You are on page 1of 45

Britannia XXXVII (2006), 213-57

The Deposition of Iron Objects in

Britain During the Later Prehistoric
and Roman Periods: Contextual
Analysis and the Significance of Iron

It is perhaps in the Roman period that the prejudice of archaeologists against a ritual
interpretation is felt most strongly. Its material remains seem so practical, so ‘modern’ in
many respects, that there is an almost instinctive reaction against anything ‘other-wordly’
in this context — except of course in its proper place, the temple precinct, where it is
acceptable. Yet this attitude is historically indefensible, for … religion in the Roman world
pervaded every human activity … (Merrifield 1987, 7)


entral to this paper is the meaning of the actions that led to iron objects being found
in archaeological contexts of later prehistoric and Roman date. It is argued that the
placing of iron objects within the physical landscape reflects the changing nature of
society at this time. In Iron Age studies, many deposits in rivers, bogs and in the pits, ditches
and post-holes of settlements are now interpreted as ‘special’ material buried for significant
reasons,2 through acts that are often called ‘structured deposition’. This approach has had a deep
influence on the excavation and post-excavation of Iron Age settlements and is now coming to
influence the study of the deposition of artefacts on Roman sites.3 This paper develops the idea
that much of the later prehistoric and Roman ironwork found on settlements and elsewhere
was deliberately deposited for what might loosely be called ‘ritual’ or ‘religious’ motives; for
much of this period the proportion of the artefacts lost accidentally was possibly quite small.4
Artefacts in other materials also require comparable study, but, while work that integrates the
examination of items made from different materials on individual sites is important,5 this paper
focuses upon iron due to its potential significance as a highly symbolic medium.

1 I am very grateful to the following for their advice and assistance: Philip and Nina Crummy, Professor Barry
Cunliffe, Dr Simon Esmonde Cleary, Professor Colin Haselgrove, Dr Sam Lucy, Professor Bill Manning, Christina
Unwin, and the two anonymous referees.
2 See Bowden and McOmish 1987; Hill 1989; 1995a; Hingley 1990a. For the early application of comparable
ideas to the Roman period, see Merrifield 1987.
3 Clarke 2000; Fulford 2001; Woodward and Woodward 2004.
4 See Haselgrove et al. 2001; Hill 1995a; Hingley 1990a for this argument in the context of Iron Age society.
For recent discussions of the meaning of ‘ritual’ that have helped to inform this paper, see Bradley 2005, 28–36 and
Insoll 2004, 10–12.
5 e.g. Hill 1995a; Clarke 2000.

© World copyright reserved. Exclusive Licence to Publish: The Society for the Promotion of Roman Studies 2006

Applying a long-term perspective to the evidence for the deposition of iron should enable a
consideration of the degree to which the practice of iron deposition changed as a result of the
Roman conquest and control of Britain. The idea that iron was often deposited for ritual reasons
is pursued through the analysis of 91 individual finds of Iron Age and Roman date, including
‘hoards’ and significant individual finds that derive from settlements, rivers, caves, and other
contexts. Three sites (Danebury, Baldock, and Wavendon Gate) are examined in greater detail in
order to explore the contexts of the deposition of individual finds within settlements.
A number of distinctive types of context in which iron objects were left are defined and, from
the evidence studied, these appear to have changed through time. An initial focus upon the
settlement boundary during later prehistory is replaced by a concentration upon deep pits and
wells during the Roman period. The contexts in which objects were deposited indicate that iron
was a significant material to many people in Britain from c. 400 B.C. to A.D. 400 and this, in turn,
justifies the increased attention that is being paid to this information in recent excavations and


A common practice in archaeology is to divide ‘hoards’, which are usually considered to

represent sealed collections of objects, from single finds. Archaeologists do not agree about
the motivation for the deposition of the collections of objects that they call ‘hoards’. Two
dominant positions — the pragmatic and the symbolic — are often defined as in opposition
to each other.6 The term hoard is defined by the Chambers English Dictionary as: ‘a store: a
hidden stock: a treasure: a place for hiding anything’,7 suggesting a deposit that was separated
off and deliberately left in a sealed context with some intention of future action. The idea of a
store, stock, or treasure suggests a collection of objects. An influential view for the hoarding of
precious metal in the Roman period has recently been restated — that these collections of objects
were deposited for primarily practical reasons, representing the deposition of valuables with the
intention of retrieval at a later date.8 Those who take such an approach to iron hoarding have
argued that people buried objects as a resource to be re-used at a later time. Often, such hoards
contain a variety of items that have been specifically collected together, while individual objects
are sometimes broken or worn. The damaged objects would have retained value, since they
represented raw material. Some iron hoards also include tools that were used in the production
of iron objects, while billets and ‘currency bars’ are also common. The presence of these objects
has been used to suggest that the hoards constituted collections of raw material and old or
unwanted artefacts intended for recycling.9 This suggests that they had a pragmatic function, as
‘smiths’ hoards’, or collections of material that had a role in the production of new objects.10
In his seminal article of 1972, Bill Manning argued, however, that many of the contexts in
which Iron Age and Roman iron hoards occur suggest that they were deposited for ritual or
religious reasons. The nature of these contexts — including bogs, rivers, wells, and human burial
— has been taken to indicate their ritual significance.11 Many metal objects were deposited

6 For a discussion of Bronze Age hoards that covers some of these issues, see Bradley 1998, xviii–xix. For
Roman hoards, see Johns 1995 and Millett 1995.
7 Chambers English Dictionary 1990, 675.
8 Johns 1995; 1996, 6–11.
9 For the idea of smith’s hoards, and the problems with such an interpretation, see Manning 1972, 239.
10 Manning observed that individual hoards did not contain all the tools required by the smith but this would
not prevent individual hoards representing material collected by smiths for reworking. See also, Bradley’s (1998,
xviii–xxi) reassessment of his earlier arguments about the significance of Bronze Age smith’s hoards.
11 Manning 1972; Haselgrove and Hingley in press; Hingley 1990a.

in bogs, lakes, and rivers during later prehistory,12 and the ironwork collections discussed
by Manning appear to represent one part of a broader tradition of hoarding various types of
objects as offerings to the supernatural or to ancestors.13 The fact that iron (and other metals)
can be recycled should lead us to question why people placed or discarded iron items in a broad
variety of domestic and landscape contexts in such a way that they survive to be found. To view
collections of objects as hoards, buried with the intention of retrieval, would require that the
individuals who deposited them died or were killed, were driven away from the area, forgot
where the objects were buried, or, for some reason, lost interest in or were prevented from
retrieving the material. These explanations may apply to some deposits of iron objects, but the
fact that iron tends to deteriorate if buried in the ground makes the large-scale hoarding of these
objects with the intention of retrieval difficult to explain.14
This raises several problems with the idea of the ‘hoard’; in this study different types of
information are assessed in an attempt to get beyond the limitations of the concept.15 Hoards
need not have been closed deposits, left at one particular time in sealed situations.16 Many
significant collections of material may have built up gradually in situations that were not sealed.
Examples considered below include some of the finds from rivers and bogs, the so-called
‘massacre deposits’ found in late Iron Age contexts at several hillforts, collections of ‘votive’
objects at temples or shrines, and the Roman deposits in deep pits and wells. Significant iron
items occur in all these types of collections, although the motivation for the deposition of these
objects may have been very different from that for the deposition of collections of iron items.
In addition, single items of iron could represent significant deposits of material. Manning did
not examine the occurrence of single iron items, emphasising that many such objects represented
rubbish disposal.17 Discoveries of individual iron objects on archaeological sites have often
been addressed in a pragmatic manner, considered to indicate the locations in which particular
activities took place. The distribution of various types of discarded iron objects, together with
the products of smelting and smithing, are often used to define the function of individual areas
and buildings. At the Roman site of Gorhambury (Herts.), for example, about 600 iron objects
were used in an attempt to inform the function of individual buildings and activity areas across
the site.18 In the recent publication of the information from a Roman site at Great Holts Farm
(Essex), the author expresses disappointment that the analysis of the distribution of the iron
objects did not aid in the identification of particular activity areas.19
Reacting against this type of ‘pragmatic’ perspective, several authors have proposed that

12 Bradley 1998; Brunaux 1988; Field and Parker Pearson 2003, 179–88; Fitzpatrick 1984; Haselgrove and Wigg-
Wolf 2005; Hunter 1997; Wait 1985, 15–50.
13 Fulford 2001; Millett 1995; Manning 1972, 239–49.
14 Manning 1972, 240. Deposition of iron in waterlogged deposits, by contrast, preserves the objects very well,
as finds from the Walbrook (London), Vindolanda (Northumbria), and elsewhere demonstrate (e.g. Merrifield 1997).
15 In Haselgrove and Hingley (in press) the term ‘hoard’ is avoided entirely. In this study, the term is used, but
the focus of analysis is on the complexity of the character of iron deposition, which the simple term hoard does not
accurately convey.
16 Many ‘hoards’ may not have formed sealed deposits and they will often have had objects added and removed
from time to time (Johns 1996, 7; Isserlin in press).
17 Manning 1972, 239.
18 Neal et al. 1990, 138–55. See, in particular, the discussion of the concentration of hipposandals and other
objects within the aisled hall (140–4). The Gorhambury report contains a very detailed publication of the iron and
should not be criticised for failing to adopt approaches to deposition that have only become common in recent
19 Major 2003, 77 notes that it was hoped that the distribution of iron items in Phase II.2 and II.3 contexts would
help to define activities associated with each of the individual buildings on the site, but that this was not possible,
apart, perhaps, from the case of Building 417.

certain individual objects in later prehistoric contexts were probably hoarded.20 Merrifield made
the observation (above) about the instinctive opposition of Roman archaeologists to the idea that
artefacts in settlement contexts might have had ritual associations. Roman artefacts may often
appear rather contemporary, while ‘industry’ has been reconstructed with a particular emphasis
upon industrial production, the distribution of the objects, and consumption — ideas that are
informed by the reconstruction of the Roman economy as distinct from other aspects of society
and belief.21 We need, therefore, to address the context of both single items and collections,
rendering hundreds of thousands of iron finds potentially of relevance.


To explore Merrifield’s proposal further, we can look to the writings of Classical authors, which
illustrate that the ancient world was imbued with beliefs about the nature of the world that
permeated everyday life, including aspects of production, consumption, and deposition that are
often reconstructed by archaeologists as elements of the economy.22 Some Classical writers
believed that both the sea and the land had a sacred character; Pliny the Elder remarked that
mining for wealth-generation represented a dangerous and morally dubious activity, one that
might disturb the ancestors.23 The products of mining and industrial production, however, could
also be very useful in various ritual activities.24 Addressing people who have curative powers in
the Natural History, Pliny records that:
It is said that difficult labour ends in delivery at once, if over the house where is the lying-
in woman there be thrown a stone or missile that has killed with one stroke each three
living creatures — a human being, a boar and a bear. A successful result is more likely
if a light-cavalry spear is used, pulled out from a human body without the ground being
touched. The result indeed is the same if the spear is carried indoors. So, too, as Orpheus
and Archelaus write, arrows drawn out of a body and not allowed to touch the ground act as
a love-charm upon those under whom when in bed they have been placed. Moreover, add
these authorities, epilepsy is cured by food taken from the flesh of a wild beast killed by the
same iron weapon that has killed a human being.25

In these contexts, weapons that had been used to kill took part in rituals connected with the
human regenerative cycle and the health of individuals. Pliny’s comment about the role of iron
weapons in curing epilepsy recalls the observations of another Classical author, Dioscorides,
who, in a lengthy discourse upon herbal remedies, recounted the medical value of different types
of metal slag and various metallic stones.26 Other significant activities involving iron included
the clavus annalis, ‘annual nailing’, ritual of the Republican city of Rome.27 While little direct
evidence survives to tell us about the specific significance of ironworking and iron objects for
the people of prehistoric and Roman Britain, the evidence for a smith god in the Roman period
indicates the potential association of ironworking with divinities, spirits, and superstition.28

20 Bradley 1987, 351; Fitzpatrick 1984, 178–9; Haselgrove 1987, 132–8; Haselgrove and Hingley in press;
Hingley 1990a, 108–9.
21 Greene 2002, 247–53.
22 Clarke and Jones 1996, 122; Merrifield 1987, 7; Woodward 1992, 51.
23 Beagon 1992, 40–2, 100; Schrüfer-Kolb 2004, 108–9.
24 Fell 1990 draws attention to the writings of Pliny and Dioscorides.
25 Pliny, Nat. Hist. 268.33–4. Spurious works of a medical nature were often attributed to the mythical Orpheus
in antiquity.
26 Herbal 5.84–183.
27 Dungworth 1998, 153.
28 See Leach 1962; Ross 1967, 196 and 476.

The problem with viewing the use of iron objects and their subsequent deposition purely
in terms of practical past activity is that this draws too directly upon the idea that abandoned
materials represent pragmatic objects; deposits of iron objects are taken to represent either
valuable materials for recycling or mundane rubbish that reflect directly upon past industrial,
economic, or domestic behaviour.29 A focus upon the symbolic and ritual significance of iron
deposition leads to a different understanding. With regard to production, for example, fragments
of metalwork found at the Iron Age hillfort of South Cadbury (Somerset) were closely associated
with iron production in the eastern part of the interior of the site. Previously, these deposits
would have been interpreted as rubbish disposal derived from the creation of iron objects. The
recent publication of this evidence stresses, however, that the production and use of iron at this
time is likely to have been highly symbolic and ritualised.30 Even fragments of broken tools
and slag in archaeological contexts could represent ritually significant materials, deposited for
significant reasons.31
This is not to argue that we should downplay the ‘industry’ of later prehistoric and Roman iron
production. People produced objects in a manner that drew upon past experiences of success
and failure. A form of scientific knowledge will have developed, derived from experimentation
and observation. The ways in which this inherited experience was conceptualised, however,
will have drawn upon beliefs that would appear alien to a modern ironworker in the West. In
many cultures, ironworking is a mystical and highly-charged process.32 In pre-modern and non-
Western societies, the exact character of the transformation of iron ore into metal items is not
understood in terms of the chemical reactions and, often, iron production is imbued with beliefs
about the social and ritual meaning of the act of creation. In this context, it has been argued
that iron production in the past cannot have been an entirely pragmatic industrial process.33
Smithing, and particularly smelting, are impressive and dangerous processes that transform
raw materials into cultural items. The production of iron often involves recycling, as old iron
artefacts can be resmelted.34
Smelting and smithing required implements that would not melt at the temperatures involved;
iron was used for many of the tools that were used in these processes — hammers and anvils,35
and iron files were used to further refine the artefacts produced. Relevant objects depicted in
representations of the smith god, and their regular occurrence in archaeological contexts, perhaps
drew upon the ritual associations of iron production.36 Currency bars and billets, which represent
partly-processed iron used in the ironworking process, also occur in a variety of archaeological
contexts.37 Iron was used to produce powerful weapons — swords, spearheads, and axes
— vital to communities both for defence and for attacking others.38 Such objects are found in
deposits derived from the whole of the period considered in this study. Agricultural tools, such
as plough and ard shares, were made out of iron.39 The deposition of items connected with the

29 A recent discussion of ironworking (Schrüfer-Kolb 2004) pays close attention to the evidence for ritual, but still
maintains a clear division between ritual and pragmatism. This, often unquestioned, division of life into two opposed
areas requires challenging (Bradley 2005, 16). Why, for example, should the placing of ironworking structures with
regard to local topography, wind direction, and ‘economic reasons’ (Schrüfer-Kolb 2004, 112) preclude the idea that
such locations were ritually ‘good’ places to work iron?
30 Barrett et al. 2000, 114–16, 239.
31 For suggestions about the ritual potential of slag see Hingley 1997, 15 and Schrüfer-Kolb 2004, 111.
32 Aldhouse-Green 2002; Gillies 1981; Herbert 1993.
33 Aldhouse-Green 2002; Budd and Taylor 1995; Hingley 1997.
34 Manning 1972.
35 Leach 1962; Manning 1972.
36 Although these objects also acted to portray the god’s occupation (Bill Manning pers. comm.).
37 Hingley 1990a; 2005.
38 Hingley 1997, 13–14.
39 Rees 1979, 48; Hingley 1997.

