can also be indicated by cut
fragments of iron stock and hoards of blacksmiths’
Experimental iron working at Plas Tan y Bwlch,
Gwynedd: removing a bloom from a furnace. tools – for example at Waltham Abbey, Essex
[photo © David Starley] (Manning 1991) – while the microstructure of
finished objects provides information about the
smiths’ techniques (Salter and Ehrenreich 1984).
Important information on the use and trade of
Chalcolithic (Allen et al 2012). Bronze (an alloy different types of iron stock can be obtained from
of copper and tin) appears with the early Bronze currency bars, for example the hoard found at
Age in the final quarter of the third millennium Danebury, Hampshire (Cunliffe 1984), and from
BC. Copper alloy artefacts were produced by more rare smithed blooms and billets.
casting and smithing. Fragments of clay moulds
and crucibles have been found on many Bronze Many Iron Age settlement sites have yielded some
Age occupation sites and a few have produced clay mould or crucible fragments for casting
large quantities of these objects, for example copper alloys but a few sites, including Gussage
Dainton, Devon (Needham 1980), Jarlshof, All Saints, Dorset (Wainwright 1979) and Grimsby,
Shetland (Hamilton 1956) and Springfield Lyons, Lincolnshire (Foster 1995), have produced
Essex (Brown and Medlycott 2013), however, finds large assemblages. Coin manufacture can be
of this type are rare in early Bronze Age contexts. demonstrated at a number of oppidum sites, such
The main evidence for bronze smithing can be as Verulamium (St Albans), Hertfordshire (Frere
found in the microstructure of finished tools and 1983), and there was possible silver production at
weapons (Allen et al 1970). Hengistbury Head, Dorset (Northover 1987).

Some evidence for iron working has been found Those parts of Britain that were not within
in contexts that are culturally assigned to the Late the Roman Empire kept Iron Age traditions of
Bronze Age (Collard et al 2006). metalworking although some also incorporated
elements of ‘Roman’ techniques.

< < Contents 3

1.3 Roman all Roman sites, and occasionally blacksmiths’
workshops are found (Buxton and Howard-Davis
A great variety of evidence for Roman 2000; Hammer 2003).
metalworking has been found throughout Britain.
Any substantial excavation of a Roman period site A number of large, circular, stamped copper ingots
is likely to recover some evidence. have been found, particularly in Wales (Kelly
1976), although no evidence of contemporary
Roman sites with large numbers of furnaces and copper mines, furnaces or slag involved in
huge quantities of iron smelting slag have been their production has yet been discovered.
discovered in the Weald of Kent and Sussex Specialised crucibles for brass production have
(Hodgkinson 2008). Other major iron smelting been identified on a few urban sites (Bayley
centres existed in the Forest of Dean (Jackson 1984). Clay moulds and crucible fragments are
2012), Northamptonshire and Lincolnshire relatively common finds on many Roman sites
(Schrüfer-Kolb 2004) but iron smelting evidence and occasionally the evidence is particularly
occurs in many other areas (Fyfe et al 2013; abundant, for example at Castleford (Bayley
Griffith and Weddell 1996; Paynter 2006). Iron and Budd 1998). Stone and metal moulds
smithing slags are routinely discovered on almost (Bayley et al 2001) are also known, but are far

Figure 2
Reconstruction of a Roman workshop, based on
excavated features and finds from Verulamium.
[illustration © Michael Bayley]

< < Contents 4

less common. A number of workshops have at Ramsbury, Wiltshire (Haslam 1980) both
been discovered in which a variety of structures non-tapping and tapping furnaces were found.
and occupation layers have been preserved, Virtually every settlement site will produce at
for example at Caerleon (Zienkiewicz 1993). least small quantities of iron smithing slag and
Where workshop remains are well preserved larger amounts are not uncommon, for example
there is often evidence for a range of both at Deer Park Farms, Antrim (Lynn and McDowell
ferrous and non-ferrous metalworking. 1988) and Coppergate, York (Ottaway 1992).
Metalworking tools are found, both in burials,
The best known evidence for Roman lead for example at Tattershall Thorpe (Hinton 2000),
production consists of large inscribed lead ingots and on settlements, such as Coppergate (Ottaway
although some smelting sites have been identified 1992). The variety of manufacturing techniques
(Page 2005). Large litharge cakes, showing that employed by smiths increased and a much wider
silver was extracted from lead, have also been range of structures, including pattern-welding, are
found in the Mendips (Dunster and Dungworth commonly seen in metallographic studies of iron
2013) and Welsh borders (Bayley and Eckstein artefacts (Blakelock and McDonnell 2007; Gilmour
1998). Small litharge cakes, produced during the and Salter 1998).
extraction of silver from debased alloys, are also
often found on urban sites. A range of non-ferrous metals was widely used
(Bayley 1991b) and evidence for refining, casting
The only evidence for tin mining in the Roman and smithing is common on many types of sites.
period is the occasional inscribed ingot. Tin Examples include urban sites, such as Coppergate,
smelting slag has been recovered (Lawson-Jones York (Bayley 1992) and Armagh (Gaskell Brown
2013) but no furnaces have yet been identified. and Harper 1984), monastic sites, such as
Palaeo-environmental evidence suggests tin Hartlepool, Tyne and Wear (Daniels 1988), and
exploitation from the late 1st century AD to the some other high status centres, for example Dinas
end of the Roman period (Meharg et al 2012). The Powys (Alcock 1963) and Dunadd (Youngs 1989).
casting of pewter is fairly well known from stone Typical finds are small crucibles, cupels, litharge
moulds that have been recovered from both urban cakes, bar ingots, scrap and waste metal. Ingot-
and rural sites (eg Lee 2009). and object-moulds were made from stone, clay
and antler. Crucibles, scrap metal and clay moulds
Roman-period gold mining is known from for small objects are common.
Dolaucothi, Dyfed (Burnham and Burnham 2004).
Parting vessels, for separating silver from gold,
have been found on a few urban sites (Bayley 1991a). 1.5 Medieval

From the medieval period onwards there was an

1.4 Early medieval increasing tendency for some metal industries to
be concentrated in towns, and often in particular
Both urban and rural settlements produce a areas of towns, although iron smithing also took
great variety of evidence for the working of many place in many rural settlements. Bell-casting was
different metals. The finds are not all the same in often, although not always, carried out where
the different cultural areas of the British Isles. the bell was to be used (Dungworth and Maclean
2011). Metal smelting was still carried out near the
A variety of iron smelting technologies, which ore sources (Pickin 2010).
produced distinctive types of slag, were in
use. Large slag blocks have been found at a An important development of this period was the
number of sites, including Mucking, Essex and introduction of water power to operate bellows or
Aylesham, Norfolk (Tylecote 1986, Fig 81), while trip hammers; however, this is poorly understood.

< < Contents 5

Water power was used for at least some bloomery onwards mining sites are often accompanied by
smelting (Mott 1961; Young and Poyner 2012) and engine houses used to provide water pumps and
it was an essential component of the blast furnace lifting gear.
and related fineries and forges introduced at the
end of the 15th century. In the iron industry, blast furnaces, both charcoal-
fuelled and (later) coke-fuelled, are well known
Urban excavations frequently recover evidence for archaeologically (Crossley 1990). Most is known
secondary working of a range of metals (Bayley about charcoal-fuelled furnaces as these have
1996). The scale of metalworking increases in this often been preserved in remote woodland
period and the size of assemblages is often larger, areas. The coke-fuelled furnaces were situated
although the range of finds is similar to that of in areas that remained intensively used into the
earlier periods. This change in scale is particularly 20th century and so are less well preserved. The
noticeable in crucibles whose size increases processes and monuments connected with the
(Fig 36), and large clay moulds for castings such conversion of cast iron into malleable iron (finery,
as cauldrons and bells became common (Richards chafery, puddling, etc) are less often identified
1993) Mass-production also led to changes in and are incompletely understood. Excavation
mould technology. Multi-part clay moulds for has identified cementation (Belford and Ross
casting dozens of objects at one time were 2007) and crucible steel furnaces and a few
developed (Armitage et al 1981) and reusable upstanding monuments are known (Cranstone
limestone piece moulds were made for casting 1997). Bloomery smelting declined in importance
pewter trinkets (eg Margeson 1993, Fig 127). but continued into the 17th century and the best
surviving examples tend to be those in rather
remote locations (Photos-Jones et al 1998).
1.6 Post-medieval
Non-ferrous smelting was initially concentrated
During this period a wide range of both ferrous in areas with good access to suitable ore sources,
and non-ferrous metalworking took place, and however, with the development of the coal-fuelled
technologies evolved rapidly, often with several reverberatory furnace in the late 17th century
complete changes in practice within the period these industries increasingly moved to locations
(Crossley 1990; Day and Tylecote 1991). With with good access to fuel (especially the coal fields).
the increasing separation of ‘industry’ from
agricultural and domestic life, many sites and The archaeological investigation of post-medieval
field monuments become primarily industrial metalworking sites can face significant obstacles.
in function and can be immediately identified The scale at which earth and rock was moved and
as such. This situation is less true, however, of slag produced could be enormous. This usually
craft workshops, small-scale urban industry, resulted in strategies to manage and dispose
and experimental laboratories and workshops. of such waste which need to be considered in
Throughout the period their archaeology remains any excavation. Waste material would often be
poorly understood, even into the 20th century removed from the production site which would
(Brooks 2000; Hull 2003; Martinón-Torres 2012). otherwise be swamped, leaving relatively little
Documentary sources, including maps and plans, material for the archaeologist. Heaps of waste
form an increasingly useful tool for studying the can be located some distance from the original
archaeology of recent metallurgical industries. production site (especially after the introduction
of the railways). Nevertheless, such waste heaps
Physical evidence of post-medieval metal mining can provide useful metallurgical evidence (which
is frequently on a large scale. These sites will might be largely absent from the production
often comprise large heaps of waste rock as well site) but might also bury and preserve earlier
as extensive washing floors. From the 17th century metalworking features and structures.

< < Contents 6

2 Standards and
Good Practice for
Archaeometallurgy
This section sets out the relationship between archaeometallurgy and
other aspects of archaeological projects. The resources provided for
archaeometallurgical remains should reflect the importance of such
evidence. This is most effectively achieved by the appointment of a
suitably qualified/experienced archaeometallurgical specialist.

Summary
Stage Archaeological Archaeometallurgical
Action Action
Initiation Curator identifies need for Respond to any request for input to brief
project and produces brief

Planning Contractor contacts Provide input to Written Scheme of Investigation or Project Design.
specialist Plan excavation and sampling strategy for metalworking features

Fieldwork Survey Identify features located and estimate scale of activity

Excavation Advise on identification of metalworking features. Suggest

sampling strategies. Advise on cleaning and packaging

Assessment Provide information on Assess all (or a sub-set) of the finds in an assemblage in the light
metalworking features and of the archaeological information. Write assessment report, which
debris (spatial distribution should include recommendations for further work (including a
and phasing) methods statement and estimate of time/cost for analysis phase)

Analysis Liaise with specialist(s) Undertake the work identified at the assessment stage. Identify
metalworking processes. Quantify debris by context, phase,
area, etc
Dissemination Incorporate Write archaeometallurgical report(s) which place the activity in
archaeometallurgical a wider social and economic context, for inclusion in excavation
reports into excavation report and/or specialist publication
report

< < Contents 7

Archaeological projects might be initiated for ■■ Start-up (Project Proposal)
many different reasons and the main drivers
are some threat to the archaeology and a ■■ Review 1
desire to better understand that archaeology.
While research will be an essential part of any ■■ Initiation (Project Design, Risk Log)
archaeological project, many are initiated
because the archaeological resource might be ■■ Review 2
at risk due to change in land use and especially
if this involves some construction work. In this ■■ Execution Stage(s)
case the damage to the archaeological record is
mitigated by recording it. Archaeological projects ■■ Data Collection (Survey, Excavation, etc)
are also carried out by higher education bodies ■■ Assessment of potential
and community groups and in these cases the ■■ Data Analysis and production of
main driver is usually an improved understanding Publication Report
of the archaeology.
■■ Review 3 (Updated Project Design)
Many archaeological projects are initiated as
part of the planning process: the development ■■ Closure (End of Project Report)
of a site for new uses is assessed to determine
whether it will potentially impact archaeological The tasks to be undertaken as part of the Project
remains. If the archaeology is deemed sufficiently Execution will vary depending on the nature
important then the recording of the affected of the project. Each phase of a project should
archaeology will be made a condition of the have clear objectives, and these should be
planning permission. The principles are laid regularly reviewed. Archaeometallurgy is an
out in the National Planning Policy Framework integral part of archaeological investigations
and are implemented at the level of local and plans should be made for its inclusion, even
government. Having decided that a site needs in small-scale evaluations, where sites have
some level of recording, the curator produces a archaeometallurgical potential. An experienced
brief for the work (or sometimes an information specialist can provide invaluable advice.
section within the planning condition). Developer
appointed archaeological practices then
respond with a written scheme of investigation. 2.1 Project planning and the
Excavations which are not driven by threats to formulation of research designs
the archaeological record also need project
documentation (eg project design). Before any fieldwork is undertaken the
archaeometallurgical potential of a particular
The successful completion of archaeological site can be anticipated to some extent from a
projects depends on careful planning and consideration of the general nature of the site
implementation. This applies whether the main (see Section 1). Previous work in the locality
driver is threat or research and for both large and will be recorded in the Historic Environment
small projects. The relevant principles are set Records held by the Local Authority and
out in the Management of Research Projects in will provide additional information on the
the Historic Environment (MoRPHE). MoRPHE nature of any archaeometallurgical evidence.
provides a flexible structure which can be adapted Regional Research Frameworks will also contain
to particular situations: information on the range of metalworking that
might be expected (see Section 9). Slags and
other archaeometallurgical finds are frequently
discovered and contractors should approach

< < Contents 8

appropriate specialists at the project planning 3
stages (Start-up and Initiation). Suitable
specialists can contribute to the project proposal
and project design and can help to prepare
excavation strategies. If the site is thought to have
been primarily metallurgical in function, then
archaeometallurgy should be a major aim of the
project design.

2.2 Fieldwork: survey

4
Much can be learnt about metal working sites prior
to, or in the absence of, excavation. Information
is sometimes gained about the types of processes
carried out and the scale of the craft or industry.
The survey methodologies employed will depend,
to a large extent, on the current land use.

Aerial photography is a relatively inexpensive

means of characterising well-preserved industrial
landscapes, such as mining and smelting features
in upland regions that are now under pasture
Figures 3 and 4
(Gerrard 2000).
3 Earthwork survey of the Iron Age slag dumps at
Sherracombe, Devon.
Metric surveys can determine the extent of 4 Iron Age bloomery furnace at Crawcwellt West,
metalworking debris that survives as earthworks, Gwynedd.
and so indicate the scale of metalworking [photo © Peter Crew]

activity (Fig 3). The interpretation of upstanding

metalworking remains from either aerial
photography or from metric survey requires input
from a specialist (Bowden 2000; Cranstone 1994).
2.3 Fieldwork: excavation
Geophysical survey, especially using magnetic
techniques, is often well suited to detecting the Many kinds of metalworking structures and
remains of archaeometallurgical processes. Many debris are distinctive in appearance, and with
slags (in particular iron smithing slags) have experience or training these can be recognised
higher magnetic susceptibilities than topsoil. in the field (Fig 4). Early consultation with a
Both primary (smelting) and secondary (smithing) metalworking specialist and a site visit will
sites will have fired structures such as furnaces enable the evidence to be better understood.
and hearths that can produce strong magnetic The specialist can provide training, sampling
anomalies (see p 62-63 for further details). strategies, put together a site reference
Geochemical surveys (especially using portable collection, and advise on cleaning and packaging
instruments) have considerable potential to procedures. Some knowledge of the relevant
identify and characterise many metalworking sites metalworking processes is greatly advantageous.
(Dungworth et al 2013).
The three metalworking processes most
frequently encountered by archaeologists during

< < Contents 9

fieldwork are iron smithing, iron smelting and the of such sites and the recovery of evidence for
casting of non-ferrous metals (copper alloys, silver early mining requires a range of specialist skills
and gold). (Cranstone 1994; Dutton and Fasham 1994;
Timberlake 2003b; Timberlake and Prag 2005).
Metalworking evidence can be divided
into structures and finds. Structures (and Some of the most useful contexts are those within
archaeological features) include mines, pits, buildings or areas where metalworking was
water channels, dams, buildings, furnaces, practised (primary deposits). More frequently,
and hearths. Finds can include slags, ceramic however, metalworking debris is recovered from
materials, tools, stock metal and metal residues. secondary deposits such as dumps, middens,
The excavation of metalworking sites should pits and ditches, or from where it was used
include the examination of associated features, for surfacing paths (the scale of dumping will
such as domestic dwellings, in order to place the depend on the nature of the metalworking, for
technology in its social and economic context. example iron smelting will produce much more
waste than iron smithing). The excavation of the
2.3.1 Structures and context two types of deposit needs to be approached in
Mine sites will display a range of structures, features slightly different ways, since the type of evidence
and deposits depending on the type of ore being recovered and its interpretation is different.
sought and the methods employed in its extraction.
Some of the earliest evidence for early mining is In primary deposits, metalworking structures
contained in historic records of mining in the 18th (furnaces, hearths and pits) might be
and 19th century and their references to ‘Old Men’s encountered, and the distribution of the residues
Workings’ (Timberlake 2003a). The excavation within a building can be crucial in identifying

Figure 5
Plot of magnetic susceptibility readings, with darker
tones indicating higher values (corresponding to higher
concentrations of hammerscale), within the medieval
smithy at Burton Dasssett, Warwickshire. The building
is 12m long.

< < Contents 10

and separating different activities. For example, reference collection should be established by the
on an iron-smelting site, charcoal production, metalworking specialist. This will form the basis
ore roasting and bloom smithing might also on which all slags and residues will be classified.
have been carried out. The excavation of areas
where metalworking was done requires gridding Slag, ores, crucible and furnace fragments are
and careful sampling, both for hand recovered usually large enough to be easily recognised;
material and soil samples for micro-residues, in some residues, however, are so small that they
particular hammerscale (see p 36). appear only as coloured ‘soil’ deposits. An
example is hammerscale (Figs 5 and 30) which
The dimensions and layout (plans and sections) is so small that it can easily be missed during
of structures should be recorded. Sometimes it trowelling. Nevertheless it can be detected
might be necessary to ‘unpeel’ them layer by layer using a magnet. Soil samples should be
to understand how they were repaired or modified taken from contexts containing hammerscale,
during use. The relationships between furnaces particularly primary contexts. A workshop
or hearths and other features (buildings, pits, etc) floor surface comprising a single context
should also be carefully recorded. It is possible should be sampled throughout (at 0.1–0.5m
that waist-high or above-ground hearths existed intervals) in order to examine the distribution
but do not survive. It is sometimes possible, of hammerscale. A 0.2 litre sample is adequate
however, to reconstruct their positions from an for magnetic susceptibility screening and
examination of the distribution of metalworking quantification of hammerscale, as at Burton
debris (Fig 5). Dassett (Fig 5; Mills and McDonnell 1992).
Samples should also be taken from contexts
Secondary deposits are contemporary with or spatially and chronologically removed from
later than the metalworking activity that produced the iron-working areas, for comparison.
the debris. Careful recording of the residues can
indicate the direction from which the material was All charcoal associated with metalworking
dumped, and so suggest where the metalworking features and debris should be collected for
activity was located. Large features often contain species identification and tree age – this can
larger, and therefore more representative, provide important evidence on the management
deposits of metalworking debris. Where a process and exploitation of wood resources for
produced large quantities of slag this might have metalworking. Radiocarbon samples should
all have been dumped some distance from the be processed in the usual manner to avoid
hearths or furnaces. The proportion of features contamination (Historic England 2015a).
left unexcavated should be recorded to provide a
means to estimate the total quantity of slag. The identification of metalworking finds and
debris usually requires that they are cleaned.
2.3.2 Finds and sampling Some materials, however, are delicate and can be
Finds include ores, slags, fragments of hearth or damaged; any cleaning procedures must be agreed
furnace structure, crucibles, moulds, metal stock, with the metalworking specialist and/or conservator
scrap and waste, and iron or stone metalworking (English Heritage 2008b). Materials that should not
tools (hammers, tongs, etc). Three-dimensional be washed (except by, or under the supervision of,
recording of bulk finds, such as slags, is not the metalworking specialist) include crucibles,
usually feasible or desirable, but crucibles, scrap moulds, hearth and furnace linings.
metal, etc should be treated as ‘registered finds’.
Sampling strategies should be tailored to the size Some minerals and metal production
and nature of the debris recovered. Best practice waste are toxic; those handling or cleaning
is to initially retain all excavated bulk finds and these materials should complete risk
soil samples. Where circumstances permit, a site assessments and/or COSHH assessments.

< < Contents 11

Bulk finds, such as slag, should be packaged produced, often remained virtually unchanged
in tubs or heavy-grade plastic bags. Woven for very long periods. Therefore dating is most
polythene bags can provide a useful alternative commonly achieved by using associated material
to conventional plastic bags. In most cases culture, radiocarbon dating, dendrochronology
bulk finds are extremely robust and do not and archaeomagnetic dating. Mining and smelting
require specialised storage conditions. sites have often yield very little datable material
Slags with a high metallic iron content (test culture, although this might, in part, be due to a
by magnet), however, should be treated as focus on the obviously ‘technological’ aspects of
metal finds, ie stored under conditions of such sites (hearths, furnaces, slags heaps, etc).
low relative humidity (English Heritage 2013). The excavation of ancillary areas should increase
Debris recorded as ‘registered finds’ should the recovery of datable artefacts. Most metalworking
be packaged individually and particular care activities made use of charcoal fuel that can be
should be taken with delicate materials, such radiocarbon dated. Samples should be of clean,
as ceramic moulds. All debris must be kept, for short-lived charcoal (Historic England 2015a).
examination by a metalworking specialist. Waterlogged metalworking sites (especially mines
and sites that used water-power) can yield timbers
2.3.3 Dating that can be dated using dendrochronology
The date of the archaeometallurgical activity (English Heritage 1998). The final use of fired clay
on a particular site will affect its significance. structures, such as hearths and furnaces, can be
It is not currently possible to date slag directly. dated archaeomagnetically (see p 63).
Metallurgical processes, and the debris they

Figure 6
Plan of the excavated features at the Roman site of
Shepton Mallet, Somerset where iron smelting (yellow)
and smithing (red) were taking place. Note the partial
spatial separation of the two activities.

