MARTIN PITTS
POTS AND PITS: DRINKING AND DEPOSITION IN LATE IRON
AGE SOUTH-EAST BRITAIN
Summary. This paper considers the role of pottery in the Late Iron Age to
Roman transition in south-east Britain. Traditional concern with the
significance of Continental imports is rejected in favour of a more holistic and
bottom-up approach giving equal emphasis to locally made forms and imports
in complete assemblages. Several stages of inter-site correspondence analysis
are conducted on a range of sites and assemblages in the region. Patterning
pertaining to the use and deposition of both imported and local pottery vessels
can be seen to contradict simplistic models for ‘Romanization before conquest’.
The main conclusions include evidence for the selective disposal of drinking
vessels and table wares in pits, the likely widespread consumption of beer as
opposed to wine, and the implied importance of indigenous social practices
such as feasting and communal drinking.
introduction
South-east Britain in the Late Iron Age to Roman transition is characterized by the first
events of British history – the campaigns of Julius Caesar, the full-scale invasion of Claudius,
and the Boudican revolt. Given this historical background of external influence, it is not
surprising that the archaeological study of this region has often focused largely on the high
visibility of imported commodities and material culture in the Late Iron Age. Indeed, during the
latter part of the twentieth century, imports were central players in the context of popular theories
involving emphasis on ‘prestige goods’ and ‘core-periphery’ models, whereby mass-produced
Continental material culture facilitated the creation of higher levels of hierarchy and complexity
in native society (e.g. Haselgrove 1987; Cunliffe 1988). Furthermore, the presence of such
imports, especially pottery, has even been interpreted as Romanization before conquest
(Haselgrove 1984), inevitably providing the ‘raw material for Chapter One of the story of the
Roman conquest and of Roman Britain’ (Willis 1995, 141). Although these models and
perspectives have since been roundly criticized and have fallen out of fashion (e.g. Woolf 1993;
Willis 1995), there have been few convincing new studies to explain trajectories of social
development in this transitional period. Therefore, a more rounded approach is needed to
contextualize such historical constructs with the archaeological record.
It is the aim of this paper to employ archaeological evidence from Late Iron Age southeast
Britain in a more holistic manner that reconciles imported artefacts with the larger body of
locally produced material culture. The principal medium for study is pottery, an artefact class
OXFORD JOURNAL OF ARCHAEOLOGY 24(2) 143–161 2005
© Blackwell Publishing Ltd. 2005, 9600 Garsington Road, Oxford OX4 2DQ, UK
and 350 Main Street Malden, MA 02148, USA. 143
which occurs in significant quantities in the study area in both local and imported fabrics.
Multiple pottery assemblages from a selection of sites in Essex and Hertfordshire were chosen
for analysis. Particular emphasis is given to the role of different vessel forms in the articulation
of the everyday social practices of eating and drinking, and the manner of their eventual disposal.
However, before moving onto specific case-studies, it is first necessary to assess briefly existing
perspectives on the role of pottery in this intensely studied region and period.
pottery in late iron age south-east britain: existing models and critique
Despite a plethora of Late Iron Age pottery studies, the understanding of this particular
class of archaeological material has been somewhat hindered by the Late Iron Age to Roman
transition lying at the interface of prehistoric and classical studies. A direct consequence of this
prehistoric-classical divide concerns the ways in which the pottery itself is recorded and
published, which continues to have serious ramifications for the study of ceramic assemblages
today. Whereas locally produced vessels are usually studied by a single Late Iron Age pottery
specialist, the presence of imports often necessitates splitting up assemblages and sending
different components (e.g. Gallo-Belgic wares, samian and amphorae) to Roman pottery
specialists. The end result usually comprises a published report with separate sections on local
pottery and the various classes of imports, with little integration. This encourages the study of
pottery fabrics in isolation, which facilitates the creation of a further conceptual divide between
the use of imports and that of indigenous vessels, further encouraging a fragmentary and entirely
artificial construction of the past. This situation is felt most acutely in south-east Britain, where
attention has been traditionally focused on the quantity and range of imports.
Agood example of the over-privileging of imports in this region is that of Mediterranean
wine amphorae. The relatively large quantities of wine implicated in a small number of high status
graves of this period (e.g. Stead 1967; Foster 1986) prompted countless papers on the wine trade
in this period (Peacock 1971; Dannell 1979; Williams 1981; 1989; Fitzpatrick 1985; Carver
2001). The classic example from south-east Britain is that of Haselgrove (1982, 81), who argued
that access to prestige goods (including wine) for use on critical occasions and in rites of passage
was vital for individuals to gain or maintain social power. Through comparisons with the rest of
north-west Europe it later became apparent that the quantity of wine entering Late Iron Age
Britain was probably insufficient to support social elites directly (Haselgrove 1996, 168–75;
Fitzpatrick 1985, 311). A better alternative is that the imports were being used as luxuries, not
being essential for the maintenance of the social hierarchy (Woolf 1993, 18). However,
irrespective of these positions, the question of whether imports were prestigious continues to
obscure our understanding of Late Iron Age society. In all such overarching models, the role of
natively produced pottery is marginalized. There have been few attempts to view amphorae as
components in wider depositional contexts, except in the instances of imported ‘drinking kits’ in
funerary assemblages (e.g. Stead 1967; Stead and Rigby 1986, 51–61). Hence, this paper will
now consider the role of imported eating and drinking vessels in their broader contexts of
association and deposition in Late Iron Age south-east Britain.