agricultural cycle — including scythes, sickles, plough and ard shares — could have drawn
upon their symbolic association through a reference to the fertility of the soil and the creation of
agricultural surplus.40 Iron was also used to manufacture nails, artefacts that played a significant
part in the ritual actions of Roman society.41 In Britain, nails were used to attach defixiones to the
walls of temples.42 Van Driel Murray has explored the potential symbolic significance of Roman
shoes, which often have hobnails in the sole.43 In this context, the large collection of nails from
Inchtuthil (Tayside), and other nails from archaeological contexts, might have held a distinct
ritual meaning.44 Latch-lifters, keys, and locks were also produced from iron, relating closely to
ideas of boundedness and security.
A useful approach to the study of iron production would be to combine scientific method and
theory with an acceptance that the ironworking industry of later prehistoric and Roman Britain
was conceived in very different ways from the contemporary understanding of the science of
metalworking.45 This is not to suggest that the artefacts produced did not have a distinct range of
specific functions in activities such as warfare, agriculture, ritual, and iron production. Indeed,
iron was a fundamentally important material, since it was used to create a variety of ‘powerful’
objects. Nevertheless, stressing the symbolic significance of artefacts does not determine that
they were specifically votive, since their practical function represents the reason for their
symbolic power. The potential metaphorical associations of individual types of iron objects
demonstrate that different types of iron items may often have had varying associations and
this may suggest that the investigation of differential patterning in the deposition of individual
categories of finds may produce useful results.46


The quantities of iron objects found vary through time. This suggests that we should not assume
that iron had a single meaning over the long period under study. Initially, during later prehistory,
iron may well have been a valuable and symbolically-charged material.47 The scarcity of finds
suggests that it may not have been very common prior to the third to second centuries B.C.;
alternatively, perhaps much of the iron was being recycled. From this time on, fairly substantial
hoards are more usual and often include currency bars. Although they have been thought to
represent an entirely middle Iron Age phenomenon,48 new evidence may indicate that currency
bars continued to be buried during the first century A.D.49 The association of deposits of currency
bars with settlement boundaries has been used to argue that iron was deposited to symbolise the
identity and status of the individual or community who made the offering.50 A more mundane
interpretation is often envisaged for the individual finds of iron items, although at least some of

40 Hingley 1990a; 1997; 2005.

41 Dungworth 1998, 153.
42 Dungworth 1998.
43 van Driel-Murray 1999.
44 Dungworth 1998. See p. 229.
45 Budd and Taylor 1995; Pearce 1998, 52–3.
46 Haselgrove and Hingley in press.
47 Aldhouse-Green 2002; Hingley 1997.
48 Allen 1967; Hingley 1990a.
49 Hoards at Ditches (Trow 1988, 37) and Camerton (Jackson 1990, 14, 18) were probably deposited during the
first century A.D. At Totterdown Lane a hoard of currency bars was found in a small feature just outside what appeared
to be a later (second- and third-century A.D.) enclosure ditch. The excavators felt it probable that an earlier enclosure
ditch had lain in this area (Pine and Preston 2004, 45), although it is possible that the currency bar hoard was placed
in this context rather later.
50 Hingley 1990a, 107–8; 1997; 2005.

these objects are likely to have formed elements in the ‘special deposits’ that occur in many Iron
Age pits and ditches.51
The importance of iron to native British communities is indicated by the writings of Herodian
of Antioch, an author of the third century A.D. Herodian, writing about Septimius Severus’
preparations for the invasion of areas of free Britain beyond the northern frontier, remarked:
Strangers to clothing, the Britons wear ornaments of iron at their waists and throats;
considering iron a symbol of wealth, they value this metal as other barbarians value

Herodian was writing about a society that had been subjected to Roman interference for well
over a century and one that was under immediate threat of invasion.53 This situation may have
placed a particular premium on iron, while Herodian was evidently concerned to emphasise
the savagery and warlike character of the Britons. His comments may, however, have some
relevance to the significance of iron to native peoples during pre-Roman times and in those areas
that remained on the edge of imperial control.
During the Roman period, iron artefacts become far more common, occurring in large
quantities on military, urban, and rural sites within the province. Large-scale military production
is evident across Roman Britain,54 while the manufacture of iron in civil contexts was also
significant.55 Although this does not necessarily mean that iron immediately became universally
available in large quantities, by the later Roman period artefacts made from this material are
very common on many rural sites.56 Iron was used, often in bulk, for numerous purposes,
including the manufacture of bars to bridge the mouths of bath-house furnaces.57 Stray finds
occur in various contexts and this has often led to the idea that they were dumped as rubbish,
or lost during use and not retrieved.58 It is likely that iron was subject to changing perceptions
of meaning over the period covered by this paper, and that these influenced the production and
deposition of objects. We should not, however, define iron objects simply as rubbish without
careful consideration. Particular types of objects may have retained symbolic significance, and
we cannot write off even substantial bodies of Roman ironwork as ‘pragmatic’ rubbish disposal
without further analysis.


The nature of the available information raises serious issues. Iron objects have often been
viewed as mundane material and, as a result, individual finds and even substantial collections
have not always been well recorded or published. Until recently, publications of archaeological
excavations rarely discussed the character of deposition, or even provided details of context.
Careful excavation, which focuses upon questions of depositional context, is required before
a full assessment of the evidence is possible. Some old excavations included the recording
of relevant evidence, but detailed excavation, recording, post-excavation and publication of
individual findspots has only become regular practice since the 1980s. Improved standards of

51 For example, Sellwood 1984; Cunliffe and Poole 1991b, 333, 354; Hill 1995a.
52 Herodian, History 3.14.7.
53 Haselgrove and Hingley in press.
54 Sim and Ridge 2002.
55 For a thorough and useful review of the evidence for military and civil involvement in the iron industry, see
Schrüfer-Kolb 2004.
56 Manning 1972, 239. An exception is discussed below (p. 231).
57 Bill Manning pers. comm.
58 See p. 215.

excavation and recording in recent years, together with an increased interest in the significance
of iron objects, allow patterns to be observed on some fully published sites.
The Appendix has been compiled from a variety of published sources. Only deposits including
iron objects that have a fairly good record of context are included. Where a site has produced
more than a single significant deposit, a number is given after the site name in the text,
corresponding with the information presented in the Appendix. The context types listed in the
appendix are discussed in the analysis below. It is important to stress that this is certainly not a
complete database, nor does it attempt to be. There is so much information for the deposition of
iron on sites of Iron Age and Roman date that a survey of all such finds would be an absolutely
immense undertaking. The likely biases that result from the methods of data collection adopted
need to be emphasised. The aim behind the collection of data was to include a wide variety of
deposits so that tentative patterns can be established in the types of contexts represented. A
deliberate attempt has been made to collect information from a wider range of contexts than
those usually defined as ‘hoards’.59 The data included deposits from the start of the Iron Age
to the end of the Roman period and information has been gathered from England, Scotland,
and Wales. Individual records in the Appendix include assemblages that are made up almost
entirely of iron objects and also collections of objects and finds of a variety of materials that
include some iron items.60 Deposits called ‘hoards’ by their publishers, other collections of
items from particular contexts, and some single finds are included. It is likely that many of
the collections of objects in various materials, such as those from rivers, the hillfort ‘massacre
deposits’, and objects from wells and pits, did not form sealed collections and may have been
deposited over a period of time,61 but examples defined as ‘hoards’ may also not have been
totally sealed.
As we have seen, single finds have often been excluded from discussion, since they are usually
assumed to represent casual losses or discards. The published accounts of single iron objects from
Asthall, Billingborough, Nadbury, and South Cadbury 2 considered that they were deliberately
deposited. The inclusion of single items is likely to add bias to the data included in the Appendix,
since these objects have been selected through their recognition as deliberately deposited items.
A vast number of additional individual complete and broken finds could have been included in
this study, but this would have made the Appendix unmanageable. These selected individual finds
help to challenge the simplistic divisions that have been drawn between collections of objects
and single examples. Since the information for single finds in the Appendix is so selective, a
later section of this article examines the evidence for the deposition of individual iron objects on
three particular sites. It is only through detailed studies of single sites that distinctions between
patterns of deposition for single objects and groups are likely to be possible.
The reader may feel that the information included in this study has been selected to fit
particular ideas, but I have attempted to address this issue by following a fairly open approach to
the collection of information. Hopefully, the tentative general patterns that are drawn out from
the material reflect in a useful manner on the entire body of data. The patterning suggested here
may, in future, be explored and assessed through more detailed analysis of individual sites, areas,
or periods. The database is certainly more complete in some areas than in others. In particular,
the author’s ongoing work on the deposition of currency bars means that the information for
deposits containing these objects should be reasonably comprehensive.62 In other areas, the

59 For discussion, see p. 215.

60 Manning 1972, 224 followed a comparable approach.
61 See note 16.
62 Hingley 1990a; 2005. I am aware of one additional find and one possible find that are not recorded in my 2005
article (Lambrick and Allen 2004; Pine and Preston 2004).

information is far less reliable. This database, inevitably, does not represent anything more than
a small percentage of the total assemblage of Iron Age and Roman finds.
In order to examine the motivations behind deposition, this paper addresses two types of
information: first, the reoccurring types of context in which the objects were buried; and
second, the frequency with which particular types of object were deposited in certain types of
context.63 The main focus of attention in this paper is upon the former. These changing contexts
of deposition help to provide an understanding of where people were choosing to deposit iron
objects.64 The analysis of the deposition of objects in these contexts may enable an interpretation
of the meaning of their actions. Concerning the types of objects deposited, different types of
objects may often have held different associations, as has already been suggested.65 As a result,
it is worth considering whether the different uses and metaphorical meanings of different artefact
types have been used to structure the acts of deposition. Both types of analysis are attempted
below, although the study of the context of individual types is restricted to currency bars,
weapons, and agricultural tools, due to the limited character of the material assembled in the


The material has been divided into three general periods to assess change through time (FIGS
1–4). A variety of contexts of deposition are defined, including:
1 = ‘natural’ = cave, wetland, river
2 = ‘shrine’ = temple, or shrine, or possible shrine
3 = ‘enclosure’ = in, or close to, the ditch or bank of an enclosed settlement
3a = the ditch or bank of the enclosure
3b = entrance area to the enclosure
3c = close proximity to enclosure earthworks (within 10m)
4 = wells or deep pits (over 2m deep) dug into, or very close to (within 10m), the boundary
of a settlement or site
5 = wells or deep pits (over 2m deep) dug within the area of a settlement or site
6 = other contexts within the area of a settlement or industrial site
7 = other features on the edges of a settlement or in the landscape
With regard to Contexts 3c and 4, ‘close to’ a boundary means within a maximum of 10m; most
examples are far closer to the boundary. The deep pits included in Contexts 4 and 5 are over 2m
FIGS 1–4 demonstrate that contexts associated with site boundaries appear to be significant
(marked as ‘enclosure’ and ‘well a’), constituting around half of the total number of iron
deposits. These include boundary contexts and wells/deep pits on, or close to, boundaries. Other
contexts within settlements or sites (marked as ‘settlement’) form the next most common type on
FIG. 1. ‘Natural’ contexts (wetlands and caves) and shrines also form significant categories.
Between the fifth and the first centuries B.C., iron objects occur in a broad range of contexts,

63 For the use of context to study motivations for burial, see Esmonde Cleary 2000, 127; Hingley 1990a; Manning
64 Haselgrove and Hingley in press.
65 See p. 217.
66 In the case of Brampton, although the pit was shallower, the excavator argued that it had been severely

FIG. 1. All deposits of ironwork classified against context type; n = 91 (each deposit
including iron is counted once). For key to context types see p. 221.

approximately 75 per cent (34 out of 45 examples) of which are connected with domestic sites.
There is a considerable depositional focus upon the enclosure boundary (FIG. 2a). Items also
occur in wetland contexts, at possible shrines, and in a variety of other contexts on settlement
sites. An examination of the deposits that occur in settlement boundary contexts (FIG. 2b)
indicates that almost one half (12 examples) are from the boundary itself (the ditch or bank),
while four examples come from entrances to enclosures. Nine examples (approximately 40 per
cent) are from contexts in very close association with the boundary.67 From the first century B.C.

FIG. 2. All deposits from the fifth to the first century B.C. (Each deposit including iron is counted once. For key to
context types see p. 221). a. Contexts (n = 45). b. Details of boundary contexts (n = 35).

to the first century A.D., the range of contexts broadens (FIG. 3a). The number of examples from
wetlands and shrines remains comparable, while the number from boundary contexts decreases
slightly. If we add the finds from wells/deep pits close to settlement boundaries to those from
enclosure boundaries, these represent 37 examples (59 per cent of all deposits).68 A significant

67 Details of these contexts are given in the Appendix and many are in quarry hollows or just at the back of
68 This is mainly a result of the pits from Newstead, the majority of which may well be of the first century A.D.
(see p. 228).


FIG. 3. All deposits from the first century B.C. to the first century A.D. (Each deposit including iron is counted once.
For key to context types see p. 221). a. Contexts (n = 63). b. Details of boundary contexts (n = 31).

number (7 examples) occur in other settlement contexts. Regarding the boundary contexts (FIG.
3b), the number of finds from enclosure boundaries increases slightly, with a slight decrease
in the number of deposits ‘close to’ the boundary. For the period from the second to the fifth
century A.D. (FIG. 4), the finds from ‘natural’ contexts and shrines stay broadly comparable. Finds
from enclosure boundaries fall dramatically, as the number in wells/deep pits close to settlement
boundaries increases. Finds from other contexts on settlement sites are also common at this time.
For this period, too few finds come from boundary contexts to make the proportions analysed for
earlier periods meaningful.

FIG. 4. All deposits from the second to the fifth century A.D. Contexts (n = 37)
(each deposit including iron is counted once). For key to context types see p. 221.

Rivers, bogs and caves (‘natural’ contexts)

Looking at the material in greater detail, a division is sometimes made between ‘natural’ contexts
and those that are associated with human settlement.69 These natural contexts include rivers,
bogs, wetlands, caves, and rocky outcrops. This distinction is, however, too general. First, the

69 For example, Hingley 1990a.


societies who created and deposited objects in such contexts comprehended their environments
differently to the present-day understanding of culture and nature.70 In fact, human activity,
including the construction of houses and sites, and the placement of artefacts, may have drawn
landscape features such as caves, rivers, and wetlands into the occupied landscape, although
it is true that they would have remained beyond the settlement and infield areas of individual
communities. Excavation of the wetland deposits at Fengate and of those in the river at Fiskerton
indicates that timber platforms or tracks were used as places from which to cast individual
objects into the water.71 Other Iron Age finds from rivers and bogs may have been associated
with similar constructions. Three early Roman hoards from lakes and bogs have been found in
Scotland,72 while Roman objects have also been found at the significant Iron Age causeway at
Fiskerton.73 Deposits in bogs, lakes, and rivers appear to be primarily an Iron Age and early
Roman practice and ironwork does not appear to have been left in such contexts from the second
to third centuries A.D. onwards.74 Although several occurrences are noted on FIG. 4, none of
these appear to postdate the second century at the latest. A few iron objects come from caves;
these locations are likely to have had some special significance, since they often contain unusual
prehistoric and Roman finds.75 Many, perhaps all, of the deposits considered in this paper were
probably intimately connected with human occupation and the adaptation of the landscape.