< < Contents 12

2.3.4 Site archive available. Metal and fired clay objects – such as
The product of the fieldwork phase of the project ingots, billets, bar stock, scrap, waste, unfinished
is the site archive, which should include all the artefacts, metalworking tools, crucibles and
fieldwork data and a brief statement of the nature moulds – are particularly important.
of the stratigraphic, artefactual and environmental
record and finds (Brown 2007; Cool et al 1993). The metalworking specialist will make an
assessment of the archaeological value of the
metalworking evidence, which is dependent on
2.4 Assessment of potential for a number of factors. The most important is the
analysis current state of knowledge of that metalworking
process (eg Bayley et al 2008). For example,
Detailed analysis of the records and materials evidence for medieval or earlier copper smelting
recovered during archaeological fieldwork must in England is extremely limited, so any early
be preceded by an assessment. An assessment smelting is important. At some periods, some
report contains a summary of the data and a processes are relatively well known and such
statement of the potential for that data to address sites would be particularly important only where
the original project aims and objectives. It should primary deposits survive in good condition. The
contain specific recommendations for further specialist will note any important or unique
work, storage and curation. This phase is also an features of the excavation record and recovered
opportunity to update the project design in the finds and debris. The site should be compared
light of the discoveries made. It is important that with other broadly contemporary sites locally,
all the evidence for metalworking is considered as regionally and nationally.
a whole and where possible all relevant material
remains should be seen by a single specialist (where This information will enable an assessment to
this is not possible provision should be made for be made of the significance of the evidence and
different specialists to share data and results). of the requirements for the analysis phase. The
assessment report should set out the procedures
The metalworking specialist will classify the for further work and specify any scientific analysis
debris into different types depending on relatively required (chemical analysis, micro-structural
simple characteristics (colour, density, size, shape, examination, etc). The specialist will also be able
surface morphology, etc – see Sections 3-6). Many to advise where the evidence for metalworking
of the recognisable types of debris are diagnostic does not justify further work.
of particular processes. In addition, the total
quantity of debris should be determined. For
large assemblages of metalworking debris, the 2.5 Analysis and report preparation
assessment can be carried out on a sub-sample
of the available material. The sub-sample should The analysis phase consists of the examination
include examples of all the different types of of the records and materials identified during
artefacts, and debris, recovered, and should also the assessment phase, and the production of
reflect the full range of contexts excavated. The a publication text that reflects the importance
selection of a sub-sample should be agreed with of the results. The analysis phase can provide
the metalworking specialist. information on the range of metals worked, the
technologies used, the social and economic
It is extremely important that the metalworking importance of these activities, trade and
specialist is provided with a brief summary of exchange, and cultural affinities.
the site, including stratigraphic and contextual
data. Information on related features and finds The metalworking specialist will provide reports
assessed by other specialists should be made on features and/or groups of material that have

< < Contents 13

been identified as having potential for analysis metalworking waste. The number of fragments
and that are linked to specific objectives in the of waste will often be influenced by taphonomic
updated project design. All metalworking debris factors (and the method of excavation) which will
must be made available to the specialist for reveal little about the metalworking process or its
study during the analysis phase of a project. The social and economic importance.
entire assemblage should be visually examined,
classified and identified as far as is possible (see Achieving a reliable estimate of the total quantity
Sections 3-6). The finds should be weighed and/ of debris present in any partly excavated, or
or counted and recorded by context. Dimensions unexcavated large feature (such as a slag heap)
should be recorded where appropriate – for is difficult, but can indicate the scale of activity
example diameters and depths of furnace or hearth on the site. The volume of the features should be
cakes, size of crucibles, diameter of hole in tuyère estimated and the proportion of slag determined.
mouths or blowing holes. The evidence should be The proportion of slag within a context might
compared with the stratigraphic record in order vary considerably between different features and
to examine spatial and chronological patterns in sites and can best be determined by excavating a
metalworking activities (see Figs 6 and 7). section. The total volume of the context/feature
(in cubic metres) should be multiplied by the
2.5.1 Quantification bulk density of the slag (varies depending on the
It is essential that the metalworking evidence is nature of the slag and context) to give the total
quantified by type and archaeological context. weight in tonnes.
It might in some cases be appropriate to record
the material by numbers of fragments, eg large The quantity of metalworking evidence recovered
fragments of ceramic mould, metal off-cuts, can be used to provide data on resource
and moderately complete examples of crucibles exploitation, such as labour required for charcoal
(or when the total quantity of material is very production and woodland management.
small). In most cases, however, weight provides Assessing resource implications depends on the
a more practical and useful quantification of accurate quantification of diagnostic debris, a full
understanding of the metallurgical process and
the precise nature of debris (ore, slag, charcoal,
etc). Bloomery iron working is currently the only
process that is sufficiently well understood for
such analyses to be possible. The ratios of ore,
charcoal, slag and bloom have been explored
through experimental reconstructions of iron
smelting and smithing (eg Crew 1991; Sauder
and Williams 2002). In one experiment (XP27,
smelting a phosphorous-rich bog ore in a low,
non-slag tapping shaft furnace, Crew 1991),
7.6kg of bog ore was smelted and yielded a
1.7kg bloom of iron. This was then smithed into
a 0.45kg bar and the whole process required
61kg of charcoal and produced 6.1kg of slag.
Sauder and Williams (2002) have succeeded in
producing blooms of ~13kg from ~40kg of ore
Figure 7
(and ~90kg of charcoal) and the forging of these
blooms has yielded bars of ~5kg. The ratios of
The histogram shows the proportions of different
types of slag for each phase of occupation at medieval raw materials, waste and finished product are
Wigmore Castle, Hereford and Worcester. likely to vary considerably depending on the type

< < Contents 14

and quality of ore, the technology used and the range of elements that can be detected. Some
skills of the metalworkers (Crew 2013). A certain types of chemical analysis are quantitative,
amount of information on these variables can be providing precise information about composition
gained from chemical and mineralogical analyses in percentages or parts per million; others give
of representative samples of ore, slag and qualitative results, identifying the main elements
charcoal. Such analyses can be integrated with an or compounds present, and provide a rough idea
examination of the wider landscape and its use of relative concentrations. Some methods require
(eg Fyfe et al 2013; Mighall and Chambers 1997; small samples that will be destroyed by the
Crew and Mighall 2013). analysis, but in other cases surface analysis can
be performed without damage to the artefact.
2.5.2 Scientific techniques
In order to fully understand the available Analysis contributes to an understanding of
metalworking residues a specialist might employ potential sources of metal (Paynter 2006; Wilson
a range of scientific techniques, including physical and Pollard 2001), and the prevalence of recycling
and chemical analytical methods to determine (Bray and Pollard 2012; Caple 2010). It has been
a range of properties, such as chemical or used to revise typological classifications of
mineralogical composition, melting point, density, artefacts (Bayley and Butcher 2004), and has shed
etc (see p 62–68). This should only be carried out, light on the ways in which copper alloys reflect
however, where there is a specific archaeological wider processes in society such as Romanisation
question that has been identified in the updated (Dungworth 1997).
project design that is likely to be answered by
scientific techniques.
2.6 Dissemination
The scientific techniques to be employed depend
on the nature of the material to be analysed and The dissemination of the results of analytical
the questions that have been asked. In some work should reflect the importance of those
cases there might only be one technique which results. In many cases the results should be
can address a particular question while in integrated into the excavation report. The
others several techniques might be available. format and approximate length of reports
Understanding the fabrication of metal artefacts should be agreed before work is started. In some
usually requires some understanding of the metal’s circumstances it might be appropriate to publish
microstructure (metallography). Metallography archaeometallurgical data separately (with a
has been applied to a range of artefacts to show summary in the excavation report). In some
the wide variety of techniques used by early projects, dissemination can also be through
metalworkers (eg Allen et al 1970; Ottaway 1992; temporary or permanent displays in a museum.
Tylecote and Gilmour 1986; Wilthew 1987).
Strategies for the storage of metalworking debris
Chemical analysis of metals and metalworking need to be flexible and take into account the size
debris are frequently used to understand the and significance of the assemblage. Deposition of
nature of the metals and the processes which the material evidence should pose no problems
produced them. Before chemical analysis is if the excavation and sampling strategies have
undertaken a consideration needs to be made of been agreed in advance by the excavator, the
the nature of the question being asked, aspects of specialist and the museum. A full copy of all data
the analytical techniques (including cost) and the produced must be supplied for inclusion in the
damage (if any) to the artefact or waste material site research archive (Brown 2007; Museums and
to be analysed. Analytical techniques vary Galleries Commission 1992; Society of Museum
widely in terms of the degree of quantification, Archaeologists 1993).
accuracy, precision, detection limits and the

< < Contents 15

3 Archaeometallurgical
Processes and Finds:
Iron and its alloys

3.1 Background The presence of small amounts of carbon in iron

can produce a range of different alloys. Steel is
Iron (Fe) is the fourth most abundant element in an alloy of iron which contains moderate levels of
the earth’s crust. Iron ore suitable for smelting carbon (typically 0.3–1% carbon). Nevertheless,
occurs in many locations, so archaeological the distinction between plain iron and low-carbon
evidence for smelting is geographically widespread. steels is not clear, especially for bloomery iron/
The methods of producing iron and its alloys, and steel. Steel is an ideal material for cutting edges
the extent to which the alloys were used, changed on tools and weapons because its strength and
with time. Several different alloys of iron have hardness can be manipulated by a combination of
been used and these have distinct properties. quenching and tempering. Steel can be produced
during smelting, owing to the presence of the
Plain iron is very pure: it contains less than 0.1% carbon-rich fuel, or afterwards, by heating the
of other elements. It is often described as ferritic iron in the presence of a carbon-rich material,
iron because structurally it is made up of many such as charcoal. Higher levels of carbon (2–5%)
crystals of a type known as ferrite. Its melting give rise to an alloy (cast iron) which has a much
temperature is extremely high, about 1545°C, so lower melting temperature (1150–1300°C) but is
rather than being melted it was forged into shape. brittle when solid making it unsuitable for forging
Alloys of iron melt at lower temperatures than in a blacksmith’s workshop. Iron and steel can be
plain or ferritic iron and have different properties. obtained by melting cast iron and removing some
Early plain iron was made using the bloomery or all of the carbon.
process. During smelting the metal never melted
and so it is typically heterogeneous and a mixture Phosphoric iron contains up to 1% phosphorus,
of alloys can be present in one object. Plain iron which makes it harder than plain iron. The
was also the product of conversion processes phosphorus enters the metal from the ore during
(fining or puddling) which accompany the blast smelting. Due to the nature of most British
furnace and was traditionally known as wrought iron ores, phosphoric iron is abundant in most
iron. These two types of plain iron (bloomery and periods before the 19th century. The presence
wrought) can be distinguished from each other of phosphorus in iron influences the uptake and
using scientific techniques to characterise the distribution of carbon, and phosphoric iron and
tiny fragments of slag trapped in the metal ores were sometimes selected or avoided for
(slag inclusions). specific applications.

< < Contents 16

 Iron in summary
Plain iron contains less than 0.1% of other elements and is often known as ferritic iron and sometimes as wrought iron.
It has a melting temperature of 1545°C. Alloys include steel (~ 0.3 to just over 1% carbon), phosphoric iron (up to1%
phosphorous), low carbon iron (up to 0.3% carbon), and pig or cast iron (~ 2 to 5% carbon). Cast iron describes a carbon-
rich iron alloy which has a relatively low melting temperature (~ 1150–1300°C); cast iron straight from the blast furnace was
usually called pig iron.

Process Description Archaeological debris

Bloomery An inhomogeneous solid bloom of metal was Fuel, ore, vitrified furnace lining and slag.
smelting produced, as the metal did not melt during Usually large amounts of slag will be
(8th C BC – the process. The main product of these recovered, including tap slag or large slag
16th C AD and furnaces was plain or ferritic iron but other blocks. The bases of furnaces and tapping
later in some alloys were commonly produced as well. The pits sometimes survive. Hammerscale
areas) impurities present in the ore reacted with some can also be found if the iron bloom was
of the iron oxide to form iron-rich slags. consolidated on the smelting site. There is
sometimes later evidence for waterpower.

Blast furnace These furnaces operated at higher temperatures and Ore and fuel. Large quantities of blast
smelting produced a form of liquid iron. The iron was cast on furnace slag were produced. The furnace
(15th C AD a bed of sand to produce bars of pig iron or ingots. rarely survives to its full height. Remains
onwards) The slag is silica rich and contains little iron. of associated buildings, possibly with
casting pits or mould fragments. Evidence
of waterpower should be expected.

Finery forge Pig iron from a blast furnace was fined in a finery Rarely pig iron feedstock, but more commonly
(end 15th to forge, commonly with two finery hearths, a chafery waste slag which is iron rich and can closely
early 19th C hearth (for reheating) and a helve hammer. The resemble many forms of bloomery iron
AD) product was plain bar iron (wrought iron). smelting slag. Hammerscale may be found.

Puddling Pig iron from a blast furnace was melted in a running Rarely pig iron feedstock, but more commonly
(1790s to out furnace to make refined iron (or finers metal). waste slag which is iron rich and can
1960s) This was remelted in a (reverberatory) puddling closely resemble some forms of bloomery
furnace. A hammer was then used to shingle iron smelting slag (especially tap slag).
(consolidate) the puddled ball. A rolling mill then Hammerscale (millscale) may also be found.
rolled it into wrought iron bars.

Smithing Most iron alloys were shaped, by smithing or forging, Smithing hearth cakes, hammerscale and
while solid. The metal was heated and then shaped vitrified hearth lining. Ground level hearths
or welded. might survive. Evidence of waterpower might
be found on later sites.

Steel Steel could be made in several different ways. Evidence of early steel production is in the
Bloomery furnaces could be used to produce form of objects, bars, billets or blooms
blooms of steel. Plain or wrought iron could be containing steel.
converted into steel by carburisation. The steel bars
were often welded together and forged to improve
the quality. From the 1740s, some blister steel was
melted in crucible furnaces to make homogenous
crucible steel.

Casting Pig iron could be tapped straight from the blast The most commonly used moulding material
furnace into moulds to produce objects. Cast iron was sand and this is rarely recognised as such.
could also be re-melted in purpose built furnaces Cupolas were often designed in sections to
(reverberatory furnaces or cupolas). make them moveable.

< < Contents 17

3.2 Smelting smelt and changes the colour of the ore (Fig 8).
Ores could also be sorted, washed and broken up
The bloomery and blast furnace processes to reduce the proportion of impurities, collectively
were the two main methods of smelting iron. In known as gangue, which entered the furnace.
England bloomery iron smelting was probably Crushing the ore would increase its surface area
introduced in the 8th century BC (there is some and hence the rate of reaction, although if ore
evidence for earlier smithing (Collard et al is crushed too finely the particles could clog the
2006), but not yet for smelting of this date) and furnace. Small particles, known as ore fines,
continued in use until the 16th century AD – and are found in areas where the ore was roasted,
later in some areas – when it was superseded crushed or stored, and sometimes in and around
by the blast furnace process. The temperatures furnace structures.
achieved during the bloomery process did not
usually exceed 1250°C, which is well below the 3.2.1 The bloomery process
melting point of plain iron (and low carbon and Although the historical evidence for early iron
phosphorus alloys). Therefore the metal does not smelting in Britain is strongest for just a few areas
melt during the process. The bloomery process (eg the Forest of Dean and the Weald of Surrey
is sometimes referred to as the Direct Method of and Sussex), archaeology has frequently shown
forgeable iron production because it produced,
in a single process, iron alloys (possibly including
steel) that could be forged by a smith.

In contrast, blast furnaces, introduced to Britain

c1500AD, produced cast iron. The lower melting
temperature of this alloy meant that the furnace
produced molten metal, which was cast to shape.
Cast iron was brittle, however, and not suitable
for all applications. Refining processes had to
be used to convert it into tougher, forgeable iron
alloys when this was required. For this reason
blast furnace smelting, and the subsequent
refining, is sometimes referred to as an Indirect
Method of forgeable iron production.

Bog ore was probably a major source of iron Figure 8

ore, especially for the bloomery process. It is Iron ores vary in colour and can be difficult to spot,
formed by the precipitation of iron compounds, particularly if they have not been roasted, as they do
in lakes, bogs and other poorly drained locations, not necessarily have a strong colour or high density.
Unroasted ores can be black, red, brown or orange
and could simply be dug out. Other recognised
while roasted ores are commonly red, purple or orange,
sources of high quality iron ore include limonite because they are oxidised. Ore fines are small particles
(hydrated iron oxide), siderite (iron carbonate) of roasted ore that sometimes respond to a magnet and
and haematite (iron oxide), and these were have high magnetic susceptibility. Pieces of reduced
ore, sometimes partially slagged, are sometimes found
extracted by mining. From the medieval period
among the debris from the bottom of the furnace, and
onwards, the iron ores found in the Coal Measures these are commonly grey. The minerals present in iron
became increasingly important (Challis 2002; ores can be determined using X-ray diffraction, and the
Dungworth 2010). Raw, or untreated, ores rarely iron content can be determined by chemical analysis (p
66). The ores recovered during archaeological fieldwork
occur in any quantity on archaeological sites.
need not be representative of the ores smelted because,
Iron ores would usually have been roasted before for example, they might have been discarded because
being smelted. Roasting makes the ore easier to they were of poor quality.

< < Contents 18

that bloomery iron smelting was widespread not provide a reducing atmosphere which would
(Crew 1986; 1998b; Dungworth 2011; Dungworth transform iron ore into metallic iron.
and Mepham 2012; Halkon and Millet 1999;
Paynter 2007a). The bloomery process required Furnaces rarely survive to their full height (Fig 9),
the production of a fluid slag (typically at so their likely structure and mode of operation
temperatures around 1200°C) and so the slag is have been reconstructed by supplementing
usually iron-rich (there is no evidence that any the archaeological evidence with a detailed
additional fluxes were used in bloomery iron examination of waste slag (Paynter 2007a),
smelting). Therefore, although a variety of ore ethnographic data and experimental work (eg
sources were employed, most of these were Crew 1991; 2013; Girbal 2013; McDonnell 2013a;
fairly iron rich. At least some bloomery iron Smith 2013). Furnaces were constructed from
smelting took place on a fairly modest scale and clay, although some stone and tile were
could use ore sources that were small enough to occasionally used. The clay was often modified
have escaped the attention of later industrialists with large amounts of temper, especially sand
and geologists. but also small stones, pieces of slag and possibly
organic material.
Charcoal was exclusively used as the fuel for
bloomery smelting. Coal could not be used as The form and size of bloomery furnaces varies
most contains sulphur, which would be absorbed considerably and their partial survival probably
by the iron, making it brittle and unsuitable for masks even more variation. In a few rare cases,
forging. There are no known charcoal production furnaces were constructed against a bank and
sites prior to the medieval period, but at early have survived to a height of 1m (Fig 9). In most
sites charcoal might have been made in small cases very little superstructure survives and it is
pits adjacent to furnaces, as observed in other usually difficult to know the original height of the
parts of Europe. Wood was unsuitable as it could furnace with certainty (Fig 4).

9 10

Figure 9 Figure 10
Section through a late Iron Age furnace at Stockbury, Experimental bloomery furnace reconstruction
Kent. The furnace is built of clay which has a (external diameter 0.7m). The tapping arch, through
characteristic orange-red, oxidised-fired outer surface which the liquid slag enters the tapping pit, can be
with a grey-black inner surface, which displays some seen at the front of the furnace. Cf Figs 1 and 4.
vitrification. Scale bars are 1m and 2m in length.
[photo © Kent Archaeological Projects and supplied by
Tim Allen]

< < Contents 19

 A Two types of reaction occurred in the furnace: the
iron ore was reduced to metallic (but solid) iron
and the impurities in the ore reacted to produce
a fluid slag. The high temperatures needed for
these reactions (and a fluid slag) were provided
by the combustion of fuel. The use of charcoal
also provided the reducing atmosphere which
ensured that iron ore was reduced into metallic
iron. The furnace was heated and charges of
roasted ore and charcoal added to the top. In
most cases extra air was needed to provide a high
enough temperature. Furnaces are known with
small holes near the base through which air could
be introduced (Fig 11). It is likely that bellows
were used to blow air into the furnace, however,
these do not survive in the archaeological record
B (probably because they were made of wood
and leather). As the ore descended through
the burning charcoal fuel, some of it would be
reduced into a lump of metal. The bloomery
furnace usually operated bellow the melting
temperature of iron and so the iron formed a
spongy mass or bloom, usually just below the
hole through which air was introduced

The formation of slag inside the furnace from the

reaction of gangue minerals had to be carefully
managed: if too much slag accumulated then it
would block the flow of air through the furnace.
The ways in which the smelters managed the
formation of slag can be reconstructed from
furnace architecture. Where a furnace has been
heavily truncated it might still be possible to
reconstruct both the form of the furnace and
the ways in which the slag was managed from
Figure 11
the nature of the slag (Dungworth 2011; Paynter
Vitrified clay lining with a blowing hole, from the
Roman site at Ribchester, Lancashire. Vitrified furnace 2007a). Slag management can be divided into
lining is produced by a high temperature reaction two main categories. In the first case, most of
between the clay lining of the furnace and the alkaline the slag was periodically removed from the
fuel ashes or slag. The outer parts are usually orange furnace by opening the furnace near the base
(oxidised-fired) ceramic, while the inner zone is grey or
black (reduced-fired) and often vesicular with a glassy and letting molten slag flow out (tapping,
surface. Furnace linings might have been repaired see Fig 16), while in the second case the slag
repeatedly or replaced, and can show a sequence of was allowed to accumulate in the furnace
vitrified layers. Although furnace walls were relatively (often in a pit at the base of the furnace,
thick, usually only the inner surface survives, or is
noticed, as the heat of the furnace will not have fired see Fig 14). Nevertheless, these distinctions
the outer part. The hottest area of the furnace was near are not always clear cut and considerable
the blowing hole (see photograph), and consequently variation existed within each category.
vitrified clay lining containing the preserved outline of
the hole is often recovered.

< < Contents 20

In the hottest zone of the furnace, near the 12
blowing holes, the temperature was probably
slightly in excess of 1250°C. Here the liquefied
slag separated from the solid iron metal and
flowed to the bottom of the furnace. The metallic
iron coalesced and eventually formed a spongy
lump known as a bloom. The bloom usually
attached to the furnace wall just below the
blowing hole and grew until it started to interfere
with the air blast, at which stage smelting was
stopped. The bloom was then removed, either
through the top of the furnace (Fig 1) or through 5 cm
the tapping arch. Since the iron did not melt
during the process, the bloom contained a lot
of trapped slag and was usually compositionally 13
heterogeneous. Therefore, although the main
product of bloomery furnaces was plain iron,
the blooms commonly included regions of other
alloys as well, such as steel and phosphoric iron.

The main aim of bloomery smelting was the

production of a lump or bloom of iron (Figs 12
and 13). The smelting took place below the
melting temperature of the iron and so the iron
formed in the solid state as a spongey mass, or
bloom. Blooms very rarely survive in the state
in which they were extracted from the furnace.
The blooms would often have been smithed
into billets or bars for transport and trade and
examples of these are fairly well known (Fig 13;
Crew 2013). During the smithing of a bloom
small fragments (often referred to as gromps)
could easily become detached, and these are Figures 12 and 13
occasionally found in smelting slag assemblages. 12. Iron blooms are rare finds on archaeological sites:
here an ethnographic example is shown. Blooms are
The preservation of prehistoric bloomery smelting made up of many small particles of iron coalesced
into a spongy lump. They are often badly corroded
furnaces has generally been rather poor – very and fragmentary but are strongly magnetic.
few survive as little more than burnt patches 13. Partially consolidated blooms, billets and bars
with traces of clay wall (Fig 4), usually ~0.3m in of iron. They vary in shape and size, are often
diameter. The original height of these furnaces is badly corroded and fragmentary, but have a high
density and are strongly magnetic. Top: Three
uncertain and some might simply be the truncated bun-shaped blooms from Dean, 2–2.5kg, probably
remains of shaft furnaces (Cleere 1972; Dungworth Roman. Centre left: Square-section Roman bars,
2011; Girbal 2013; Tylecote 1986, 140). Prehistoric with the typical diagonal facets from hot-cutting of
bloomery smelting slag takes a variety of forms longer bars, from Houghton Down, Hampshire and
Hangerbury Hill, Dean. Centre right: Cuboid billet
but generally includes little or no tap slag. In of second century date from Banc y Coed, near
most cases the slag was allowed to accumulate Ffestiniog 1.9kg. Bottom: An early 12th-century bar
within the base of the furnace (or a pit at the base (possibly a “piece”) from Winchester, Hampshire,
of the furnace) as a large slag cake or furnace 1.3kg. Scales in cm.
[Photo © Peter Crew]

< < Contents 21

bottom: >0.3m diameter and 10kg or more in disparate and fragmentary. In southern and
weight (Fig 14, Paynter 2007a). Slag cakes (and eastern England the recovery of large smelting
fragments of slag cakes) often have impressions slag blocks (McDonnell 1993) indicates the use
of charcoal and other organic material. Organic of slag-pit furnaces of a kind that is well known
material could have been used to pack the pit from northern and central Europe (Pleiner 2000,
at the base of the furnace. In some cases such 149–162). These generally date to the early Anglo-
slag can contain such an abundance of charcoal Saxon period and it is tempting to see this as
impressions that it has a rather loose and porous a technology imported by immigrants. Broadly
structure (Dungworth 2011; Tylecote 1986, 137). similar smelting which made use of non-tapping
Some sites have also yielded quantities of slag furnaces and yielded furnace bottoms or slag
prills or small flows of slag (Fig 15) where the cakes is known from Millbrook, Sussex (Tebbutt
direction of flow is vertical (or at least fairly steep) 1982), Burlescomb, Devon (Reed et al 2006),
– such flows appear to have occurred within the Clearwater, Gloucestershire (Pine et al 2009) and
furnace (Dungworth 2011). The quantities of slag Ramsbury, Wiltshire (Haslam 1980). The use of
found on prehistoric bloomery smelting sites furnaces with provision for tapping slag is also
is often rather small, suggesting production to known from the late Saxon period at Ramsbury
satisfy a low level of demand. The size of iron (Haslam 1980, see Bowyer and Keys 2013 for a
billets and currency bars suggests that a smelt re-examination of the dating), Stamford (Burchard
would typically yield around 2kg of smithed metal 1982) and West Runton, Norfolk (Tylecote 1967).
(Crew 2013).
Slag tapping from shaft furnaces became the
Bloomery iron smelting is widespread in the most frequent method of bloomery iron smelting
Roman period and some regions saw quite through the rest of the medieval period. Although
intensive metal production (Hodgkinson 2008; a good deal of slag was tapped from the furnace
Jackson 2012; Schrüfer-Kolb 2004). At least some (eg Crew and Charlton 2007), there are sites
of the furnaces survive to a height of 1m and these where a proportion of slag was also allowed to
are most commonly cylindrical with an internal accumulate at the base of the furnace, eg the
diameter close to 0.3m – examples are also known rather porous slag cakes or furnace bottoms from
with walls that taper inwards (Fig 9). Roman the 14th–15th-century bloomery at Minepit Wood,
iron smelting sites are generally characterised Sussex (Money 1971, 105). A similar pattern can
by the presence of tap slag (Fig 16), however, also be seen in the slag from the 13th-century
there are sites where at least a proportion of the bloomery at Stanley Grange, Derbyshire (Challis
slag appears to have remained as slag cakes or 2002). The size and shape of medieval furnaces,
furnace bottoms in the furnace (Fulford and Allen and the provision of air blast, is rather poorly
1992). Some Roman bloomery smelting sites yield known as relatively few have been excavated and
small quantities of slag and appear to represent these have often been severely truncated. In many
the continuation of some prehistoric modes of respects the furnaces and associated slags of this
production. Some sites, however, have very large period resemble those of the Roman period.
slag heaps associated with them, suggesting a
high level of production to meet non-local needs Documentary sources suggest that there were
(Hodgkinson 2008). A few Roman period sites significant developments in smelting technology
have indicated that larger furnaces (up to 1m in in the medieval period (Tylecote 1986, 188-9).
diameter) were used for bloomery iron smelting From the 13th century onwards, terms such as
(eg Crew 1998a). The few Roman iron billets that mill are increasingly used to refer to ironworking
survive display a range of weights within the range sites. This suggests that water power was being
2–15kg (Crew 2013). harnessed for bloomery iron smelting. From
later practice it is known that water power
There is evidence for iron production in the was ultimately used to drive both bellows and
early medieval period, however, this is often hammers, however, the earliest applications of

< < Contents 22

 14 15

16

18

17

Figures 14–18
14. Section through a prehistoric furnace bottom. of a few hundred grams to accumulations weighing
Furnace bottoms are dense, dark-coloured slags 10kg or more. Hot, fluid slag can also form long,
that solidified in the furnace and can retain the thin runs.
shape of the furnace base, sometimes with part of 17. Frothy tap slag from a late bloomery at Goscote,
the baked clay structure attached. Furnace bottoms West Midlands. The section at the front reveals
are of varied size (some are as small as 0.3m in the high porosity of this slag. Although this type of
diameter while others exceed 0.5m), and will slag has increasingly been noted from sites which
often contain pieces of reduced ore and fuel. The appear to have made use of water power, it has also
smaller varieties are sometimes referred to as slag been noted from bloomeries where water power
cakes, with furnace bottom reserved for the larger was not used. The frothy texture suggests high air
examples (scale = 0.15m). pressure inside the furnace while the slag was liquid
15. Slag prills resemble very small pieces of tap slag but a relatively quick solidification when the slag
(cf Figure 16) but usually display more evidence for was tapped, which would prevent the escape of the
vertical (rather than close to horizontal) flow. gas which had been dissolved in the slag).
16. Tap slag has a characteristic shape, resembling 18. Undiagnostic slags (from Ribchester) are small
a flow of lava, with rivulets of slag on the upper or fractured pieces of slag that have the dark
surface and a rough under surface which can have colour of iron-rich slags, but do not have any
adhering sand or clay. Tap slag is dense with few diagnostic surface morphology. Therefore, although
relatively large bubbles, as it flows out while hot indicative of iron-working, they cannot be used to
and fluid. It is dark in colour, usually grey to black, distinguish between smithing and smelting. They
sometimes with a liverish or maroon upper surface. are sometimes the largest proportion of slags in an
The size of tap slags can vary from individual runs assemblage.