putting imports into context: a multivariate approach
Complete ceramic assemblages were selected from a number of sites in the study region
of modern day Essex and Hertfordshire, roughly corresponding to the tribal areas of the
POTS AND PITS
OXFORD JOURNAL OF ARCHAEOLOGY
144 © Blackwell Publishing Ltd. 2005
Trinovantes and Catuvellauni, and also equating to a large part of the ‘core’ area of Haselgrove’s
(1982) core-periphery model for south-east Britain. A chronological span of c.50 BC–AD 80
was chosen to account for patterns of pottery use and deposition in the generations immediately
preceding and following the Roman conquest inAD 43. The close dating of ceramic assemblages
from the different sites permitted this period to be split into three separate sub-phases for analysis
(see Table 1). The full list of sites (with references) is provided in Appendix 1. All the
assemblages had been quantified by minimum number of vessels (MNV), with the exception of
Elms Farm, Heybridge and Head Street, Colchester, which had both been quantified by estimated
vessel equivalent (EVE). These values were subsequently converted to percentages to minimize
the error in comparing two different quantification measures.
Analysis was conducted using correspondence analysis, a multivariate statistical
technique that provides a quantitative means of summarizing the associations of find-types (in
this case pottery) in multiple assemblages (sites and features). Each phase of correspondence
analysis produces two plots – one showing find-types by their relative occurrence in different
contexts, and the other showing contexts by their composition of find-types. The values that can
be read off from correspondence analysis outputs should be regarded as statistical abstractions
rather than having any independent quantitative meaning in an archaeological sense. Instead,
interpretation is dependent on cross-referencing between the two different plots, with the area
of the first plot directly equating to the equivalent area of the second, albeit represented in terms
of different variables. For example, the position of a particular pottery type on the first plot will
correspond most closely to the position of the context(s) containing the greatest proportion
of that pottery type on the second plot, and vice-versa. The use of this method with pottery
assemblages has already been successfully applied at both intra-site (Pitts 2002; 2004;
forthcoming b) and inter-site resolutions (Pitts forthcoming a; Pitts forthcoming b), and has had
further applications in the analysis of small finds assemblages (e.g. Cool, Lloyd-Morgan and
Hooley 1995; Lockyear 2000; Cool and Baxter 2002).
Three tiers of inter-site analysis were performed. The first comprised the analysis of
vessel form, with extra subdivisions for imported material. This was intended to provide an
overview of which sites were receiving imports, and the general combinations of vessels
favoured at different locations. The second tier of analysis was similar to the first, albeit without
the distinctions for imported vessels. In contrast, the last set of analysis considers the functional
composition of pottery assemblages from different classes of excavated feature. This was
proposed in order to investigate potential depositional patterning which may relate to the social
practices of the preparation, serving and consumption of food and drink. These tiers of analysis
are each discussed in turn below.
1. Inter-site analysis: vessel form (with imports)
Prior to analysis, the main jar forms were removed. Jars were present on all sites in this
period in such quantity that the correspondence analysis plots would be significantly distorted.
However, this material was re-included for the final stage of analysis of depositional traits.
Furthermore, amphorae were also removed due to major discrepancies in quantification with the
bulk of other pottery vessels (Peacock and Williams 1986, 18–19; Sealey forthcoming). This
can be justified on the basis of the differential primary function of amphorae relative to other
classes of pottery (i.e. as containers for the bulk transportation of foodstuffs rather than direct
involvement in consumption) and their comparative scarcity in the study area (the presence of
MARTIN PITTS
OXFORD JOURNAL OF ARCHAEOLOGY
© Blackwell Publishing Ltd. 2005 145
amphorae in the assemblages used in this paper is discussed in Pitts forthcoming b). The results
of analysis are presented in Figures 1a and 1b. Figure 1a shows the sites (with codes and phases
detailed in Table 1) by their relative proportions of different vessel forms, whereas Figure 1b
shows the vessel forms by their relative occurrence at the different sites. Points on the first plot
can be directly referenced with equivalent areas of the second, and vice-versa.
A number of patterns were apparent after examining both plots. Most striking was the
clustering of imported vessels in the upper-left quadrant of Figure 1a. The lower portion of this
cluster, largely composed of Gallo-Belgic imports, corresponded to composite assemblages from
a number of sites, most notably post-conquest Sheepen and several sub-sites and phases from
the Braughing complex. In comparison, the upper portion of the cluster corresponded to Head
Street, Colchester, a site located within the pre-Boudican colonia. Perhaps not surprisingly, the
planted Roman veteran settlement was quite distinct from all the other sites, largely on the basis
of high proportions of imported samian and typically ‘Roman’ serving and preparation vessels
such as flagons and mortaria. Such a clear contrast between pottery from the imposed colonia
POTS AND PITS
OXFORD JOURNAL OF ARCHAEOLOGY
146 © Blackwell Publishing Ltd. 2005
Figure 1a
Correspondence analysis – sites and individual phases plotted by their proportions of different vessel forms.
MARTIN PITTS
OXFORD JOURNAL OF ARCHAEOLOGY
© Blackwell Publishing Ltd. 2005 147
Figure 1b
Correspondence analysis – vessel form classes plotted by their occurrence at selected sites and sub-phases. Prefixes
denote imported fabric types (CG = Central Gaulish ware, GB = Gallo-Belgic ware, TSG = samian ware and
I = miscellaneous imports).
table 1
Dating of composite assemblages for analysis
Phase Date-range Description Site assemblages
1 c.50–1 BC Late Iron Age Baldock (Bdk), Elms Farm (EF), Skeleton Green (SG)
2 c.AD 1–40 Pre-conquest Baldock (Bdk), Elms Farm (EF), Ermine Street (ES), Great Dunmow
(GD), Nazeingbury (Naz), Sheepen (Cam), Skeleton Green (SG)
3 c.AD 40–80 Transitional Ardleigh (Adl), Baldock (Bdk), Elms Farm (EF), Folly Lane (FL), Great
Dunmow (GD), Head Street (HS), Lockleys (Ls), Nazeingbury (Naz),
Park Street (PS), Sheepen (Cam), Wickham Kennels (WK), Woodham
Walter (WW)
and that of indigenous centres of power (including the nearby Sheepen site) demonstrates the
extent to which the occupants of the former may have been an alien entity. This contradicts
traditional models in Roman archaeology of coloniae as civilizing centres for the Romanization
of the province (e.g. Frere 1987, 189). At the most obvious level, a relatively high proportion
of mortaria (and to a lesser extent flagons) at Head Street implies fundamentally different social
practices of food preparation as opposed to consumption practices at neighbouring Sheepen (and
the rest of the region).