Contexts associated with settlement and temple boundaries

A number of studies of the boundaries that surround Iron Age hillforts and enclosed settlements
have emphasised the ritual and symbolic significance of these structures.76 In the past the
military function of ramparts has often been stressed, but it is increasingly realised that the
boundaries to sites had other metaphorical associations which may explain the common
occurrence of currency bars and other finds in these contexts.77 The association between Iron
Age currency bars and enclosure boundaries (FIG. 5a; Table 1) mirrors a slightly less clear-cut
association for fifth- to first-century B.C. ironwork deposits (FIG. 2a); this is unsurprising, since
many of the latter contain currency bars. Hoards and objects were deposited in different contexts
in, and close to, the boundaries of hillforts and enclosed settlements at this time (FIGS 2b and 5b).
Some of the iron objects came from the boundary earthworks themselves, from the ditches and
banks of enclosures; the recorded details at Ditches and Stanway suggest deliberate deposition.
At Stanwick 2, a set of iron shears was placed in the ditch that surrounded a house within the
boundary of the so-called oppidum.78 At Park Farm, the currency bar came from the upper fill
of the enclosure ditch, possibly suggesting that it was originally placed within, or on, ramparts,
later being redeposited into the ditch. At Nadbury, the single currency bar was deposited in its
pit at a time at which the Iron Age rampart had apparently gone out of use, while the two acts
of deposition at South Cadbury (1 and 2) dated to phases of use of the long-term rampart. At

70 Bradley 1998.
71 Field and Parker Pearson 2003; Pryor 2001.
72 Piggott 1955; Hunter 1997, 116. Piggott (1995) and Manning (1972, 242) argued for a Roman military origin
for these hoards, but Hunter (1997, 116–17) has examined the evidence that they may have been native in origin.
73 At Fengate Powerstation, a Roman road (the Fen Causeway) ran c. 100m to the north of the earlier timber
track/platform and followed a roughly comparable course for some distance (Pryor 2001, figs 4.1, 5.1). Perhaps the
course of this Roman road perpetuates the memory of an earlier routeway. The Fengate excavations produced some
Roman finds, although it is not clear how these came to be deposited.
74 Although, see the thoughtful study by Merrifield (1997) of the material from the Walbrook (London).
75 Hingley 1990a, 98; Branigan and Dearne 1992.
76 Bowden and McOmish 1987; Hill 1995a; 1995b; Hingley 1990a; 2005.
77 Hingley 1990a; 2005; Haselgrove and Hingley in press.
78 Perhaps the object was originally stored in the roof of the building.


FIG. 5. Contexts containing currency bars. (Each deposit including iron is counted once. For key to context types see
p. 221). a. Contexts (n = 22). b. Details of boundary contexts (n = 15).


Site Context Number of objects

Bearwood 3a 4
Bredon Hill 6 3a 2
Danebury 1 3c 21
Danebury 2 3c 1
Ditches, Glos. 3b c10
Fiskerton 1 1
Gretton 7 48
Hayling Island 2 2
Hinchingbrooke Park 3a 2
Hod Hill 4 3c 1
Kingsdown 3a 2
Madmarston 3a 12
Nadbury, Warks. 3a 1
Orton Meadows 1 9
Old Down Farm 6 1
Park Farm, Barford, 3a 1
South Cadbury 1 3a 1
Spettisbury 3a 1
Stanway 3a 2
Totterdown Lane 3c c10
Wookey Hole 1 1
Worthy Down 7 13

Madmarston, the collection of iron artefacts was placed on the back of the rampart, possibly
within a building or paved area.79
Between the first century B.C. and the first century A.D., the emphasis upon enclosure
entrances appears to be particularly marked (FIG. 3b). The smaller proportion of currency bars
from enclosure entrances (FIG. 5b) may indicate that this fixation on the entrance is a late Iron
Age trend. The entrance contexts include several so-called ‘massacre deposits’.80 It has often
been supposed that the disarticulated human remains, associated with weapons and personal
objects, at the sites of Bredon Hill, Ham Hill, Maiden Castle, South Cadbury, and Spettisbury
are the results of battles occurring in and around the hillforts during the final phases of the Iron
Age, or at the time of the Roman conquest of these areas. At Bredon Hill, South Cadbury, and
Maiden Castle the deposits were made in the entrance to the hillforts, while the Spettisbury finds
had been placed in the ditch outside the rampart, and those at Ham Hill may have been placed
in a hollow behind the rampart. The damage on some of the human bones from these deposits
has often been used to support the idea that these remains are those of war victims, although
the occurrence of multiple individuals, together with weapons and other objects, on sanctuary
sites in France (at Ribemont-sur-Ancre and Gournay-sur-Aronde) suggests that these ‘massacre
deposits’ might equally well form ritual deposits associated with sacred locations.81 In fact,
individual iron weapons also occur in boundary contexts on hillfort sites; the ditch ends of
entrances at Stanwick 1 and Bredon Hill 1 produced, respectively, a very well preserved sword
in a sheath and a ‘flamboyant’ spearhead. The Stanwick sword was closely associated with a
decapitated human skull.
The distinct objects in these deposits may originally have been placed in highly visible
locations before they found their way into hollows and ditches. The human skull at Stanwick 1
may originally have been displayed on a pole at the entrance,82 while the number of skulls along
the line of the timber gate at the hillfort at Bredon Hill has been taken to indicate that they had
been set up on top of the entrance.83 Presumably, weapons and other objects could also have
been placed on display above the gateway, or on top of the rampart. Alternatively, bodies and
weapons may have been displayed on the sloping outer or inner face of ramparts.
A variety of other finds have been made close to ramparts and enclosure ditches, defined in
an earlier article as having a ‘loose association’.84 On FIGS 2b, 3b, and 5b these are classified as
‘close to’ the enclosure boundary. To associate all these finds with boundaries may be to push
the evidence too far. After all, many past excavations have focused on the boundaries of Iron
Age enclosed sites.85 It is necessary, however, to consider the significance of areas immediately
inside and outside the line of the ramparts, as boundary zones are unlikely to have included only
the physical structures that defined sites. At Balksbury Camp (Hants.), the symbolic significance
of this marginal internal space inside a late Bronze Age hillfort has been stressed,86 and the
quarry hollows and periphery of later hillforts might have had comparable associations. At
Danebury, the four distinct ironwork hoards identified by the excavators all occurred in contexts

79 A recently discovered hoard from the East Riding of Yorkshire, which included five iron swords with scabbards
and spearheads, was found in a ditch which may have defined the boundary of a settlement (Fenn 2003). This is not
included in the Appendix, since its exact findspot has not been published.
80 Wheeler 1943; Sharples 1991; Barrett et al. 2000, 105–16.
81 Brunaux 1988.
82 Wheeler 1954, 53.
83 Hencken 1939, 23.
84 Hingley 1990a.
85 Cunliffe 1978, 243.
86 Ellis and Rawlings (2001) review the evidence for the midden that had built up behind the ramparts of an
apparently relatively sparsely occupied hillfort.

that were just behind the ramparts, within roundhouses situated within the quarry hollows. The
extent of the excavation at this site suggests that the context of the burial of the iron objects close
to the rear of the rampart is significant, as none of the excavated houses within the interior of the
site produced hoards.87 Finds of iron objects in comparable contexts at Hod Hill 4 and, possibly,
at Bredon Hill 2 and 3 may indicate that this type of context was common. The act of deposition
of the ironwork may only have been visible to those within or immediately outside the house in
each case, stressing the importance of the immediate context of the act.88
The areas between defence-works and immediately outside the boundary also appear to have
been significant. At Bredon Hill 6, the currency bars were found in a feature positioned between
the inner and outer defences of the hillfort, while the currency bars at Totterdown Lane may
have been deposited just outside a boundary ditch.89 The other contexts defined as ‘close to’ an
enclosure boundary on FIGS 2b, 3b, and 5b and in the Appendix have comparable associations to
the physical boundaries of individual sites.
Not all significant Iron Age deposits occur in the boundaries around enclosed settlements
(FIG. 2a), and it is likely that the attention paid to these contexts has biased the recovery of
data. Occasionally, currency bars are located in pits in the interior of settlement enclosures,90
and sometimes they may occur in unenclosed settlements.91 These exceptions do not, however,
detract from the strong overall pattern. The vast majority of significant deposits of Iron Age iron
come from enclosed settlements. Although unenclosed, or ‘open’, settlements are fairly common
in the middle Iron Age settlement record, collections of currency bars are rarely found in these
contexts.92 Many of the larger collections of currency bars come from settlement boundary
contexts, emphasising the significance of these deposits.93
The evidence from a number of later Iron Age and Roman temples and sanctuaries strengthens
the association.94 At Uley, a large number of spearheads was found in a ditch and pit that defined
the east side of the late Iron Age shrine; while at Hayling Island, fragments of currency bars,
spearheads, and linch pins, together with finds such as human bone, were located on the eastern
boundary of the enclosure around the early temple building. At Baldock, in a third-century A.D.
context, Pit A13 produced 33 iron spearheads; this feature was one of a number located close to
the northern boundary of what may well have been a shrine or sacred area. Other Roman shrines
and temple sites have produced spearheads, for example, at Bancroft, where eighteen examples
were found in and around a circular building which replaced the mausoleum.95
Boundary contexts also appear significant at certain Roman settlements and forts.96 Relevant
forts at Newstead and Bar Hill housed army units with origins outside Britain. At Newstead, in

87 In addition, the distribution of particular types of pit fills at Danebury (Poole 1995, 255–9) shows no focus on
the peripheral areas of the enclosure.
88 Hingley 2005, 201.
89 Pine and Preston 2004.
90 Hingley 1990a, 103.
91 The recent publication of two possible fragments of currency bars from the settlement at Gravelly Guy, Oxon.
(Lambrick and Allen 2004, 364–5) illustrates that some unenclosed settlements may have produced fragmentary
examples that have not been recognised (these finds are not included in the Appendix). It should be noted, however,
that the Gravelly Guy settlement lay on the edge of a substantial territory that was defined by a linear ditch that
bounded the settlement, so this may also represent another type of boundary context.
92 Hingley 2005.
93 The massive currency bar hoard from Meon Hill (Warwicks.), for example, probably came from a boundary
context, although the exact findspot is uncertain (Hingley 1990a) and it has not been included in the Appendix or FIGS
94 These are defined as ‘shrine’ contexts rather than enclosure contexts in the Appendix and on relevant FIGS.
95 Five examples were also found during the villa excavation (Skinner 1994, 339).
96 Manning 1972, 243. For the significance of boundaries in the Roman period, see Hingley 1990b and Esmonde
Cleary 2000, 137–8.

the late first or second century A.D., a substantial number of deep pits across the fort and annexes
contained significant finds, including iron objects.97 Some of these pits appear to have served as
wells, the majority of which were located in the south annexe of the fort where recent work has
indicated that they supplied a variety of timber buildings, many involved in industrial production.
A number of the pits had significant associations with features defining the boundaries of the
fort.98 Although many of the pits were actually within the interior of the fort and its annexes,99
Clarke and Jones noted that the boundary pits contained more in the way of special objects than
the average pits across the site and that they were often deeper.100 Seven pits with significant iron
objects are mainly within, or near to, the defences of the fort and its associated enclosures.101
Hoard 1 was buried in a pit that was partly sealed by the ‘rampart’ that surrounded the bath-
house in the western annexe. Hoards 2, 3, and 4 were buried in the defensive system of the fort,
perhaps, in the case of Pit 58 (Hoard 4), associated with the defences of the initial fort. Hoards
5 and 7 came from pits located approximately 10m south of the outer line of the later fort ditch,
along with a number of other pits that contained significant finds. Bill Manning is currently
studying the material from the Newstead pits and has pointed out that, while Curle argued that
much of the material was associated with the Antonine fort, most of the pottery appears to be
Flavian.102 This may mean that the seven significant pits are not quite as closely associated with
the boundary earthworks as would currently appear to be the case, but, since the ramparts of the
earlier fort appear to have been followed approximately by those of the later fort, the pits would
still appear to occur just beyond its boundary.103
The exact details of the context of the boundary pits excavated early in the twentieth century
at Newstead are unclear, but one, partly excavated by R.F.J. Jones and Peter Rush in 1990, was
cut into the side of the ditch of the eastern annexe and had been deliberately filled. This was
followed by the deliberate back-filling of the annexe ditch. Clarke and Jones interpreted the pit
and its fill as representing part of a termination ritual at the end of the life of the annexe.104 It
should be noted, however, that this pit was not excavated deeper than the top fills and we do not
know whether it contained significant objects. In addition, finds of individual items of ironwork
from Newstead were not restricted to the boundaries of the site. Some classes of iron finds,
particularly swords, do appear, however, to have a particular association with the boundaries,
while, interestingly, spearheads appear to have been more widespread.
Hoard 6 at Newstead is the only exception to the boundary context of these collections, placed
in a central location in the courtyard of the principia. A significant collection of material from
the Roman fort on Bar Hill (1) came from a well in the principia, while a second pit (2), which
held a complete wheel with iron tyre, was located immediately inside the rampart, just to the
west of the southern entrance. The evidence from Newstead and Bar Hill suggests that units of
the Roman army may have indulged in ritual practices broadly comparable to those that occurred
on native sites in Roman Britain. Evidently, more information is required on the ritual practices
performed by these people in their native lands before we can place the British information in

97 Curle 1911; Clarke and Jones 1996.

98 Clarke and Jones 1996, 120.
99 Clarke 1997, 79.
100 ibid.
101 Including contexts that contained iron weapons and/or those that contained more that five individual iron
102 Bill Manning pers. comm.
103 See Curle 1911, 185.
104 Clarke and Jones 1996, 113, 115, 120.
105 See van Enckevort and Willems 1996 and Bonnamour and Dumont 1996 for ritual deposits of military
metalwork within forts on the Continent.

Manning’s ongoing research on the material from the Newstead pits has led him to conclude
that many of these ‘hoards’ form closing deposits, relating to the abandonment of the fort. The
Inchtuthil nails might have had a comparable relevance.106 Such a consideration does not,
however, detract from the idea that these deposits might have had a ritual dimension. Finds from
other sites may hint at acts of deposition commemorating an old settlement feature, or an entire
settlement, that was going out of use, or being obscured. At Gretton, the small pit containing
currency bars apparently cut into an earlier and largely silted-up pit that formed part of a pit
alignment. At Billingborough, close to an Iron Age settlement, a poker was placed in a shallow
slot dug into a late Bronze Age ditch, while the feature with the currency bar at Nadbury appears to
have been excavated when the hillfort rampart was collapsing. At Kilvertstone, during the third to
fourth century A.D., the pit containing the iron objects and pewter vessels was dug on the edge of a
settlement, where it was cut into an earlier and disused boundary ditch. Deposits from the Roman
forts of Inchtuthil, Newstead, and Bar Hill are often associated with the idea of abandonment and
the clearing of the site,107 acts that are likely to have had a deeply ritual significance.
Other Roman-period finds also come from boundaries, although this context type appears
to become less common with time (compare FIGS 3 and 4). This trend may, however, be partly
an aspect of the way in which the material has been categorised in this study. At Barton Court
and Dalton Parlours, wells were sunk close to the edges of the enclosure systems of the Roman
settlements. At Dorchester on Thames, the late Roman hoard was placed on a road immediately
inside the rampart of the town, recalling an earlier tradition of the deposition of iron objects on
hillfort sites.108

Contexts within settlements, including wells

A number of Iron Age iron deposits come from buildings and other features within settlements
(classified as ‘settlement’ on FIGS 2a and 3a). At Houghton Down, Old Down Farm, and,
possibly, at Worthy Down, finds were made from pits inside an enclosed settlement,109 while
collections of objects also sometimes come from contexts within roundhouses in the interior of
hillforts, as at Hod Hill 1, 2, and 3. The evidence may suggest that finds of the first century B.C. to
fourth century A.D. come in increasing numbers from contexts within settlements and industrial
sites (compare FIGS 2a, 3a, and 4). This is just the type of information that could indicate that the
meaning of hoarding was changing, with some of the late Roman deposits perhaps taking on a
more mundane significance, as deposits of useful material that was being saved for reuse.110 This
is uncertain, however, since it has been proposed that certain iron deposits within settlements held
ritual significance. At Haddon, a small pit contained iron shears, a chain segment, and a Roman
coin; this was interpreted as a ‘foundation deposit’ connected with the construction of a malting
oven.111 At Wavendon Gate, a deposit from a third-century A.D. pit within an enclosed settlement
may also suggest that at least some of the iron objects were deposited for similar reasons (see
p. 233). Finds from other settlement contexts, including the oven at Asthall, the foundations of
a building at Great Holts Farm, and a stone drain inside a farm building at Stanwick villa, may