< < Contents 23

water power are not well understood. During the wall smelting huts have been found on prehistoric
14th and 15th centuries the size of blooms rapidly sites (Crew 1998b) and large, square post-built
increased from 10–20kg to around 100kg and this shelters are known on Roman (Hammer 2003;
could be due in part to the application of water Paynter 2007b) medieval sites (eg Money 1971).
power. While a few water-powered bloomery sites
have been archaeologically investigated (Crossley 3.2.2 The blast furnace
and Ashurst 1968; Dungworth 2010; Tylecote Documentary evidence suggests that blast
1960; Young and Poyner 2012) many questions furnaces were introduced to this country towards
remain. Tylecote (1960) and Dungworth (2010) the end of the 15th century (Awty and Whittick
noted especially porous or frothy tap slag (Fig 17) 2002; Crossley 1990). The blast furnace process
at putative water-powered sites, however, similar was fundamentally different to the bloomery
material has also been discovered on slightly process and over the next two centuries it
earlier sites with no evidence of water power gradually replaced the earlier process.
(Young and Poyner 2012). The blast furnace was substantially larger than
the bloomery furnace: the early examples were
Some later medieval bloomery sites have yielded stone-built, tower-like structures, 5–6.5m square
small amounts of slightly greenish slag which and probably over 6m high (Crossley 1990, 158).
contain relatively low levels of iron (Dungworth The height of the furnace allowed the production
2010; Money 1971). This slag shows some of a more reducing atmosphere which could
similarities to the blast furnace slag (see below) produce an iron-carbon alloy, ie cast iron, which
and suggests that late bloomeries were on would melt (Fig 19). Blast furnaces were provided
occasion run at higher than usual temperatures with large bellows which were powered by water
and with a more reducing atmosphere. wheels. This gave a powerful air blast allowing
higher temperatures to be reached. The reducing
While bloomery smelting slag takes on many atmosphere and the high temperature meant that
different forms (Figs 14–18) there is relatively more of the iron was extracted from the ore and
little variation in its chemical and mineralogical the resulting slag was rich in silica (Fig 21). Blast
make up. Most bloomery smelting slag contains a furnaces could even smelt bloomery-furnace
significant proportion of the iron silicate mineral slags, since these contained fairly large amounts
fayalite (2FeO.SiO 2) and such slags are often of iron that could be extracted by the new, more
referred to as fayalitic. The slags formed during efficient process. By the 17th century, limestone
the smithing of iron (Fig 32) are also iron-rich, fayalitic was also added to the blast furnace charge,
slags and as such can resemble smelting slags. Much however, it is unclear whether this was normal
of the slag on a site might not be diagnostic practice for the earliest furnaces.
of any particular iron-working process, being
fragmentary, corroded or possessing intermediate A blast furnace comprised a hearth where the metal
characteristics, and is simply referred to as and slag could collect with a shaft above this. It was
undiagnostic slag (Fig 18). normal practice to provide two arches, in adjacent
sides (Fig 19). One arch was for the air blast from
There is little evidence for either the tools or the bellows and the other was for casting the iron
bellows used in the smelting process, except in and tapping the slag. The hearth itself was made of
some later literary sources. Some iron-working a refractory material, such as sandstone, although
sites have produced evidence for fire-lighting, later furnaces used fire clay. This material eroded
either as lumps of iron-pyrites, used to produce gradually with use, but this had the advantage of
sparks, or fire-drill stones with cup-shaped increasing the capacity of the hearth, and thus the
hollows, which would have been used as bearings size of the castings that could be made.
for a fire drill. Shelter would have been essential
for the storage of ore and charcoal and for The blast furnace would work continuously for
protecting the furnaces. Examples of round stake- months at a time, in production runs known as

< < Contents 24

campaigns, and was repaired between campaigns. feeding piglets, and so the castings were called
The charge put into the mouth at the top of the pigs. Most of the pig iron was sent to forges for
furnace at regular intervals would typically consist conversion to wrought iron (see pp 29–31).
of iron ore, fuel and limestone. The iron ore was
reduced as it travelled down the furnace and cast Furnaces were frequently built against a slope or
iron and slag formed at the base of the furnace bank which provided a charging platform often
(both molten, with the lower density slag floating with a bridge house leading to the mouth of the
on top). The cast iron was tapped off at intervals furnace where raw materials could be poured
and could be cast straight into objects, such as into the top of the furnace (Fig 19). There were
guns, or into pig iron. These castings were linked other structures associated with the furnace. The
to a supplying channel of metal, resembling a sow bellows were housed in the blowing house, built

Figure 19
Reconstruction of Duddon blast furnace, [illustration by kind permission of Alison Whitby and the Lake
Cumbria which was built in 1736 and is now a District National Park Authority]
Scheduled Monument.

< < Contents 25

alongside a water wheel for power. Earlier bellows into bar iron had limited success. This may have
were wedge-shaped and made of leather and been because the pig iron produced in the early
wood with iron nozzles, known as tuyères, which coal-fuelled blast furnaces, mostly grey cast iron,
fitted through custom-made holes in the stone was not of the same quality as that produced
furnace lining. Later bellows were cylindrical and in the charcoal-fuelled furnaces, mostly white
made of iron; these could be powered by a water cast iron (King 2010). The coke-fuelled blast
wheel or by a steam engine. From the 1830s, furnace pig iron may also have contained small
the blast was sometimes preheated in a stove, amounts of sulphur (from the coke) which would
ultimately firebrick Cowper stoves, which could be tend to make the bar iron brittle. Nevertheless
as large as the furnace. The casting house covered pig iron was increasingly used to produce cast
the area where castings were made, either using (rather than forged) artefacts (where traces of
moulding sand for casting pig iron and small sulphur did not overly affect the properties of
objects, or in a pit containing moulds for large the metal). Precision casting in iron played a
objects such as cannons and steam engine significant role in the development of engineering
cylinders. There would also be large buildings and munitions (and the grey cast iron of the
nearby for storing charcoal and ore (Bowden 2000). coal-fuelled furnace is better in this respect
than the white cast iron of the charcoal-fuelled
The need for water power meant that early blast furnace). The casting of large iron cylinders
furnaces were located in suitable river valleys contributed to the success of the emerging steam
(often relatively small tributaries). In addition, the engine technology which was subsequently
need for such power over several months meant used to help drain water from deep mines and
that rivers were dammed to form ponds with improve air blast for smelting (see pp 28).
leats to control and manage the water supply.
Smelting campaigns would often end during the The process of coking coal seems to have initially
summer months when the supply of water ran borrowed the methods used by charcoal burners.
low. During this time the iron smelters could The coal was simply heaped up on a stone surface
carry out repairs, reline the furnace and stockpile and a coating of coal dust and earth used to prevent
necessary materials such as ore and fuel. Until complete combustion. Specialised coking ovens
the 18th century blast furnaces used charcoal were also developed; these were usually a beehive
fuel and the size of the furnace and the duration shape and were initially front loaded (Cranstone
of the campaign required the supply of very large 1989) but later top loaded (Battye et al 1991).
quantities of charcoal. The earliest blast furnaces
in England are all located in areas with abundant By the middle of the 18th century coke-fuelled
woodland. The supply of sufficient charcoal was blast-furnace technology began to be increasingly
achieved through careful woodland management; adopted by the rest of the British industry (Hyde
commonly by coppicing woodlands cyclically. 1977). The switch from charcoal to coke resulted
in the relocation of the industry. The earlier
The reliance of the early blast furnace on charcoal industry was concentrated primarily in areas with
put pressure on the supply and cost of wood, abundant woodland such as the Weald and the
timber and charcoal. Through the 17th century Forest of Dean, while the later industry moved to
efforts were made to use coal in place of charcoal the coal fields of the Midlands, Yorkshire, South
in a number of metal industries. Around the Wales and the Scottish Midlands.
turn of the 18th century, iron was successfully
smelted using coke (Day and Tylecote 1991; King The earliest coke-fuelled blast furnaces appear
2001–2). Coke is coal which has been heated to to be comparable in size and form to the
remove some of its impurities. This technology contemporary charcoal-fuelled furnaces – indeed
was developed in Coalbrookdale, Shropshire the earliest examples were re-used charcoal-
but its adoption by the rest of the iron industry fuelled furnaces. The nature of coke fuel made
was slow. Initial attempts to convert this pig iron some changes desirable and over time furnace

< < Contents 26

design developed. Coke is a stronger material provided far fewer opportunities for iron smelters
and so can support a taller charge of fuel and so to build their furnaces against slopes or banks
furnaces tended to be built larger. By the end of which could be used as charging platforms. From
the 18th century furnaces were being built over the early 19th century inclines and hoists were
10m high and the extra height might have played increasingly used to carry raw materials to the top
a part in the dramatic increase in the production of the furnaces.
of cast iron in the 18th century. In the early 19th
century furnace shape changed from a square The success of the coke-fuelled blast-furnace
plan to a circular one (Fig 20). In addition the industry relied on the identification of suitable
furnaces were now increasingly made using bricks coals which could be converted into low-sulphur
rather than stone, although the brick structures coke – but it also benefited from the use of steam
were often held together with the aid of iron power and advances in the provision of the air
bands or sheets. From about the 1790s, they often blast. Newcomen atmospheric engines were used
had multiple tuyères, with the blast (provided from the 1740s, however, these were mainly used
using a steam engine) conducted to them through to recycle the water used to drive conventional
cast iron pipes. The topography of the coalfields waterwheels and bellows. By the last quarter of the
18th century steam engines were used to drive
blowing cylinders which provided increased and
more steady air flow. The air pressures which could
obtained inside the furnace increased considerably
which reduced the amount of fuel needed (King
2011). The higher blast rates also lead to an
improvement in the quality of the cast iron.
Higher temperatures would produce a more fluid
slag which could accept a greater proportion of
limestone; and lime-rich slags will tend to absorb
sulphur (Morton and Wingrove 1970; White 1980).

The early 19th century saw the introduction of hot

blast in the iron smelting industry (Belford 2012).
By heating the air that was drawn into the furnace
the iron smelters could further reduce the amount
of fuel needed. In addition, the heated air allowed
the use of coal rather than coke and enabled the
exploitation of the famous blackband ores of the
Scottish Midlands which were mixtures of coal and
iron ore. Initially stoves were heated with coal
(Fig 20), but from the 1860s onwards the waste
heat from the blast furnace was increasingly
directed into Cowper stoves that heated the
Figure 20
incoming air. The use of hot blast required water-
A 19th-century blast furnace of circular plan. The
cooled tuyères to prevent these overheating and
furnace is entirely built of brick and is strengthened
with iron bands. Note the pig beds in the foreground melting. At least some hot-blast slags display
where the cast iron was tapped from the furnace. The characteristics quite different to earlier blast-
small structure to the left of the furnace is an oven furnace slags – they are often grey and less
where the air was heated before being introduced into
vitreous (Fig 23), however, there has been little
the furnace. This image omits details of the blowing
mechanism as well as the charging incline (Muspratt systematic study of dated samples to determine
1860, Fig 279). the detailed relationship between smelting
technology and slag appearance.

< < Contents 27

 21

22 23

Figures 21-23
21. Blast-furnace slags are usually glassy in appearance vicinity, but this became problematic as furnaces
and range in colour from blue and green to grey produced more and more slag. Some slag was cast
or brown. They usually have abundant fracture into simple bricks and used as kerbstones and the
surfaces with little or none of the original surface like. Increasingly, slag was cast into wagons or bogies
remaining. They are less dense than bloomery- so that the it could be transported and dumped in
furnace fayalite slags, as they contain mush less slag heaps some distance from the furnaces.
iron. These slags can be found in large quantities 23. Late blast furnace slag from Bilston (sample held in
and were often reused, for example as hardcore or the National Slag Collection). The texture and grey
scattered across fields to improve soil quality. colour of many late blast-furnace slags resemble
22. Early blast furnaces could be successfully run with some concrete products (such as breeze blocks or
a policy of dumping waste slag in the immediate cinder blocks).

< < Contents 28

The layout of late blast furnace sites was quite attested archaeologically (see Crossley 1975a).
different from the earlier examples. It became Understanding such sites is often made difficult
increasingly common to construct two or more by the fact that they were often converted from
furnaces in a line and these would be water-powered bloomeries, and some were
accompanied by inclines, stoves and engine subsequently converted for other uses later on
houses. Blast-furnace complexes would also be (all had ceased to be used in the 19th century).
provided with a network of railways to move raw Archaeological evidence for finery and chafery
materials, finished metal and waste products forges can include the wooden foundations for the
which could cover an area of 20ha or more. forge hammers and the wooden support for the
anvil, plus evidence to indicate the presence of
3.2.3 Refining pig iron water power to drive the bellows for each hearth
The product of the blast furnace was pig iron and the hammer. The hearths themselves were
which is a brittle material, completely unsuited above floor-level and therefore rarely survive.
for use by the blacksmith. When forgeable iron
was required (and there was usually much Fining generated various types of debris, including
more demand for forgeable iron than cast iron), hammerscale, tap slag, large slag lumps and a
processes were used to convert the pig iron by type of porous slag, sometimes with traces of
reducing its carbon content. Several different flow on the surface. Fining slags share many
process were developed over time and these are characteristics with bloomery tap slags (Morton
described below in chronological order. and Wingrove 1970; Photos-Jones et al 2008). The
tap slag from the finery hearth usually contains
Most early blast furnaces were associated with rather low levels of sulphur (due to the use of
finery forges which could supply malleable charcoal fuel) while the large lumps of slag
wrought iron for the blacksmith. The finery used (hambones or mossers – see Fig 24) from the coal-
water power for both bellows and hammers: a fuelled chafery hearth often have high levels of
blast furnace and a finery forge would often form sulphur (Morton and Wingrove 1970).
a pair on the same water course. Pig iron was
melted in an open charcoal hearth under an air The advent of coke-fuelled blast furnaces put
blast provided by water-powered bellows (Den pressure on the finery and chafery hearths.
Ouden 1981; 1982; Dillmann et al 2012; Morton The increased productivity of the new furnaces
and Wingrove 1970). The carbon in the iron (as required ever larger volumes of cast iron to be
well as some other impurities, such as silicon) converted. While the early blast furnace was
was oxidised and removed and a bloom of low- often accompanied by a single finery, by the 18th
carbon iron would form in the hearth. The hot century the proportion of finery forges to blast
bloom was taken to a water-powered hammer for furnaces had risen (King 2011). The geographical
forging, which removed most of the trapped slag association between blast furnaces and fineries
and consolidated the bloom of iron. The bloom also waned depending on the nature of the
was drawn out into a bar using the hammer, being iron produced and the markets for it, as well as
reheated periodically in another hearth known the availability of water power. While fineries
as the chafery, which was also blown by water- proliferated in the West Midlands to supply
powered bellows. Coal or coke could not be used wrought iron to the forges of the Black Country
in the finery because of its sulphur content but it (Hayman 2005), they were in decline in the Weald
could be used in the chafery hearth where the iron which increasingly specialised in munitions and
was reheated (but not melted). other castings (Cleere and Crossley 1995).

Although the operation of finery and chafery The high demand for malleable iron encouraged
hearths is known from historical sources many in the later 18th century to search for ways
(eg Berg and Berg 2001) few examples are to increase production and lower costs, especially

< < Contents 29

 24 heat from the hearth into the main chamber. The
furnace had a low roof which would reflect heat
onto the raw material being melted. The low
wall between the coal fire and the main chamber
ensured that the coal ash would not contaminate
the raw material. The reverberatory furnace had
been adopted for copper and lead smelting in the
17th century (see p 42). Reverberatories included
‘air furnaces’ for foundry work (from 1690s);
‘balling furnaces’ for recycling scrap iron (from
1740s); some ‘potting and stamping process’;
and ‘puddling’.

A variety of processes were patented and

25 practised and these are often referred to as
‘potting and stamping’ (Hayman 2004; King 2011).
The processes varied somewhat but all seem to
have included a stage in which fragments of iron
were heated in crucibles (often with a variety of
fluxes and other materials). The crucibles were
heated in reverberatory furnaces and so were able
to use coal rather than charcoal. The fragments of
iron from this process were then heated together
and forged with a water-powered hammer into
blooms and then bars. Potting and stamping
processes and any diagnostic residues are very
poorly known from archaeological sources
(Morton and Gould 1967).

By the end of the 18th century a new process for

producing malleable iron was developed by Henry
Cort at Fontley in Hampshire (Mott and Singer
1983). The new process was called puddling and
Figures 24 and 25 used a reverberatory furnace for melting cast
24. Large lump of slag (hambone) from finery and iron and reducing the carbon content (as well
chafery forge at Colton, Cumbria (sample held in
as impurities such as silicon). The process was
the National Slag Collection).
25. Tap slag from Downside Mill. subject to a series of patents which also covered
the working of the blooms into bars using a
water-powered rolling mill (Mott and Singer 1983).
The use of both a reverberatory furnace and
through the use of coal rather than charcoal fuel. rolling mills was not new but these techniques
Most of these involved the use of reverberatory were successfully combined to produce a quality
furnaces of various kinds for different processes. of malleable iron which met with the approval
The reverberatory furnace had two chambers: a of the Admiralty. Initially it proved difficult to
hearth where coal was burnt and, separated from successfully reproduce the technology in the
it by a low wall, an area where the raw material main iron-producing districts but when combined
could be melted. At the far end of the chamber with the refinery or ‘running out hearth’ (which
was a chimney which would draw flames and reduced the silicon content and converted grey

< < Contents 30

cast iron into white cast iron) the technique was and then air pumped into the molten metal. The
fruitful. Cort’s early furnaces were provided with a oxygen in the air caused the silicon and carbon
sand base and this tended to react with some of in the cast iron to oxidise (and the heat thus
the iron to form fayalitic slag – similar slags were generated raised the temperature and kept the
also formed in reheating and refinery hearths metal molten). The later Bessemer convertors
(Phelps et al 2012). Puddling sites, such as Cort’s were installed so that molten pig iron could be
site at Fontley (Killick and Gordon 1987), and tapped directly from a blast furnace into the
the Millington’s site at Swalwell (Proctor 2011), convertor. Once the conversion was complete
will often yield quantities of tapped slag. This the mild steel (essentially iron containing small
slag closely resembles bloomery tap slag (cf Fig amounts of carbon, <0.3%) was poured from
16) but can be distinguished through scientific the convertor into ingots. The process was
analyses. Early puddling was sometimes referred initially successful because Bessemer had used
to as ‘dry puddling’ to distinguish it from the later a low-phosphorus pig iron. The first firms which
‘wet puddling’ processes (see below). took out licences for his patent had difficulties
reproducing his success with pig iron that
While many aspects of the puddling process contained appreciable amounts of phosphorus.
remained largely unchanged for the next 150 This was solved by the use of haematite pig
years, there were significant developments in the iron and the addition of speigeleisen, a high
nature of the puddling furnace. The sand base manganese alloy of iron, where the manganese
of the furnace was replaced with fayalitic slag carried impurities into the slag. Later, Gilchrist
and hammerscale which provided oxygen to help and Thomas developed the use of an alkaline (or
the removal of carbon and other impurities (Day basic) refractory lining which would contribute to
and Tylecote 1991; Flemings and Ragone 2009; the formation of a slag which had greater power
Photos-Jones et al 2008). Indeed the reaction to absorb phosphorus. Even later versions of this
between the carbon in the cast iron and the process made use of pure oxygen (rather than
oxygen in the slag and scale produced blue flames air) as the nitrogen present in the air could make
that were called ‘puddler’s candles’. This version the iron/steel slightly brittle. The contemporary
of puddling was referred to as ‘wet puddling’ or (and slightly later) technical literature contains
‘pig boiling’. Wet puddling was an efficient process some information on the nature of waste products
that allowed increased production (Flemings and but this is mostly limited to occasional chemical
Ragone 2009) and it was also possible to take analysis (eg Percy 1861–4; Bell 1884) with no
pig iron in an unrefined state. Later puddling descriptions of hand specimens.
hearths were provided with cast iron bases. In the
19th century the puddling process was adapted An alternative conversion process (the Open
to allow the formation of steel rather than plain Hearth process) was developed by Siemens in
iron. The quality could be rather variable and for the 1860s. Scrap metal, molten pig iron and iron
many applications the puddled steel was re- ore were charged into a furnace: the oxygen in the
melted before use, and so was in competition with iron ore would help to oxidise the carbon in the
crucible steel (see below). Both of these processes pig iron and so remove it. The reaction between pig
were largely superseded by the Bessemer and iron and iron ore was slower than in the Bessemer
Open Hearth processes (see below). process and so more controllable – it was
possible to stop the process at a particular
From 1860 it became increasingly popular to tap stage to produce steel or mild steel as required.
cast iron into convertors or hearths which would Iron ore was often omitted where the desired
convert pig iron into mild steel (Tylecote 1992). product was a high-carbon steel. The Open
The first example was the Bessemer convertor Hearth process was fuel efficient as it made use
which comprised an iron vessel with a firebrick of the regenerators developed by Siemens. The
lining. The pig iron was melted in the convertor furnace was provided with several underground

< < Contents 31

chambers (regenerators) which were filled with 26
bricks arranged in a chequer-board fashion,
ie with regular spaces for air to flow. These
chambers were connected via tunnels and pipes
to the air inlets, the Open Hearth furnace and
the chimney. By means of valves, the air supply
could be regulated so that hot waste gases would
heat the bricks in one chamber which could then
be used to heat the air coming in to the furnace.
This preheating of the incoming air greatly
reduced the amount of fuel required to achieve
a given temperature. Heat regenerators were 27
applied to many other types of furnace (such as
reverberatory furnaces, crucible steel furnaces,
etc). Material residues of the Open Hearth process
are not well known from contemporary accounts
or from archaeological contexts. In part, the lack of
available evidence is due to the strategies used by
iron and steel producers to dispose (and sometimes
re-use) waste materials such as slag. For much of
the 20th century iron slags were transported
(usually by rail) some distance from the production Figures 26 and 27
site and dumped in large heaps. The excavation 26. The foundations of the earliest steel cementation
of such slag heaps is challenging given their scale furnaces in England. The furnaces at Coalbrookdale,
and the difficulty of dating deposition. Shropshire were built in the 17th century.
27. Early 20th-century steel cementation furnaces in
Sheffield, South Yorkshire.

3.3 Making steel

Steel was often a more valued material than Another method of making steel was to surface
plain iron as it was harder and so could take carburise or case harden iron objects. The plain
(and retain) a sharper edge. Steel has been iron would be forged to the desired shape and
produced by various methods at different then heated in a bed of charcoal. Carbon from
periods of time. The earliest steels were probably the charcoal entered the outer surface of the
produced in bloomery furnaces: the blooms were iron, creating a shell of steel. If the object was
heterogeneous and as early as the Iron Age steely then quenched the shell became harder. There
portions of blooms were selected for certain types is evidence that this method was known in the
of tool (Fell 1993; McDonnell 2013b). Steel can Iron Age (Fell and Salter 1998) and it was widely
be deliberately produced in a bloomery furnace, employed in the medieval period (Williams 2003).
by manipulation of the smelting conditions
and types and ratios of raw materials but it is The cementation method of making steel, in
not clear how often this was done. The varied which bars of iron were converted into steel, was
properties of iron alloys were certainly recognised introduced in the 17th century. The earliest steel
and exploited during the early medieval period cementation furnaces were in Coalbrookdale,
(Gilmour and Salter 1998; McDonnell 1989). Shropshire (Belford and Ross 2007) and were
The metallographic examination of early ferrous circular and c5m in diameter (Fig 26). While the
tools and weapons shows that many were superstructure had not survived later furnaces
manufactured by welding together steel and other (eg Derwentcote, Durham, Cranstone 1997) were
iron alloys (Tylecote and Gilmour 1987). beehive- or bottle-shaped (Fig 27). Later furnaces

< < Contents 32

tended to be larger – up to 8m in diameter of surviving cementation furnaces is usually
(Barraclough 1984). vitrified. The examination of fragments of the
Coalbrookdale furnace suggests that parts were
It is possible that the earliest furnaces packed subject to temperatures in the range 1300–1500°C.
iron into charcoal-filled crucibles but the standard The chests at Derwentcote were sealed with
later practice included the use of stone chests. sand and clay (Cranstone 1997) and the Sheffield
Plain iron bars were packed in charcoal in a practice was to use the sludge from the bottom of
chest made from sandstone that was sealed cutlery grinders’ troughs (Barraclough 1984). The
and heated – the firing and cooling cycle would sealing material would be partially vitrified by the
typically last for two weeks. The plain iron used heat of the cementation furnace and this waste
in the cementation furnace had to be of a good material was known in Sheffield as ‘crozzle’ (Fig 28).
quality. Phosphoric iron would not absorb carbon
and so would not form a steel. The earliest steel In the 1740s Benjamin Huntsman developed the
cementation furnaces in Britain probably used use of sufficiently refractory materials (for both
iron from the Forest of Dean (one of the few furnace and crucible) which allowed blister steel
areas with low-phosphorus ores) while the later to be melted (Barraclough 1984). This method
industry was largely based on the use of imported involved breaking up cementation bars, placing
Swedish iron (Barraclough 1984; Mackenzie and them in crucibles, and heating them in a furnace
Whiteman 2006). The bars removed from the to melt the alloy, before casting homogenous steel
cementation furnace were called ‘blister steel’. ingots (‘crucible steel’). The process was not widely
The metallographic examination of samples of used before the last decades of the 18th century
blister steel shows that this was not a uniform (Craddock and Wayman 2000). While crucible steel
material and the proportion of carbon could had obvious advantages for the production of
vary significantly across the thickness of a bar high-quality tools, it was also used for the
(Mackenzie and Whiteman 2006). Blister steel manufacture of decorative items such as buttons
bars were often broken up and then reforged to (Craddock and Lang 2004). The ‘secret’ of the
improve their homogeneity; the resulting steel crucible steel process was undoubtedly the
was known as ‘shear steel’. The interior brickwork refractory materials used to construct the furnaces
and crucibles. A high-carbon steel (eg 1.7% carbon)
would begin to melt at about 1320°C but would
not be fully molten until 1400°C. The crucible steel
furnace was constructed from sandstone with a
cellar beneath and a chimney above to provide the
necessary draught for the coke fire. The interior
of the furnaces would over time become highly
vitrified. The crucibles were made from carefully
selected and processed raw materials. Early accounts
often stress the use of clay from Stourbridge – this
fireclay had long been prized in the glass industry
for the manufacture of their crucibles (which
needed to withstand temperatures in excess of
1300°C and the corrosive effects of molten glass).
Later recipes usually specify a blend of clays from
Figure 28 several locations. It has been suggested that the
Waste from steel cementation (crozzle) used to top earliest crucibles also contained a significant
walls in Sheffield. Crozzle comprised a mixture of steel proportion of graphite (Barraclough 1984).
swarf and sand that was used to seal steel cementation
Naturally graphitic clays had been exploited for
chests. The high temperature of the steel cementation
furnace led to the partial fusion of this mixture. some years for the production of crucibles used
[Photo © Roger Doonan] for precious metalworking (Martinón-Torres 2012).