More complicated is the relationship between the imported fine wares. Whereas samian
ware occurred in larger quantities at the colonia, the Gallo-Belgic fine wares (largely comprising
terra rubra and terra nigra) were almost completely absent from the colonia, occurring in their
highest proportions at Sheepen and Braughing. At a basic level, this seems to imply that the
colonia was being provisioned separately, probably via official Roman supply systems.
However, despite the Gallo-Belgic pottery industry being somewhat dependent on copies of
arretine samian (Hawkes and Hull 1947, 202–3), there are fundamental functional differences
between the imported samian and Gallo-Belgic vessels turning up in Britain in this period.
Whereas both imported fabrics share a number of common vessel forms, including platters,
dishes, bowls and cups, only the Gallo-Belgic suite of vessels has substantial provision for larger
drinking vessels such as butt-beakers and girth-beakers. The implication here is that indigenous
society had a need for large drinking vessels (implying communal drinking), whereas the
incoming population at the colonia did not.
A continued emphasis on drinking vessels is further evident among the remaining
clusters of sites in Figure 1a. In the lower-right quadrant of this plot is a cluster of sites
corresponding to a series of drinking vessels in local fabrics – butt-beakers, cups, girth-beakers
and tazze. With the benefit of archaeological hindsight, it is possible to see all of these particular
site-phases as forerunners to more ‘Romanized’ or ‘Gallo-Romanized’ occupations in following
generations. Whereas pre-conquest Sheepen and Great Dunmow were later characterized by
higher proportions of Gallo-Belgic imports (see the centre-left cluster discussed above), the Late
Iron Age occupations from Lockleys and Park Street were quickly superseded by first century
villas. Likewise, the ceremonial site at Folly Lane continued to be used as part of the Roman
city of Verulamium (Niblett 1999). The occurrence of relatively high proportions of drinking
vessels at these sites, particularly the local copies of Gallo-Roman forms (such as butt-beakers
and girth-beakers), suggests a selective diffusion of Continental material culture. That this
process favoured material from a Gallo-Roman cultural template is unsurprising, given the
supposed ties of kinship on both sides of the Continent (Millett 1990, 9). Moreover, the
preference for large drinking vessels probably reflects the importance of communal drinking in
indigenous society, particularly evident in the extensive provision of drinking accoutrements in
the so-called Aylesford (Whimster 1981, 160) and higher-status Welwyn (Stead 1967) burial
traditions.
The final cluster of note in this phase of intra-site analysis was the close association of
indigenous sites (Ardleigh, Woodham Walter and the various phases of Baldock and Elms Farm)
in the centre to upper-right quadrant of Figure 1a. These sites corresponded to a combination
of some of the earlier imports (most notably from central Gaul) and older indigenous forms in
grog-tempered wares (Thompson 1982), such as pedestal jars and tazze, and various other jar
forms. The relative emphasis on jar forms and first century BC vessel types perhaps indicates
that these sites formed the lowest status tier of sites analysed in this study. Nevertheless, a strong
theme of Continental contact remains, in addition to the use of drinking vessels. Whereas
POTS AND PITS
OXFORD JOURNAL OF ARCHAEOLOGY
148 © Blackwell Publishing Ltd. 2005
pedestal jars have been suggested by J.D. Hill to have been used for alcohol consumption (Hill
2002, 147), a related function has been postulated for the rarer spouted strainer bowls (Sealey
1999; see also below).
2. Inter-site analysis: vessel form (without distinctions for imports)
A second stage of correspondence analysis was conducted with all vessels grouped by
form regardless of fabric (see Figures 2a and 2b). Otherwise, the same conditions were applied
as in the first stage (removal of jars and amphorae). The results of this phase of analysis strongly
echo those outlined above, suggesting that differential access to imported material was also
accompanied by similar patterning in the use of general combinations of vessel forms. Occurring
as an outlier in the upper-left quadrant is Head Street (corresponding to mortaria and to a lesser
extent flagons), again highlighting the palpably distinct consumption practices taking place at
the colonia. Untangling the crowd of remaining sites is a little more complicated, although the
MARTIN PITTS
OXFORD JOURNAL OF ARCHAEOLOGY
© Blackwell Publishing Ltd. 2005 149
Figure 2a
Correspondence analysis – sites and individual phases plotted by their proportions of different vessel forms
(without distinctions for imports).
same groupings of sites are broadly evident as noted in Figure 1. At the top-left of the main
cluster are native centres of importance including the various sites from Braughing and postconquest
Sheepen; below this are the middle range of sites which went on to become more
‘Romanized’ (e.g. the Late Iron Age pre-villa occupations at Park Street and Lockleys),
corresponding to Gallo-Roman forms (most notably platters and butt-beakers); and finally, to
the right are the sites emphasizing the use of jar variants and older pedestalled forms (e.g.
Baldock and Elms Farm).