106 Bill Manning pers. comm.

107 Manning 1972, 241, 243, and 246.
108 At Silchester (2), the well that contained the iron hoard may be significant since it was very close to the line of
the boundary of the insula in which it was located.
109 Hooley 1931, 178–9 suggests that two Iron Age ditches at Worthy Down defined an area around a number
of pits, including the one that contained the currency bars. In my earlier work this site has been classified as an
unenclosed settlement.
110 See p. 219.
111 Hinman 2003, 52, 112.

have had a comparable significance. By contrast, it is highly likely that some other objects were
simply lost, or discarded as rubbish;112 for example, at Cannard’s Grave, several fittings found
in the make-up of a track or road probably fell off a cart.
Wells usually appear to be features of settlement and, together with waterlogged pits, are
common Roman-period contexts for depositing iron objects (FIGS 3a and 4). The collections
of finds made in deep pits and wells have often been interpreted in ritual terms,113 although
care is necessary over the idea of timeless, unchanging ‘Celtic’ religion and the role of wells
in any such tradition.114 The Iron Age wells reviewed by Jane Webster do not appear to have
contained significant deposits and the information in the Appendix contains no significant iron
objects of Iron Age date deposited in deep pits and wells. It is possible that the placing of votive
materials in wells and deep pits represents a Roman phenomenon, generally of late Roman date,
but occurring during the late first to second centuries A.D. at Newstead and perhaps elsewhere.
Woodward and Woodward have suggested that the deposition of significant materials in pits/
wells may actually be an imported ‘Roman’ tradition,115 although the general similarity of some
of the iron objects reviewed in this paper that occur in later prehistoric and Roman deposits
could suggest that they form a new type of context for what are effectively old practices.
A consideration of the character of the fills of some of these Roman-period pits and wells
points clearly to a ritual interpretation.116 The remarkable layer formation of the highly structured
deposits in the fill of the pit at Jordan Hill, and the cists half-way down and at the base, with their
apparent reference to burial (single sword and spearhead), is hard to interpret any other way. The
well at Silchester (2) also had a massive collection of ironwork sealing a cist at the bottom that
contained two complete pots. Pits and wells at Newstead (Hoards 1, 6 and 7), Dalton Parlours,
and the probable well at Appleford (2) contained human remains and other significant finds. The
majority of material discussed by Poulton and Scott in their article on pewter hoarding and the
evidence from the wells with significant ironwork suggest that the deposition of significant items
in this type of context is mainly a third- and fourth-century A.D. practice.
This does not mean that these wells were all constructed in the late Roman period, as they
are often dated using the materials that filled them. They may have been kept clean, and the
material deposited in them might therefore represent disuse or termination deposits dating from
the final stage of the disuse of the well. This raises an important issue for the iron deposits that
have been considered already. If only the final deposits placed in long-lived wells survive to be
found, perhaps the same is true for many of the later prehistoric ironwork deposits that were
made in settlement boundary contexts. Perhaps finds of these types were commonly deposited in
settlement boundaries and pits, but the deposits located by archaeologists are the final examples
that marked the end of the life of a feature. Certainly, as we have seen, when the excavated
evidence is sufficiently detailed, iron objects often appear to occur in relatively late contexts in
the sequence of site development.


As we have seen, single iron objects may often have been deposited for significant reasons. It

112 Osbourne 2004, 3 has discussed the difficulty in distinguishing when an object was dedicated, or simply
113 Ross 1974, 50–6; Poulton and Scott 1993, 122–3; Wait 1985, 51–82.
114 Webster 1997.
115 Woodward and Woodward 2004.
116 See Wait 1985, 51–82 and 321–35 for a full discussion of the evidence. The objects from many of the wells and
pits discussed by Wait are very mixed. Only examples that have produced a reasonable quantity of iron objects are
discussed in this paper. Wait’s list includes other wells and pits that have produced one or two iron items.

is, in general, difficult to study the distribution of single iron objects from excavated sites, since
the large number of finds often means that individual items are not recorded and published in
sufficient detail to enable detailed analysis. The evidence from three particular sites enables an
assessment of the distribution of iron objects. These sites have been selected since they raise
interesting issues.
At the extensively-excavated later prehistoric hillfort of Danebury (Hants.), four ironwork
hoards have been distinguished, while a plethora of additional iron artefacts were scattered
across the excavated area in a variety of types of archaeological context, particularly storage
pits.117 Analysis of the content of Iron Age storage pits at Danebury and elsewhere has indicated
that many contained ‘special deposits’,118 and it has been suggested that many of the individual
(broken and complete) iron objects at Danebury were deliberately deposited.119 It is interesting,
however, that the hoards have a clear association with houses that were built either in the quarry
hollows or very close to the rampart,120 while the other objects appear to be distributed more
randomly and may represent different practices of deposition.121
A sub-rectangular ditched enclosure to the north-east of Site A at Baldock (Herts.) formed
part of an extensive late Iron Age and early Roman site; this enclosure was initially constructed
around the middle of the second century A.D. (FIG. 6).122 It contained a round building, defined
by a ditched enclosure, and a variety of other features, including a cluster of pits close to its
northern corner. An elaborate and multi-period entrance, defined by four posts and two pits, lay
to the south-east. This enclosure featured a marked concentration of iron objects when compared
with the remainder of the site. Significant finds included a scatter of iron projectile heads and
spearheads, mainly from the northern part of the enclosure and the entrance. A particular
concentration of 33 spearheads and projectile points came from a third-century A.D. pit or well
(A13), while six further examples were found close by. Five further spearheads or projectile
heads came from the entrance area, which also produced a ritual ‘rattle’ and a fragment of a
bronze statue. Other iron finds included a hammer-head and a cock-spur, still attached to the leg
bone of the cock. Several bronze model weapons were also found. This enclosure may, perhaps,
represent a shrine or temple,123 while the distribution of iron objects across the remainder of
the excavated area may, perhaps, be less highly structured. Certain iron objects, including
spearheads and projectile heads, may have been deposited as votive objects, while some other
finds could represent casual loss.
The iron finds from the site at Wavendon Gate (Bucks.), however, indicate that we should not
assume that even single items found on Roman sites are necessarily casual losses or discards. This
site was a large settlement enclosure (FIG. 7). Built around the middle of the first century A.D.,
close to a pre-existing Iron Age settlement, it remained in use with various phases of recutting
until at least the early fourth century.124 Within this enclosure there was evidence for domestic
settlement and also a small cemetery. Although a variety of objects suggest ritual practices,
the site was interpreted as an agricultural settlement. Small finds from the site were common;
for instance, there were 108 Roman coins and numerous copper-alloy items, including eleven

117 Cunliffe and Poole 1991b, 333; Cunliffe 1995, 83; Poole 1995, 262–3.
118 Hill 1995a; Cunliffe 1995, 80–6.
119 Cunliffe and Poole 1991b, 354. Poole 1995, 264–75 reviews the complexity of the fill deposits from pits at
120 See pp. 226–7.
121 For further discussion of the distribution of individual iron objects at Iron Age sites, see Haselgrove and
Hingley in press.
122 Stead and Rigby 1986, 44.
123 See Booth 2001 for a potentially comparable enclosure at Westhawk Farm (Kent).
124 Williams et al. 1996, 83.

FIG. 6. Ditched enclosures at Baldock, Herts., showing locations of iron objects (after Stead and Rigby 1986, fig, 4)

FIG. 7. Settlement enclosure at Wavendon Gate (Bucks.), showing locations of iron objects
(after Williams et al. 1996, fig. 45).

brooches. Iron objects, however, were rarer: the extensive excavations produced 25 stratified
pieces in Roman contexts,125 along with 116 stratified Roman nails. This is a remarkably small
collection for a domestic site that was occupied throughout the Roman period.126 A waterlogged
pit within the settlement enclosure produced finds that dated to the third century A.D., including
a spearhead and four other iron objects, in addition to a significant proportion of the nails from
the site. The pit also contained other significant finds, including a large part of the bole of an
ash tree, which appears to have grown close by. The impressive wooden wheel from the pit is
reminiscent of the symbol in various representations from Britain of an iron-working god.127

125 Hylton 1996, 120–5, who notes 34 objects. Seven of these were unstratified, while two of the objects are post-
126 Hylton 1996, 125.
127 See the pottery vessels reviewed by Leach 1962. This author suggests that the wheel symbol and full-face
helmet on the sceptre from Farley Heath are symbols of a second god (ibid., 40; Goodchild 1938), but they could
well be connected with the apparent smith god shown on this object (see Green 1976, 24–5 for associations between
Taranis and Vulcan).

A series of flat stones appeared to have been positioned to create a means of access to the pit,
suggesting that people may have utilised it from time to time. Perhaps many of the objects in
the pit were offerings placed in or on a tree that formerly grew within the enclosure and became
incorporated in the pit when it died.128
Of the twenty-five iron objects from the site, five were found in the waterlogged pit, while another
seven came from the metalled surface associated with it, which may have been used as a working
area. Five objects came from various phases of the enclosure ditch that surrounded the settlement,
four from various phases of the sub-division of this enclosure, and two from a cremation burial.
One came from the ditch defining a later enclosure built to the north-east of the main enclosure;
another (a scythe fragment) from cleaning above a corn-drier. All these contexts related directly
to elements of the site that could be interpreted, in the terms addressed above, as of potential
significance. Of the nails, 32 (28 per cent) derived from inhumations and cremations, 43 (37 per
cent) from the waterlogged pit and associated contexts, and 41 (35 per cent) from the remainder
of the settlement enclosure and associated contexts. The numbers of nails and other objects of iron
appear to be remarkable low and highly-structured to certain types of context. For example, the
numerous pits and post-holes inside the settlement enclosure produced no iron objects.129
Many of the iron objects from these three sites appear to be structured in their deposition,
although this does not necessarily mean that they were all votive.130 The dramatic variations
in the quantity of iron finds on different sites, and the significant contexts in which these finds
occur, enable an informed discussion to take place about the variable motivations for the
deposition of the iron objects.131 Other Iron Age and Roman sites have produced vast quantities of
iron objects, and the reasons for their deposition require serious contextual study on an individual
basis, comparing their occurrence to other types of artefacts, to determine patterns in deposition.


The information discussed above about the contrasting context of swords and spears at Newstead,
together with the distribution of finds at Baldock, raises the issue that it may often have been the
metaphorical associations of the particular object, rather than the material that it was made out
of, which determined why it was deposited in particular contexts. To pursue this topic, one might
assess the contrasting contexts in which different iron artefact types occur to see whether these
were used in varying combinations. The types of objects deposited will not be discussed in any
great detail in this paper, since there are many other examples of the types of objects addressed
that are not included in the Appendix, and this is a serious limitation on the value of any such
analysis. It is useful to provide a provisional examination of the data assembled in this paper to
see whether general patterns emerge, however limited these may be. It may even be possible to
use this approach to examine the idea that certain finds represent casual discards or losses.132
In order to pursue this, data for the depositional contexts of five types of artefact have been
compiled (FIGS 5, 8 and 9; Tables 1–3).

128 For the sacred character of some trees, see Ross (1967, 33–4) and Woodward (1992, 51). Ross discusses the
evidence from Irish sources for the sacred character of trees, including the ash tree. Meetings may often have been
held below the boughs of sacred trees (Ross 1967, 34). This argument does not mean that all the iron finds from the
pit need to have been ritually deposited.
129 The publication does not locate the exact context of the nails within the settlement enclosure itself and some of
these objects (although not many) may have come from these pits and post-holes.
130 Indeed, structured deposition need not indicate ritual behaviour (Hill 1995a).
131 A number of other Roman settlements have produced relatively low quantities of iron finds (Bill Manning pers.
132 But see p. 219.

FIG. 8. Contexts containing selected types of weapons. (Each deposit including iron is counted once. For key to
context types see p. 221). a. Spearheads/boltheads/arrowheads (counting each example once, n = 31).
b. Swords (counting each example once, n = 15).


Site Context Type & number of object

Baldock 4 Spearheads (33)
Bancroft 2 Boltheads or spearheads (18)
Bar Hill 1 5 Arrowheads (7)
Barton Court 4 Spearhead (1)
Bredon Hill 1 3b Spearhead (1)
Bredon Hill 4 3b Spearheads (7) Sword (1)
Bredon Hill 5 3b Spearhead (1)
Carlingwark Loch 1 Sword (8) – tip only
Fengate Power Station 1 Swords (2)
Fiskerton 1 Swords (5) Spearheads (11)
Great Holts Farm 2 6 Bolthead or arrowhead (1)
Ham Hill 3c Spearheads (‘many’)
Hayling Island 2 Spearheads
Hod Hill 2 6 Spearheads (3+)
Hod Hill 3 6 Spearheads (11)
Hod Hill 5 2? Spearhead (1)
Jordan Hill 2 Swords (2) Spearheads (2)
Maiden Castle 1 3b Swords (2) Arrowheads (4) Bolthead (1)
Maiden Castle 2 2? Swords (4) Arrowhead (1) Boltheads (2)
Maiden Castle 3 3b Arrowheads (2)
Newstead 1 4 Swords (4)
Newstead 2 4 Spearhead
Newstead 3 4 Spearhead, Arrowheads (3)
Newstead 4 4 Swords (2)
Newstead 6 6 Spearheads (6)
Newstead 7 4 Sword, Spearheads (5)
Orsett ‘Cock’ 3a Spearheads (7)
Orton Meadows 1 Swords (7) Spearhead (1)
Silchester 1 5 Sword (1)


Site Context Type & number of object

Silchester 2 4 Spearhead (1)
South Cadbury 3 3b Spearheads (37) Boltheads (21)
South Cadbury 4 6 Spearheads (11)
Spettisbury 3a Spearheads (13) Sword (1)
Stanwick, North Yorks. 1 3b Sword (1)
Uley 1 2 Spearheads (19) Boltheads (12)
Waltham Abbey 1 Sword (1)
Wavendon Gate 6 Spearhead (1)
Wookey Hole 1 Spearhead (1) Arrowheads (5)

As has already been remarked, currency bars are usually deposited in settlement boundary
contexts, although significant numbers occur in wetland and other occupation contexts (FIG.
5 and Table 1). Swords and spearheads (FIG. 8 and Table 2) are highly symbolic objects, often
hoarded and sometimes included in burials. The limited number of examples of deposits that
contain swords makes interpretation difficult, but they appear, from the data assembled in
this paper, to be more common in wetland contexts than spears and less common in enclosure
boundaries than currency bars. Spearheads/boltheads/arrowheads occur in a wide variety of
context types. Ploughshares and sickles/reaping-hooks/scythes (FIG. 9 and Table 3) are relatively
common in wetland deposits but, again, the numbers of occurrences restrict interpretation in
any detail. The occurrence of ploughshares on the temple sites at Frilford and Harlow clearly
demonstrates that such objects represented ritual items in certain contexts.
Despite the limitations, these data do indicate different proportions of finds in the various
context types. This, in turn, suggests that additional detailed analysis addressing chronological
and regional variations in practices of deposition should enable a fuller interpretation and
understanding, which may help to explain motives for the differential deposition of various
types of iron objects. Other types of object that would benefit from a similar analysis include
iron-working tools, tyres, locks and keys.133 No detailed analysis has been attempted of these

FIG. 9. Contexts containing seleted types of agricultural tools. (Each deposit including iron is counted once. For key
to context types see p. 221). a. Ploughs (counting each example once, n = 12). b. Sickles, reaping hook, billhooks
(counting each example once, n = 17).

133 For Iron Age metalworking tools, see Fell 1990.



Site Context Type of object

Aldclune 3a Digging tool
Blackburn Mill 1 Plough, Sickle
Brampton 6 Plough, Scythe
Carlingwark Loch 1 Scythe, Sickle
Dalton Parlours 4 Reaping hook
Dorchester on Thames 3c Plough
Eckford 1 Plough, Sickle
Fiskerton 1 Reaping hook
Frilford 2 Plough
Great Chesterford 4 Plough, Scythe
Harlow 2 Plough
Hod Hill 1 6 Sickle
Hod Hill 4 3c Sickle
Hod Hill 5 2? Sickle
Kingsdown 3a Ard/plough
Madmarston 3a Sickle
Newstead 5 4 Sickle
Newstead 6 5 Sickle
Silchester 1 5 Plough
Silchester 2 4 Plough
South Cadbury 1 3a Reaping hook
South Cadbury 2 3a Billhook
South Cadbury 3 3b Reaping hook
Spettisbury 3a Plough
Stanwick Villa 6 Plough
Wookey Hole 1 Billhook, Sickle

find categories, but the seven keys/latch-lifters from the entrance passageway at the hillfort of
South Cadbury (3) could not have all been intended to lock a single door. One might expect
such items to have a close association with settlement boundaries, since they would have held
metaphorically contrasting associations with security and access to locked places.