< < Contents 33

The raw materials for the crucibles were carefully
processed by grinding and sieving to remove
any impurities (such as iron pyrites). Over time
the used crucibles became heavily vitrified; a
proportion of the old crucibles could be crushed
and re-used to produce more crucibles. During
melting the remains of any slag previously
trapped in the metal would rise to the surface and
combine with any flux used (Barraclough 1984).

3.4 Smithing

Iron (and many of its alloys) becomes relatively

soft when heated so that it can be smithed or
forged using a hammer into a variety of shapes.
In addition, the metal can be joined together by
forge welding. The techniques used in smithing,
and associated waste products, varied depending
on the nature of the iron (or alloy) and the size
and shape of the object being formed.

Bloomery iron is a heterogeneous material: the Figure 29

proportion of alloying elements (such as carbon Late medieval illustration showing smiths at work. Note
and phosphorus) can vary considerably even the waist-level hearth in the background and the anvil
with a single bloom. In addition, it is a composite set in a wooden block.
material and (like wrought iron) can contain
quantities of trapped slag. The bloom would
need to be refined to produce iron stock (such as
billets or bars) suitable for forging into objects. metal. In general plain iron can be forged at
The initial stages of refining the bloom involved higher temperatures than steels or other iron
hammering it while hot to consolidate the metal alloys. Plain iron with slag inclusions needs to
and expel the trapped slag; losses at this stage be forged at a sufficiently high temperature that
can be considerable (Crew 1991; Sauder and these will be at least partially molten (around
Williams 2002). This primary smithing was often 1100°C). The range of techniques employed by
carried out at the smelting site, and therefore blacksmiths to shape iron (see Andrews 1991)
smelting and refining residues can be found appear to have changed little over the last few
together (Paynter 2007b). Bloom refining residues millennia. Using hand tools and working on
can include hammerscale (Fig 30), small prills an anvil, the metal could be thinned down,
of fayalitic slag expelled from the bloom and thickened, straightened, bent, split, pierced and
smithing hearth cakes (Fig 32). otherwise shaped. Many of the tools described in
modern textbooks can be compared directly with
The iron stock, or billet, would then undergo examples from archaeological contexts (eg Hinton
secondary smithing or forging, also while hot, to 2000; Manning 1991). Secondary smithing also
produce artefacts. Forging is carried out above includes the repair and recycling of iron objects.
700°C and the metal becomes softer the hotter
it is, however, excessive heat (above 1300°C Iron and its different alloys have different
for plain iron) can damage and even burn the properties: plain iron is relatively soft and often

< < Contents 34

contains slag inclusions, steel is harder than plain (and its alloys) inevitably led to some oxidation
iron and can be made even harder by quenching of the metal at the surface. As the red-hot metal
(see below), and phosphoric iron is harder than was forged, some of this oxidised surface film
plain iron although it can be somewhat brittle would be detached (the ’sparks’ seen when a
when cold (and is unaffected by quenching). smith forges iron). Fragments of this film, often
The smiths’ skill encompassed the control and referred to as flake hammerscale, are regularly
appropriate application of these properties in found in archaeological deposits associated
forming objects. Smiths recognised that not with iron smithing (Fig 30; Dungworth and
all iron behaved in the same way, and stock Wilkes 2009; Mills and McDonnell 1992; Palmer
metal with different properties would have been 2015; Young 2011). Detecting hammerscale is
available. For example, Iron Age currency bars relatively easy as it is magnetic. Plotting the
are thought to be a form of stock iron and the distribution of hammerscale has great potential
elaborate socketed ends or welded tips on these to uncover detailed aspects of the layout and
bars are a significant feature, demonstrating use of space in blacksmiths’ workshops (Fig 5).
visibly the forging properties of the iron (Crew Flake hammerscale is often accompanied by
1995). Finds – such as blooms, billets and bars small amounts of spheroidal material which
and all forms and types of stock iron – are is also magnetic (Fig 30). These spheres can
important to further research into the trade and form in a variety of ways but especially the
use of different iron alloy types (Fig 13). joining of iron (and its alloys) by forge welding
(the other two methods responsible for the
Objects were formed from a combination of production of spherical hammerscale are iron
different iron alloys. For example, knives and
other edged tools were often made with a hard
alloy for the cutting edge (eg steel) and a softer
but tougher alloy (plain iron) for the back. The
different parts of metal were heated and then
forge welded. Forge welding requires a higher
temperature than ordinary forging – for steels this
was around 1200°C and slightly higher (1300°C) for
plain iron. Modern blacksmiths often use a flux,
such as sand, to clean the surfaces that are to be
welded, however, plain iron can usually be welded
without using a flux (due to the presence of the
trapped slag inclusions). The extent to which
fluxes were used in antiquity is unknown. The
different alloys were often visible in the finished
and polished object. This technique was also used
for the productive of some prestigious pattern-
welded weapons. Rods or bars of plain iron and Figure 30
steel were welded together and repeatedly folded Scanning electron microscope(SEM) image of flake
and twisted during forging to obtain an attractive hammerscale and spheroidal hammerscale. Flake
hammerscale consists of grey to black, fish-scale
patterned surface (Gilmour and Salter 1998). Some
like fragments, typically 1–3mm across. Its small size
iron is lost during smithing (hammerscale, see Fig means that it is rarely detected during excavation but
30), and this loss is greater during welding and it is sometimes recovered from environmental samples
complex smithing operations such as pattern welding. or from soil samples taken specifically to recover
hammerscale. Hammerscale is highly magnetic and
can be separated from soil using a magnet. Spheroidal
The smithing of iron produced distinctive forms hammerscale consists of small round slag droplets,
of slag and waste. The periodical heating of iron which can be hollow to varying degrees.

< < Contents 35

smelting and the accidental burning of iron). by hand forging, however, comparable material
The extreme heat and pressure produced during would also be produced during mechanised iron
forge welding caused the oxidised films at the forging or rolling (Young 2011).
surface of the metal to momentarily melt and
this was then ejected from the weld and formed Blacksmithing will also generate a variety of
a series of droplets (Dungworth and Wilkes 2009). lumps of slag, the most distinctive of which is
Crew (1988) has reported the presence of slag the smithing hearth cake. Some hammerscale
spheres from smelting furnaces – in almost all will fall from a heated piece of iron into the fire
respects these smelting slag spheres resemble and react with fuel ash, the hearth wall and any
spherical hammerscale produced during welding. flux used. Droplets of slag accumulate in the hot
Dungworth and Wilkes (2009) demonstrated that region near the blowing hole and coalesce to
similar spheres can also be produced by burning form a lump, known as a smithing hearth cake
iron alloys. Where large quantities of hammerscale (Fig 32). Smithing hearth cakes vary considerably
build up it can become cemented together with iron in size and form but some have a concave upper
corrosion products into smithing pan (Fig 31). surface (often displaying the greatest degree of
The archaeological investigation of hammerscale vitrification) and a convex lower surface (often
has mostly concentrated on the material produced with abundant impressions of charcoal fuel).

31 32

Figures 31 and 32
31. Smithing pan from the blacksmith’s workshop slags have not been fully incorporated. The lower
on Hythe Road, Colchester. It consists of a layer surface usually has impressions from charcoal or
of debris, largely hammerscale, trodden down the hearth lining. The size of the cake depends
and corroded together (image 20mm wide). The on the amount of iron forged, how much slag it
hammerscale visible in the surface includes several contained, whether fluxes were used and how
spherical examples. often the hearth was cleaned out. The larger
32. A partially sectioned smithing hearth cake. smithing hearth cakes can be misinterpreted as
These are normally plano-convex to concavo- furnace bottoms. Smithing hearth cakes from
convex in section and circular or oval in plan. primary smithing, or refining, will generally be
Their size and weight can vary considerably, from larger than those from secondary smithing.
100g to more than 2kg, although the majority Smithing hearth cakes are sometimes slightly
weigh 200–500g. The upper surface sometimes magnetic as they can contain fragments of iron
has a depression produced by the air blast, or broken from the bloom and some hammerscale.
is sometimes irregular, where the last formed

< < Contents 36

If left in place, a smithing hearth cake will begin to to produce more glassy smithing slags. An air
impede the flow of air from the bellows and so it blast was used to obtain high temperatures,
would be pulled out and discarded. While smithing although no evidence for the nature of the
produced some cakes of slag (Fig 32) it would also bellows survives archaeologically. Vitrified clay
produce smaller amorphous lumps of slag (Fig 18). hearth wall or hearth lining is most likely to
Smithing slags produced in a coal-fired hearth will be produced in the hottest part of the hearth,
tend to be less dense than those produced in around the blowing hole. Vitrified clay hearth
a charcoal hearth. Smithing slags can be found linings are similar to smelting furnace linings
heaped near to the smithy or they could be (Fig 11), though hearth lining is generally thinner
transported farther away for dumping or reuse, for and is found in smaller fragments and smaller
example in road construction. Recent smithing by quantities. Sometimes the outline of the blowing
Dogon smiths in Mali (Soulignac and Serneels hole is preserved. The absence of hearth features
2013) provides some indications of the rate at which and clay hearth lining fragments on some later
smithing slag might be produced by traditional sites is consistent with the use of raised hearths.
forging in a charcoal hearth. Smithing 3kg of stock
iron over a period of approximately 4 hours yielded Blacksmithing would usually take place within a
around 2kg of finished iron object, 215–675g of building or shelter to protect the hearth, and the
slag and an estimated 700–900g of hammerscale. smith, from the elements and also provide dim
lighting round the hearth (Crew 1986), allowing
Blacksmiths needed to heat their metal in a the smith to better judge the temperature of the
hearth which was capable to achieving the iron from its colour. Some smithing workshops
temperatures required for the work they were were in rather simple wooden structure such as
undertaking. Medieval and later forges were the Roman examples at Southwark (Hammer 2003)
often built of stone or brick with the hearth at while others were stone buildings (Mills and
waist height. The excavation of such workshops McDonnell 1992). Stone anvils and hammer
usually finds no direct evidence for the hearth stones, with slagged surfaces, have also been
or anvil (eg Mills and McDonnell 1992). Many found. There might also be indications of the
earlier hearths were probably situated at location of a wooden anvil or a wood block into
floor level although some raised hearths are which a small metal anvil was inserted. Metal tools
suggested by iconographic sources (Halkon 2013). such as anvils, tongs and hammers do survive, but
Archaeological excavation in Southwark found hardly ever in a workshop context. There is no
numerous blacksmiths’ workshops in use from evidence for the type of bellows used at early sites,
the late 1st century AD until the middle of the although their location can sometimes be inferred.
fourth century (Hammer 2003). The hearths were
set at ground level, showing that smithing took The forging of iron has also benefited from varying
place on the floor rather than using raised hearths degrees of mechanisation from the medieval period
and anvils. The size and form of the hearth onwards. Waterpower was harnessed to drive
varied greatly but most were simple hollows in hammers: the flow of water drove a water wheel
the ground (0.3–1m in diameter, Hammer 2003, and this was used (via a wheel with cams or teeth)
Fig 110). Hearths occasionally have a clay lining to raise and then drop a hammer (the same principle
and even elements of a simple superstructure to was also used to power bellows). Such plating forges
retain the fuel and heat and perhaps to protect were used to produce an increasing variety of iron
the bellows. The hearth was filled with a bed goods, such as frying pans, saws and cutlery, with
of fuel, predominantly charcoal, but from the some specialist centres emerging, eg Sheffield.
Roman period onwards there is growing evidence
for the use of coal (Dearne and Branigan 1995). From the 17th century water power was used
Medieval and later blacksmiths made increasing to roll metal into sheets and cut (slit) these into
use of coal (and even coke) and this would tend rods (Johnson and Bearpark 1978; King 1999).

< < Contents 37

Iron could only be worked in this way when it was Another important application of the roiling mill
red hot and so rolling and slitting mills also had was for the production of tinplate which began
furnaces where the metal could be heated. Rolling in earnest in about 1725 at Pontypool. Tinplate
used two cylinders each of which rotated under comprised a sheet of iron (or from 1880s, steel)
power from a water wheel. The distance between with a thin layer of tin on the surface. The fully
the two rollers could be adjusted to produce sheets developed process required two rolling mills,
and strips of the desired thickness (Fig 33). Slitting heating and annealing furnaces, and pots for
mills also used rollers but the rolls had blades fitted a pickling medium (acid), washing and tinning
which cut the metal. Rolling and slitting were (Minchinton 1957).
important for producing metal rods of the correct
thickness for the nail trade and the greatest Further mechanisation of iron working took
concentration was around the Black Country, place with the adoption of steam power. In the
reflecting the importance of the nail-making 18th century some of the earliest application
industry to that region (King 1999). The rolling of steam power in the iron industry was to
mill also formed an essential element of Cort’s compliment rather than replace water power.
development of the puddling process (see p 30). Water was collected in mill ponds and then

Figure 33
Rolling iron bars at the Black Country Museum.
[photo © Paul Belford]

< < Contents 38

used to power water wheels but this supply innovations in the iron industry. The use of cast
could run dry, especially in the summer months. iron cylinders for steam engines (for pumping
In many cases, therefore, steam power was water or air) used in the iron industry was only
used to pump water from the tail race back possible because of skills and expertise in iron
up to the mill pond and allow it to be re-used casting. Early moulds were made from loam (a
(Hayman 2005). As steam technology (and the mixture of clay, silt and sand) and then fired
associated engineering of gears and related but later moulds were often made using ‘green’
mechanisms) developed it became possible to sand (sand with a little clay). The ‘green sand’
use steam power in more diverse ways. Matthew moulds contained small amounts of moisture
Boulton and James Watt used steam power for and were not fired, but relied on their porosity
a rolling mill in 1781 which meant that slitting to prevent the formation of too much steam.
and rolling mills no longer had to be sited next
to suitable rivers. In 1842 Naysmith patented The early casting of iron was undertaken at the
a steam-powered hammer suitable for forging blast furnace. This was practical where large
large pieces of iron. The hammer was raised artefacts such as cannons were to be cast. Where
vertically above the forging and then dropped. a furnace was used to cast smaller objects it
The steam hammer was better able to cope with was sometimes provided with a small hearth
the largest items that needed forging. With the extension, or forehearth, from which cast iron
use of steam power to control the descent of the was ladled into small moulds. Casting direct from
hammer, it was eventually possible to control a furnace could be a rather doubtful business
great power with considerable precision. due to the uncertain quality of the metal. The
18th century saw the appearance of foundries
which bought (rather than made) their pig iron
3.5 Casting and re-melted it in ‘re-melting furnaces’. Two
kinds of re-melting furnace are known. The first
Iron with a sufficiently high carbon content was a reverberatory furnace known as an air
(2–5%) can be melted relatively easily (ie a furnace and introduced in 1690s (King 2001–2).
melting temperature of 1200°C) and so cast The second was known as the foundry cupola
into shape. Such cast iron could have been and is a variety of small blast furnace (Tylecote
accidentally formed in bloomery furnaces (if the 1991). At least some of the smaller cupolas
smelting conditions were especially reducing), appear to have been made from joining sections
however, it would be brittle and so would which could be stacked on top of each other, and
probably have been rejected by blacksmiths. The enabled the cupola to be moved on occasion.
introduction of the blast furnace at the end of Re-melting cast iron and the addition of fluxes
the 15th century provided large quantities of cast or alloying elements (such as manganese)
iron, most of which was converted in fineries to tended to improve the quality of the cast iron
malleable iron (see p 29). Nevertheless, some of and of the castings made. Cupola melting of
the cast iron was formed directly into artefacts cast iron and casting in ‘green sand’ moulds
by pouring the molten metal into moulds. has continued to the present day but has rarely
Cannons were usually made this way (Crossley been identified in the archaeological record.
1975b) as were a range of other relatively large
objects, such as firebacks and grave slabs. Over
time the range of articles which were made
using cast iron expanded (pots, pans, cooking
ranges, mangles, drain pipes, bridges, machine
and engine parts, etc). Indeed the development
of precision casting for the engineering trade
was often an essential requirement for other

< < Contents 39

4 Archaeometallurgical
Processes and Finds:
Copper and its alloys

4.1 Background are bronze (copper with tin), brass (copper with
zinc) and gunmetal (copper with tin and zinc).
Pure copper has a melting point of 1084°C, lower If lead is also added, then the alloy is described
than that of plain iron, and is a very versatile as leaded, for example ‘leaded bronze’ and so
metal. Copper and copper alloys can be melted on. Alloying increases the hardness of the metal,
and cast to shape or they can be wrought reduces the melting temperature, and can also
(hammered). Copper is very ductile and soft, and change the colour (Fang and McDonnell 2011).
so can be drawn into long wires or hammered Bronze and brass were used for wrought and
into thin sheets. The common alloys of copper cast objects, but the uses to which each alloy

Copper in summary
Copper is a soft and ductile metal, with a melting temperature of 1084°C. Alloys of copper include brass (with zinc), bronze
(with tin) and gunmetal (with tin and brass). Sometimes lead was also added and the alloys are then described as leaded.

Process Summary Archaeological debris

Smelting Ores were smelted in one or more stages. There is little evidence for early copper smelting,
Molten metal was produced. Later, although it is likely that debris such as slag and vitrified
complex smelting operations and then clay would have been produced. In later periods there
reverberatory furnaces were introduced. can be evidence for waterpower.

Casting Metal could be melted in a crucible and Crucibles, moulds, metal spills, failed castings and
cast directly into objects or into ingots surplus metal trimmed from castings (sprues, flashings
using moulds. Moulds were made from and runners).
sand, clay, metal or stone and could be
open or closed, one piece (investment
mould) or two (piece mould).

Wrought metal The solid metal was shaped, for example Scrap metal, such as turnings or offcuts, metal sheet,
working by cutting or hammering, which, if done rods, bars and wires. Small ingots or blanks, tools
at room temperature, caused the metal and anvils are rarer finds. Waterpower can be used for
to harden and become brittle. Heating mechanised processes at later periods.
(annealing) the work-hardened metal
at intervals restored its toughness and
softness.

< < Contents 40

was put tends to vary with time. Additions the slag. The evidence for copper smelting
of lead to copper alloys could improve the from Pentrwyn on the Great Orme headland
quality of castings, but was detrimental for in Gwynedd (Williams 2013) has recently been
alloys that were to be worked or gilded. dated to the late Bronze Age. Tangible remains of
Roman and medieval copper smelting is virtually
unknown and it is likely that copper was imported
4.2 Smelting and alloying from Europe in the later Middle Ages.

Copper-based artefacts were widely used in In the 16th century the Company of Mines Royal
Britain from the end of the third millennium was provided with monopolistic control of copper
BC and Britain has numerous of copper ores, production and introduced German workers who
including copper carbonates (eg malachite applied their techniques in the Lake District and
Cu 2CO 3(OH) 2) and copper sulphides (eg chalcocite
Cu 2S, chalcopyrite CuFeS 2 and ‘grey ores’
Cu 12(AsSb) 4S 13). Nevertheless relatively little
physical evidence for copper production in Britain
has been recovered that pre-dates the Industrial
Revolution. Until recently it was widely assumed
that the intensive exploitation of copper ores in the
recent past would have destroyed all traces of early
mining. Nevertheless field survey, underground
exploration and archaeological excavation have
identified at least twenty sites in Britain and Ireland
where copper ores were mined in prehistory (Crew
and Crew 1990; Dutton and Fasham 1994; O’Brien
2004; Timberlake 2003a; 2003b; Timberlake
and Marshall 2013; Timberlake and Prag 2005).
Mining techniques included firesetting (the use
of heat to weaken or shatter rock) and antler (or
bone) picks, as well as a variety of hand-held and
hafted stone tools. Suitable stones were often
carefully selected and brought considerable
distances. Stones were sometimes modified
to make them easier to haft (Pickin 1990).

Although a growing number of mining sites

can now be shown to have been exploited in
prehistory (Timberlake 2003a), there has been
a general absence of smelting evidence such
as furnaces or slags (Craddock 1990; 1994).
Replication experiments have shown that copper
carbonate and copper oxide can be smelted
directly, using charcoal fuel and an air blast to
obtain sufficiently high temperatures. The molten
metal sometimes forms prills (droplets) scattered
through the smelting slag, which forms from the Figure 34
reaction of gangue in the ore with metal oxides. Copper smelting furnaces of the 16th century as
The prills could have been recovered by crushing illustrated in Agricola’s De Re Metallica.

< < Contents 41

south Wales (Day 1991). The roasting and smelting
operations consisted of a complex sequence
which produced, first, matte (copper sulphide)
and eventually copper metal, over a period of four
months. The furnaces are not described in detail
but were probably similar to those described by
Agricola, ie simple shaft furnaces (Fig 34). The fuel
and copper ore would be charged at the top of the
furnace and slag and matte or metal tapped from
the bottom. The contemporary accounts show
that the bellows were driven by water power and
peat fuel and coal were used as far as possible,
although the later stages of copper refining
required charcoal (Day 1991). Nevertheless, the
material evidence for the mining and smelting
described in 16th-century documents has
received relatively little attention.

Copper production appears to have gone into

decline for much of the 17th century due to heavy
competition from continental producers, but was
revived in the final decades of that century by
two developments. The first was the passing of
legislation which removed the crown’s prerogative Figure 35
(and the power of the royal companies) and A reverberatory furnace for smelting copper with
encouraged private investors. The second was exterior view (above) and a cross-section (below). The
coal fire to the left is separated from the ore by a low
the introduction of the reverberatory furnace wall (G). The heat from the coal fire is drawn across the
for smelting copper. This introduction is known ore by the draught induced by the chimney (E).
from documentary (King 2001–2) rather than
archaeological sources and so is incompletely
understood. The reverberatory furnace was a
low rectangular furnace which was much longer The ore would be charged and strongly heated in
than it was wide. The fire was placed at one end oxidising conditions to calcine the ore and drive
and separated from the ore by a low wall (Fig 35). off as much sulphur as possible. The calcined
The heat from the fire was drawn across the ore ore would then be melted to form a slag and a
by the draft formed by a chimney at the far end matte (mixture of copper and iron sulphides).
of the furnace and the low ceiling would reflect The matte would then be calcined again and
heat down onto the ore. The physical separation the calcined matte melted in several steps until
of the fuel from the ore meant that smelters a pure copper was formed. Each melting stage
could use coal instead of charcoal (and thereby would produce some slag and the slag from the
save money). The atmosphere in the furnace later stages would be sufficiently rich in copper
could easily be changed from heavily reducing that it would be added to fresh ore to help oxidise
through to heavily oxidising through opening and that ore and recover some of the copper from
closing doors and vents. This encouraged the the slag. By the 19th century copper smelters
development of smelting procedures which used carefully blended ores from different mines. In this
alternating steps with oxidising and then reducing way they could combine sulphide and naturally
conditions, although archaeological evidence for oxidised ores which would tend to react together
this is virtually unknown. and remove sulphur (the ‘Welsh process’).

< < Contents 42

A variety of copper-based alloys have been whole vessels. The patterns of vitrification can be
produced. Most bronze (a copper-tin alloy) was used to reconstruct how crucibles were placed in
probably made by melting previously smelted a hearth. If a crucible was covered with charcoal
copper and tin. Similarly, leaded alloys would and blown from above then the vitrification would
have been produced by adding metallic lead to be most pronounced at the rim.
the molten alloy. Brass (a copper-zinc alloy) was
not made until the Roman period; its production The melting of copper alloys can easily lead to
is described in the section on zinc (p 58). Many the formation of small amounts of slag. If the
mixed alloys could have been produced by copper alloy is subject to even slightly oxidising
recycling mixed scrap copper alloys with relatively conditions then some of the metal will oxidise and
little thought to the exact composition of the can then react with fuel ash, fragments of broken
finished metal. crucibles and other materials to form lumps
of amorphous slag. As these slags weather the
presence of small amounts of copper produces a
4.3 Casting distinctive green colour.