Thus, the picture emerging from the first two phases of analysis is a hierarchy of
settlement elaborated on the basis of differential use of imports and combinations of vessels
deriving from several distinct cultural templates of pottery production. It must be stressed that
such distinctions have been made through emphasizing only the relative differences between
pottery assemblages. Although Baldock and Elms Farm have stood out for their comparatively
high levels of older pottery forms, both sites received significant proportions of imported vessels
before the Roman conquest. Nevertheless, this approach clearly demonstrates the value of
POTS AND PITS
OXFORD JOURNAL OF ARCHAEOLOGY
150 © Blackwell Publishing Ltd. 2005
Figure 2b
Correspondence analysis – vessel form classes plotted by their occurrence at selected sites and sub-phases (without
distinctions for imports).
studying Continental imports in their wider contexts of use. It is envisaged that many imports
were favoured because they could easily be accommodated within existing practices of
consumption, a trend clearly evident in the extensive local copying of butt-beaker and platter
forms. Although Creighton’s recent study of coin imagery in Late Iron Age south-east Britain
seems to resonate strongly with the idea of an indigenous client aristocracy in direct engagement
with the ideological development of Augustan Rome (Creighton 2000, 124), the patterning in
contemporary ceramic assemblages appears to bear witness to more deep-rooted changes in
society. A complex melting-pot of influences emerges which was neither Romanization before
conquest, nor a wholesale ‘Gallo-Belgicization’.
3. Inter-site analysis: pot function and depositional traits
This stage of analysis considered the depositional traits of the main functional classes
of pottery. The same pottery assemblages considered above were categorized according to the
feature types they had been found in. Most of the features could be classed as pits or linear
(largely ditches and gullies), with a smaller proportion of wells, shafts and other features. In
addition, the pottery was reclassified on the basis of perceived function – drinking vessels (e.g.
cups, beakers and pedestal jars), jars (universal cooking wares), preparation vessels (e.g.
mortaria), serving vessels (e.g. flagons and strainers) and table wares (bowls, platters and
dishes).
The results of correspondence analysis on these data are presented in Figures 3a and
3b. The patterning that emerges seems to point towards the consistent selective deposition of
different functional classes of pottery in particular feature types. Probably most significant is
the strong clustering of linear features on the right-hand side of the plot which corresponds to
jar forms, and the clustering of pits (in addition to wells and shafts) on the left-hand side of the
plot which corresponds with greatest intensity to drinking vessels and table wares. The main
exceptions are the pits from Elms Farm occurring with the mass of linear features. In hindsight,
this trend is likely to have been the result of quantification by EVE over-emphasizing jar forms
due to the high survival rate of their thick and robust rims (from which EVEs are calculated).
Nevertheless, closer inspection of Figure 3a reveals that within every site-phase represented by
composite assemblages from both linear and pit features (including three from Elms Farm), the
linear features all tended to have higher proportions of jar forms. Thus, it seems likely that linear
features were favoured for the deposition of jars, whereas pits, wells and shafts were preferred
for the disposal of other functional classes. In particular, the bulk of features on the left side of
Figure 3a seemed to correspond closely with drinking vessels and table wares, especially wells
and shafts (and to a lesser extent pits). The preference for the deposition of drinking vessels and
table wares (together) places particular emphasis on the actual consumption of food and drink
(as opposed to preparation and serving). Combined with the association with features
traditionally linked to Iron Age ritual deposits such as pits and other vertically defined features
(e.g. Hill 1995), it is tempting to interpret such patterning as the deliberate deposition of material
involved in feasting and/or communal drinking. This trend is even more apparent in Table 2,
which shows pits, wells and shafts across all sites quantified by MNV having tangibly higher
proportions of non-jar forms, especially drinking vessels and table wares.
However, in the absence of detailed study of both depositional processes and
assemblage condition for each individual feature, caution must be exercised. In compiling these
data from various publications and databases it was clear that few features actually contained
MARTIN PITTS
OXFORD JOURNAL OF ARCHAEOLOGY
© Blackwell Publishing Ltd. 2005 151
near-complete vessels, the kind of assemblages one would expect from the disposal of feasting
remains. Indeed, in the material from Elms Farm, Heybridge, the average estimated vessel
equivalent for pits in this period was slightly higher than for ditches (see Table 3), although this
patterning was heavily weighted towards the earliest occupation of the site in the mid- to late
first century BC. This suggests that pits were also slightly more likely to receive less broken
POTS AND PITS
OXFORD JOURNAL OF ARCHAEOLOGY
152 © Blackwell Publishing Ltd. 2005
Figure 3a
Correspondence analysis – composite assemblages from different feature types plotted by their proportions of
different functional classes of pottery.
table 2
Average functional composition of features from all sites quantified by minimum number of vessels (highest
proportions per functional class in bold)
Feature type Jars Preparation vessels Serving vessels Drinking vessels Table wares
Pits 32.18% 5.01% 6.91% 29.28% 26.63%
Wells 36.65% 3.73% 1.24% 39.13% 19.25%
Shafts 45.83% 4.17% 0.00% 29.17% 20.83%
Linear 64.31% 3.00% 1.63% 19.62% 11.44%
MARTIN PITTS
OXFORD JOURNAL OF ARCHAEOLOGY
© Blackwell Publishing Ltd. 2005 153
Figure 3b
Correspondence analysis – functional classes of pottery plotted by their deposition in different feature types at
selected sites.