The focus of this paper has been on the general context in which deposition occurred, and a
variety of specific types of context have been defined. It is important, however, to avoid over-
simplistic interpretations and the arguments above certainly should not be taken to infer that all
iron objects were votive.134 The evidence reviewed indicates a rather variable picture and the
consistency of context may be partly an illusion created by the method by which the data has

134 See p. 217.


been collected and analysed. Further work is necessary to assess the ideas developed here, but it
is possible to make some tentative observations that may help to provide an understanding of the
changing conceptual geographies of later prehistoric and Roman populations.
The strong emphasis upon settlement boundaries during the Iron Age helps to highlight a
period in which people in some areas of Britain were particularly enthusiastic about creating
clear physical boundaries around their settlements.135 Iron deposition may well have performed
a significant role in the definition, perpetuation, or abandonment of these significant physical
boundaries. Even during this time, however, the exact context of deposition in which objects
were placed in boundaries was highly variable — in ditches, in pits on the back of ramparts, on
the backs of ramparts, in quarry hollows, in roundhouses behind ramparts, and in entrances. If
this is a single practice of deposition across southern Britain, how is the variation in the exact
context in which it occurred to be explained? Many individual finds may have been moved as a
result of post-depositional factors,136 but the evidence may also hint at the varying importance
of particular elements of boundaries and differences in activity over time. The objects that were
deposited within settlement areas, into wet contexts, and at shrines hint at similar local variations
in practice, which may all be related to the meaning of a variety of types of transitional area.
Currency bars show a particularly close association with settlement boundaries, which
highlights the idea that different classes of iron objects may have been used in different ways.
Indeed, the provisional analysis of the contexts in which different types of objects were used
suggests differing associations with settlement boundaries, although much further work is
required to pursue this topic. Although currency bars are particularly common in association
with settlement boundaries, the majority of later prehistoric deposits of iron also occur in these
contexts. A particular focus on the entrance to enclosed settlements exists in the first century
B.C. / first century A.D., which is only partly explained by the ‘massacre deposits’. The evidence
for iron deposits of the second to fourth century A.D. shows rather less of a focus on settlement
boundaries, although a higher proportion of finds (86 per cent, or 31 out of 37 deposits) occur
within the areas of settlements, in contrast with the fifth to first century B.C. deposits (75 per
cent). The apparent reduction of the focusing of iron deposition on the settlement boundary in
this later phase may be exaggerated by the fact that some deposits are placed in wells and deep
pits which lie within settlement boundary zones, and by the observation that many settlements
from the late first century B.C. to the end of the Roman period possessed more complex systems
of boundaries than their later prehistoric predecessors.137 After the first century B.C. there may
have been less of a fixation on creating clear physical boundaries to the sites, although enclosed
settlements do occur and many of the towns of the province received walled circuits during the
later Roman period. Only one collection of ironwork (Dorchester on Thames) comes from a
context close to the ramparts of a town.
The main focus of deposition for significant collections of iron during the Roman period
involves wells and deep pits, with a secondary focus on deposition within the settlement, but
often in significant locations. The iron objects from wells and deep pits are matched by the
deposition of many other significant objects in such contexts. Although there appears, from the
evidence reviewed in this paper, to be a dramatic change in the contexts used for the deposition of
iron objects during the course of the period discussed, the survey of individual finds at Danebury,
Wavendon Gate, and Baldock suggests that many iron objects continued to be deposited for
ritual reasons throughout the period.
Thinking in more general terms about the relationship between the location in physical

135 Hingley 1990a; Hill 1995b.

136 As with the currency bar from the top of the enclosure ditch at Park Farm (see p. 224).
137 Hingley 1989, 55–9.

space and the objects that were deposited, accounts of the hoarding of material often state that
weapons and tools were placed in a variety of locations, such as wetlands, boundaries and wells,
which had special significance. It may well be erroneous, however, to assume such a static view
of landscape and action.138 Instead, it may have been a combination of object and place that
determined the power of the action. Objects may have been placed in particular locations, which
then took on new (or renewed) ritual and symbolic associations that drew upon the significance
of the deposited objects. The ingredients of the act may, sometimes, have been as important
as the particular context in which they were deposited. In fact, the locations where iron and
other significant objects were placed were not necessarily in themselves particularly special
(or necessarily sacred). The presence of water may have been significant for a whole variety
of contexts in which metal is found. This is relevant, not only for rivers, marshes, and lakes,
but also for the enclosure ditches of settlements that held water (for example, at Stanwick).
The watery association is also evident for the Roman wells and deep pits. Deposition may also
often have related to beliefs about particular moments in time: the establishing or the disuse and
abandonment of the structure.
The important issue when studying the deposition of iron, as with other classes of finds, is to
examine each deposit in order to attempt to address its own particular meaning. Each discovery
represents a unique act of deposition and should be studied and interpreted accordingly. Broad
patterns in the evidence may enable some conclusions to be drawn about the general contexts
of deposition and, perhaps, about aspects of ritual and religion in general,139 but variation
constitutes a vital part of the comprehension of the information. Only through careful excavation
and detailed publication can such associations be determined. The significance of the evidence
from Britain is beginning to be appreciated, as is demonstrated by some of the studies reviewed
in this paper, but further more detailed analysis of deposits of iron in Gaul, Germany and
elsewhere would also help to set the British material in context, particularly with regard to the
objects derived from military sites occupied by soldiers from overseas.140

Department of Archaeology, Durham University

This paper is published with the aid of a grant from The Council for British Archaeology

138 Clarke and Jones 1996, 120.

139 Insoll 2004, 12.
140 See van Enckevort and Williams 1996 and Bonnamour and Dumont 1996 for two relevant studies of the context
of the deposition of Roman military items. Currency bar hoards from Montans, France (Martin and Ruffat 1998)
and the cave at Le Trou de l’Ambre, Belgium (Mariën 1970) also indicate the potential of the continental material
(Hingley 2005, 184).


This material represents collections of iron objects and some single items. All have reliable contextual
information. Many have been picked out by the excavators, publishers, or other researchers as having a
significant context of deposition. Some form parts of larger collections of objects of a variety of materials.

Under ‘Content’, non-iron finds are listed as follows:

AB = animal bone Ca = cauldron Po = pottery
HB = human bone Co = coin Q = quern
B = bronze object P = pewter vessel W = wooden object

Under ‘Context’, the numbers refer to the following types:

1 = ‘natural’ = cave, wetland, river
2 = ‘shrine’ = temple or shrine or possible shrine
3 = ‘enclosure’ = in, or close to, the ditch or bank of an enclosed settlement
3a = the ditch or bank of the enclosure
3b = entrance area to the enclosure
3c = close proximity to enclosure earthworks (within 10m)
4 = wells or deep pits (over 2m deep) dug into, or very close to (within 10m), the boundary of a
settlement or site
5 = wells or deep pits (over 2m deep) dug within the area of a settlement or site
6 = other contexts within the area of a settlement or industrial site
7 = other features on the edges of a settlement or in the landscape

Site Context Content Dating Reference Con-

Aldclune, Perth Ditch around Digging tool C1–C2 Hingley et al. 3a
& Kinross ‘homestead’ (large A.D.? 1997, 437, 455
enclosed roundhouse)
Asthall, Oxon. Hearth associated with Billet Early to Booth 1997, 50 6
iron production mid-C2
Baldock, Site A, Pit, particularly Pit 33 spearheads (6 others very C3 A.D. Manning and 4
Herts. A13, close to enclosure close by & 5 fairly close) Scott 1986;
boundary. Others from Stead and
pits close by Rigby 1986.
Bancroft, Bucks. Mausoleum and shrine 18 spearheads/boltheads, Mid to Williams and 2
10 from metal detecting late C4 Zeepvat 1994,
immediately above circular A.D. 107, 339
shrine, 2 and an iron ferrule
from central pit of building
and 6 from late Roman ditches
nearby. Many fragmentary,
some possibly votive. Co, etc.
Bar Hill, Well, in open court of Bag full of what looked c. 140s Macdonald 5
Dumbarts. 1 praetorium of fort. Top like tools (highly corroded), A.D. 1906
12ft full of building including nails and possibly a
material and other pair of pliers. Additional iron
finds, including 2 altars. objects in fill of well: nails,
Collection of iron in a door latch, bridle-bit, ring,
bag inside large part of compasses, 7 arrowheads,
a complete but broken punch AB, Po, W

Site Context Content Dating Reference Con-

Bar Hill, Pit 1, located just to Complete chariot wheel with c. 140s Macdonald 3c
Dumbarts. 2 the W of the S entrance iron tyre. Wheel bushed inside A.D. 1906, 496–9
within the fort with decorated iron disc. Placed
at base of pit, with 3 oak stakes
driven into base of pit, one
of which passed between two
spokes of the wheel. W

Barton Court, Well 832, to E of ‘villa’, Lower deposit which built up C4–C5 Miles 1986, 15 4
Oxon. a couple of metres during use included biological A.D.?
inside outer boundary material, iron-bound wooden
ditch to site. Within fill bucket, iron hook, latch lifters,
under upper dump of spearhead and shoes with nails.
stone, animal bone, and Po
waterlogged material

Bearwood, Ditch of enclosed 4 currency bars and another c. 150–50 Jarvis 1985 3a
Dorset hillslope settlement site iron object B.C.?
– at bottom
Billingborough, Found in a shallow Blacksmith’s poker, broken in C3–C1 Chowne et al. 7
Lincs. recut or pit in top of half and parts deposited side- B.C. 2001, 23–4, 95
Bronze Age enclosure by-side
ditch. Suggestion of
digging of a pit in a still
recognisable boundary
for a formal votive
deposit. This was just
NE of a small Iron Age

Blackburn Mill, Found before 1852 by 65 pieces: cauldron rim, patera C 1–C2 Piggott 1955, 1
Borders labourers cutting a drain, without handle, ingot or cake; A.D.? 4–5, 40–50
lying on blue clay under iron: tyre frag., linch-pin,
peat. Probably deposited terrett, cap or linch-pin, link
in a lake. In 2 cauldrons, or bit x 2, key, handle x 4, tub
one inverted on other handles x 2, cauldron chain,
chain-junction, knife rough-out,
hipposandal x 2, rod x 2, loop,
eyed hook, lampstand, staple,
hinge-staple, shears, shield
boss?, ploughshare, knife, peat
spade, sickle, bar, pick, field-
anvil, adze-blade, socketed
gouge x 2, hinge frag., spike,
mounting, chisel, punch x3,
bar x 2, disc and a few other
objects. Ca

Site Context Content Dating Reference Con-

Brampton, Pit on tile-making site, Iron plough-share, hoe, rake Early C2 Manning 1965; 6
Cumbria 2ft diameter x 3ft deep prong, scythe x 2, axe, adze, A.D.? Hogg 1966,
(probably originally 10ft morticing chisel, chisel x 2, fig. 2 for a plan
deep). May have been spoon auger, large auger, L- of site
to provide water for tile- ended tool, anvil-spike?, tool
making in surrounding handle, tanged hook, chain x
tile works. Pit just to W 2, hub-lining, hub-rim x 2, tyre
of top of hill, with tile frags x 2, strip, mounting for
kilns around, but mostly bucket handle x 6, L-shaped
to W binding, couch leg x 2, drop
hinge, staple, wall hook x
2, spiked collar x 4, spring
padlock case, T-shaped tumbler
lock key, holdfast x 3, various
other fragmentary objects
Bredon Hill, Hillfort ditch, N terminal Large and ornate (‘flamboyant’) C3–C1 Hencken 1939, 3b
Worcs. 1 to Phase 1 entrance in iron spearhead B.C.? 13, 42, 75
inner rampart. Top fill.
The butt end of this ditch
appears to have been
backfilled before later IA
modifications to entrance
Bredon Hill, On Site A, just N of Almost complete iron horse-bit C3 Hencken 1939, 3c
Worcs. 2 inner entrance at rear B.C.–C1 32, 71
of rampart, on original A.D.
ground surface, where
scraped to provide bank
Bredon Hill, On Site I, toward N end Deposits included iron pin, C3 Hencken 1939, 3c
Worcs. 3 of E rampart, within clamps and harness hammers B.C.–C1 30
possible building at back from horse-bit. Miniature stone A.D.
of rampart axe, B, HB
Bredon Hill, On final stone floor to 7 spear/javelin heads, an iron Early C1 Hencken 1939, 3b
Worcs. 4 entrance passageway in sword, scabbard, 2 small iron A.D.? 21–4, 76–84
inner rampart of hillfort hammers and a knife. Other
(entrance to S of fort). objects recorded from ‘Final
Found among human phase Inner entrance’ include a
bones scattered in piles sickle. B, HB
in front of position of
gate and for 4m down
corridor outside gate
Bredon Hill, North-east entrance to 1 spearhead. Excavator also C3 Hencken 1939, 3b
Worcs. 5 outer rampart of hillfort records ‘a number of iron rods B.C.–C1 63, 71, 76
in a highly corroded condition’ A.D.
in north-west entrance
Bredon Hill, From a feature between Fragments of two currency bars C3–C1 Allen 1967, 3a
Worcs. 6 the inner and outer B.C. 331
ramparts. Found in 1959
Burrington From cave. Fairly well Shackles, key, possible tankard C3–C1 Palmer 1921; 1
Coombe, recorded finds scattered handle, clamp, spike, hook, B.C.? Field and
‘Keltic Cavern’, around interior frag. of adze or axe, frag. of Parker Pearson
Somerset sickle, nails x 4, bar and other 2003, 188
objects. B, HB

Site Context Content Dating Reference Con-

Cannards Grave, In trackway (Context 5 objects derived from a cart C4 A.D. Birbeck 2002, 6
Somerset 292), part of a settlement (U-shaped binding, L-shaped 64
hook, 2 guides or mounts)
Carlingwark Found in 1866 by two Mirror handle, two-link bit x C1–C2 Piggott 1955, 1
Loch, Dumfries fisherman near the Fir 2, tyre frag., linch-pin, cleat or A.D. 3–4, 28–40
& Galloway Island in the loch. No staple, chain-junction, handle-
further details available, loop x 2, padlock spring,
but objects in a large latch-lifter, saw, scythe-blade
bronze cauldron which frag. x 2, knife-blade x 2,
may have been dragged sword blade tip x 8, ferrule
up from the bottom of x 2, loop or bucket, sickle,
the loch circular mounting, tanged
blade x 2, hammer-head x 8,
adze-hammer, axe-head, staple
x 3, hook x 5, draw-knife, loop
x 3, punch x 4, file x 2, bar,
gridiron, tripod, chain mail and
some other frags. B, Ca, W
Dalton Parlours, Well, to SE of Building Toward base (deposit L), Late Wrathmell and 4
West Yorks. M on the E edge of remains of several iron- C3–C4 Nicolson 1990
the settlement. There bound wooden buckets, A.D.
does not appear to be sledgehammer, masonry pick,
a ditched boundary in spade sheath, reaping hook,
front of the building/well prong, ox goad x 2, spatula,
knife x 3, teathering ring,
L-shaped lift-key x 2, ring x 5,
strip, binding frag. x 2, bar x 2.
AB, HB B, Po, W
Danebury, In pit within house close 21 complete currency bars C4–C1 Cunliffe 1984, 3c
Hants. 1 to rampart. Sectioned in B.C. 170, 357;
1969 trench (CP 7) Cunliffe and
Poole 1991b,
Danebury, On floor within house Cauldron hook x 2, latch lifter, C4–C1 Cunliffe and 3c
Hants. 2 close to rampart rod handle, punch?, CB frag., B.C. Poole 1991b,
bar x 2 (CP 7) 354
Danebury, In post-hole within Collection of cart fittings C4–C1 Cunliffe and 3c
Hants. 3 house close to rampart B.C. Poole 1991b,
(CP 7) 354
Danebury, In shallow hollow within 2 iron linch pins. Not C4–C1 Cunliffe 1995, 3c
Hants. 4 house close to rampart necessarily a hoard B.C. 86
Ditches, North At north-eastern entrance Hoard of currency bars, Mid-C1 Trow 1988, 33, 3b
Cerney, Glos. to outer enclosure, at possibly around 10. Possibly A.D.? 37, 40
base of ditch just S of deposited in a bundle
entrance. At this point
there was a 0.25m
deepening in base of
ditch. Objects were
possibly hidden under
limestone slabs