Copper has a melting temperature which is In some cases the molten metal was cast into
sufficiently low that it can be melted and so open moulds to produce suitable pieces of metal
objects can be shaped by casting into suitable to begin forging (see below) but in other cases
moulds. Melting small amounts of copper (or more complex moulds were used to produce a
copper alloys) does not necessarily require largely finished object. Large and/or complex
a custom-built hearth, although a reducing moulds were usually provided with a funnel-
atmosphere is needed to prevent the metal from shaped opening – the in-gate or sprue cup (Fig 41).
oxidising – this was usually achieved by using The molten metal was poured into the in-gate and
charcoal fuel. In order to achieve sufficiently ran down through channels (runners) into the
high temperatures to melt the copper it would actual shape to be cast (the matrix).
be necessary to use bellows to increase the flow
of air. Where crucibles were placed in simple Crucibles come in various shapes and sizes, from
charcoal-filled hearths, it was usual for the bases thimble-sized to larger than pint beer-mug sized
to be rounded or pointed to help keep them (Fig 36; Bayley and Rehren 2007). Prehistoric
upright. Wheel-thrown crucibles often have flat crucibles were made by hand and come in a
bases which are better suited to placing in a variety of sizes and shapes. Some of the earliest
hearth with a flat base (often with the fuel in a (Bronze Age) examples (Needham 1980) are fairly
separate compartment, such as a reverberatory large shallow bowls (typically 200mm across) and
furnace or a muffle furnace). Crucibles could circular in plan. These crucibles were provided
be used to produce alloys (eg bronze) from with three legs and although 100mm tall overall,
different metals (tin and copper): the copper the legs formed almost half of this height. The
would be melted first and then the tin added. crucibles had thick walls (around 20mm thick) and
Alloys generally have lower melting temperatures capacities of around 150ml (thus capable of holding
making them easier to cast: bronzes have lower perhaps 1kg of copper alloy). Iron Age crucibles
melting temperatures than pure copper. Crucibles tend to be triangular (Wainwright 1979); and were
could also be used to melt scrap metal. The probably formed as shallow bowls and then had
crucibles had to be made from a refractory clay, ie three corners made by slightly pinching the clay
one which could withstand the high temperatures while still wet. These are commonly 60–90mm
needed to melt copper alloys. The high degree across and 35–50mm deep with walls 5–15mm
of firing makes crucibles fairly durable although thick and capacities of around 50ml (ie capable of
most are recovered as fragments rather than melting around 300g of molten copper alloy).

< < Contents 43

Figure 36
Drawings of common crucible forms of Iron Age to 4 and 5: Anglo-Saxon, 6: early Christian,
post-medieval date. 1: Iron Age, 2 and 3: Roman, 7: late medieval, 8: post-medieval.

From the Roman period onwards some crucibles Few early or middle Saxon crucibles have been
are wheel-thrown, but handmade crucibles recovered; the examples from Bloodmoor (Blakelock
continued to be used into medieval times. Roman 2005; Cowgill 2009) are simple bowls or beaker
wheel-thrown crucibles typically have external shapes in a variety of sizes. The vessels were made
diameters of 50–150mm and are 50–100mm high by hand from a quartz-rich fabric. The smallest are
with wall thickness of 3–6mm (giving capacities only 20mm across and 20mm high, while the largest
of 100–500ml, equivalent to 0.5–3kg of molten are 100mm across and perhaps up to 100mm deep.
copper alloy). Many (although not all) Roman The crucible walls are 5–15mm thick and the
wheel-thrown crucibles also have an extra layer largest vessels could have held 100ml of molten
or coating of clay on the outside (Fig 37; Bayley metal (a little under 1kg). A silica-rich outer
1991c). This coating could have provided a degree surface appears to have been produced by rolling
of insulation which would have helped keep the in sand while it was still damp. A few crucibles are
metal molten for longer and prevent failure of also known from contemporary sites in the west
the crucible proper through thermal shock. It is and north, such as Dinas Powys (Alcock 1963) and
rare for handmade crucibles to have these extra Dunadd (Lane and Campbell 2000). These are all
outer layers. It is possible that while potters made simple and rather small handmade forms with
wheel-thrown crucibles (largely following existing bowls being among the most common.
beaker and cup forms) these had walls which
were too thin for a metallurgical application. The Many more sites have provided evidence for
metalworkers could have added the extra outer melting copper alloys in the late Saxon period.
layer of clay themselves to provide a suitably The examples from Winchester are made by
thick wall. In addition, the extra outer layer of hand using not very refractory clays (Bayley
clay produced a more rounded base profile which and Barclay (1990) while those from York
would be more stable in a charcoal-filled hearth. (Bayley 1992) and Lincoln (Bayley 2008) are

< < Contents 44

usually wheel-thrown in Stamford ware. Other
sites often yield both handmade and Stamford
ware crucibles, eg London (Bayley et al 1991).
Stamford ware was used for the production of
a range of domestic forms but the clays used
were well suited for the melting of copper alloys
due to their refractory characteristics (Bayley
et al 1991). The Stamford ware crucibles have
maximum diameters of 60–90mm, rim diameters
of 30–70mm and heights of 50–70mm (the
bases were pushed out). The walls are typically
2–5mm thick giving capacities of 25–200ml
(200g–1.5kg of molten metal). A small proportion
of the crucibles were provided with an extra
outer layer which would have provided greater
insulation and protection from thermal shock.

Evidence for later medieval crucibles is less

abundant although the assemblage from the
Guildhall, London (Bowsher et al 2007) shows that
by the end of the 13th century crucibles usually Figure 37
were often larger. The Guildhall crucibles have Roman crucible from Dorchester, Dorset (120mm
vertical sides with rounded bases. The diameters high). Crucibles are invariably grey or black as a result
of being reduced-fired. Crucible clay was usually
are usually 100mm and the walls 10–20mm thick; tempered with fine sand or, occasionally, organic
although none from this site survived to their full matter. Crucibles can become vitrified because of
height, the capacity is estimated at 400–500ml the high temperatures at which they are used, either
(3–4kg of molten metal). developing a thin external ‘glaze’ or becoming glassy
and bubbly throughout their entire thickness. Some
crucibles have an added outer layer of less refractory
The 15th century saw the appearance of clay, to improve heat insulation and to increase the
distinctive triangular crucibles (Cotter 1992; robustness of the vessel, and this usually becomes
Martinón-Torres 2012; Martinón-Torres et al heavily vitrified. Small quantities of the metal being
melted can become chemically bound in the crucible
2008). These were rather small crucibles made surface, or physically trapped as droplets of metal.
from highly refractory clays (originally from Copper can be seen as green corroding droplets or
Hesse, but with later production elsewhere). as bright red patches where it has reacted with the
The crucibles were wheel thrown with flat bases glassy surface of the crucible. Chemical analysis (see
p 65), however, is often the only way of determining
but the rim was pulled to produce a triangular the process in which the crucible was used.
form (similar round vessels are known in
virtually the same fabric). The crucibles were
produced in a range of sizes with heights from
20mm to 150mm (giving volumes of 5–500ml). of these crucibles suggests that most heating
The crucibles were made from highly refractory took place in reverberatory or muffle furnaces.
clays with very low alkali and iron content.
In addition, some were tempered with quartz By the 16th century fairly large and thick-walled
and some with graphite. These crucibles were crucibles often with flat bases became the norm
probably used for a range of metallurgical for casting copper alloys (eg Bayley and White
activities including casting molten metal as 2013). These crucibles could be up to 240mm
well as testing ore quality and more alchemical high with capacities of 1.5l (equivalent to over
procedures (Martinón-Torres 2012). The flat bases 10kg of molten metal). These crucibles were

< < Contents 45

made with a high proportion of grog (old pots) as Casting copper alloys requires a mould in order to
temper and the absence of any rilling (as well as produce a casting of the desired size and shape.
the coarseness of the grog) suggests they were Moulds might be open or closed and might be
probably made by mould forming. made from a variety of materials: sand, clay, metal

38 39

40 41

Figure 38–41
38. Part of an investment mould from Beckford, 40. Part of the cope (outer part) from a cauldron mould
Worcestershire. It has no mating surfaces since it from Prudhoe castle, Northumberland, Note the
was made in one piece. Note the in-gate at the top inner surface in reduced-fired (black) but the outer
and the runner down to the circular object. surface is oxidised-fired (red).
39. Complete clay piece mould for a trumpet brooch 41. Sprue with two runners from Wicklewood, Norfolk,
from Prestatyn, Clwyd. The in-gate is by the foot of cut from a copper alloy casting.
the brooch. The locating marks round the edges of
the two halves (valves) of the mould which would
have aided correct assembly, can clearly be seen.
Fragments of luting clay, which was used to seal the
join, is also sometimes found.

< < Contents 46

or stone. Moulds for small objects were usually Stone moulds were more durable than clay but
made of either fired clay or, less commonly, required greater effort for their manufacture.
fine-grained stone. Clay moulds are not common Only some types of stone are able to withstand
finds, partly because they are fragile and so do the high temperatures involved in casting copper
not survive well. The clay used to make moulds alloys. Stone was used to make moulds in the
was usually carefully selected and processed and Bronze age (especially in the north and west)
was usually tempered with fine sand or organic but are rarer in later times. Bronze age stone
matter. Clay moulds are invariably grey or black moulds are often simple one-piece open moulds
(reduced-fired) on their inner surfaces, which for the casting of simple flat axes, although more
were in contact with the cast metal, and orange- complex objects such as spearheads were cast
red (oxidised-fired) on the outer surfaces. Clay using stone moulds (Tylecote 1986). Later stone
moulds were usually broken open to recover the moulds are usually restricted to ‘ingot’ moulds,
casting, so identification of the objects cast is that is simple rectangular cavities. Some Bronze
often difficult. When clay moulds survive well, the age metal moulds are also known and these were
way they were made and used can be determined used mainly for casting palstaves and socketed
(Figs 38 and 39). Often the largest and most easily axes. Bronze moulds were used during the Roman
identifiable fragments of ceramic moulds are the period for the manufacture of brooches (Bayley et
funnel-shaped in-gates. al 2001). Nevertheless, it is likely that most copper
casting made use of clay moulds.
Two main types of clay moulds are found,
investment (lost-wax) moulds (Fig 38) and piece Large objects such as cauldrons and bells were
moulds (Fig 39). Investment moulds were made also cast in moulds. The process of making these
by first modelling an object in wax and coating it moulds (Fig 40) is well known from medieval
thickly in clay. The clay/wax assembly was then documents such as Theophilus’ De Diversis
fired and the wax melted or burnt out to leave a Artibus (Hawthorn and Smith 1979). Sometimes
fired clay mould. Molten metal was poured into a tallow model was used, the mould was formed
the mould and allowed to solidify, then the mould around it, and then the tallow was melted out.
was broken to remove the casting. Another method was to shape the inner part
of the mould (the core) first, then to make the
Piece moulds were formed in two or more outer part of the mould (the cope) around
sections. An original object, or a pattern made it. The cope was then removed, in pieces if
in the desired shape, was impressed into a lump necessary, and the core trimmed down. When
of clay and locating marks made round the the mould was reassembled there was a void
edge. Another piece of clay was pressed over left between the cope and the core to receive
the pattern. The two valves of the mould were the molten metal. These moulds were broken to
then separated, the pattern was recovered, and remove the casting (Blaylock 2001; Dalwood and
the mould reassembled and sealed (luted) with Edwards 2004; Dungworth and Maclean 2011).
more clay. The mould was then fired and used.
Although the valves of clay piece moulds could As well as the moulds themselves, corroded
be taken apart, they were fragile and therefore spillages of metal can be found. Castings were
are not likely to have been used more than cleaned up (fettled), with surplus metal such as
once. Multiples clay piece moulds could also be flashings (the metal that ran between the valves
assembled using a single pattern which allowed of a piece-mould), runners and sprues trimmed
the casting of several identical objects at once off, and these are also sometimes found (Fig 41).
(Armitage et al 1981; Bowsher et al 2007). Patterns Failed castings, where the molten metal failed to
in wood or lead for making piece-moulds are completely fill the mould, are also found.
known, if rare.

< < Contents 47

4.4 Wrought metalworking particular object, the metal must be heated
between successive bouts of working otherwise
Wrought metalworking describes the processes it will eventually break. This heating stage is
of shaping the solid metal, for example by known as annealing, and it causes the crystalline
hammering or cutting. Unlike ferrous alloys, structure of metal to regenerate, restoring its
copper alloys can be easily worked at room original toughness and softness so that working
temperature. The most commonly used tool can continue. Annealing takes place at temperatures
would be the hammer, although comparison with that could be achieved in a domestic hearth: less
modern practice shows a variety of other tools than 800°C (Dungworth 2013).
could be used, such as strakes. Wrought work
most commonly produced sheets, strips and Large ingots of metal are not usually found on
wires. Larger items could be made from sheet by wrought metal working sites. The metal workers
riveting components together. Early rivets were used small ingots or blanks as their starting
simply small fragments of copper alloy. These point, producing sheets, bars, rods and wires
could be initially cast into shape but were also of metal, which were then worked further to
formed by rolling up a short length of strip or produce finished objects using hammers, files,
sheet. The rivet was inserted into the sheets to gravers, chisels, dies and punches. Anvils made of
be joined and hammered from both sides to form various materials, such as bone, wood and iron,
‘heads’. Most cauldrons made before the later are occasionally found. Evidence for wrought
medieval period were made in this way. metalworking comprising small pieces of scrap
metal, such as turnings and sheet and wire
When copper alloys are hammered at room offcuts, is occasionally found. Metal filings and
temperature the metal becomes work-hardened. offcuts were collected for recycling, sometimes
This increased hardness is often desirable, but in boxes set into workshop floors (Fig 2 and
it also leads to increased brittleness. If a large Zienkiewicz 1993, Figs 13–14). Whetstones and
amount of working is required to produce a abrasives were used to create a good surface
on metal objects, which was then polished.
Alternatively the surface could be burnished with
a hard material such as steel or agate.

Water was used to power hammers for the

production of sheet metal and hollow ware,
such as basins, from at least the 16th century.
This formed a significant industry in some areas,
such as Bristol (Day 1973). Water power was also
used to draw wire; and in the 16th century it was
harnessed to power rolling mills.

The archaeological evidence for the wrought

working of copper alloys is often most apparent
in the objects themselves. Visual examination can
indicate how an object was wrought to shape (eg
early cauldrons). Where the outward form of an
object does not provide obvious clues as to the
forming process, this can be uncovered through
an examination of its microstructure (see p 64).
Figure 42 Aspects of the later, water-powered industry can
Flow chart showing how the product of one be determined from a study of the surviving mill
metalworking process is the raw material of the next. buildings (Day 1973).

< < Contents 48

5 Archaeometallurgical
Processes and Finds:
Lead

5.1 Background 5.2 Smelting

Lead has a low melting point of 327°C and lead The common lead ore is galena (lead sulphide,
ores can be reduced to lead metal below 800°C. PbS) which often contains minor amounts of
Lead is very soft and is easily formed into sheets. silver. The silver content was often the main
It has a tendency to creep, that is, to distort slowly economic reason for mining and smelting the
over long periods of time. Because of its high lead (see p 53–55). There is relatively little
density, lead was often used to make weights. archaeological evidence for early lead smelting
and few of these sites have been studied in
Alloys of lead and tin were used as soft solder and, sufficient detail to fully understand the lead
from the Roman period onwards, they are also used smelting technology. A great many sites are
for casting objects – which are described as pewter. known but not dated (eg Murphy and Baldwin

Lead in summary
Lead is a very soft, dense metal with a low melting point of 327°C. Lead ores were often mined and smelted for the silver
that they contained (p 54–55).

Process Summary Archaeological debris

Smelting Lead ores can be smelted at less than 800°C, Shallow clay depressions have been found
so simple structures could be used, which from the Roman period. Later structures
rarely survive. Early furnaces (bole hills) made were sometimes stone built. Sparse
use of natural draughts. Later, bellows-blown vegetation can indicate lead contamination.
furnaces (ore hearths) were developed, which Some slag and evidence of waterpower can
were subsequently adapted for waterpower. be found. The flues of reverberatory furnaces
Reverberatory furnaces (cupolas) developed in often survive.
the 17th century and were coal fired. Smelting
produced molten lead metal and liquid slags.
The lead-rich slags from early processes were
often re-smelted later.

Lead working Owing to the low melting temperature of lead, Ingots are quite common. Lead sheet, offcuts
domestic pots could be used instead of crucibles and lead-melting dross are sometimes found.
when melting lead. Limestone, wood or antler Moulds, failed castings and sprues indicate
moulds could be used instead of clay ones for that lead was cast.
casting lead.

< < Contents 49

Figure 43
Lead smelting slags are known in small amounts from of lead carbonate on the surface. While some lead slags
the Roman period onwards. They are often glassy, very have a flowed surface, similar to iron-smelting tap slag,
dense and black, green or grey in colour, although the some lack any diagnostic morphology.
surface can be creamy coloured due to the formation

2001; Smith 2006; Smith and Murphy 2010). The recovery of over a hundred inscribed Roman
Later lead smelting is described in a number of lead ingots suggests lead production occurred on
documentary sources but marrying this historical a large scale. Excavations have identified Roman
evidence to the available archaeological evidence lead smelting at a number of sites, including
is a continuing challenge. Llangynelin, Dyfed (Page 2005), Flint (Petch and
Taylor 1924) and Scarcliffe Park, Derbyshire
The common lead ore, galena (PbS), is much (Lane 1973), however, insufficient information
easier to smelt than most other metals. If galena is currently available to determine details of the
is heated in air then it tends to oxidise to the smelting technologies employed. In most cases all
sulphate (PbSO 4) and ultimately the oxide (PbO). that remains of the smelting furnace is a bowl-
Both of these compounds will readily react with shaped depression with signs of burning. No clear
galena to form metallic lead and sulphur oxides evidence for superstructure has been recovered.
(the latter will be dispersed as a gas). Any gangue In contrast with the largely undated bole smelting
minerals (especially silica) will react with some of sites (see below) the Roman smelting does not
the lead oxide to form lead-rich slags (Fig 43).

< < Contents 50

appear to have made use of prevailing winds structures known as ore hearths. Air was blown
to aid the smelting process. into the hearth with bellows and this forced draft
made the process more efficient at extracting
Field survey (plus a small amount of excavation) lead. In Derbyshire these hearths were fuelled
has recognised a large number of early lead with kiln-dried wood, and the kilns can be found
smelting sites in upland regions with associated near the ore hearth remains, but in the northern
lead ores (Murphy and Baldwin 2001; Smith 2006; Pennines peat was used. Ores rejected by bole
Smith and Murphy 2010). These sites are known smelters, as well as bole slags, could be smelted
as boles or bales and some have been traced in ore hearths. The ore hearth began to make use
through place name evidence suggesting that of water-powered bellows technology and so most
some were in use in the medieval or early modern were sited in river valleys adjoining the mining
period. Bole sites are often sited on exposed districts. Water-powered ore hearths continued to
hilltops – it is likely that such sites were chosen be used until the late 19th century (Tylecote 1986).
to exploit prevailing winds. Some sites have also By the 17th century smelters were re-smelting
been identified during field survey through the the slag from ore hearths in structures called slag
presence of lead-tolerant flora or in extreme cases hearths (Paynter et al 2010) which were usually
the complete absence of plants. Few boles have water-powered and often fuelled by coal or coke.
been systematically excavated (eg Anguilano et al
2010; Timberlake 2002). In the later 17th century the cupola was introduced
(King 2001–2). These coal-fired reverberatory
Although parallels for boles are often sought in furnaces consisted of a chamber containing the
the historical literature (Blanchard 1981; Kiernan ore and another containing the coal fire. The
1989), the documentary evidence (largely Tudor heat from the fire was drawn into the smelting
in date) indicates large, if simple, hearths that do chamber. The chimney provided an induced
not match the slight archaeological remains (eg draught which removed the need for bellows and
Anguilano et al 2010). The few boles that have been the provision of doors allowed the smelters to
excavated usually comprise an area approximately manipulate the atmosphere (oxidising-reducing)
1m in diameter with some evidence for burning inside the furnace. The advantages of this
with a scattering of lead slag (Anguilano et al process were yet greater smelting efficiency and
2010). The late 16th-century description of a fuel economy. From the mid-18th century this
Derbyshire bole (Kiernan 1989, 40–43, Figure 4) technology was rapidly adopted and towards
indicates a rectangular furnace 6m long and 3m the end of the century the associated flues
wide with stone walls on three sides (the fourth became very long and complex, with condensing
side being exposed to the prevailing wind. The chambers to collect metal-rich residues. As the
documentary sources suggest that bole smelting stone from these constructions has often been
used wood rather than charcoal and that each robbed, frequently all that remains are the trenches
smelt would consume 30 tons of wood and 40 leading from furnace to chimney. Such furnaces
tons of ore but yield 18 tons of lead. It is likely that were used into the 20th century (Crossley 1990).
the bole smelting would be a seasonal activity
undertaken only a few times a year. Archaeological evidence for many of the latter
processes and furnaces (such as blackwork ovens,
Documentary sources indicate that bole slags ore hearths, slag hearths and cupolas) is limited.
were being re-smelted using charcoal fuel in a While samples of waste materials have been
foot-operated, bellows-blown hearth known as a collected during fieldwork it is sometimes unclear
blackwork oven in Devon by the late 13th century whether these represent raw materials (ie slag
(Claughton 1992) and in other areas somewhat from an earlier process, such as bole smelting) or
later. Historical sources show that during the 16th are waste products.
century, lead smelters changed from boles to

< < Contents 51

5.3 Lead working

Newly smelted lead was cast into ingots, often

known as pigs, which are quite common,
particularly from the Roman period. The main
evidence for lead working, however, is lead
melting dross. This is the oxidised layer of metal,
which forms on the surface of the melt and is
skimmed off before the metal is poured. Other
evidence of melting is harder to detect because
any domestic pot could be used, instead of a
crucible, owing to the metal’s low melting point.
The low melting point also means that the metal
can easily be melted accidentally, so the presence
of melted lead is not necessarily an indication
of lead working. Much lead was used as sheets
and sheet offcuts are common finds. Lead from
buildings was frequently recycled, being easily
melted down and re-cast.

For casting lead or pewter objects, fine limestone,

wood or antler moulds could be used instead of
clay because the moulds did not have to stand
high temperatures (Fig 44). Roman pewter plates
were cast in stacking stone piece-moulds (eg Lee
2009). Antler burrs were carved to act as moulds Figure 44
for late Saxon brooches (eg Newman 1993). As
A later medieval piece-mould made of fine-grained
with copper alloy casting, sprues and failed stone with holes for locating pegs at the corners from
castings are sometimes found. Hereford (length 57mm).

< < Contents 52

6 Archaeometallurgical
Processes and Finds:
Other metals

6.1 Silver and gold (see p 49). Precious metals have similar melting
temperatures to those of copper alloys and were
Unlike most other metals, the main source of melted in clay crucibles. The metals could be
gold is native gold, rather than an ore. Gold cast to shape or, more commonly, worked as
mining occurred in Wales during the Roman solid metals. Both silver and gold are very soft.
period (Burnham and Burnham 2004). Silver was They were alloyed with each other and with other
mainly obtained from argentiferous, or silver- metals, commonly copper, and the alloys have the
rich, lead and the mining of lead was often advantage of being harder than the pure metals
undertaken for the silver that it contained (Bayley 1991a).

Silver and gold in summary

Native gold is the principal source of gold. Silver is mainly obtained from lead ores (p 49). Silver and gold are soft metals
with similar melting temperatures to those of copper alloys. They were commonly alloyed with each other, and with copper
and other metals.

Process Summary Archaeological debris

Refining silver To separate silver from base metals the cupellation process Early cupels are ceramic (heating
and gold was used. This involved melting the silver alloy with added trays). Later ones were made
lead and oxidising the melt. Cupellation could also be used to from bone ash. Litharge cakes
test the purity of silver (assaying). Shallow dishes (cupels) were are formed during large-scale
used for small-scale cupellation and assaying, but large-scale cupellation.
cupellation took place in hearths. Gold refining and assaying
usually did not use lead.

Parting silver To part silver from gold, the silver was removed by reacting it Ceramic parting vessels.
and gold with salt. Later, strong mineral acids were used.

< < Contents 53

 45 46

47 48

Figures 45–48
45. Primary litharge the waste material produced 47. Ceramic cupel or heating tray from Bainbridge
during the extraction of the small amounts of silver Roman fort and settlement. The vitrified upper
that occur naturally in most lead. The litharge has surface is rich in lead and highly coloured. There
a red colour but the surface usually weathers to is a central depression where the assayed metal
a cream colour (lead carbonate). Primary litharge solidified. Sometimes droplets of silver or gold
often occurs in large lumps and lacks the greenish that failed to coalesce became trapped in the area
colour often seen in secondary litharge. surrounding the depression (Gardner 2009).
46. Secondary litharge the waste material produced 48. Bone ash cupels from the Tower of London: the top
during the recovery of silver from debased silver and right examples are unused while the left and
alloys. Secondary litharge is red but often weathers bottom examples have been used. Unused cupels
to a greenish colour due to the presence of small are pale coloured and powdery. The absorbed lead
amounts of copper. Secondary litharge generally in used cupels makes them noticeably heavy for
occurs in smaller lumps than primary litharge. their size. Note also the circular impression in the
centre of the used cupels where the silver formed.