assemblages than linear features. However, given the relatively low average EVE for vessels in
pits at Elms Farm (at 0.16), it is unfeasible to conceptualize the function of these features as
‘feasting deposits’. It must be considered that ‘different types of feature may have different
characteristic pottery assemblages because of the different taphonomic pathways through which
their contents passed’ (Hill 1995, 21). The apparent disparity in assemblage composition
table 3
Average estimated vessel equivalents in pits and linear features at Elms Farm, Heybridge
Feature type Late Iron Age (1) Pre-conquest (2) Transitional (3) Overall mean
Linear 0.10 0.15 0.14 0.13
Pits 0.19 0.17 0.15 0.16
% Difference Pits + 31% Pits + 6% Pits + 3% Pits + 10%
between ditches and other features may in part be accounted for by the regular cleaning or
recutting of the former (e.g. Chadwick 1999, 163). After all, a ditch can only function effectively
as a boundary if it is to some degree open and visible. Nevertheless, it is envisaged that the
effect of such activity on the assemblages under scrutiny is minimal, given they were specifically
selected for their low-levels of residuality. Another possibility is that the lower proportions of
drinking vessels and table wares in ditches could be accounted for by the often liminal nature
of linear features, in contrast to the primary locus of consumption within households.
To summarize this stage of analysis, it has been possible to isolate what appears to have
been a depositional bias in the treatment of functional classes of pottery probably used for
different stages of the consumption process across various sites in Late Iron Age to Roman
south-east Britain. Whereas table wares and drinking vessels occurred in higher proportions in
pits, wells and shafts, cooking forms (i.e. jars) were predominantly disposed of in linear features.
Preliminary consideration of vessel completeness at Elms Farm, Heybridge suggests that pit
assemblages were further distinguished from ditch assemblages on the basis of having
higher levels of vessel completeness. However, further research into individual deposit and
site formation processes is ultimately required to enhance the understanding of such phenomena.
discussion
Consuming passions in Late Iron Age to Roman south-east Britain
A recurrent theme in the analyses above is an emphasis on consumption in later preRoman
Iron Age pottery assemblages. A concern with eating and especially drinking vessels is
particularly manifest at indigenous sites in both locally produced and imported pottery groups,
in a manner quite distinct from that of the first official Roman settlement at Colchester.
Moreover, it seems likely that such ‘consumption technology’ was intentionally deposited in
pits, wells and shafts, as opposed to cooking vessels, which were more likely to be disposed of
in linear features. The deposition of table wares and drinking vessels in sub-surface domestic
features echoes the more explicit social practices evident in contemporary funerary contexts
where whole vessels and sometimes meal remains were commonplace. Indeed, Whimster (1981,
155) has noted that a characteristic feature of both richly and poorly furnished cremations
was a preference for flat-grave burial in this region and period. The widespread use of pits as
the final destination for cremated remains and associated grave-goods suggests that the
accompanying ceremony would have taken precedence over the need to create a lasting
monumental memorial to the deceased. Thus, if such funerary practice represents a form of ritual
which can be defined as the explicit reproduction of social structures (Hill 1995, 99), the
evidently similar treatment of eating and drinking vessels in domestic contexts would seemingly
represent the more implicit reproduction of the same social structures through more routine and
mundane everyday practice (Hill, ibid.).
Whilst the meaning of such explicit and implicit associations of consumption and subsurface
features in the Late Iron Age remains uncertain, it is possible that a form of this practice
continued into the Roman period. A similar preference for the use of below-ground features may
have been incorporated into the design of early Roman houses. In particular, Perring (1989) has
noted the extensive occurrence of cellars and sunken rooms (some even containing similar
provisions for table wares and drinking vessels) in south-east Britain in the period c.AD 70–155.
It is unclear whether this represents a continuity of social practice, although parallel research
POTS AND PITS
OXFORD JOURNAL OF ARCHAEOLOGY
154 © Blackwell Publishing Ltd. 2005
into early Roman practices of consumption and deposition in Essex and Hertfordshire reveals
similar patterns to those evident in the Late Iron Age (Pitts forthcoming b).
In vino veritas?
Given the relatively small quantities of wine amphorae entering Britain, it is highly
unlikely that such imported commodities can account for what appears to be an almost universal
emphasis on drinking vessels at a range of site-types and identities in Late Iron Age south-east
Britain. Even if wine was being imported in a manner that was largely invisible archaeologically,
using barrels alongside amphorae (Fitzpatrick 2003, 23), one would expect much higher levels
of accompanying imports, coupled with a more even distribution of the latter to match the
provision of locally produced drinking vessels.Amore convincing explanation is that indigenous
beverages were being consumed in greater quantity, with access to wine being restricted to an
elite few. However, the survival of evidence for such local forms of alcohol in Late Iron Age
Britain is extremely limited.
Nevertheless, Sealey (1999) has argued that it may be possible to identify pottery
vessels specifically used for the serving of locally made ‘Celtic beer’ in the same region and
period. The vessel in question is the Late Iron Age spouted strainer bowl, which roughly
resembles a teapot in appearance, with the added inclusion of a strainer panel next to the spout.
Although it is found with wine amphorae in some contexts (including a bronze version from
the Welwyn Garden City grave, c.25–15 BC) this association would appear to be secondary, as
there are no direct parallels in the Roman world (Sealey 1999, 119–24). Clues to the possible
use of strainers in the serving of beer have recently been suggested by the finding of wormwood
flavouring in a bronze strainer at the Doctor’s grave at Stanway, Colchester (Paul Sealey, pers.
comm. June 2004). Such flavourings would thus be filtered out by the strainer before pouring
the drink into a vessel for subsequent consumption. To investigate this phenomenon, ten
stratified assemblages from the study sample featuring bronze and pottery spouted strainers were
subjected to correspondence analysis (see Figures 4a and 4b). This included funerary material
from Welwyn Garden City and the King Harry Lane cemetery at Verulamium, several
assemblages from Sheepen, one from the colonia at Colchester (Head Street), assemblages from
conquest period Elms Farm and Ardleigh near the Essex coast, and finally, two early second
century AD deposits from Roman Colchester and Chelmsford. Where possible, occurrences of
amphorae were included to examine the possible association of strainers and wine.