Site Context Content Dating Reference Con-

Dorchester on On road just behind Plough-coulter, crowbar- C4 A.D. Frere 1964, 3c
Thames, Oxon. rampart to south of town like implement (possibly 119, pl. 7a
(Site A). Either buried in ploughshare) and other objects
silt on road (p. 119) or
‘lying on street’ (caption
to pl. VII)
Eckford, Found in 1883 digging in Iron hanging-lamp or C1–C2 Piggott 1955, 3, 1
Borders a small patch of ground lampstand, gridiron, chariot A.D. 20–8
which may have been a or cart pole, linch-pin,
dried up loch fragmentary tyre, rod, pot-
hook x 2, hook, ploughshare,
socketed reaping knife or
sickle, door hinge, small pick
x 2, adze-hammer, mattock,
farrier’s buttress, cold chisel,
field anvil. B
Fengate Power Deposited into lake or Sword x 2, hook, ring x 3, rod, C4 B.C.– Pryor 2001, 1
Station, Cambs. marsh along line of post cylinder, strip, looped socketed C1 A.D. 298–300
alignment, mostly on or axe, nail and other objects of
to S of line. Alignment IA date
itself constructed c.
1300–900 B.C. and in use
for long period to deposit
Fiskerton, Lincs. Substantial IA and 5 swords (& bronze sword C3 Field and 1
Roman votive deposit fittings), 11 spearheads, 2 B.C.–C1 Parker Pearson
in a river, associated hammers, 6 metalworking files, A.D. 2003
with a timber causeway. 3 axe heads, 2 woodworking
Possible that IA artefacts files, a possible bench anvil,
deposited in one go punch, tool, an iron pull saw, a
as a ‘closing’ deposit linch-pin, reaping hook, cleaver
marking the end of the ?, bar, stud, nail and various
causeways initial phase frags; Roman ironwork: spoon,
of use (p. 175), but there bar, nail, rod x 2, linch-pin. AB,
are also some Roman HB, B, W
Frilford, Oxon. At base of post-hole in Plough share C3 Bradford and 2
‘ritual’ structure B.C.–C1 Goodchild
A.D. 1939, 13
Gorhambury, Against S wall of Blade, hooked mount – cart Early C3 Neal et al. 6
Herts. Building 40. Other iron fitting?, steelyard, lifting key A.D.? 1990, 63,
objects from Building 138–52
39/40, including axe
from post-hole (find 363)
and key from post-hole
(592). Other iron objects,
including over yard N of
Building 39/40

Site Context Content Dating Reference Con-

Great Pit capped with chalk. Anvil, 5 ‘anvil pegs’, 2 axle C3–C4 Neville 1856; 4
Chesterford, 6ft deep, close to other or pole guards, 2 chains, 5 A.D. Brinson 1963
Essex pits and ditches. Pit coulters or ploughs, 10 felloe
with ironwork was bands, 7 hammers, 4 hoops,
from outside town 4 holdfasts, 7 hinges, 3 keys,
walls and adjoining the 4 locks, 1 pivot for millstone,
churchyard. Many pits 1 pail handle, 2 pail hoops, 1
and shafts on this site long pair of shares, 8 shackles,
(Brinson 1963). Neville’s 1 saw, 12 scythes, 1 square
writings suggest that he girder, 1 turf cutter, 2 wall pegs,
found as many outside 1 small wheel, 5 iron bars
ramparts as in
Great Holts In Oven (320), close 21 hobnails, suggesting that C1–mid- Germany 2003; 3c
Farm, Essex 1 to boundary ditch of a shoe or shoes had been C3 A.D. Major 2003, 77
enclosure E 1 (to side ‘discarded’ in it
of developing villa
Great Holts From foundation trench Small socketed bolthead or Late Germany 2003, 6
Farm, Essex 2 (575) of bath-house arrowhead C3–mid- 49; Major
414. This was the C4 A.D. 2003, 83
construction trench for
the praefurnium, but the
area had been robbed
Great Holts From well (567), within Remains of at least 2 leather A.D. 220. Germany 2003, 5
Farm, Essex 3 Building 416. Only top shoes, 1 with hobnails Slightly 20, 191
of feature excavated later?
Great Holts From post-hole forming Scrap sheet x 26, prick iron, Mid-C3– Major 2003, 77 6
Farm, Essex 4 part of wall of Barn 417 steelyard, possible weight, lateC4
(6162) knife A.D.
Gretton, From a small pit cut into Approximately 48 currency C3–C1 Jackson 1974 7
Northants. the edge of a pit (39) bars packed close together B.C.?
which forms part of a pit
alignment. The small pit
was probably dug into
the edge of a shallow
depression left after pit
39 had been largely filled
Gussage all Pit inside entrance area Variety of scraps and C3–C1 Spratling 1979 3b
Saints, Dorset of enclosure. Some fragments, including heavy B.C.
iron objects from cutting implement, punches,
neighbouring pit (438). rods, nails. Also some evidence
Close to other features for bronze working
with metalworking
deposits, including pit
with much bronze-
working (209)

Site Context Content Dating Reference Con-


Haddon, Small pit, then cut by Iron shears, chain segments and Early C4 Hinman 2003 6
Peterborough, a malting oven. Report Co (bronze follis of Diocletian, A.D.
Cambs. claims that the pit was A.D. 298–9)
integral to the location
of the oven and calls
deposits a ‘foundation
deposit’ (pp. 52, 112).
The flue cut the top of
the pit fill. About 5m
away was an earlier
cistern, abandoned c.
A.D. 200

Ham Hill, Probably just behind Arrowheads/spearheads and C1 Colt Hoare 3c

Somerset rampart. Spot shown cart wheel tyre B, HB. Colt B.C.–C1 1827. See
on Colt Hoare’s pl. 7 as Hoare argues that a battle A.D.? Smith 1991
‘Roman remains found’. occurred here but a difficult for a recent
On W side of hillfort, place to use chariots? account of the
found a few years before hillfort
1827 during quarrying.
In a ‘chink’ or ‘gulley’ in
the rock.
Harlow, Essex On site of a Roman Iron tools, including a plough C1 Bartlett 1988 2
temple, close to a round share and large quantities of B.C.–C1
building, associated with iron strips. B, Co A.D.
various ‘votive’ deposits

Hayling Island, Mainly around E 2 currency bars, spearheads. B, C1 King and Soffe 2
Hants. enclosure boundary to S Co, HB B.C.–C1 1998
of entrance A.D.
Hinchingbrooke Boundary ditch, possibly 2 currency bars placed side by 100–50 Cambridge 3a
Park Road, on edge of settlement, side in ditch B.C. Archaeological
Huntingdon, replacing two earlier Field Unit 1997
Cambs. phases of boundary.
Earlier phases of
ditch include various
significant deposits

Hod Hill, From floor of hut 43. 2 iron sickles C3 Richmond 6

Dorset, 1 This is toward the SE B.C.–C1 1968, 21, 39
area of interior of hillfort A.D.
Hod Hill, Hut 36A, within Outside door, iron catapult C1 Richmond 6
Dorset, 2 rectangular enclosure head, inside iron spearhead B.C.–C1 1968, 22, 39,
toward SE area of in scoop on SW side of A.D. fig. 14
hillfort interior entrance, second spearhead
in interior together with iron
frag. – currency bar? 3 Roman
ballista bolts found in and
around this building and more
to SE around Hut 37
Hod Hill, Hut 37, just SE of Around building, 11 catapult C1 A.D.? Richmond 6
Dorset, 3 rectangular enclosure heads. Outside iron ferrule 1968, 39, fig.
including Hut 36A 14

Site Context Content Dating Reference Con-


Hod Hill, Hut 60, at inner edge of Currency bar and socketed C3 Richmond 3c
Dorset, 4 quarry pits on SE of fort knife from bottom of wall; B.C.–C1 1968, 19–20,
close by l fragment of currency A.D. 40, fig. 12
bar, clamp and sickle blade;
knife from pit just outside
entrance to S (post-Roman?). B
Hod Hill, Pits 15 A–C, in SE of Pit 15A contained iron objects, C3 Richmond 2?
Dorset, 5 hillfort interior, within including strip, spade binding?, B.C.–C1 1968, 26–8,
Enclosure 15. Enclosure two points, blade, uncertain A.D. 40–1
is a scooped area with object; also AB & B. Pit 15B
ditch around and E contained iron bucket handle,
entrance – unknown bucket hoop, latch-lifter; also
function, but unusual AB, HB. Pit 15C contained
range of finds (shrine?) massive spearhead, socket
and B, AB. Near these pits
were catapult head, sickle and
socketed frag.

Houghton Pit (P 340) from interior Hooked billet, deposited near C5–C4 Cunliffe and 6
Down, of site base of pit, with small fragment B.C. Poole 2000,
Stockbridge, of forging waste 107, 124
Hants. 1
Houghton Pit (P 331) from interior Incomplete iron adze or chisel, C5–C1 Cunliffe and 6
Down, of site, associated with together with incomplete saw, B.C. Poole 2000, 111
Stockbridge, large body of ‘slag’ at knife and tapering frag. of rod
Hants. 2 base of pit (p. 124),
but this may be just the
result of a fire
Inchtuthil, In square pit in fabrica. Nearly 10 tons of nail (at c. A.D. Pitts and St 6
Tayside In small room to S of least 875,428 examples) and 9 83–86 Joseph 1985,
W entrance. Measured wheel-tyres. Some nails used. 109–13,
3.07 by 2.74m and Tyres used, not all complete 289–92
3.66m deep. Sealed
by 1.83m of gravel ‘to
prevent later recovery’
(p. 109). Evidence for
ironworking in same
Jordan Hill, Shaft/pit in small Fill of pit consisted of 16 layers Late C4 RCHM 1970, 2
Preston, Dorset rectangular temple of ash and charcoal alternating A.D. 616–17
building. Building with double layers of roofing
set within rectangular slabs arranged in superimposed
enclosure filled with pairs, between each of which
animal bones and other were bones of a single bird and
evidence of feasting. a small bronze coin. At bottom,
Shaft oblong, 4ft x 3ft, a small stone cist containing
12ft deep. Sides lined sword, spearhead, knife,
with roofing slabs set steelyard, bucket handle, crook
in clay and other iron objects. Similar
cist half-way down which
enclosed sword, spearhead. AB,
Co, P

Site Context Content Dating Reference Con-

Kilverstone Pit (F 221) just W of Against NE side of pit – long- C3–C4 Garrow et al. 7
Thetford, small circular smithy handled tongs, iron hammer- A.D. forthcoming
Norfolk 1 buildings. Pit was cut head, anvil frag., with knife,
into a sequence of earlier frag. of padlock bolt, nails.
boundary ditch of Phase Covered by burnt deposit and
3 – earlier Roman. These scorched oak planks. P
may well have defined
the edge of the settlement
in earlier phases. Other
pits nearby (F 169, F
962, F 968) with iron
objects, also extensive
metalworking dumps
Kilverstone, Pit in ‘smithy’ appears Iron axe. Some slag in the C3–C4 Garrow et al. 7
Thetford, to be built into earlier ditch. B A.D. forthcoming
Norfolk 2 (discussed?) Roman
enclosure, facing toward
pits. This enclosure
neatly enclosed an
earlier substantial 4-post
structure which produced
IA pottery from the
Kingsdown In inner enclosure 2 currency bars, iron ‘spud’ C3–C1 St George Gray 3a
Camp, Somerset ditch to small enclosed (ard?), horse-bit. B, HB, Co B.C.? 1930
settlement. Generally
scattered, but apparently
particular concentration
on E, to S of NE entrance
Lesser Garth, Found removing topsoil Iron bridle-bit, linch-pin, chisel, C1 Savory 1966 7
Pentyrch, from quarry. On gently billet (tapering), cauldron ring, B.C.–C1
Glamorgan sloping hill above steep staple, cauldron hanger and A.D.
slope 200 yds from cave cauldron chain. B
opening. Exact context
Madmarston, Laid on clay which 12 currency bars, a square- C3–C1 Fowler 1961, 3a
Oxon. forms back of rampart sectioned bar, axe head, sickle, B.C. 11, 13, 41
at S of hillfort (Cutting ‘poker’, 2 pairs of bridle
8H). Area then sealed by bits. On top of stone layer
a stone layer associated – socketed chisel and flanged
with post-holes, with 2 plate
additional iron objects
on its edge
Maiden Castle, In hillfort E entrance. 2 swords or daggers frags, 1 C1 A.D. Wheeler 1943, 3b
Dorset 1 Associated with battle by bolt head, blade, 4 arrowheads 62–3, 277, 281
excavator (additional example from
outworks), linch pin
Maiden Castle, Site L, within hillfort 4 swords or daggers frags & 2 C1 A.D. Wheeler 1943, 2?
Dorset 2 on and around E end of bolt heads, an arrowhead. B 277–8 and 281
Neo/BA long mound
and close to earlier IA
occupation and later
Roman temple and ritual
structures (shrine?)

Site Context Content Dating Reference Con-

Maiden Castle, ‘In the northern portal Fragments of ‘several’ iron C3–C4 Wheeler 1943, 3b
Dorset 3 of the eastern entrance’. horseshoes found. 14 are A.D. 120–1, 290–1
A gate and metalled illustrated. Co.
road had been built into
this portal and on this
surface were found iron
boot-nails , ‘several’
horseshoes and 2 coins.
Stone base just inside
entrance for an altar?
Nadbury, From small pit (F5) cut Currency bar. AB C3–C1 McArthur 3a
Warwicks. into back of rampart on B.C.? 1990, 8
N boundary of hillfort.
By the time this was cut,
the rampart had slumped
and may have been
Newstead, Pit (57), in W annexe, 4 complete or frag. swords Late Curle 1911, 97, 4
Borders 1 sealed by cobble (1 bent double), strigil, C1–late 128
foundations surrounding hipposandal, hub rims x 5, C2 A.D.
bath-house (possibly a lamp. B, HB, Po
Newstead, Pit (54), dug in area of S Stylus, spearhead, key, knife, Late Curle 1911, 127 4
Borders 2 defences of fort hook x 2. AB, B, Po, Q C1–late
C2 A.D.
Newstead, Pit (55), dug in areas of Spearhead with broken point, Late Curle 1911, 128 4
Borders 3 S defences of fort knife, arrowhead x 3, socket. C1–late
B, Po. C2 A.D.
Newstead, Pit (58), dug in area 2 swords (1 bent in half), linch Late Curle 1911, 4
Borders 4 of N defences of fort. pin, ingot. AB, B, Co, Po C1–late 129–30, 158,
Suggested that this was C2 A.D. pl. LXV
early, predating main
phase of fort
Newstead, Pit (22), in S annexe just Helmets and frags. x 3, sickle, Late Curle 1911, 4
Borders 5 to S of defences armlet, bridle bit; armour. AB, C1–late 121–2
B, Po, Q C2 A.D.
Newstead, Pit (1), in principia Bar, knife, linch pin, sickle, Late Curle 1911, 5
Borders 6 (central court) armour, rim of bucket, C1–late 116–17
arrowheads x 5, shield umbo, C2 A.D.
holdfast. AB, HB, Co, P, Q
Newstead, Pit (16), in S annexe just Sword, rib of shield, spearhead Late Curle 1911, 4
Borders 7 S of defences x 5, axes x 5, stirrup, shod, C1–late 119–20
hammers x 5, tongs x 2, ‘drift’, C2 A.D.
anvil, staple mandrils x 3,
chisel x 2, gouges x 2, mower’s
anvil, scythe x 4, chain, door
fittings ?, harness frag., linch
pin, nave bands for wheel x 24,
hub lining x 3, pieces x 20. AB,
HB, B, Po, Q, W