6.1.1 Refining oxidising conditions – copper and other base

Gold and silver were often refined before use, or metals would be oxidised and form a dross or
reuse, especially if the metal had previously been slag while the gold (and any silver) would remain
debased. The purity of gold could be determined unchanged. Silver was removed from gold-silver
by using a touchstone, which was a black stone alloys by a parting process (see below).
used to obtain a smear of metal, the colour of
which was an indication of its purity. Gold could Silver was extracted from lead (and recovered
be refined somewhat in a crucible by using from debased silver) by a process known

< < Contents 54

as cupellation. Slightly different cupellation that the gold was simply melted in strongly
processes were used depending on the scale of the oxidising conditions to burn out the base metal
refining and the nature of the starting material. impurities, perhaps with a flux of some sort.
Nevertheless, the processes were all fairly similar. Ceramics used for gold assaying are usually made
The raw material (lead or debased silver) was of harder, more refractory, fabrics than those used
placed in a hearth with fuel and a strong air blast for silver.
used to oxidise the base metals which were then
absorbed into the hearth lining. The silver (as well Later processes were developed to improve the
as any gold) remained at the centre of the hearth. extraction of silver from lead. In the Pattinson
Lead-rich hearth lining (litharge) is the most process (1833) the lead was allowed to cool
commonly recovered evidence for this process. slowly; the first metal to solidify was rich in lead
which left the remaining molten lead enriched
The use of the cupellation process for the in silver. The silver was then recovered from
extraction of silver from freshly smelted lead this enriched lead using cupellation. The Parkes
(primary cupellation) is known from a few Roman process (1852) used the addition of zinc to extract
sites (eg Bayley and Eckstein 1998; Dunster and silver from lead: zinc and lead are immiscible but
Dungworth 2013). The litharge from these sites silver readily dissolves in the zinc. The zinc-silver
tends to occur as fairly large pieces up to 60mm compounds could be easily separated from the
thick and probably from heaths around 0.6m in lead and then heated to drive off the zinc.
diameter (Fig 45). The other elements present
in litharge cakes suggest that the hearth lining 6.1.2 Parting
comprised clay, possibly mixed with wood ash. Cupellation could not be used to separate, or part,
silver from gold, so a different technology was
The cupellation process used to recover silver developed. Two different parting processes are
from debased silver alloys is well known from a known from historical and archaeological sources:
number of urban sites from the Roman period salt parting (Bayley 1991a) and sulphur parting
onwards (Girbal 2011). In order to ensure that the (Rehren 1996), although the former appears to
hearth lining effectively absorbed the oxidised have been more common.
base metals it was common to add lead to the
debased silver. The lining of the hearth could be For salt parting, the gold-silver alloy was hammered
made from clay (and wood ash) or from bone ash. into thin sheets, packed into a pot interleaved
with a ‘cement’ of crushed brick or tile mixed with
Before the cupellation process was used to common salt (NaCl). The pot was then sealed up
recover silver from debased silver alloys it was and heated, but to a temperature below the melting
common to carry out a test first to determine the point of the metal. The salt would react with the
proportion of silver present. This test comprised silver, forming silver chloride, which would then
simply of a very small-scale cupellation: the be absorbed by the cement and the walls of
weight of silver was compared to the weight of the the pot (Fig 49). When the pot cooled, the gold
test piece to determine silver content. From Roman could be removed and re-melted and the cement
and Saxon times small ceramic dishes, often called smelted to recover the silver (Bayley 1991a).
heating trays, were used as cupels and makeshift
varieties were sometimes made from potsherds Sulphur parting could employ sulphur or
(Fig 47). The reaction of the litharge with these metallic sulphides (such as antimony sulphide)
ceramics produced a glassy surface. By 1600 AD cupels which would react with the silver (to form silver
made from absorbent bone ash were being used (Fig 48). sulphide) and leave the gold unchanged. With the
introduction of distillation in the later medieval
Analysis of some heating trays used for gold period, the method of parting changed to one
assaying has failed to detect lead. This suggests using strong mineral acids.

< < Contents 55

 washing (it is much denser than the sediment).
Stream tin was probably exploited before lode
tin, however, archaeological evidence for either
activity is rare before the 18th century.

Tin ores would usually have to be crushed and

sorted to reduce the proportion of gangue
minerals (Gerrard 2000). From the 14th century
the crushing (or stamping) of ore increasingly
used water-powered hammers. Stamp mills can
be positively identified by the presence of mortar
stones with saucer-shaped hollows on which the
Figure 49
ore was crushed. The partly crushed material
Parting vessels were not always purpose-made and
a wide variety of vessels were used; all were lidded
from the stamp mill was ground to a fine powder
or would have been sealed with clay. They are in a crazing mill. Around the mid-16th century
the only metal-working vessels that are normally most crazing mills were abandoned because
oxidised fired. They are readily identifiable as they of the introduction of the more efficient wet
usually have a pale pink-purple colour on the inside
rather than the orange-brown normally associated
stamping process. Water was used to separate
with oxidised-fired ceramics. Sometimes areas of the dense cassiterite from the lighter gangue.
lemon-yellow colour, specular haematite crystals Some of the clearest remains of mining are the
(as here on a fragment from Lincoln), or even flecks heaps of waste gangue minerals that surround
of gold are visible. Some parting vessels show no
surface vitrification, while others have a thick,
mine site.
exterior glaze that can be turquoise or deep green.
6.2.2 Smelting
There is little archaeological evidence for early
tin-smelting processes. A number of tin ingots
have been found with a roughly plano-convex
6.2 Tin shape, however, most of these have not been
recovered from an archaeological context and so
Tin is a soft, white metal with a melting point cannot be dated. High grade cassiterite can be
of 232°C. Its primary use was in alloying reduced to metallic tin at fairly low temperatures
copper to form bronze, however, it was also in a bowl furnace (Timberlake 1994). Lower grades
alloyed with lead to form pewter and from of cassiterite can be smelted at low temperatures,
the 18th century it was applied to iron sheets however, the presence of abundant gangue
(tinplate). Tin ores are found in only a few minerals would prevent the smelted tin from
locations globally and the important cassiterite collecting at the base of the furnace. If higher
(SnO 2) deposits in Devon and Cornwall were temperatures were employed (eg >1000°C) then
exploited from the Bronze Age onwards. the gangue minerals would combine to form a
slag. This slag would tend to contain some tin
6.2.1 Mining depending on the smelting temperature and
Cassiterite can be found as veins (lodes) that the atmosphere (oxidising-reducing) inside the
extend to considerable depths but much of this furnace. The rather amorphous shapes of the few
ore was not exploited until the 18th century when examples of early tin smelting slag that have been
steam power made deep mining possible. Streams reported indicate that it would have been fairly
have also eroded the exposed veins of lode tin viscous (Lawson-Jones 2013; Malham et al 2002).
and deposited this in sediments in the valleys that Most reported tin slag occurs as relatively small
radiate from the area of lode tin. These deposits fragments and these could have been broken up
could be excavated and the stream tin sorted by to recover some of the trapped tin.

< < Contents 56

 (Lee 2009) and became popular again in the
In the 14th century water wheels began to be used late medieval and post-medieval periods (Egan
to power the bellows in what quickly became 2005). Its low melting temperature means that
known as ‘blowing houses’ (Greeves 1994). The pewter does not need to be melted in refractory
furnaces were largely built of stone blocks (often crucibles. Medieval and later texts often refer to
granite) although they were probably lined with the use of iron pans or pots for melting pewter;
clay and the preferred fuel was charcoal. The earlier pewter melting may have used domestic
molten tin was tapped from the furnace hearth pottery. The archaeological record provides
into a trough and was then ladled into smaller little direct evidence for the casting of pewter.
moulds or troughs (Gerrard 2000). From the Stone moulds for casting plates are known from
beginning of the 18th century the coal-fuelled the Roman period (Lee 2009) and it is likely that
reverberatory furnace was used for smelting clay was used for at least some types of object.
tin ores (Smith 1996). Smelting debris from a Documentary sources suggest that copper alloy
number of sites (Farthing 2005; Malham et al 2002) moulds were used from at least the 15th century.
including manually-powered bellows smelting,
blowing houses and reverberatory furnaces has Tin has occasionally been used to coat the surface
been examined. It is striking that the chemical of another object (copper alloys and iron) to give
composition of slags from sites with manual a silvery white appearance. From the early 18th
bellows and water-powered bellows show few century increasing quantities tin were used in the
differences; the reverberatory furnace slags, manufacture of tinplate (p 38).
however, have quite distinct compositions (higher
iron and calcium content and usually less tin).
6.3 Zinc
6.2.3 Alloying
Tin has been used with copper since the Bronze Zinc has low melting (420°C) and boiling
Age to produce the alloy bronze. The vast majority temperatures (907°C) and so any attempt to
of bronzes contain 8–12wt% tin which yields a reduce zinc ore using charcoal fuel at around
substantially harder metal than pure copper but 1000°C, would result in the zinc metal being
which retains enough toughness to be serviceable produced as a vapour and so lost as fumes.
for the manufacture of tools and weapons (Allen Consequently zinc was not generally available in
et al 1970). Bronzes with greater amounts of tin Europe until the post-medieval period. Zinc was
have found occasional specialist uses: 20–25% tin manufactured in India and small quantities were
produces a silvery white metal which was used in imported from the 16th century. Zinc manufacture
Roman times for the manufacture of mirrors, and was introduced into Britain in the 18th century.
the brittleness of the alloy made it suitable for the Nevertheless, zinc is found in much earlier copper
manufacture of church bells (a lower tin content alloys, such as brass (copper-zinc alloy).
would give a tougher metal which would not ring
when struck). From the 1st century BC brass was produced by a
cementation process. Ground zinc ore was mixed
From the Roman period onwards tin was alloyed with copper and charcoal in a crucible which
with lead to produce soft solder and pewter. was then sealed and heated to between 950°C
Alloys of tin and lead have very low melting and 1000°C (below the temperature at which
temperatures (<200°C) and so can be used to copper or brass melts). The charcoal provided
join other metals together. Soft solder can be a reducing atmosphere in the crucible, so the
used by bronze smiths as well as by plumbers zinc ore was reduced to metallic zinc metal,
and window makers who need to join lead which was absorbed by the solid copper (Bayley
components together. Pewter was used for the 1998). Crucibles for brass cementation have been
production of tableware during the Roman period recognised from Roman (Bayley 1998) and post-

< < Contents 57

Figure 50
Base fragment of a 16th-century brass cementation
crucible from Taynton, Gloucestershire. The intense
blue-purple colouration is due to the reaction between
the crucible and zinc.

medieval contexts (Dungworth and Wilkes 2010). deliberate additions; instead the ores used
Prolonged use of crucibles for brass cementation would contain impurities which could be carried
can cause the clay to turn blue or purple (Fig 50). through to the finished metal. The extent to
which finished metal will contain such impurities
During the 18th century the direct manufacture of (sometimes referred to as trace elements) will
metallic zinc was introduced to Britain. Zinc ore depend on the nature of the ore and the smelting
and a reducing material (charcoal or coal) was and refining processes employed. Common
packed into a ceramic retort. The retorts were impurities in gold include platinum group
placed inside a conical furnace (similar to a steel metals while copper alloys can often contain
cementation furnace). The holes at the base of arsenic, nickel and antimony. While impurities
the retorts were connected by pipe to buckets can be removed by refining, this will often mean
outside the furnace. The heat of the furnace a lower yield metal (eg losses in refining slag).
reduced the zinc ore to metallic zinc which then In some cases the presence of impurities may
descended the pipes and collected in the water- have beneficial properties. The arsenic in the
filled buckets. The retorts were made from similar earliest copper artefacts (Allen et al 1970) simply
clays to the brass cementation crucibles and reflects the nature of the ores used. Nevertheless,
usually have the same intense blue-purple colour early copper smiths often seem to have been
(Dungworth and White 2007). aware of the physical characteristics of copper
with or without arsenic and used these alloys
appropriately (McKerrell and Tylecote 1972). In
6.4 Other metals the medieval period the copper alloy used to
manufacture cauldrons (Dungworth and Nicholas
The metals described above are those that 2004) was particularly rich in impurities (it was a
are most frequently encountered in the bi-product of extracting silver from argentiferous
archaeological record. The chemical analysis copper ores). In this case the impurities had no
of these metals will often reveal the presence detrimental effect on the finished metal goods
of small amounts of other metals. In most (the interior of the cauldron would be coated in
cases these other metals would not have been tin) and the impurities ‘bulked out’ the metal.

< < Contents 58

7 Non-Metallurgical
Residues and
Materials
A variety of materials are sometimes confused with the debris from
metallurgical processes. These include naturally-occurring materials
(especially geological material), debris from non-metallurgical
industries, conflagration debris, heat magnetised residues and plastics
and tarmac.

7.1 Geological materials 7.2 Debris from non-metallurgical

industries
Some types of iron ore (especially bog iron
ore), pyrite (iron sulphide) nodules, pieces of Few early non-metallurgical industries reached
puddingstone, and lava can all be mistaken temperatures high enough to vitrify materials
for slag. and so produce debris that can be recovered
archaeologically. Occasionally pottery, brick and
Iron-rich concretions are often recovered during tile kilns became too hot and the ceramics inside
archaeological excavations. These comprise were over-fired. Pottery wasters have slumped and
lumps of soil or sediment joined together by iron distorted shapes and can be glassy and blistered
compounds (Fig 53). They form as a result of the (Historic England 2015b).
re-deposition of iron compounds in a similar
manner to the natural phenomenon of iron Any industry which used wood or another organic
panning. The process is sometimes enhanced by fuel to provide heat would produce quantities
the presence of iron objects or scrap metal. of ash. If the temperatures were sufficiently high
then this ash would vitrify and produce a durable
residue – vitrified fuel ash (Fig 54). Some vitrified
fuel ash can form as a result of the fires used in
some metallurgical processes, however on its own
it is not diagnostic (Biek and Bayley 1979).

< < Contents 59

 51 52 53

54 55

Figures 51–55
51. Section through a puddingstone boulder; the poorly to a magnet but do not have the typical
rounded exterior can be mistaken for iron slag. vitrified surfaces of metal working debris.
52. Pyrites nodule. The weathered outside (right) can 54. Vitrified fuel ash from Furzton, Buckinghamshire.
look like iron slag but the interior (left) has a silver They are lightweight, vesicular and fragile, and are
colour and radial structure. usually off-white to green or mid-grey in colour,
53. An iron concretion consisting of pebbles and sand generally much paler than iron-working slags.
grains bound together by iron compounds. They 55. Vitrified coal ash (clinker) from the 17th-century
are amorphous orange-brown lumps that respond coal-fuelled glass furnace at Silkstone, South
Yorkshire.

The working of glass requires very high which can be difficult to distinguish from clinker
temperatures and glass production sites yield a formed in different contexts (eg blacksmithing or
variety of diagnostic residues (English Heritage heating steam engines).
2011a). This will often include recognisable glass
waste as well as crucibles which crucially have The manufacture of salt usually made use of
most vitreous material inside the crucible while ceramic containers (briquetage) in which brine
the vitrification of crucibles for melting metals was heated. Briquetage is usually fairly soft and
is usually restricted to the exterior. Glasshouses oxidised-fired (orange-red) but does not have any
of the 17th century and later often yield large vitrified surfaces.
quantities of vitrified coal ash (clinker, Fig 55)

< < Contents 60

7.3 Conflagration residues 56

Where a structure has burnt down (deliberately

or by mischance) durable residues can form
and these can be mistaken for slags. The ashes
from burnt thatch or structural timbers can
react with daub walls to form vitrified fuel ash
and/or vitrified clay. This material is not readily
distinguishable from the vitrified fuel ash
described above.

Buildings that contain a greater variety of

building materials can produce more complex
conflagration debris (Fig 56). Lead fittings within
a building (window lead, plumbing or roof lead)
57
will easily melt, oxidise and react with brick, tile
or stone to form glassy residues. A severe fire
can even melt window glass which will react with
other materials such as brick and tile.

7.4 Heat-magnetised residues

Soils samples taken for the recovery of plant

macrofossils (English Heritage 2011b) often
contain magnetic residues which are mistaken for
hammerscale (Keys 2012). A variety of naturally-
occurring materials and sediments containing at
least some iron can acquire a weak magnetism as
a result of exposure to heat.
Figures 56–57
56. Daub and ceramic tiles from a Roman building in
7.5 Plastics and tarmac London destroyed by fire, which are stuck together
by an accidentally-formed lead-rich glass.
57. Partially melted lumps of plastic. These superficially
Modern plastics and tarmac that have been resemble blast furnace and other vitreous slags
subject to high temperatures will easily melt
and can take on colours and forms which
resemble a range of metalworking slags (Fig 57).
These materials can usually be distinguished
from slags by their low density, softness and
low melting temperature.

< < Contents 61

8 Scientific Techniques
Applied to
Metalworking
This section provides an introduction to a few of the scientific
techniques that have been applied to the study of early metalworking,
including geophysics, microscopy and various methods of chemical
analysis. The data obtained can be used to explore a wide range of
issues, such as resource exploitation, economy, trade and exchange
and cultural affinities.

The scientific study of early materials can provide a wealth of

information about the raw materials and manufacturing techniques
used. Only the most commonly used methods are described.

8.1 X-radiography 8.2 Geophysics

X-radiography is an imaging technique that Geophysical techniques have considerable

is particularly useful for examining and potential in the study of early metalworking sites
recording archaeological metalwork and some and are useful tools for assessing the scale, date,
types of debris. The main archaeological preservation and significance of sites (English
applications are the identification of objects Heritage 2008; Vernon et al 1999). The two
and examination of their morphology, methods geophysical techniques most commonly applied
of construction and condition (English Heritage on metal working sites are magnetometry and
2006a). X-radiography has been used to magnetic susceptibility.
identify inlays, stamps, weld lines and pattern-
welding in iron artefacts, examine crucibles Magnetometery with a fluxgate gradiometer or a
and moulds (where metallic particles might total field instrument (eg alkali vapour) is usually
be trapped in the ceramic fabric), distinguish carried out as a prospection technique, as these
slag from corroded iron artefacts, and detect instruments can take readings continuously,
hammerscale and other debris in soil samples. making it possible to survey large areas quickly.
Gradiometers record localised variations in the

< < Contents 62

gradient of the earth’s magnetic field. These of non-ferrous metalworking sites should detect
variations can be caused by fired structures and hearths and areas of burning, and possibly large
magnetic materials (metallic iron and some slags) dumps of crucibles, moulds or other debris.
as well as by underlying geology. High-resolution Domestic hearths, however, can give similar signals.
gradiometer surveys, in which the data is gathered
at smaller intervals than the norm (for example
0.25m), are used for distinguishing features such 8.3 Archaeomagnetic dating
as furnaces, typically 0.5m in diameter.
Archaeomagnetic dating is a technique that can
Magnetic susceptibility is a measure of the be used to date the fired clay of furnaces, hearths
degree to which a body becomes magnetised. and slag that have cooled in situ (English Heritage
Human activity can enhance the susceptibility 2006b). Materials such as clay, which contain
of surrounding soils. Magnetic susceptibility is a significant proportion of magnetic minerals,
rarely used for survey of large areas, but detailed acquire a remanent magnetisation when they are
work can be very informative. This technique fired. This magnetisation is in the same direction
has the advantage of only measuring to a depth as that of the Earth’s magnetic field at the time.
of about 100mm below the coil (depending The precise direction of the Earth’s field varies
on the size of the coil), therefore reducing the over time; hence, if a fired clay feature is found
amount of interference from nearby features with that has not moved since it was last fired, it is
strong responses. It can provide an estimate of, possible to date the firing using the direction of
for example, the amount of hammerscale in a magnetisation recorded in the feature. Dating
sample because this can be related to the signal ironworking features can be challenging due to
magnitude. Measurements are made either on the the formation of magnetic materials, not least an
soil in situ or on samples recovered from a site iron bloom. Samples for archaeomagnetic dating
(including cored samples). should be taken by, or under the supervision of, a
relevant specialist.
In situ smelting furnaces result in distinctive
dipolar features in magnetometer surveys, which
can be further emphasised if the data is not clipped 8.4 Microscopic examination
and is plotted on a coloured scale. Magnetic
susceptibility survey can also indicate, by a high Optical and electron microscopes can provide
response, the location of iron working. Bloomery invaluable information on the surface condition
iron slag typically produces a higher magnetic and internal microstructure of a wide range of
response than topsoil. Magnetic surveys of slag rich materials, including metals and metalworking
areas usually produce a very ‘noisy background’, debris. The principal types of microscope used
with extreme peaks. Large dumps of slag can are low and high power optical microscopes, and
be so strongly magnetic that they distort the scanning electron microscopes.
magnetic field for several metres around, masking
responses from adjacent occupation features. Low power (x1–x20 magnification) optical
examination (eg Fig 31) can reveal traces of
Survey of iron smithing sites can reveal strong metal on crucibles, traces of silvering or other
magnetic responses in areas (workshops) where decoration on a metal artefact, or tools marks
hammerscale is concentrated. A ground-level hearth and other features diagnostic of the method of
should also provide a significant response, although manufacture (eg casting seams). It should be used
waist-high hearths rarely survive in situ. The before other analytical or investigative techniques
position of such a hearth (or of an anvil) can be in order to evaluate what further analysis will be
indicated by a low response in an area surrounded useful, whether there are any features in particular
by high values (Mills and McDonnell 1992). Survey that require analysis, for example decorative inlay,

< < Contents 63

 58 59

Figure 58 Figure 59
58.Optical microscope image of high-carbon steel 59. A back-scattered electron image of prehistoric
from an experimental bloom smelt. The long white iron working slag (Shooter’s Hill, London) showing
laths are iron carbide (cementite) and the areas several different phases. The white areas are wüstite
inbetween are a mixture of cementite and ferrite (and iron oxide) while the larger grey crystals are
(plain iron). fayalite (and iron silicate).

and also to ensure that any data obtained material has been heat treated or quenched, for
is from representative areas. example to increase the hardness of the metal, this
will also be apparent. Steel and iron were sometimes
High power optical microscopes (x50–x1000 welded together to form composite artefacts. Such
magnification) can only be used on flat, structures are frequently found in edged tools and
polished specimens to determine the internal weapons. Techniques for combining different
microstructure of materials. Scott (1991) provides alloys might have important cultural implications.
a good introduction to the structure of metals, For example, in many Saxon knife blades a steel
metallography and the phase diagrams that help edge was butt welded to an iron back, while Anglo-
explain the microstructures it reveals. Metallography Scandinavian smiths favoured ‘sandwiching’ the
requires the removal of a small sample, which is steel between two low carbon sides.
then mounted in a resin block and polished.
The shape of the metal crystals in non-ferrous
Polished metallic samples can be etched to alloys will show how the object was produced,
reveal the crystal structure of the metal (Fig 58). for example cast alloys generally have the
From this an assessment can be made of the characteristic dendritic structure. An additional
type of alloy, its mechanical properties and the tool frequently used in metallography is hardness
ways in which it was treated during manufacture testing, which gives a direct measurement of the
and use (eg Allen et al 1970; Blakelock and mechanical properties of small samples.
McDonnell 2007). Metallography can also identify
the methods used to apply surface treatments, Scanning electron microscopes (SEM) use a
such as gilding, silvering and tinning. The beam of electrons, rather than light, to examine a
shape of any non-metallic inclusions often sample. The advantages of electron microscopes
shows the way the artefact has been wrought. are that a much greater magnification and
depth of focus can be obtained. Images can be
Different iron alloys (plain iron, steel and phosphoric obtained using a variety of detectors, of which
iron) can be identified using a microscope. If a the secondary electron detector and back-

< < Contents 64

scattered electron detector are most widely used. 60
Secondary electron detectors provide an image
of the topography or shape of a sample (see
Fig 30). Back-scattered electron detectors show
the compositional differences across a sample
(usually a flat, polished sample), since areas with
different compositions are seen as varying shades
of grey (Fig 59). Sample preparation techniques
vary depending on the mode in which the SEM is
to be used. The SEM can be used in conjunction
with analytical techniques (EDS and WDS), which
are described below. 61

8.5 Chemical analysis

A variety of different analytical techniques are

available depending on the questions that are
being asked, the nature of the material, and
constraints associated with sampling, costs and
time. The most common analytical techniques
determine either the chemical or mineralogical
composition of a material (Pollard et al 2007). The
chemical analysis of a material can be qualitative
(simple presence or absence) or quantitative
(proportions of different elements in percentages
or parts per million). Many archaeological Figures 60 and 61
materials are heterogeneous and corroded; 60. An EDXRF spectrum obtained from a crucible used
therefore, analysis of very small samples or of for melting copper alloys, from Mucking, Essex.
61. SEM image (bottom left) and five X-ray maps for
surface layers can be misleading.
various elements in a litharge cake showing several
different phases. The red (silver) and green (copper)
X-ray fluorescence (XRF) is one of the most widely maps show a silver-copper droplet near the centre,
used methods of chemical analysis in archaeology. the white (lead) map show the distribution of lead
throughout the sample. The light blue (calcium) and
A beam of X-rays is directed onto an object, or
yellow (phosphorus) maps show the presence of
sample, which then emits an X-ray spectrum. The calcium phosphates which derive from the bone ash
spectrum (Fig 60) contains peaks for each of the hearth lining.
elements present in the object or sample (this
does not include light elements associated with
organic materials). XRF spectra are detected in
one of two ways: energy-dispersive detectors EDXRF can be used qualitatively on whole
(EDXRF) allow the simultaneous detection of the artefacts (so long as they can be fitted into the
whole x-ray spectrum, while wave-dispersive sample chamber – typically 100mm across) and
detectors (WDXRF) measure the intensity of each causes no damage. Used in this way, EDXRF
characteristic peak individually. EDXRF is relatively permits the identification of the range of elements
cheap and quick, and can determine the presence present in a material, for example the technique
of most elements within a few seconds. WDXRF is can determine if a crucible was used for melting
more expensive and slower, but is more accurate copper alloys or silver. EDXRF is effectively a
and can detect smaller amounts of each element. surface technique: analysis of corroded objects

< < Contents 65

will usually provide data on the surface corrosion 8.6 Isotope analysis
rather than the uncorroded core. EDXRF can be
used quantitatively, but only where samples are The proportions of different isotopes of the
removed, mounted in resin and polished. same element can provide insights into the
origin of archaeological materials. The principal
XRF instruments are also available as small units application in archaeology is the analysis of lead
which can be taken into the field – variously isotopes in lead, copper alloys and silver. The
referred to as portable XRF (pXRF) or Handheld relative abundance of these isotopes characterise
XRF (HH-XRF). All of these field instruments are the ore source, but the lead isotopes in different
EDXRF and will provide good limits of detection British ore sources are similar (Rohl and Needham
for metals but reduced sensitivity for light 1998). Most isotope analysis has been undertaken
elements (Mg–P). Portable or handheld XRF are using Thermal Ionisation Mass Spectroscopy
well suited to the analysis of large numbers of (TIMS), however ICP-MS has seen increasing use
objects or samples which cannot be brought to a for isotope analysis.
laboratory for analysis (Shugar and Maas 2012).
These instruments can be used to carry out
geochemical surveys, especially of metalworking 8.7 X-ray diffraction
sites (eg Dungworth et al 2013).
X-ray diffraction can determine the structure
Similar XRF spectra are also generated in a SEM. of a compound, as opposed to the chemical
This can be fitted with an energy dispersive (EDS) composition. A small powdered sample is required.
analyser. Alternatively an SEM can incorporate XRD is useful because many materials contain
wavelength dispersive spectrometry (WDS) and, the same elements but have different structures,
if dedicated to analysis using WDS, is referred to for example iron ores. This technique can only
as a microprobe, and the technique as electron identify crystalline materials. This technique
probe microanalysis (EPMA). is also useful for analysing corrosion products,
precipitated salts, pigments and soil samples.
Most analytical SEMs permit great flexibility.
Multiple element analysis can be undertaken of a
single spot (down to a few microns in diameter)
or of larger predetermined areas. Line scans and
maps can be used to show the distribution of
individual elements in one or two dimensions (Fig
61). This is particularly useful for the analysis of
such heterogeneous materials as slags and iron.