Several patterns are evident in this final set of correspondence analysis. Firstly, the early
Roman assemblages from Chelmsford and Colchester occurred as distinct outliers,
corresponding with typical ‘Roman’ forms such as mortaria, flagons and bowls. The peripheral
position of these later assemblages suggests they were quite different from the main body of
Late Iron Age material. A second, more complicated, cluster comprising the Welwyn Garden
City burial and deposits from both Roman (Head Street) and Gallo-Roman (Sheepen) parts of
transitional Colchester all corresponded to wine amphorae, cups, platters and pedestal vessels.
This relationship seems to be a hybrid of Late Iron Age, Gallo-Roman and Roman-influenced
suites, and clearly emphasizes the Roman links implicated generations before the conquest at
Welwyn. However, the final relationship is most telling. The spouted strainers have an almost
perfect association with butt-beakers, a very common (and large) drinking vessel which was
ideal for the consumption of beer. The nearest corresponding sites all date to within a generation
of the conquest, reinforcing the idea of importance of drinking in this period. Furthermore, 11
MARTIN PITTS
OXFORD JOURNAL OF ARCHAEOLOGY
© Blackwell Publishing Ltd. 2005 155
POTS AND PITS
OXFORD JOURNAL OF ARCHAEOLOGY
156 © Blackwell Publishing Ltd. 2005
of the 13 spouted strainers under scrutiny occurred in pits, reinforcing the association of drinking
vessels with this class of feature. Although this is far from conclusive evidence for the
widespread consumption of beer in Late Iron Age south-east Britain, it sits nicely alongside the
other trends noted above. Indeed, the strong contextual association of spouted strainers with a
common drinking vessel suggests that beer consumption may have been more prevalent and
central to Late Iron Age consumption practices than hitherto assumed.
conclusions
A number of significant patterns have emerged in the consideration of ceramic
assemblages in Late Iron Age south-east Britain, c.50 BC–AD 80, and are outlined here:
1. The identification of multiple trajectories of cultural influence (i.e. Roman military, GalloRoman
and late La Tène) within this transitional period. Most notable was the marked
contrast of consumption patterns evident at the Roman colony established at Colchester with
other indigenous sites throughout Essex and Hertfordshire. Thus, it is no longer tenable
simply to interpret the presence of imports in Late Iron Age society as direct indicators of
Romanization.
Figure 4a
Correspondence analysis – assemblages containing spouted strainer bowls plotted by their proportions of different
vessel forms (without distinctions for imports).
2. The use and copying of a selection of imports by Late Iron Age society, especially GalloBelgic
drinking vessels (i.e. butt-beakers), suggest that such artefacts were chosen for their
ability to fit into existing indigenous social practices of feasting and communal drinking.
This is supported by circumstantial evidence for the consumption of native beer in the
contextual associations of such drinking vessels with a pottery type (spouted strainer bowls)
thought to have functioned specifically for the serving of this beverage (Sealey 1999).
3. Further verification of the social importance of drinking and eating in this period was
provided by the tendency of ceramic consumption refuse (i.e. drinking vessels and table
wares) to be deposited together in pits, wells and shafts (as opposed to the disposal of
cooking wares in linear features). This trend appears to represent a day-to-day equivalent
of the more ostentatious contemporary mortuary practice of placing feasting remains in
cremation pits.
A recurrent theme of this paper has been the important role of commensal practices
(e.g. Dietler 1999) such as feasting in Late Iron Age society. From this perspective, it is now
possible to suggest an alternative model for the consumption of imports, especially wine, in this
period. Instead of viewing the imports as prestige goods which generated social distinctions in
their own right, it is increasingly likely that the wine was entering a society which was already
using communal drinking and feasts for the empowerment of individuals and groups. In this
MARTIN PITTS
OXFORD JOURNAL OF ARCHAEOLOGY
© Blackwell Publishing Ltd. 2005 157
Figure 4b
Correspondence analysis – vessel form classes plotted by their occurrence in assemblages containing spouted
strainer bowls (without distinctions for imports).
context, wine would undoubtedly have offered a few individuals the chance for greater prestige,
but at the same time would not have altered one of the potential mechanisms for maintaining
power, the capacity to dispense largesse through feasting. This is evidenced in the increasing
emphasis on drinking vessels in the period, and particularly in the vast majority of accompanied
cremation burials which occurred with large drinking vessels but no wine amphorae. Indeed,
the importance of alcohol in this period may have been enough for one of the major personalities
of this period (Cunobelin) to inscribe his gold coinage with ears of barley. Frere (1987, 31) has
argued that this signified native beer, and was a form of propaganda against the pro-Roman
Verica who had vine-leaves on his gold coins. However, in the context of this paper, the meaning
of such symbolism may have had more deeply rooted significance. Perhaps it was not the
imported goods themselves which were empowering, but rather the dynamic social practices in
which they were implicated. Therefore, the available evidence seems to suggest that the
importance of feasting and communal drinking is better understood through a model of prestige
practice than through a model of prestige goods.
Acknowledgements
I am very grateful to Joyce Compton (Essex County Council Field Archaeology Unit), Philip
Crummy and Howard Brooks (Colchester Archaeological Trust), and Jane Timby for providing
unpublished material for the analyses conducted here. I thank the Arts and Humanities Research Board
for funding my doctoral research, without which this paper would not have been possible. Some aspects
of this paper were originally aired at the Iron Age Research Student Seminar 2004 (York) and a research
forum at the Department of Archaeology, University of York. Last but not least, I am further indebted to
Dom Perring and Steve Roskams for supervising my research (and to the latter for commenting on this
paper), and to Paul Sealey (Colchester and Essex Museum) for his correspondence on Late Iron Age
alcohol.