Site Context Content Dating Reference Con-

Old Down Pit (2420) in the internal Socketed gouge, linch pin x 2, C3–C1 Davies 1981, 6
Farm, Andover, W area of enclosure. frag. Sickle blade, 1 complete B.C. 124
Hants. Covered with layer of and 1 frag. currency bar. B
loom weights which
might suggest deliberate
Orsett ‘Cock’, Fill of inner enclosure 7 spearheads (no further C1 A.D. Carter 1998, 3a
Essex ditch at SE corner, with artefacts recorded) 83, 168
one nearby in middle
Orton Meadows, From old bed of the 7 currency bars, 2 currency bar C3–C1 Stead 1984; 1
Northants. River Nene. May be part frags, 7 swords, 1 spearhead, B.C. Field and
of a much larger deposit latch lifter and ‘ladle’ Parker Pearson
made in a fast-flowing 2003, 185–6
river channel. Object
possibly deposited from
top of a pre-existing
Park Farm, From top fill of Currency bar C3–C1 Cracknell and 3a
Barford, enclosure ditch on SW of B.C. Hingley 1994,
Warwicks. enclosure 19–20
Silchester, 1890 hoard. Pit N. Pit On top of pit, iron sword and 2 C3–C4 Fox and Hope 5
Hants. 1 roughly walled in flint iron bars. In pit, axe, hammers, A.D. 1891; Evans
and toward the middle of gouges, chisels, plough coulter 1894, 742
an insula x 2, anvil, tongs, files, rasp,
hippo-sandal, lamp, gridiron,
carpenter’s plane. Immediately
under, short cist of pieces of
chalk, 2 complete pots. B.
Silchester, 1990 hoard. Well 2, 2 striking hammers, 10 small C3–C4 Reid 1901, 4
Hants. 2 which is on E boundary hammers, 2 pairs tongs, 2 A.D. 246–9
of Insula XXIII and states, 1 drift, 1 chisel, 1 hand
may overlie earlier (IA) wringer or hand leaver, 2
enclosure ditch. Well pairs dividers or compasses, 2
21ft 7in deep, lined with instruments for making nails, 4
12ft of square timber iron bars, axehead, 3 socketed
framing. Upper part, for chisels, 1 adze, 1 centre bit,
7ft downward ‘blocked’ anvil or shoemakers hobbling-
with ironwork. Rested foot, 3 plough coulters, 1
on deposit of black ash coulter, 2 forks (?) and 8
which filled the rest mower’s anvils. Also knives,
choppers, bucket handles, 2
files, 2 saws, 1 spearhead, huge
padlock and frag of another.
B, Po
South Cadbury, Pit D630A, behind main Upper two thirds of a currency C3–C1 Barrett et al. 3a
Somerset 1 rampart and sealed by bar, axe, saw, 4 knives (some B.C. 2000, 62, 82–3
late phase incomplete), saw frag., adze,
4 reaping hooks, 3 awls. Some
iron objects appear to have
been wrapped in straw and
there was burnt material in pit.
B, W

Site Context Content Dating Reference Con-

South Cadbury, Pit D521A. Cut into Billhook wrapped in straw C3–C1 Barrett et al. 3a
Somerset 2 back of rampart, quite B.C. 2000, 59
close to above pit
South Cadbury, Entrance passage at SW Scatter of objects in C1 Barrett et al. 3b
Somerset 3 of enclosure. Dispersed passageway and W guard B.C.–C1 2000, 114–16,
deposit built up over chamber. Included distinct A.D. 123–32, 239
time hoard of iron 7 latch lifters
and 3 keys. Also deposit of HB
and other objects, including 37
spearheards, 21 bolt heads, 1
reaping hook, armour, etc.
South Cadbury, E of interior of hillfort in Fragments of weapons C3–C1 Barrett et al. 6
Somerset 4 occupation layers (including 11 spearheads or B.C. 2000, 239, 295,
frags). Associated with metal 298–301
production. B, Ca
Spettisbury, ‘Burial pit’ cut into Sword and possibly 4 currency C1 Gresham 1940 3a
Dorset enclosure ditch, on E bars, 9 spearheads, javelin B.C.–C1
of hillfort. Pit about head, knives, shears, key and A.D.?
35ft long, by 15ft wide plough share. AB, HB, B. It is
and from 4 to 9 or 10ft not clear if all these finds came
deep. Possibly a second from the ‘burial pit’
comparable pit found
along railway line in
1858 (which follows
ditch), but this was not
Stanway, Enclosure ditch on 2 currency bars placed side by C1 B.C. Crummy 3a
Colchester, SE side. Settlement side on inner face of ditch 1997a, 1997b
Essex enclosure and pers.
Stanwick Villa, In stone-capped drain 2 unequal-armed plough shares C4 A.D. Neal 1989, 165 6
Northants. within subrectangular and large wheel hub from cart
building on N of interior.
In a pit slightly further
N a hoard of 4 pewter
vessels and offcuts of
leather shoes
Stanwick, North Ditch of Phase II Sword in wood scabbard with C1 A.D. Wheeler 1954, 3b
Yorks. 1 enclosure, in N terminal bronze fittings. HB, W. 44, 52, 53
of entrance. Ditch at
terminal deliberately
deepened and held water
Stanwick, North Gully 1, shallow ditch Largely complete shears C1 Wheeler 1954, 3a
Yorks. 2 surrounding roundhouse. B.C.–C1 50; Haselgrove
Fairly close to ditch that A.D.? et al. 1990, 69
supposedly defines early
enclosure – part of a
complex of settlement
features shown by

Site Context Content Dating Reference Con-

Totterdown Scoop (small feature) Hoard of currency bars C3 Pine and 3c
Lane, Horcott, close to later enclosure (impossible to work out how B.C.–C1 Preston 2004,
Glos. boundary. Boundary many, possibly around 10) A.D. (or 45, 69–70
probably on line of later?)
earlier enclosure
Uley, Glos. 1 Mainly from ditch Collection of spearheads (12) C1 A.D. Woodward and 2
(F 264) which partly and catapult bolt heads (19). Leach 1993,
encloses Building XVI Includes military type objects 131–5
(to its SE) and pit (251)
dug into this ditch
Uley, Glos. 2 Demolition rubble over Hobnails – offerings of shoes to C2–C4 Woodward and 2
Structure XIV and bank Mercury? A.D. Leach 1993,
material of Structure 184
Waltham Abbey, Found gravel digging, 11 blacksmith’s tools (5 pairs of C1 Manning 1977 1
Essex but close to river, tongs, 3 anvils, possible sledge B.C.–C1
probably deposited in a hammer, file and poker), 6/7 A.D.
bag or box dropped into carpentry tools, frag. of a cart
shallow water at river tyre, broken socketed hook,
edge? billhook and a sword. Many
pieces damaged
Wavendon Gate, In Pit 835, cut into partly Spearhead with remains of Mid-C3 Williams et 6
Bucks. silted up hollow 900, wooden shaft and three pronged A.D. al. 1996, 65;
within enclosure on fork. Also 3 small iron objects Hylton 1996,
settlement site and 18 nails (37% of those 120–8
from whole site), some from 3
hobnailed shoes. W
Wookey Hole, Within cave. Little 5 spears, a dagger, a dagger C3 Balach 1911; 1
Somerset detailed info. handle, arrowheads, billhook, B.C.–C1 Field and
sickle, knife, saw x 2, drill/awl A.D. Parker Pearson
x 6, linch pin, currency bar x 2003, 188
3 and various other iron finds.
AB, HB, P, B, etc.
Worthy Down, From the top filling of a 13 currency bars. AB, HB, P C3–C1 Hooley 1921, 7
Hants. pit within an unenclosed, B.C. 1931
or open settlement


Dioscorides, The Greek Herbal of Dioscorides. Translated by R.T. Gunter, 1934, Oxford
Herodian, Herodian of Antioch’s History of the Roman Empire from the Death of Marcus Aurelius to the
Accession of Gordian III. Translated by E.C. Echols, 1961, Berkeley
Pliny, Natural History VIII. Translated by W.H.S. Jones, 1963, London

Aldhouse-Green, M. 2002: ‘Any old iron! Symbolism and ironworking in Iron Age Europe’, in M.
Aldhouse-Green and P. Webster (eds), Artefacts and Archaeology: Aspects of the Celtic and Roman
World, Cardiff, 8–19
Allen, D. 1967: ‘Iron currency bars in Britain’, Proceedings Prehistoric Society 33, 307–35
Balch, H.E. 1911: ‘A late Celtic and Romano-British cave-dwelling at Wookey-Hole, near Wells, Somerset’,
Archaeologia 62, 565–92
Barrett, J.C., Freeman, P.W.M., and Woodward, A. 2000: Cadbury Castle Somerset, the Later Prehistoric
and Early Historic Archaeology, London
Bartlett, R. 1988: ‘The Harlow Celtic temple: a Celtic shrine underlies the famous Roman temple’, Current
Archaeology 10, 163–6
Beagon, M. 1992: Roman Nature: the Thoughts of Pliny the Elder, Oxford
Birbeck, V. 2002: ‘Excavations on Iron Age and Romano-British settlements at Cannards Grave, Shepton
Mallett’, Somerset Archaeology and Natural History 144, 41–116
Booth, P. 1997: Asthall, Oxfordshire: Excavations in a Roman ‘Small Town’, 1992, Eynsham
Booth, P. 2001: ‘The Roman shrine at Westhawk Farm, Ashford: a preliminary account’, Archaeologica
Cantiana 121, 1–23
Bonnamour, L., and Dumont, A. 1996: ‘Les armes romaines de la Saône: état des découvertes et données
récentes de fouilles’, Proceedings of the Ninth International Roman Military Equipment Conference,
Leiden, 1994 = Journal of Roman Military Equipment Studies 5, 141–54
Bowden, M., and McOmish, D. 1987: ‘The required barrier’, Scottish Archaeological Review 4, 76–84
Bradford, J.S.P., and Goodchild, R.G. 1939: ‘Excavations at Frilford, Berks’, Oxoniensia 4, 1–70
Bradley, R. 1987: ‘Stages in the chronological development of hoards and votive deposits’, Proceedings
Prehistoric Society 53, 351–62
Bradley, R. 1998: The Passage of Arms: An Archaeological Analysis of Prehistoric Hoards and Votive
Deposits (new edn), Oxford
Bradley, R. 2005: Ritual and Domestic Life in Prehistoric Europe, London
Branigan, K., and Dearne, M. 1992: Romano-British Cavemen: Cave Use in Roman Britain, Oxford
Brinson, J.G.S. 1963: ‘Chesterford, Great’, in The Victoria County History of the Counties of England: A
History of Essex, III, 72–88
Brunaux, J.-L. 1988: The Celtic Gauls: Gods, Rites and Sanctuaries, London
Budd, P., and Taylor, T. 1995: ‘The faerie smith meets the bronze industry: magic versus science in the
interpretation of prehistoric metal-making’, World Archaeology 27, 133–43
Cambridge Archaeological Field Unit 1997: Preliminary Summary: Middle and Late Iron Age Settlements
Adjacent to Hinchingbrooke Park Road, Huntingdon, Cambridgeshire, Cambridge
Carter, G.A. 1998: Excavations at the Orsett ‘Cock’ Enclosure, Essex 1976, East Anglian Archaeology 86,
Chowne, P., Cleal, R.M.J., Fitzpatrick, A., and Andrews, P. 2001: Excavations at Billingborough,
Lincolnshire 1975–8: a Bronze–Iron Age Settlement and Salt-Working Site, East Anglian Archaeology
94, Salisbury
Clarke, S. 1997: ‘Abandonment, rubbish disposal and “special” deposits at Newstead’, in K. Meadows, C.
Lemke and J. Heron (eds), TRAC 96: Proceedings of the Sixth Annual Theoretical Roman Archaeology
Conference Sheffield 1996, Oxford, 73–81
Clarke, S. 2000: ‘In search of a different Roman period: the finds assemblage at the Newstead military
complex’, in G. Fincham, G. Harrison, R. Holland and L. Revell (eds), TRAC 99: Proceedings of the
Ninth Annual Theoretical Roman Archaeology Conference Durham 1999, Oxford, 22–9
Clarke, S., and Jones, R. 1996: ‘The Newstead pits’, Proceedings of the Ninth International Roman Military
Equipment Conference, Leiden, 1994 = Journal of Roman Military Equipment Studies 5, 109–24

Colt Hoare, R. 1827: ‘Account of antiquities found at Hamden Hill, with fragments of British chariots’,
Archaeologia 21, 39–42
Cracknell, S., and Hingley, R. 1994: ‘Park Farm, Barford: excavation of a prehistoric settlement site, 1988’,
Trans. Birmingham and Warwicks. Arch. Soc. 98 (1993–4), 1–30
Crummy, P. 1997a: City of Victory: the Story of Colchester, Colchester
Crummy, P. 1997b: ‘Colchester: the Stanway burials’, Current Archaeology 153, 337–42
Cunliffe, B.W. 1978: Iron Age Communities in Britain, London
Cunliffe, B.W. 1984: Danebury: An Iron Age Hillfort in Hampshire: Vol. 2. The Excavations 1969–1978:
the Finds, London
Cunliffe, B.W. 1995: Danebury: An Iron Age Hillfort in Hampshire: Vol. 6. A Hillfort Community in
Perspective, London
Cunliffe, B.W., and Poole, C. 1991a: Danebury: An Iron Age Hillfort in Hampshire: Vol. 4. The Excavations
1979–1988: the Site, London
Cunliffe, B.W., and Poole, C. 1991b: Danebury: An Iron Age Hillfort in Hampshire: Vol. 5. The Excavations
1979–1988: the Finds, London
Cunliffe, B.W., and Poole, C. 2000: The Danebury Environs Programme: Vol. 2, Part 6, Houghton Down,
Stockbridge, Hants, 1994, Oxford
Curle, J. 1911: A Roman Frontier Post and its People: the Fort of Newstead in the Parish of Melrose,
Davies, S.M. 1981: ‘Excavations at Old Down Farm, Andover: Part II: prehistoric and Roman’, Proc.
Hampshire Field Club and Arch. Soc. 37, 81–163
Dungworth, D. 1998: ‘Mystifying Roman nails: clavus annalis, defixiones and minkisi’, in C. Forcey,
J. Hawthorne and R. Witcher (eds), TRAC 97: Proceedings of the Seventh Annual Theoretical Roman
Archaeology Conference Nottingham 1997, Oxford, 148–59
Ellis, C.J., and Rawlings, M. 2001: ‘Excavations at Balksbury Camp, Andover 1995–97’, Proc. Hampshire
Field Club and Arch. Soc. 56, 21–94
Esmonde Cleary, S. 2000: ‘Putting the dead in their place: burial location in Roman Britain’, in J. Pearce,
M. Millett and M. Struck (eds), Burial, Society and Context in the Roman World, Oxford, 127–42
Evans, J. 1894: ‘On some iron tools and other articles formed of iron found at Silchester in the year 1890’,
Archaeologia 54, 139–56
Fell, V. 1990: Pre-Roman Iron Age Metalworking Tools from England and Wales: their Use, Technology,
and Archaeological Context, unpub. M. Phil. thesis, University of Durham
Fenn, C. 2003: ‘Important Iron Age weapon cache found in East Riding’, Treasure Hunting March 2003, 6–9
Field, N., and Parker Pearson, M. 2003: Fiskerton: An Iron Age Timber Causeway with Iron Age and
Roman Votive Offerings, Oxford
Fitzpatrick, A. 1984: ‘The deposition of La Tène Iron Age metalwork in watery contexts in Southern
England’, in B.W. Cunliffe and D. Miles (eds), Aspects of the Iron Age in Central Southern Britain,
Oxford, 178–90
Fowler, P.J. 1961: ‘Excavations at Madmarston Camp, Swalcliffe’, Oxoniensia 25 (1960), 3–48
Fox, G.E., and St John Hope, W.H. 1891: ‘Excavations on the site of the Roman city at Silchester, Hants’,
Archaeologia 52, 733–58
Frere, S.S. 1964: ‘Excavations at Dorchester on Thames’, Arch. Journ. 119 (1962), 114–49
Fulford, M. 2001: ‘Links with the past: pervasive “ritual” behaviour in Roman Britain’, Britannia 32,
Garrow, D., Lucy, S., and Gibson, D. forthcoming: Excavations at Kilverstone, Norfolk: the Neolithic
Pits, Later Prehistoric, Romano-British and Anglo-Saxon Occupation and Later Activity, East Anglian
Germany, M. 2003: Excavations at Great Holts Farm, Boreham, Essex, 1992–94, East Anglian Archaeology
105, Chelmsford
Gillies, W. 1981: ‘The craftsman in early Celtic literature’, Scottish Archaeological Forum II, 70–85
Goodchild, R.G. 1938: ‘A priest’s sceptre from the Romano-Celtic temple at Farley Heath, Surrey’, Antiq.
Journ. 18, 391–6
Green, M.J. 1976: The Religions of Civilian Roman Britain, BAR 24, Oxford
Greene, K. 2002: ‘Pots and plots in Roman Britain’, in M. Aldhouse-Green and P. Webster (eds), Artefacts
and Archaeology: Aspects of the Celtic and Roman World, Cardiff, 235–56