A number of analytical techniques generate

characteristic spectra in the visible spectrum
rather than as X-rays. The most widely used
instrument is inductively coupled plasma mass
spectrometry (ICP-MS). For this technique a small
powdered sample (typically 40mg) is taken, for
metallic samples this is usually done by drilling.
Figure 62
The sample itself is destroyed during analysis as
XRD spectrum of fragment of a steel cementation
it is dissolved in acid. ICP-MS can give very good
furnace (Coalbrookdale, Shropshire). The range of
accuracy with limits of detection limits 1ppm (or mineral phases present gives some clues as to the
less) for many elements. It is a bulk analysis method temperature attained.
and so cannot distinguish between different
components or phases of a composite material.

< < Contents 66

9 Where to Get Advice
Historic England provides advice on all aspects of archaeometallurgy.
In addition, specialist interest groups, in particular the Historical
Metallurgy Society, provide a forum for the discussion of
archaeometallurgical research. A variety of resources are
available and many of these are now available online.

9.1 Historic England latest research, and explore a wide range of

metallurgical landscapes and locations. The
Advice on archaeometallurgy can be sought Society publishes a peer-reviewed journal
from Historic England which employs Science (Historical Metallurgy) as well as occasional
Advisors ( www.HistoricEngland.org.uk/advice/ papers, including Metals and Metalworking: a
technical-advice/archaeological-science/science- Research Framework for Archaeometallurgy
advice/ ) who provide support and advice on all ( http://hist-met.org/publications/hms-occasional-
aspects of archaeological science. Advice can publications.html ) (Bayley et al 2008). The
also be obtained from David Dungworth and Historical Metallurgy Society also provides
Sarah Paynter who work in the Historic England Archaeological datasheets ( http://hist-met.org/
archaeological science laboratories ( www. resources/datasheets.html ) and the illustrated
HistoricEngland.org.uk/research/approaches/ catalogue of the National Slag Collection
research-methods/Archaeology/ancient-technology/ ). ( http://hist-met.org/resources/national-slag-
Historic England occasionally provides training collection.html ).
days for archaeologists on how to recognise and
deal with slags and other industrial debris.
9.3 Regional Research Frameworks

9.2 The Historical Metallurgy Society A number of Regional Research Frameworks also
provide a review of current knowledge relating
The Historical Metallurgy Society ( http://hist-met. to archaeometallurgy. Topics that would benefit
org/ ) is dedicated to the exchange of information from further research in a given region are
and research in all aspects of the history of metals highlighted. The Regional Research Frameworks
and associated materials from prehistory to the can be accessed through the Association of Local
present. The Society holds several conferences Government Archaeological Officers website
and meetings each year which showcase the ( www.algao.org.uk/england/research_frameworks ).

< < Contents 67

9.4 Published archaeometallurgical 9.6 ADS Grey literature
research
The rapid increase in the recording and reporting
Published reports (both in specialist journals and of archaeological remains that followed from
within archaeological site reports) form a vital 1990 (due to changes in the legal framework
resource for understanding the archaeological for funding archaeology in the UK) has not
evidence for early metalworking and the been followed by an equally rapid expansion
bibliography contains many relevant examples. of conventional publication. The requirement
Many of these publications are now available (as part of the planning process) to record and
online as publishers move to make all content report on archaeological remains that would
available electronically. be lost as a result of construction work has
led to the proliferation of ‘grey literature’. This
comprised detailed archaeological reports which
9.5 Historic England Reports were provided to Historic Environment Records
although many of these were not published as
Historic England has undertaken specialist such. In order to make this data more accessible,
examination of metalworking assemblages for reports are now deposited with the Archaeological
over 40 years (eg Bayley 1991c; 2008; Bayley and Data Service ( http://archaeologydataservice.ac.uk/
Eckstein 1998; Blakelock 2005; Dungworth et al archives/view/greylit/index.cfm )
2013; Dunster and Dungworth 2012; Girbal 2011;
Mills and McDonnell 1992; Phelps et al 2011). The
specialist reports are often made available ahead
of the production of the relevant site reports
and often contain supplementary detail that is
not repeated in the final publication. The entire
library has been digitised and made available
online ( http://research.historicengland.org.uk/ ).

< < Contents 68

10 Glossary
Alloy the properties of pure metals can be dramatically Mine in order to obtain ores it is usually necessary to
changed by combining them or adding non-metallic dig into the earth. In many cases this might consist
elements to form alloys. For example, steel is an alloy of little more than a pit or quarry. The term mine
of iron and carbon; bronze is an alloy of copper and tin. is usually reserved for the more complex system of
tunnels and shafts that are used to extract ore.

Bloom the lump of iron that forms inside an early Mould one technique for shaping metals is to melt and
iron smelting furnace. The bloom formed below pour them into a container. Once the metal solidifies
the metaling temperature of the metal and so has it takes on the shape of the container. Moulds were
a spongy appearance. usually made from clay, but could also be made from
metal, stone, sand or bone. Moulds were not usually
exposed to high enough temperatures to vitrify them.
Crucible is a vessel to hold a metal while it is melted.
Metals are melted to refine them or before casting Non-Ferrous the principal non-ferrous metals used
them in moulds. Crucibles were usually made from before the 20th century were copper, tin, lead, zinc,
refractory ceramics and, because they were exposed silver, gold and mercury, and alloys of these metals.
to high temperatures, the clay was sometimes
partially vitrified.
Ore many rocks and minerals contain metallic
Ferrous the principal ferrous metals used before the elements but not all are ores. A rock containing
20th century were cast iron, steel, phosphoric iron metallic elements can only be regarded as an ore if the
and plain iron. technological, social and economic conditions enable
people to extract the metallic element(s) by smelting.

Furnace is a structure used to hold the ore as the metal Refine the initial product of most smelting processes
is extracted from it by smelting. Furnaces were usually is an impure metal, which is then refined. The refining
made from clay and, because they were exposed to process depends on the nature of the metal and the
high temperatures, the clay was sometimes partially available technology. Copper was often refined by
vitrified. The archaeological remains of furnaces and melting and partially oxidising it to remove impurities.
hearths are often similar. Bloomery iron, because of its high melting point, was
often smithed to squeeze out any slag still trapped inside.
Hardness is a measurement of the strength of a
material (its ability to resist plastic deformation). Refractory materials are those which can stand high
Hardness is measured by making an indentation in a temperatures without vitrifying.
polished sample of metal, usually with a diamond and
a known weight.
Slags are vitreous waste products of many
Hearth is a structure used to obtain the temperatures metalworking activities. Slags can be produced during
necessary to work metal, the exact temperature smelting, refining, smithing and even during casting of
depending on the metal being worked and on the metals. Most ores contain unwanted components (eg
process used. Hearths were used to melt non-ferrous silica) and these are removed during smelting as a slag.
alloys in crucibles, anneal copper alloys and heat iron The size, shape and composition of slags are related to
before smithing. Hearths were usually made from clay the processes that produced them.
and, because they were exposed to high temperatures,
the clay was sometimes partially vitrified. The
Smelt the process of extracting metal from ores
archaeological remains of hearths and furnaces are
is smelting. This is usually carried out at high
often similar.
temperatures in a furnace, using a fuel such as charcoal.

< < Contents 69

Smith most metals can be shaped while solid by Toughness is a measure of the energy required to
hammering (smithing). In some cases (eg iron) the break a material. It is difficult for a crack to grow in a
metal needs to be heated in a hearth to make it tough material, whereas a crack in a brittle material,
sufficiently soft to allow easy smithing. In some cases such as a glass or ceramic, will grow very rapidly.
(eg copper alloys) a metal is made much harder by
smithing. This work-hardening can be removed by
heating (annealing) the metal. Vitrification is the change into a glassy (vitreous) state,
brought about by heating a material. The temperature
at which this change takes place can be reduced by
Strength the strength of a material is a measure of the
the presence of fluxes, which can be accidentally or
stress (load per unit area) it can support before failing.
deliberately added.

< < Contents 70

11 Bibliography
Alcock, L 1963 Dinas Powys. Cardiff: University of Bayley, J 1984 ‘Roman brass making in Britain’.
Wales Press Historical Metallurgy 18, 42–43

Allen, I M, Britton, D and Coghlan, H H 1970 Bayley, J 1991a ‘Archaeological evidence for parting’,
Metallurgical Reports on British and Irish Bronze Age in Pernicka, E and Wagner, G A (eds) Archaeometry ’90.
Implements and Weapons in the Pitt Rivers Museum . Heidelberg: Birkhäuser Verlag, 19–28
Oxford: Oxford University Press

Bayley, J 1991b ‘Anglo-Saxon non-ferrous

Allen, M J, Gardiner, J and Sheridan, A (eds) 2012 Is metalworking: a survey’. World Archaeology 23, 115–130
there a British Chalcolithic? People, place and polity in
the late 3rd millennium. Oxford: Oxbow
Bayley 1991c Analytical Results for Metal- and Glass-Working
Crucibles from Frere’s Excavations at Verulamium, Herts.
Ancient Monuments Laboratory report 68/1991. London:
Andrews, J 1991 Edge of the Anvil. A resource
English Heritage
book for the blacksmith. London: Intermediate
Technology Publications

Bayley, J 1992 Non-ferrous Metalworking from 16–22

Anguilano, L, Timberlake, S and Rehren, T 2010 Coppergate. Archaeology of York 17/7. London: CBA
‘An early medieval lead-smelting bole from Banc
Tynddol, Cwmystwyth, Ceredigion’. Historical
Bayley, J 1996 ‘Innovation in later medieval urban
Metallurgy 44, 85–103
metalworking’. Historical Metallurgy 30, 67–71

Armitage, K H, Pearce, J E and Vince, A G 1981 ‘A Bayley, J 1998 ‘The production of brass in antiquity with
late medieval ‘bronze’ mould from Copthall Avenue, particular reference to Roman Britain’, in Craddock, P T (ed),
London’. Antiquaries Journal 61, 362–364 2000 Years of Zinc and Brass. (Revised edition). London: British
Museum Press, 7–26

Atkinson, D and Taylor M V 1924 ‘Flint excavation

report’. Flintshire Historical Society 10, 1–29 Bayley, J 2008 Lincoln: evidence for metalworking on
Flaxengate and other sites in the city. Research Department
report 67/2008. London: English Heritage
Awty, B and Whittick, C 2002 ‘The Lordship of
Canterbury, iron-founding at Buxted, and the
continental antecedents of cannon-founding in the
Bayley, J and Barclay, K 1990 ‘The crucibles, heating
Weald Sussex Archaeological Collections 140, 71-81
trays, parting sherds and related material’, in Biddle,
M Object and Economy in Medieval Winchester. Oxford:
Clarendon Press, 175-197
Barraclough, K C 1984 Steel Before Bessemer. 2 vols.
London: Metals Society
Bayley, J and Budd, P 1998 ‘The clay moulds’, in Cool,
Battye, K, Doncaster, R, Mitchell, I and Newing, D 1991 H E M and Philo, C (eds) Roman Castleford. Excavations
‘Summerley Colliery coke ovens’. Industrial Archaeology 1974–85. Volume I. The small finds. Wakefield: West
Review 13, 152–161 Yorkshire Archaeology Service, 195–222

< < Contents 71

Bayley, J and Butcher, S 2004 Roman Brooches in Blakelock, E 2005 Qualitative Analysis of Crucibles from
Britain: A technological and typological study based Bloodmoor Hill, Carlton Coalville, Suffolk. Centre for
on the Richborough Collection. London: Society of Archaeology Report 22/2005. London: English Heritage
Antiquaries

Blakelock, E and McDonnell, G 2007 ‘A review of

Bayley, J, Crossley, D and Ponting, M 2008 Metals metallographic analyses of early medieval knives’.
and Metalworking. A Research Framework for Historical Metallurgy 41, 40–56
Archaeometallurgy. London: Historical Metallurgical
Society
Blanchard, I S W 1981 ‘Lead mining and smelting
in Medieval England and Wales’, in Crossley, D W
(ed) Medieval Industry. London: Council for British
Bayley, J and Eckstein, K 1998 Metalworking Debris
Archaeology, 72–84
from Pentrehyling Fort, Brompton, Shropshire. Ancient
Monuments Lab Rep 13/1998. London: English Heritage

Blaylock, S R 2001 ‘Excavation of an early post-

Bayley, J, Freestone, I, Jenner, A and Vince, A 1991 medieval bronze foundry at Cowick Street, Exeter,
‘Aspects of Saxon and Norman London: Metallurgy’, in 1999–2000’. Devon Archaeological Society 58, 1–92
Vince, A G (ed) Aspects of Saxon and Norman London.
2: Finds and environmental evidence. London: London
and Middlesex Archaeological Society, 389–405. Bowden, M (ed) 2000 Furness Iron: the Physical Remains
of the Iron Industry and Related Woodland Industries
of Furness and Southern Lakeland. Swindon: English
Bayley, J, Mackreth, D F and Wallis, H 2001 ‘Evidence
Heritage
for Romano-British brooch production at Old
Buckenham, Norfolk’. Britannia 32, 93–118

Bowsher, D, Dyson, D, Holder, N and Howell, I 2007

Bayley, J and Rehren, T 2007 ‘Towards a functional The London Guildhall. An archaeological history of a
and typological classification of crucibles’, in La Niece, neighbourhood from early medieval to modern times.
S and Craddock, P (eds) Metals and Mines. Studies in London: Museum of London
Archaeometallurgy. London: Archetype, 46–55

Bowyer, P and Keys, L 2013 ‘Saxon iron smelting in

Bayley, J and White, H 2013 ‘Evidence for workshop Bermondsey? Archaeological investigations at 150-156
practices at the Tudor mint in the Tower of London’, Abbey Street’. Surrey Archaeological Collections 97,
in Saunders, D, Spring, M and Meek, A (eds), 43–58
The Renaissance Workshop. London: Archetype
Publications, 138–143
Bray, P and Pollard, A M 2012 ‘A new interpretative
Belford, P 2012 ‘Hot blast iron smelting in the approach to the chemistry of copper-alloy objects:
early 19th century: a re-appraisal’. Historical Metallurgy source, recycling and technology’. Antiquity 86, 853–867
46, 32–44

Brooks, H 2000 ‘Excavations at 79 Hythe Hill, Colchester

Belford, P and Ross, R 2007 ‘English steelmaking 1994–5’. Essex Archaeology and History 31, 112–124
in the seventeenth century: the excavation of two
cementation furnaces at Coalbrookdale’. Historical
Brown, D 2007 Archaeological Archives. A guide to best
Metallurgy 41, 105–123
practice in creation, compilation, transfer and Curation.
Reading: Institute for Archaeologists

Bell, I C 1884 Principles of the Manufacture of Iron

and Steel. Brown, N and Medlycott M 2013 The Neolithic and
Bronze Age Enclosures at Springfield Lyons, Essex:
Excavations 1981–91. Gressenhall: East Anglian
Biek, L and Bayley, J 1979 ‘Glass and other vitreous
Archaeology
materials’. World Archaeology 11, 1–25

< < Contents 72

Burchard, A 1982 ‘A Saxo-Norman iron-smelting site Craddock, P 1990 ‘Copper smelting in Bronze Age
at ‘Co-op’ in High Street’, in C Mahany, Burchard, A Britain: problems and possibilities’, in Crew, P
and Simpson, G Excavations in Stamford Lincolnshire and Crew, S (eds) Early Mining in the British Isles,
1963–1969. London: Society for Medieval Archaeology, Maentwrog: Snowdonia National Park Study Centre,
105–144 69–71

Burnham, B C and Burnham, H 2004 Dolaucothi- Craddock, P 1994 ‘Recent progress in the study of early
Pumsaint: Survey and excavation at a Roman gold- mining and metallurgy in the British Isles’. Historical
mining complex (1987–1999). Oxford: Oxbow Books Metallurgy 28, 69–83

Buxton, K and Howard-Davis, C 2000 Bremetenacum: Craddock, P 1995 Early Metal Mining and Production.
Excavations at Roman Ribchester 1980, 1989–1990. Edinburgh: Edinburgh University Press
Lancaster: University of Lancaster Archaeological Unit

Craddock, P T and Lang, J 2004 ‘Crucible steel – bright

Caple, C 2010 ‘Ancestor artefacts – ancestor materials’. steel’. Historical Metallurgy 38, 35–46
Oxford Journal of Archaeology 29, 305–318
Craddock, P T and Wayman, M L 2000 ‘The
Challis, K 2002 ‘A medieval iron smelting site at Stanley development of European ferrous metallurgy’ in
Grange, Derbyshire’. Historical Metallurgy 36, 33–42 Wayman, M L (ed) The Ferrous Metallurgy of Early Clocks
and Watches: Studies in Post-medieval Steel. London:
BM Press, 13–27
Chapman, D 1997 ‘Great Orme smelting site,
Llandudno’. Archaeology in Wales 37, 56–57 Cranstone, D 1987 ‘Early coke ovens: a note’. Historical
Metallurgy 23, 120–122
Claughton, P 1992 ‘Medieval silver-lead smelting
in Devon’, in Willies, L and Cranstone, D (eds) Boles Cranstone, D 1994 ‘Early surface features of metal
and Smeltmills. Matlock Bath: Historical Metallurgy mining: towards a typology’, in Ford, T D and Willies,
Society, 12–15 L (eds) Mining before Powder. Matlock Bath: Historical
Metallurgy Society, 144–147

Cleere H 1972, ‘The classification of early iron-smelting

furnaces’. The Antiquaries Journal 52, 8–23. Cranstone, D 1997 Derwentcote Steel Furnace.
Lancaster: Lancaster University Department
of Archaeology
Collard, M, Darvill, T and Watts, M 2006 ‘Ironworking
in the Bronze Age? Evidence from a 10th-century BC
settlement at Hartshill Copse, Upper Bucklebury,
Crew, P 1986 ‘Bryn y Castell hillfort – a late prehistoric
West Berkshire’. Proccedings of the Prehistoric Society
iron working settlement in North-West Wales’, in Scott,
72, 367–422
B G and Cleere, H (eds) The Crafts of the Blacksmith.
Belfast: UISSP CPSA/Ulster Museum, 91–100
Cool, H E M, Cowgill, J, Jennings, S, Jones, C E E,
Summerfield, J, Swain, H and Tribe, A 1993 Guidelines
for the Preparation of Site Archives and Assessments of
Crew, P 1988 ‘Bryn Y Castell hillfort, Gwynedd, North
All Finds Other Than Fired Clay Vessels. Roman Finds
Wales: a preliminary analysis of the iron working
Group and Finds Research Group AD 700–1700
debris’, in Ellis Jones, J (ed) Aspects of Ancient Mining
and Metallurgy. Bangor: University College of North
Cotter, JP 1992 ‘‘The Mystery of the Hessian Wares’: Wales, 129–135
post-medieval triangular crucibles’, in D Gaimster and
M Redknap (eds) Everyday and Exotic Pottery from
Crew, P 1991 ‘The experimental production of
Europe c. 650–1900. Oxford: Oxbow, 256–272
prehistoric bar iron’. Historical Metallurgy 25, 21–36

Cowgill, J 2009 ‘The crucibles and moulds’, in Lucy, S, Crew, P 1995 ‘Currency bars in Great Britain. Typology
Tipper, J and Dickens, A The Anglo-Saxon Setlement and function’, in Mangin, M (ed) La Sidérurgie Ancienne
and Cemetert at Bloodmoor Hill, Carlton Colville, Suffolk. de l’Est de la France dans son Contexte Européen. Paris:
Cambridge: Cambridge Archaeology Unit, 259–265 Belles Lettres, 345–351

< < Contents 73

Crew, P 1998a ‘Laxton revisited: A first report on the Daniels R 1988 ‘The Anglo-Saxon monastery at Church
1998 excavation’. Historical Metallurgy 32, 49-53 Close, Hartlepool’. Cleveland Archaeological Journal
145, 158–210

Crew, P 1998b ‘Excavations at Crawcwellt West,

Merioneth 1990–1998: a late prehistoric upland iron- Day, J 1973 Bristol Brass: the history of the industry.
working settlement’. Archaeology in Wales 38, 20–35 Newton Abbot: David & Charles

Crew, P 2013 ‘Twenty-five years of bloomery Day, J and Tylecote, R F 1991 The Industrial Revolution
experiments: perspectives and prospects’, in in Metals. London: Institute of Metals
Dungworth, D and Doonan, R C P (eds) 2013 Accidental
and Experimental Archaeometallurgy. London:
Historical Metallurgy Society, 25–50 Dearne, M J and Branigan, K 1995 ‘The use of coal in
Roman Britain’. Antiquaries Journal 75, 71–105
Crew, P and Crew, S 1990 Early Mining in the British
Isles. Plas Tan y Bwlch: Snowdonia National Park Den Ouden, A 1981 ‘The production of wrought iron
Study Centre in finery hearths, part 1: the finery process and its
development’. Historical Metallurgy 15, 63–87

Crew, P and Charlton, M 2007 ‘The anatomy of a

furnace ... and some of its ramifications’, in La Niece, Dillmann, P, Perez, A, Vega, E, Arribet-Deroin, D, Aranda,
S, Hook, D and P Craddock (eds) Metals and Mines: R, L’Heritier, M, Neff, D and Bellot-Gurlet, L 2012
Studies in Archaeometallurgy. London: Archetype, ‘Understanding the Walloon method of iron refining:
219–225 archaeological and archaeometric experiments, phase
1.’ Historical Metallurgy 46, 19–31
Crew, P and Mighall, T M 2013 ‘The fuel supply and
woodland management at a 14th century bloomery Dungworth, D B 1997 ‘Roman copper alloys: analysis
in Snowdonia: a multi-disciplinary approach’, in of artefacts from northern Britain’. Journal of
J Humphris and T Rehren (eds) The World of Iron. Archaeological Science 24, 901–910
London: Archetype, 473–482

Crossley, D 1975a The Bewl Valley Ironworks. London: Dungworth, D B 2010 ‘The possible water-powered
The Royal Archaeological Institute bloomery at Goscote (Rushall), Wallsall, west Midlands’.
Historical Metallurgy 44, 15–20

Crossley, D 1975b ‘Cannon-manufacture at Pippingford,

Sussex: the excavation of two iron furnaces of c. 1717‘. Dungworth, D B 2011 ‘The metalworking debris’, in
Post-Medieval Archaeology 9, 1–37 Nowakowski, J A and Quninnell, H Trevelgue Head,
Cornwall: the importance of C K Croft Andrew’s 1939
excavations for prehistoric and Roman Cornwall. Truro:
Crossley, D 1990 Post-Medieval Archaeology in Britain. Cornwall County Council, 220–244
Leicester: Leicester University Press
Dungworth, D B 2013 ‘An experimental study of some
Crossley D and Ashurst D 1968, ‘Excavations at Rockley early copper smithing techniques’, in D Dungworth
Smithies, a water-powered bloomery of the 16th and and R C P Doonan (eds) Accidental and Experimental
17th centuries’, Post-Medieval Archaeology 2, 10–54 Archaeometallurgy. London: Historical Metallurgy
Society, 149–152.

Cunliffe, B 1984 Danebury. An Iron Age Hillfort in Dungworth, D, Comeau, B and Lowerre, A 2013 St
Hampshire. Volume 2. The Excavations 1969–1978: the Algar’s Farm, Selwood, Somerset. Geochemical survey.
Finds. London: CBA Research Report 28/2013. London: English Heritage

Dalwood, H and Edwards, R 2004 The Excavations Dungworth, D and Maclean, P 2011 ‘Analysis of bell-
at Deansway, Worcester, 1988–89. The Excavations casting and other metalworking debris’, in Rodwell,
at Deansway, Worcester, 1988–89. Romano-British W and Atkins, C (eds) St Peter’s, Barton-upon-Humber,
small town to late medieval city. London: Council for Lincolnshire - A Parish Church and its Community.
British Archaeology Volume 1: History, Archaeology and Architecture. Oxford:
Oxbow, 835–838

< < Contents 74

Dungworth, D and Mepham, L 2012 ‘Prehistoric iron English Heritage 2011b Environmental Archaeology.
smelting in London: evidence from Shooters Hill’. A guide to the theory and practice of methods, from
Historical Metallurgy 46, 1–8 sampling and recovery to post-excavation. Second
edition. Swindon: English Heritage

Dungworth, D B and White H E 2007 ‘Scientific

examination of zinc-distillation remains from Warmley, English Heritage 2013 Guidelines for the Storage and
Bristol’. Historical Metallurgy 41, 77–83 Display of Archaeological Metalwork. Swindon: English
Heritage

Dungworth, D and Wilkes, R 2009 ‘Understanding

hammerscale: the use of high-speed film and electron Fang, J-L and McDonnell, G 2011 ‘The colour of copper
microscopy’. Historical Metallurgy 43, 33–46 alloys’. Historical Metallurgy 45, 52–61

Dunster, J and Dungworth, D 2013 St Algar’s Farm, Farthing, D J 2005 ‘Differentiation of British blast
Frome, Somerset. The analysis of lead-working waste. and reverberatory furnace slag using morphology,
Research Report 15/2012. London: English Heritage mineralogy, and chemistry’, in Kars, H and Burke, E
(eds) Proceedings of the 33rd International Symposium
on Archaeometry, 22–26 April 2002, Amsterdam.
Dutton, L and Fasham, P 1994 ‘Prehistoric copper Amsterdam: Vrije Universiteit, 335–338
mining on the Great Orme, Llandudno, Gwynnedd’.
Proceedings of the Prehistoric Society 60, 245–286
Fell, V 1993 ‘Examination of four Iron Age ferrous
hammer heads from Bredon Hill (Hereford and
Worcester), England’. Historical Metallurgy 27, 60–70
Egan, G 2005 Material Culture in London in an Age of
Transition: Tudor and Stuart period finds c1450–c1700
from excavations at riverside sites in Southwark.
Fell, V and Salter, C 1998 ‘Metallographic examination
London: Museum of London
of seven Iron Age ferrous axeheads from England’.
Historical Metallurgy 32, 1–6

English Heritage 1998 Dendrochonology. Guidelines on

producing and interpreting dendrochronological dates. Flemings, M C and Ragone, D V 2009 ‘Puddling: a
London: English Heritage metallurgical perspective’. Historical Metallurgy 43,
103–108

English Heritage 2006a Guidelines on the X-radiography

of Archaeological Metalwork. Swindon: English Heritage Foster, J 1995 ‘Metalworking in the British Iron Age: the
evidence from Weelsby Avenue, Grimsby’, in Raftery, B
with Megaw, V and Rigby, V (eds) Sites and Sights of the
English Heritage 2006b Archaeomagnetic
Iron Age. Oxford: Oxbow, 49–60
Dating. Guidelines on producing and interpreting
archaeomagnetic dates. Swindon: English Heritage

Frere, S S 1983 Verulamium Excavations. Volume 2.