Department of Archaeology,
University of York
The King’s Manor
York YO1 7EP
appendix 1
Ardleigh – Brown 1999; Baldock – Stead and Rigby 1986; Colchester (Insula 7) – Hull 1958; Chelmsford
– Going 1987; Elms Farm – Atkinson and Preston, forthcoming; Ermine Street – Potter and Trow 1988;
Folly Lane – Niblett 1999; Great Dunmow – Lavender 1997; Head Street – unpublished information from
Jane Timby and Howard Brooks; King Harry Lane – Stead and Rigby 1989; Lockleys – Ward-Perkins
1938; Nazeingbury – Huggins 1978; Park Street – O’Neill 1945; Sheepen – Niblett 1985; Skeleton Green
– Partridge 1981; Welwyn Garden City – Stead 1967; Wickham Kennels – Partridge 1982; Woodham Walter
– Buckley and Hedges 1987.
references
atkinson, m. and preston, s.j. forthcoming: Heybridge: a late Iron Age and Roman settlement,
excavations at Elms Farm, 1993–5 (Chelmsford, East Anglian Archaeology Report).
brown, n.r. 1999: The archaeology of Ardleigh, Essex: Excavations 1955–1980 (Chelmsford, East
Anglian Archaeology Report No. 90).
POTS AND PITS
OXFORD JOURNAL OF ARCHAEOLOGY
158 © Blackwell Publishing Ltd. 2005
buckley, d.j. and hedges, j.d. 1987: Excavation of a cropmark enclosure complex at Woodham Walter,
Essex, 1976 (Chelmsford, East Anglian Archaeology Report No. 33).
carver, e. 2001: The visibility of imported wine and its associated accoutrements in later Iron Age Britain
(Oxford, BAR British Series 325).
chadwick, a. 1999: Digging ditches, but missing riches? Ways into the Iron Age and Romano-British
cropmark landscapes of the north midlands. In Bevan, B. (ed.), Northern exposure: interpretive devolution
and the Iron Ages in Britain (Leicester), 149–71.
cool, h.e.m. and baxter, m.j. 2002: Exploring Romano-British finds assemblages. Oxford Journal of
Archaeology 21(4), 365–80.
cool, h.e.m., lloyd-morgan, g. and hooley, a.d. 1995: Finds from the fortress. The archaeology of
York, The small finds 17/10 (York, York Archaeological Trust and Council for British Archaeology).
creighton, j. 2000: Coins and power in late Iron Age Britain (Cambridge).
cunliffe, b. 1988: Greeks, Romans and Barbarians (London).
dannell, g.b. 1979: Eating and drinking in pre-conquest Britain: the evidence of amphora and samian
trading, and the effect of the invasion of Claudius. In Burnham, B.C. and Johnson, H.B. (eds.), Invasion
and response: the case of Roman Britain (Oxford, BAR British Series 73), 177–84.
dietler, m. 1999: Rituals of commensality and the politics of state formation in the ‘princely’ societies
of early Iron Age Europe. In Ruby, P. (ed.), Les princes de la protohistoire et l’émergence de l’état (Naples,
Collection de l’Ecole Française de Rome), 135–52.
fitzpatrick, a.p. 1985: The distribution of Dressel 1 amphorae in north-west Europe. Oxford Journal of
Archaeology 4(3), 305–40.
fitzpatrick, a.p. 2003: Roman amphorae in Iron Age Britain. In Plouviez, J. (ed.), Amphorae in Britain
and the western Empire (Oxford, Journal of Roman Pottery Studies 10), 10–24.
foster, j. 1986: The Lexden Tumulus: a re-appraisal of an Iron Age burial from Colchester, Essex (Oxford,
BAR British Series 156).
frere, s. 1987: Britannia: a history of Roman Britain (London).
going, c.j. 1987: The mansio and other sites in the south-eastern sector of Caesaromagus: the Roman
pottery (London, CBA Research Report 62).
haselgrove, c.c. 1982: Wealth, prestige and power: the dynamics of late Iron Age political centralisation
in south-east England. In Renfrew, C. and Shennan, S. (eds.), Ranking, resource and exchange
(Cambridge), 79–88.
haselgrove, c.c. 1984: Romanization before the conquest: Gaulish precedents and British consequences.
In Blagg, T.F.C. and King, A.C. (eds.), Military and civilian in Roman Britain (Oxford, BAR International
Series 136), 187–217.
haselgrove, c.c. 1987: Culture process on the periphery: Belgic Gaul and Rome during the late Republic
and early Empire. In Rowlands, M., Larsen, M. and Kristiansen, K. (eds.), Centre and periphery in the
ancient world (Cambridge), 104–24.
haselgrove, c.c. 1996: Roman impact on rural settlement and society in southern Picardy. In Roymans,
N. (ed.), From the sword to the plough. Three studies on the earliest Romanisation of Northern Gaul
(Amsterdam), 127–88.
hawkes, c.f.c. and hull, m.r. 1947: Camulodunum. First report on the excavations at Colchester
1930–1939 (Oxford).
hill, j.d. 1995: Ritual and rubbish in the Iron Age of Wessex (Oxford, BAR British Series 242).
hill, j.d. 2002: Just about the potter’s wheel? Using, making and depositing middle and later Iron Age
pots in East Anglia. In Woodward, A. and Hill, J.D. (eds.), Prehistoric Britain. The ceramic basis (Oxford),
143–60.