Gresham, C. 1940: ‘Spettisbury Rings, Dorset’, Arch. Journ. 96 (1939), 114–31

Haselgrove, C. 1987: Iron Age Coinage in South-Eastern Britain: the Archaeological Context, BAR Brit.
Ser. 174, Oxford
Haselgrove, C., Armit, I., Champion, T., Creighton, J., Gwilt, A., Hill, J.D., Hunter, F., and Woodward, A.
2001: Understanding the British Iron Age: An Agenda for Action, Salisbury
Haselgrove, C., and Hingley, R. in press (2005): ‘Iron deposition and its significance in pre-Roman
Britain’, in G.G. Bataille and J.-P. Guillaumet (eds), Les dépôts d’objects métalliques aux âges du fer
(Collection Bibracte), Glux-en-Glenne
Haselgrove, C., Lowther, P., and Turnbull, P. 1990: ‘Stanwick, North Yorkshire, Part 3: excavations on
earthwork sites 1981–86’, Arch. Journ. 147, 37–90
Haselgrove, C., and Wigg-Wolf, D. 2005: ‘Introduction’, in C. Haselgrove and D. Wigg-Wolf (eds), Iron
Age Coinage and Ritual Practices, Studien zu Fundmünzen der Antike (SFMA) 20, 9–22
Hencken, T.C. 1939: ‘The excavation of the Iron Age Camp on Bredon Hill, Gloucestershire, 1935–1937’,
Arch. Journ. 95 (1938), 1–111
Herbert, E.W. 1993: Iron, Gender and Power: Rituals of Transformation in African Societies, Bloomington
Hill, J.D. 1989: ‘Re-thinking the Iron Age’, Scottish Archaeological Review 6, 16–23
Hill, J.D. 1995a: Rituals and Rubbish in the Iron Age of Wessex, BAR Brit. Ser. 242, Oxford
Hill, J.D. 1995b: ‘How should we understand Iron Age societies and hillforts: a contextual study from
Southern Britain’, in J.D. Hill and C.G. Cumberpatch (eds), Different Iron Ages: Studies of the Iron Age
in Temperate Europe, BAR Int. Ser. 602, Oxford, 45–66
Hingley, R. 1989: Rural Settlement in Roman Britain, London
Hingley, R. 1990a: ‘Iron Age “currency bars”: the archaeological and social context’, Arch. Journ. 147,
Hingley, R. 1990b: ‘Boundaries surrounding Iron Age and Romano-British settlements’, Scottish
Archaeological Review 7, 96–103
Hingley, R. 1997: ‘Iron, ironworking and regeneration: a study of the symbolic meaning of metalworking in
Iron Age Britain’, in A. Gwilt and C. Haselgrove (eds), Reconstructing Iron Age Societies, Oxford, 9–18
Hingley, R. 2005: ‘Iron Age “currency bars” in Britain: items of exchange in liminal contexts?’, in C.
Haselgrove and D. Wigg-Wolf (eds), Iron Age Coinage and Ritual Practices, Studien zu Fundmünzen
der Antike (SFMA) 20, 183–206
Hingley, R., Moore, H.L., Triscott, J.E., and Wilson, G. 1997: ‘The excavation of two later Iron Age fortified
homesteads at Aldclune, Blair Atholl, Perth & Kinross’, Proc. Soc. Antiq. Scotland 127, 407–66
Hinman, M. 2003: A Late Iron Age Farmstead and Romano-British Site at Haddon, Peterborough,
Hogg, R. 1966: ‘A hoard of Romano-British ironwork from Brampton, Cumberland’, Trans. Cumberland
& Westmorland Antiq. & Arch. Soc. 66, 1–36
Hooley, R. 1921: ‘Note on a hoard of iron currency-bars found on Worthy Down, Winchester’, Antiq.
Journ. 1, 321–7
Hooley, R. 1931: ‘Excavation of an early Iron Age village on Worthy Down, Winchester’, Proc. Hampshire
Field Club and Arch. Soc. 10 (1926–30), 178–92
Hunter, F. 1997: ‘Iron Age hoarding in Scotland and northern England’, in A. Gwilt and C. Haselgrove
(eds), Reconstructing Iron Age Societies, Oxford, 108–33
Hylton, T. 1996: ‘Iron objects/iron nails’, in R.J. Williams, P.J. Hart and A.T.L. Williams, Wavendon Gate:
A Late Iron Age and Roman Settlement in Milton Keynes, Aylesbury, 120–5
Insoll, T. 2004: Archaeology, Ritual and Religion, London
Isserlin, R. in press: ‘Deposit or withdrawal?’, in R. Hingley and S. Willis (eds), Roman Finds: Context
and Theory, Oxford, 183–95
Jackson, D. 1974: ‘Two new pit alignments and a hoard of currency bars from Northamptonshire’,
Northants. Arch. 9, 13–45
Jackson, R. 1990: Camerton: the Late Iron Age and Early Roman Metalwork, London
Jarvis, K. 1985: ‘A late Iron Age site with a currency bar hoard at Bearwood, Poole, Dorset’, Proc. Dorset
Nat. Hist. & Arch. Soc. 106 (1984), 138–42
Johns, C. 1995: ‘Romano-British precious-metal hoards: some comments on Martin Millett’s paper’, in
S. Cottam, D. Dungworth, S. Scott and J. Taylor (eds), TRAC 94: Proceedings of the Fourth Annual
Theoretical Roman Archaeology Conference 1994, Oxford, 107–18

Johns, C. 1996: ‘The classification and interpretation of Romano-British treasures’, Britannia 27, 1–16
King, A., and Soffe, G. 1998: ‘Internal organization and deposition at the Iron Age temple on Hayling
Island’, Hampshire Studies 53, 35–47
Lambrick, G., and Allen, T. 2004: Gravelly Guy, Stanton Harcourt, Oxfordshire: the Development of a
Prehistoric and Romano-British Community, Oxford
Leach, J. 1962: ‘The Smith God in Roman Britain’, Arch. Ael.4 40, 35–46
Macdonald, G. 1906: ‘The Roman fort on the Bar Hill, Dumbartonshire’, Proc. Soc. Antiq. Scotland 4 4
(1905–6), 403–546
Major, H. 2003: ‘Iron objects’, in M. Germany (ed.), Excavations at Great Holts Farm, Boreham, Essex,
1992–94, East Anglian Archaeology 105, Chelmsford, 77–83
Manning, W.H. 1965: ‘Excavation of the Roman auxiliary tilery, Brampton’, Trans. Cumberland &
Westmorland Antiq. & Arch. Soc. 65, 133–68
Manning, W.H. 1972: ‘Ironwork hoards in Iron Age and Roman Britain’, Britannia 3, 224–50
Manning, W.H. 1977: ‘Blacksmiths’ tools from Waltham Abbey, Essex’, in W.A. Oddy (ed.), Aspects of
Early Metallurgy, London, 87–96
Manning, W.H., and Scott, I.R. 1986: ‘Iron objects’, in I.M. Stead and V. Rigby, Baldock: the Excavation
of a Roman and Pre-Roman Settlement, 1968–72, London, 145–62
Mariën, M.E. 1970: Le Trou de l’ambre au bois de Wérimont Eprave, Brussels
Martin, Th., and Ruffat, H. 1998: ‘Un dépôt de lingots de fer du début de La Tène III à Montans (Tarn)’, in
M. Feugère and V. Serneels (eds), Recherches sur l’économie du fer en Méditerranée nord-occidentale,
Montagnac, 110–15
McArthur, C. 1990: ‘Excavations at Nadbury Camp, Warwickshire’, Trans. Birmingham & Warwicks.
Arch. Soc. 95 (1987–88), 1–16
Merrifield, R. 1987: The Archaeology of Ritual and Magic, London
Merrifield, R. 1997: ‘Roman metalwork from the Walbrook – rubbish, ritual or redundancy?’, Trans.
London Middlesex Arch. Soc. 46 (1995), 27–44
Miles, D. 1986: Archaeology at Barton Court Farm, Abingdon, Oxfordshire, York
Millett, M. 1995: ‘Treasure: interpreting Roman hoards’, in S. Cottam, D. Dungworth, S. Scott and J.
Taylor (eds), TRAC 94: Proceedings of the Fourth Annual Theoretical Roman Archaeology Conference
1994, Oxford, 99–106
Neal, D. 1989: ‘The Stanwick Villa, Northants: an interim report on the excavations of 1984–88’, Britannia
20, 149–68
Neal, D., Wardle, A., and Hunn, J. 1990: Excavation of the Iron Age, Roman and Medieval Settlements at
Gorhambury, St Albans, London
Neville, R.C. 1856: ‘Description of a remarkable deposit of Roman antiquities at Great Chesterford,
Essex’, Arch. Journ. 13, 1–13
Osborne, R. 2004: ‘Hoards, votives, offerings: the archaeology of the dedicated object’, World Archaeology
36, 1–10
Palmer, L.S. 1921: ‘Some early British remains from a Mendip Cave’, Journ. Royal Anthropological
Institute 51, 200–15
Pearce, M. 1998: ‘Reconstructing prehistoric metallurgical knowledge: the Northern Italian Copper and
Bronze Ages’, European Journ. Arch. 1, 51–70
Piggott, S. 1955: ‘Three metal-work hoards of the Roman period from Southern Scotland’, Proc. Soc.
Antiq. Scotland 87 (1952–53), 1–50
Pine, J., and Preston, S. 2004: Iron Age and Roman Settlement and Landscape at Totterdown Lane, Horcott
near Fairford, Gloucestershire, Reading
Pitts, L., and St Joseph, J.K. 1985: Inchtuthil: The Roman Legionary Fortress, London
Poole, C. 1995: ‘Pits and propitiation’, in B.W. Cunliffe, Danebury: an Iron Age Hillfort in Hampshire: Vol.
6. A Hillfort Community in Perspective, London, 249–75
Poulton, R., and Scott, E. 1993: ‘The hoarding, deposition and use of pewter in Roman Britain’, in E. Scott
(ed.), Theoretical Roman Archaeology Conference: First Conference Proceedings, Aldershot, 115–32
Pryor, F. 2001: The Flag Fen Basin: Archaeology and Environment of a Fenland Landscape, London
Rees, S. 1979: Agricultural Implements in Prehistoric and Roman Britain, BAR, Oxford
Reid, C. 1901: ‘Excavations on the site of the Roman city at Silchester, Hants in 1900’, Archaeologia 57,

Richmond, I. 1968: Hod Hill: Volume Two: Excavations Carried Out Between 1951 and 1958 for the
Trustees of the British Museum, London
Ross, A. 1967: Pagan Celtic Britain, London
RCHM 1970: An Inventory of Historical Monuments in the County of Dorset, Volume 2, South-east, Part
3, London
Savory, H.N. 1996: ‘A find of early Iron Age metalwork from the Lesser Garth, Pentyrch (Glam.)’, Arch.
Camb. 115, 27–44
Schrüfer-Kolb, I. 2004: Roman Iron Production in Britain; Technological and Socio-economic Landscape
Development along the Jurassic Ridge, BAR 380, Oxford
Sellwood, L. 1984: ‘Tribal boundaries viewed from the perspective of numismatic evidence’, in B.W.
Cunliffe and D. Miles (eds), Aspects of the Iron Age in Central Southern Britain, Oxford, 191–204
Sharples, N. 1991: Maiden Castle: Excavations and Field Survey 1985–6, London
Sim, D., and Ridge, I. 2002: Iron for the Eagles: the Iron Industry in Roman Britain, Stroud
Skinner, C. 1994: ‘Iron objects’, in R.J. Williams and R.J. Zeepvat, Bancroft: a Late Bronze Age/Iron Age
Settlement, Roman Villa & Temple-Mausoleum, Aylesbury, 322–47
Smith, G. 1991: ‘Excavations at Ham Hill, 1983’, Somerset Archaeology and Natural History 134 (1990),
Spratling, M. 1979: ‘The debris of metalworking’, in G.W. Wainwright, Gussage All Saints: an Iron Age
Settlement in Dorset, London, 125–49
St George Gray, H. 1930: ‘Excavations at Kingsdown Camp, Mells, Somerset, 1927–9’, Archaeologia 80,
Stead, I.M. 1984: ‘Iron-Age metalwork from Orton Meadows’, Durobrivae 9, 6–7
Stead, I.M. 1998: The Salisbury Hoard, Stroud
Stead, I.M., and Rigby, V. 1986: Baldock: The Excavation of a Roman and Pre-Roman Settlement, 1968–
72, London
Trow, S. 1988: ‘Excavations at Ditches hillfort, North Cerney, Gloucestershire, 1982–83’, Trans. Bristol &
Glos. Arch. Soc. 106, 19–86
van Driel Murray, C. 1999: ‘And did those feet in ancient times ... feet and shoes as a material projection’,
in P. Baker, C. Forcey, S. Jundi and R. Witcher (eds), TRAC 98: Proceedings of the Eighth Annual
Theoretical Roman Archaeology Conference Leicester 1998, Oxford, 131–40
van Enckevort, H.L.H., and Willems, W.J.H. 1996: ‘Roman cavalry helmets in ritual hoards from the
Kops Plateau at Nijmegen, The Netherlands’, in Proceedings of the Ninth International Roman Military
Equipment Conference, Leiden, 1994 = Journal of Roman Military Equipment Studies 5, 125–40
Wait, G.A. 1985: Ritual and Religion in Iron Age Britain, BAR, Oxford
Webster, J. 1997: ‘Texts, expectations: the archaeology of “Celtic” ritual wells and shafts’, in A. Gwilt and
C. Haselgrove (eds), Reconstructing Iron Age Societies, Oxford, 134–44
Wheeler, R.E.M. 1943: Maiden Castle, Dorset, Oxford
Wheeler, R.E.M. 1954: The Stanwick Fortifications, North Riding of Yorkshire, Oxford
Williams, R.J., Hart, P.J., and Williams, A.T.L. 1996: Wavendon Gate: a Late Iron Age and Roman
Settlement in Milton Keynes, Aylesbury
Williams, R.J., and Zeepvat, R.J. 1994: Bancroft: a Late Bronze Age/Iron Age Settlement, Roman Villa &
Temple-Mausoleum, Aylesbury
Woodward, A. 1992: Shrines & Sacrifice, London
Woodward, A., and Leach, P. 1993: The Uley Shrines: Excavation of a Ritual Complex on West Hill, Uley,
Gloucestershire: 1977–9, London
Woodward, P., and Woodward, A. 2004: ‘Dedicating the town: urban foundation deposits in Roman
Britain’, World Archaeology 36, 68–86
Wrathmell, S., and Nicholson, A. 1990: Dalton Parlours: Iron Age Settlement and Roman Villa, Yorkshire
Archaeology 3, Wakefield