English Heritage 2008a Geophysical Survey in London: Society of Antiquaries
Archaeological Field Evaluation. Second edition.
Swindon: English Heritage
Fulford, M G and Allen, J R L 1992 ‘Iron-making at the
Chesters villa, Wollaston, Gloucestershire: survey and
excavation 1987–91’. Britannia 23, 159–215
English Heritage 2008b Investigative Conservation.
Guidelines on how the detailed examination of artefacts
from archaeological sites can shed light on their
Fyfe, R, Bray, L, Juleff, G, Woodbridge, J and Marshall,
manufacture and use. Swindon: English Heritage
P 2013 ‘The environmental impact of Romano-British
ironworking on Exmoor’, in Humphris, J and Rehren, T
(eds) The World of Iron. London: Archetype, 462–472
English Heritage 2011a Archaeological Evidence for
Glassworking. Guidelines for best practice. Swindon:
English Heritage

< < Contents 75

Gaskell Brown, C and Harper, A E T 1984 ‘Excavations Hammer, F 2003 Industry in North-West Roman
on Cathedral Hill, Armagh, 1968’. Ulster Journal of Southwark. London: Museum of London
Archaeology 47, 109–161

Hanworth, R and Tomlin, D J 1977 Brooklands,

Gerrard, S 2000 The Early British Tin Industry. Weybridge: the Excavation of an Iron Age and
Stroud: Tempus Medieval Site, 1964–5 and 1970–71. Guildford: Surrey
Archaeological Society

Giles, M 2007 ‘Making metal and forging relations:

ironworking in the British Iron Age’. Oxford Journal of Haslam, J 1980 ‘A middle Saxon iron smelting site at
Archaeology 26, 395–413 Ramsbury, Wiltshire’. Medieval Archaeology 24, 1–68

Gilmour, B and Salter, C 1998 ‘Ironwork: technological Hawthorn, J G and Smith, C S 1979 Theophilus. On
examination of the knives, spearheads, and sword/ Divers Arts. New York: Dover
weaving batten’, in Malim, T and Hines, J The
Anglo-Saxon Cemetery at Edix Hill (Barrington A),
Cambridgeshire. London: CBA, 250–56 Hayman, R 2004 ‘The Cranage brothers and eighteenth-
century forge technology’. Historical Metallurgy 38,
113–120
Girbal, B 2011 Roman and Medieval Litharge Cakes:
A scientific examination. Research Report 51/2011.
London: English Heritage
Hayman, R 2005 Ironmaking. The history an archaeology
of the iron industry. Stroud: Tempus
Girbal, B 2013 ‘Experimenting with bowl furnaces’,
in Dungworth, D and Doonan, R (eds) Accidental and Hinton, D A 2000 A Smith in Lindsey: the Anglo-Saxon
Experimental Archaeometallurgy. London: Historical grave at Tattershall Thorpe, Lincolnshire. Leeds: Society
Metallurgy Society, 83–92 for Medieval Archaeology

Griffith, F and Weddell, P 1996 ‘Ironworking in the Historic England 2015 Guidelines on the Use of
Blackdown Hills: results of recent survey’, in Newman, Radiocarbon Dating and Chronological Modelling in
P (ed) Mining and Metallurgy in South-West Britain. Archaeology. London: Historic England
London: Historical Metallurgy, 27–34

Hodgkinson, J 2008 The Wealden Iron Industry. Stroud:

Greeves, T 1994 ‘Stamping and smelting at Upper Tempus
Merrivale tin mill: preliminary results 1991–1993’, in
Ford, T D and Willies, L (eds) Mining Before Powder.
Hull, G 2003 ‘The excavation and analysis of an 18th-
London: Historical Metallurgy Society, 76–80
century deposit of anatomical remains and chemical
apparatus from the rear of the first Ashmolean Museum
(now Museum of the History of Science), Broad Street,
Halkon, P 2013 ‘Iron, myth and magic: perceptions of Oxford’. Post-Medieval Archaeology 37, 1–28
smiths and metal workers through cultural evidence
from Britain and beyond’, in Humphris, J and Rehren, T
Hyde, C K 1977 Technological Change and the British
(eds) The World of Iron. London: Archetype, 188–200
Iron Industry 1700–1870. Princeton, NJ: Princeton
University Press

Halkon, P and Millett, M 1999 Rural Settlement

and Industry: Studies in the Iron Age and Roman Jackson, R 2012 Ariconium, Herefordshire: An Iron Age
Archaeology of Lowland East Yorkshire. Leeds: Yorkshire settlement and Romano-British ‘small town’. Oxford:
Archaeological Society Oxbow Books

Hamilton, J R C 1956 Excavations at Jarlshof, Shetland. Johnson, B and Bearpark, P 1976 ‘Work in progress at
Edinburgh: HMSO Lymm Slitting Mill’. Historical Metallurgy 10, 83

< < Contents 76

Jones, S 1999. ‘Great Orme, Bronze Age smelting site, Lee, R 2009 Production, Use and Disposal of Romano-
Llandudno’. Archaeology in Wales 39, 79. British Pewter Tableware. Oxford: Oxbow Books

Kelly, R S 1976 ‘Metalworking in N Wales during the Lewis, A 1990 ‘Underground exploration of the Great
Roman period’. Bulletin of the Board of Celtic Studies Orme copper mines’, in Crew, P and Crew, S (eds) Early
27, 127–147 Mining in the British Isles, Maentwrog: Snowdonia
National Park Study Centre, 5–10

Keys, L 2012 ‘Slag and high-temperature debris’, in

Biddulph, E, Stafford, E, Foreman, S, Stansbie, D and Lynn, C J and McDowell, J A 1988 ‘A note on the
Nicholson, R A (eds) London Gateway: Iron Age and excavation of an Early Christian period settlement
Roman Salt Making in the Thames Estuary, Excavation in Deer Park farms, Glenarm, Co Antrim, 1984–1987’.
at Stanford Wharf Nature Reserve, Essex. Oxford: Oxford Journal of the Glens of Antrim Historical Society 16, 2–16
Archaeology Monograph

Kiernan, D 1989 The Derbyshire Lead Industry Macfarlane, W 1917 The Principles and Practice of
in the Sixteenth Century. Chesterfield: Derbyshire Iron and Steel Manufacture. Fifth edition. London:
Record Society Longmans, Green and Co

Killick, D and Gordon, R B 1987 ‘Microstructure of Mackenzie, R J and Whiteman, J A 2006 ‘Why pay more?
puddling slags from Fontley, England and Roxbury, An archaeometallurgical investigation of 19th-century
Connecticut, USA’. Historical Metallurgy 21, 28–36 Swedish wrought iron and Sheffield blister steel’.
Historical Metallurgy 40, 13–149

King, P 1999 ‘The development of the iron industry in

South Staffordshire in the early seventeenth century: Malham, A, Aylett, J, Higgs, E and McDonnell, J G 2002
history and myth’. Transactions of the Staffordshire ‘Tin smelting slags from Crift Farm, Cornwall, and the
Archaeological and Historical Society 38, 59–76 effect of changing technology on slag composition’.
Historical Metallurgy 36, 84–94

King, P 2001–2 ‘Sir Clement Clerke and the adoption

of coal in metallurgy’. Transactions of the Newcomen Manning, W H 1991 ‘Blacksmiths’ tools from Waltham
Society 73, 33–52 Abbey, Essex’, in Oddy, W A (ed) Aspects of Early
Metallurgy. London: British Museum Press, 87–96

King, P 2011a ‘The choice of fuel in the 18th

century iron industry: the Coalbrookdale accounts Margeson, S 1993 Norwich Households: The Medieval
reconsidered’. Economic History Review 64, 132–156 and Post-medieval Finds from Norwich Survey
Excavations 1971–1978. Norwich: Norwich Survey

King, P 2011b ‘Iron in 1790: production statistics

1787–96 and the arrival of puddling’. Historical Martinón-Torres, M 2012 ‘Inside Solomon’s House: an
Metallurgy 45, 102–133 archaeological study of the Old Ashmolean Chymical
Laboratory in Oxford’. Ambix 59, 22–48

Lane, A and Campbell, E 2000 Dunadd: an early

Dalriadic capital. Oxford: Oxbow. Martinón-Torres, M, Freestone, IC, Hunt, A and Rehren,
Th 2008 ‘Mass-produced mullite crucibles in medieval
Europe: manufacture and material properties’. Journal
Lane, H C 1973 Field Surveys and Excavation
of the American Ceramic Society 91, 2071–2074
of a Romano-British Rural Native Settlement at
Scarcliffe Park, East Derbyshire. Matlock: Derwent
Archaeological Society
May, J 1996 Dragonby. Report on Excavations at an Iron
Age and Romano-British Settlement in Lincolnshire.
Oxford: Oxbow
Lawson-Jones, A 2013 ‘Discoveries along the North
Land’s End pipeline’. Cornish Archaeology 52, 197–222

< < Contents 77

McDonnell, J G 1989 ‘Iron and its alloys in the fifth to Money, J H 1971 ‘Medieval iron-workings in Minepit
eleventh centuries AD in England’. World Archaeology, Wood, Rotherfield, Sussex’. Medieval Archaeology 15,
20, 373–382 86–111

McDonnell, G 1993 ‘Slags and ironworking residues’, Morton, G R and Gould, J 1967 ‘Little Aston Forge:
in Hamerow, H Excavations at Mucking. Volume 2: the 1574–1798’. Journal of the Iron and Steel Institute 205,
Anglo-Saxon settlement. London: English Heritage, 82–85 237–244

McDonnell, G 2013a ‘Temperature profiles and cast iron Morton, G R and Wingrove, J 1970 ‘The charcoal finery
production in experimental iron-smelting furnaces’, and chafery forge’. Bulletin of the Historical Metallurgy
in Dungworth, D and Doonan, R (eds) Accidental and Group 5, 24–28
Experimental Archaeometallurgy. London: Historical
Metallurgy Society, 61–68
Mott, R A 1961 ‘English bloomeries (1329–1586)’.
McDonnell, G 2013b ‘The metallurgical and vitrified Journal of the Iron and Steel Institute 198, 149–161
material’, in Armit, I and Mckenzie, J (Eds) An Inherited
Place. Broxmouth Hillfort and the South-East Scottish
Mott, R A and Singer, P 1983 Henry Cort: The Great Finer,
Iron Age. Edinburgh: Society of Antiquities of Scotland,
creator of Puddled Iron. London: The Metals Society
382–393

McKerrell, H and Tylecote, R F 1972 ‘The working of Murphy, S and Baldwin, H 2001 ‘Early lead smelting
copper-arsenic alloys in the Early Bronze Age and sites in the Swaledale area of Yorkshire’. Historical
the effect on the determination of provenance’. Metallurgy 35, 1–21
Proceedings of the Prehistoric Society 38, 209–218

Museums and Galleries Commission 1992 Standards in

Meharg, A A, Edwards, K J, Schofield, J E, Raab, the Museum Care of Archaeological Collections.
A, Feldmann, J, Moran, A, Bryant, C L, Thornton,
B and Dawson, J J C 2012 ‘First comprehensive
Needham, S 1980 ‘An assemblage of Late Bronze Age
peat depositional records for tin, lead and copper
metalworking debris from Dainton, Devon’. Proceedings
associated with the antiquity of Europe’s largest
of the Prehistoric Society 46, 177–215
cassiterite deposits’. Journal of Archaeological Science
39, 717–727
Northover, P 1987 ‘Non-ferrous metallurgy’, in Cunliffe,
Mighall, T M and Chambers, F M 1997 ‘Early ironworking B Hengistbury Head, Dorset. Volume 1: The prehistoric
and its impact on the environment: palaeoecological and Roman settlement, 3500 BC–AD 500. Oxford: OUCA,
evidence from Bryn y Castell hillfort, Snowdonia, 186–196
North Wales’. Proceedings of the Prehistoric Society 63,
199–219.
O’Brien, W 2004 Ross Island. Mining, Metal and Society
Mighall, T M, Timberlake, S, Foster, I D L, Krupp, E, in Early Ireland. Galway: National University of Ireland
and Singh, S 2009 ‘Ancient copper and lead pollution
records for a raised bog complex in Central Wales, UK’.
Ottaway, J, 1992 Anglo-Scandinavian Non-ferrous
Journal of Archaeological Science 30, 1–12
Metalworking from 16–22 Coppergate. The Archaeology
of York 17/6. London: CBA

Mills, A and McDonnell, J G 1992 The Identification

and Analysis of the Hammer Scale from Burton Dassett, Page, N 2005 ‘Industrial complex and timber trackway
Warwickshire. Ancient Monuments Laboratory Report at Llangynfelin’. Archaeology in Wales 45, 103–104
47/92. London: English Heritage

Paynter, S 2006 ‘Regional variations in bloomery

Minchinton, W E 1957 The British Tinplate Industry: a smelting slag of the Iron Age and Romano-British
history. Oxford: Clarendon Press periods’. Archaeometry 48, 271–292

< < Contents 78

Paynter, S 2007a ‘Innovations in bloomery smelting Pine, J, Allen, J R L and Challinor, D 2009 ‘Saxon
in Iron Age and Romano-British England’, in LaNiece, iron smelting at Clearwell Quarry, St. Briavels,
S, Hook, D and Craddock, P (eds) Metals and Mines: Lydney, Gloucestershire’. Archaeology in the Severn
Studies in Archaeometallurgy. London: Archetype, estuary 20, 9–40
202–210

Paynter, S 2007b ‘Romano-British workshops for Pollard, A M, Batt, C, Stern, B and Young, S M M 2007
iron smelting and smithing at Westhawk Farm, Kent’. Analytical Chemistry in Archaeology. Cambridge:
Historical Metallurgy 41, 15–31 Cambridge University Press

Paynter, S, Claughton, P and Dunkley, T 2010 ‘Further Proctor, J 2011 ‘Archaeological investigations at
work on residues from lead/silver smelting at Combe Swalwell ironworks, Tyne and Wear’. Industrial
Martin, North Devon’. Historical Metallurgy 44, 104–111 Archaeology Review 33, 18–39

Penhallurick, R D 1997 ‘The evidence for prehistoric Reed, S J, Juleff, G and Bayer, O J 2006 ‘Three late
mining in Cornwall’, in Budd, P and Gale, D (eds) Saxon iron-smelting furnaces at Burlescombe,
Prehistoric Extractive Metallurgy in Cornwall. Truro: Devon’. Proceedings of the Devon Archaeological Society
Cornwall Archaeology Unit, 23–33 64, 71–122

Percy, J 1861–4 Metallurgy: the art of extracting metals Rehren, T 1996 ‘Alchemy and fire assay — an analytical
from their ores, and adapting them to various purposes approach’. Historical Metallurgy 30, 136–142
of manufacture. Two volumes. London: John Murray

Richards, J D 1993 The Bedern Foundry. The

Petch, J A and Taylor, M V 1924 ‘Report on the Archaeology of York 10/3. London: CBA
excavation carried out at Pentre, Flint’. Flintshire
Historical Society Journal 10, 5–29
Salter, C and Ehrenreich, R 1984 ‘Iron Age iron
metallurgy in central southern Britain’, in Cunliffe, B
and Miles, D (eds) Aspects of the Iron Age in Central
Phelps, M, Paynter, S and Dungworth, D 2011 Downside
Southern Britain. Oxford: OUCA, 146–161
Mill, Cobham, Surrey. Analysis of the metalworking remains.
Research Report 43-2011. Portsmouth: English Heritage

Sauder, L and Williams, S 2002 ‘A practical treatsie on

Photos-Jones, E, Atkinson, J A, Hall, A J and Banks, I the smelting and smithing of bloomery iron’. Historical
1998 ‘The bloomery mounds of the Scottish Highlands. Metallurgy 36, 122–131
Part 1: The archaeological background’. Historical
Metallurgy 32, 15–32
Schrüfer-Kolb, I E 2004 Roman Iron Production
in Britain. Technological and Socio-economic
Landscape Development along the Jurassic Ridge.
Photos-Jones, E, Dalglish, C, Coulter, S, Hall, A, Ruiz-
Oxford: Archaeopress
Nieto, R and Wilson, L 2008 ‘Between archives and
the site: the case of the 19th-century iron and steel
industry in the Monklands, Central Scotland, and one
small forge within, Moffat Upper Steam Forge’. Post- Scott, D A 1991 Metallography and Microstructure
Medieval Archaeology 42, 157–180 of Ancient and Historic Metals. Los Angeles: Getty
Conservation Institute

Pickin, J 1990 ‘Stone tools and early metal mining in

England and Wales’ in Crew, P and Crew, S (eds) Early Sharples, N 1999 Scalloway: a Broch, Late Iron Age
Mining in the British Isles, Maentwrog: Snowdonia Settlement and Medieval Cemetery in Shetland.
National Park Study Centre, 39–42 Oxford: Oxbow

Pickin, J 2010 ‘Early lead smelting in southern Shugar, A N and Mass, J L (eds) Handheld XRF for Art
Scotland’. Historical Metallurgy 44, 81–84 and Archaeology. Leuven: Leuven University Press

< < Contents 79

Smith, R 1996 ‘An analysis of the processes for tin Timberlake, S and Prag, A J N W 2005 The Archaeology
smelting’, in Newman, P (ed) Mining and Metallurgy of Alderley Edge. Oxford: Archaeopress
in South-West Britain. London: Historical Metallurgy
Society, 91–99
Tylecote R F 1960, ‘The location of Byrkeknott: a
fifteenth century iron smelting site’, Journal of the Iron
and Steel Institute 194, 451–458
Smith, R 2006 ‘A typology of lead-bale slags based on
their physio-chemical properties’. Historical Metallurgy
40, 115–128
Tylecote, R F 1967 ‘A bloomery site at West Runton,
Norfolk’. Norfolk Archaeology 34, 187–214
Smith, R and Murphy, S 2010 ‘Lead bale slags in
Nidderdale’. Historical Metallurgy 44, 73–80 Tylecote, R F 1986 The Prehistory of Metallurgy in the
British Isles. London: Institute of Metals
Smith, T 2013 ‘A report on the Wealden Iron Research
group smelt’, in Dungworth, D and Doonan, R (eds) Tylecote, R F 1991 ‘Iron in the Industrial revolution’, in
Accidental and Experimental Archaeometallurgy. Day, J and Tylecote, R F (eds) The Industrial Revolution
London: Historical Metallurgy Society, 99–106 in Metals. London: The Institute of Metals, 200–260

Society of Museum Archaeologists 1993 Selection, Tylecote, R F 1992 A History of Metallurgy. Second
Retention and Dispersal of Archaeological Collections. edition. London: The Institute of Materials

Soulignac, R and Serneels, V 2013 Tylecote, R F and Gilmour, B J J 1986 The Metallography
of Early Ferrous Edge Tools and Edged Weapons. Oxford:
BAR
Tebbutt, C F 1982 ‘A Middle-Saxon iron smelting
site at Millbrook, Ashdown Forest, Sussex’. Sussex
Archaeological Collection 120, 19–35
Vernon, R W, McDonnell, J G and Schmidt, A 1999
‘Medieval iron and lead smelting works: a geophysical
comparison’, in Pollard, A M (ed) Geoarchaeology:
Timberlake, S 1994 ‘An experimental tin smelt at Flag
Exploration, Environments, Resources. London:
Fen’. Historical Metallurgy 28, 121–128
Geological Society, 15–34

Timberlake, S 2002 ‘Medieval lead smelting boles near Wainwright, G J 1979 Gussage All Saints. An Iron Age
Penguelan, Cwmystwyth, Ceredigion’. Archaeology in settlement in Dorset. London: HMSO
Wales 42, 45–59

White, J R 1980 ‘Historic blast-furnace slags:

Timberlake, S 2003 Excavations on Copa Hill, archaeological and metallurgical analysis’. Historical
Cwmystwyth (1986–1999). An Early Bronze Age copper Metallurgy 14, 55–64
mine within the uplands of Central Wales. Oxford:
Archaeopress
Williams, A R 2003 The Knight and the Blast Furnace.
Leiden: Brill
Timberlake, S 2007 ‘The use of experimental
archaeology/ archaeometallurgy for the understanding Williams, RA 2013 ‘Linking Bronze Age copper smelting
and reconstruction of Early Bronze Age mining slags from Pentrewyn on the Great Orme to ore and
and smelting technology’, in La Niece, S, Hook, D metal’. Historical Metallurgy 47, 93–110
and Craddock, P (eds) Metals and Mines: Studies in
Archaeometallurgy. London: Archetype, 27–36.
Wilson, L and Pollard, A M 2001 ‘The provenance
Timberlake, S and Marshall, P 2013 ‘The beginnings hypothesis’, in Brothwell, D R and Pollard A M (eds)
of metal production in Britain: a new light on the Handbook of Archaeological Sciences. Chichester:
exploitation of ores and the dates of Bronze Age mines’. Wiley, 507–517
Historical Metallurgy 47, 75–92

< < Contents 80

Wilthew, P 1987 ‘Appendix. Metallographic examination
of medieval knives and shears’, in Cowgill, J de
Neergaard, M and Griffiths, N Knives and Scabbards.
London: HMSO, 61–74

Young, T 2011 ‘Some preliminary observations

on hammerscale: implications for its genesis,
interpretational potential and the understanding of
welding’. Historical Metallurgy 45, 26–41

Young, T and Poyner, D 2012 ‘Two medieval bloomery

sites in Shropshire: the adoption of water power for
iron smelting’. Historical Metallurgy 46, 78–97

Youngs, S M (ed) 1989 ‘The Work of Angels’ –

Masterpieces of Celtic Metalwork, 6th–9th Centuries AD.
London: BM Press

Zienkiewicz, J D 1993 ‘Excavations in the Scamnum

Tribunorum at Caerleon: The Legionary Museum site
1983–5’. Britannia 24, 27–140

< < Contents 81

 This page is left blank intentionally

< < Contents 82

Contact Historic England

East Midlands South East

2nd Floor, Windsor House Eastgate Court
Cliftonville 195-205 High Street
Northampton NN1 5BE Guildford GU1 3EH
Tel: 01604 735400 Tel: 01483-252000
Email: eastmidlands@HistoricEngland.org.uk Email: southeast@HistoricEngland.org.uk

East of England South West

Brooklands 29 Queen Square
24 Brooklands Avenue Bristol BS1 4ND
Cambridge CB2 2BU Tel: 0117 975 0700
Tel: 01223 582700 Email: southwest@HistoricEngland.org.uk
Email: eastofengland@HistoricEngland.org.uk

Fort Cumberland Swindon

Fort Cumberland Road The Engine House
Eastney Fire Fly Avenue
Portsmouth PO4 9LD Swindon SN2 2EH
Tel: 023 9285 6704 Tel: 01793 414700
Email: fort.cumberland@HistoricEngland.org.uk Email: swindon@HistoricEngland.org.uk

London West Midlands

1 Waterhouse Square The Axis
138-142 Holborn 10 Holliday Street
London EC1N 2ST Birmingham B1 1TG
Tel: 020 7973 3000 Tel: 0121 625 6820
Email: london@HistoricEngland.org.uk Email: westmidlands@HistoricEngland.org.uk

North East Yorkshire

Bessie Surtees House 37 Tanner Row
41–44 Sandhill York YO1 6WP
Newcastle Upon Tyne Tel: 01904 601901
NE1 3JF Email: yorkshire@HistoricEngland.org.uk
Tel: 0191 269 1200
Email: northeast@HistoricEngland.org.uk

North West
Suites 3.3 and 3.4
Canada House
3 Chepstow Street
Manchester M1 5FW
Tel: 0161 242 1400
Email: northwest@HistoricEngland.org.uk

< < Contents 83

We are the public body that looks after
England’s historic environment. We champion
historic places, helping people understand,
value and care for them.

Please contact:
guidance@HistoricEngland.org.uk
with any questions about this document.

HistoricEngland.org.uk

If you would like this document in a different

format, please contact our customer services
department on:

Tel: 0370 333 0607

Fax: 01793 414926
Textphone: 0800 015 0174
Email: customers@HistoricEngland.org.uk

Please consider the environment before printing

this document

HEAG003
Publication date: v1.0 September 2001 © English Heritage
Reissue date: v1.2 April 2015 © Historic England
Design: Historic England