MARTIN PITTS
OXFORD JOURNAL OF ARCHAEOLOGY
© Blackwell Publishing Ltd. 2005 159
huggins, p.j. 1978: Excavation of Belgic and Romano-British farm with Middle Saxon cemetery and
churches at Nazeingbury, Essex, 1975–6. Essex Archaeology and History 10, 29–117.
hull, m.r. 1958: Roman Colchester (Oxford, Reports of the Research Committee of the Society of
Antiquaries of London, No. 20).
lavender, n.j. 1997: Middle Iron Age and Romano-British settlement at Great Dunmow: excavations at
Buildings Farm 1993. Essex Archaeology and History 28, 47–92.
lockyear, k. 2000: Site finds in Roman Britain: a comparison of techniques. Oxford Journal of
Archaeology 19(4), 397–423.
millett, m. 1990: The Romanization of Britain (Cambridge).
niblett, r. 1985: Sheepen: an early Roman industrial site at Camulodunum (London, CBA Research
Report 57).
niblett, r. 1999: The excavation of a ceremonial site at Folly Lane, Verulamium (London, Britannia
Monograph Series No. 14).
o’neill, h.e. 1945: The Roman villa at Park Street, near St. Albans, Hertfordshire: Report on the
excavations of 1943–45. The Archaeological Journal 102, 21–110.
partridge, c. 1981: Skeleton Green. A late Iron Age and Romano-British site (London, Britannia
Monograph Series No. 2).
partridge, c. 1982: Braughing, Wickham Kennels 1982. Hertfordshire Archaeology 8, 40–51.
peacock, d.p.s. 1971: Roman amphorae in pre-Roman Britain. In Hill, D. and Jesson, M. (eds.), The Iron
Age and its hillforts (Southampton), 161–88.
peacock, d.p.s. and williams, d.f. 1986: Amphorae and the Roman economy (London).
perring, d. 1989: Cellars and cults in Roman Britain. Archaeological Journal 146, 279–301.
pitts, m. 2002: Pottery consumption, power and identity at Elms Farm, Heybridge, Essex, 50 BC–AD
200 (University of York, unpublished MA thesis).
pitts, m. 2004: ‘I drink, therefore I am?’ Pottery consumption and identity at Elms Farm, Heybridge,
Essex. In Croxford, B., Eckardt, H., Meade, J. and Weekes, J. (eds.), TRAC 2003. Proceedings of the
thirteenth annual theoretical Roman Archaeology conference, Leicester 2003 (Oxford), 16–27.
pitts, m. forthcoming a: Regional identities and the social use of ceramics. TRAC 2004. Proceedings of
the fourteenth annual theoretical Roman Archaeology conference, Durham, 2004.
pitts, m. forthcoming b: Consumption and identity in Essex and Hertfordshire, c. 50 BC–AD 200: a
ceramic perspective (University of York, forthcoming PhD thesis).
potter, t.w. and trow, s.d. 1988: Puckeridge-Braughing, Herts.: The Ermine Street excavations,
1971–1972. The late Iron Age and Roman settlement. Hertfordshire Archaeology 10, 1–189.
sealey, p.r. 1999: Finds from the cauldron pit. The spouted strainer bowls. In Brown, N.R., The
archaeology of Ardleigh, Essex: Excavations 1955–1980 (Chelmsford, East Anglian Archaeology Report
No. 90), 117–24.
sealey, p.r. forthcoming: Amphoras. In Atkinson, M. and Preston, S.J., Heybridge: a late Iron Age and
Roman settlement, excavations at Elms Farm, 1993–5 (Chelmsford, East Anglian Archaeology Report).
stead, i.m. 1967: A La Tène III burial at Welwyn Garden City. Archaeologia 101, 1–62.
stead, i.m. and rigby, v.a. 1986: Baldock: the Excavation of a Roman and Pre-Roman Settlement,
1968–72 (London, Britannia Monograph Series No. 7).
stead, i.m. and rigby, v.a. 1989: Verulamium: the King Harry Lane site (London, English Heritage
Archaeological Report No. 12).
thompson, i. 1982: Grog tempered ‘Belgic’pottery of south-eastern England (Oxford, BAR British Series
108).
POTS AND PITS
OXFORD JOURNAL OF ARCHAEOLOGY
160 © Blackwell Publishing Ltd. 2005
ward-perkins, j.b. 1938: The Roman villa at Lockleys, Welwyn. The Antiquaries Journal 18, 339–76.
whimster, r. 1981: Burial practices in Iron Age Britain (Oxford, BAR British Series 90).
williams, d.f. 1981: The Roman amphora trade with late Iron Age Britain. In Howard, H. and Morris,
E.L. (eds.), Production and distribution: a ceramic viewpoint (Oxford, BAR International Series 120),
123–32.
williams, d.f. 1989: The impact of the Roman amphora trade on pre-Roman Britain. In Champion, T.C.
(ed.), Centre and periphery: comparative studies in archaeology (London), 142–50.
willis, s. 1995: Roman imports into late Iron Age societies: towards a critique of existing models. In
Cottam, S., Dungworth, D., Scott, S. and Taylor, J. (eds.), TRAC 94. Proceedings of the 4th annual
theoretical Roman archaeology conference, Durham (Oxford), 141–50.
woolf, g. 1993: European social development and Roman imperialism. In Brun, P., Leeuw, S. van der
and Whittaker, C.R. (eds.), Frontières d’Empire. Nature et signification des frontières romaines (Nemours,
Mémoires du Musée de Préhistoire d’Ile-de-France No. 5), 13–20.
MARTIN PITTS
OXFORD JOURNAL OF ARCHAEOLOGY
© Blackwell Publishing Ltd. 2005 161