A Quantitative Analysis of the Finds from the Roman Fort of Newstead – some preliminary findings

by Simon Clarke

Introduction
For the last five years Bradford University has been excavating the Roman Fort at Newstead
(Jones et al. 1989 to 1993), a site originally made famous by the discoveries of James Curie at
the beginning of this century (Curie 1911). This late first and second century fort complex is
one of the largest and most important Roman period settlements in Northern Britain (see
Figure 8.1). The most extensive of a succession of forts occupied just under 6 ha, easily room
for one thousand troops. In addition extramural occupation is now known to have existed on all
four sides of the fort, raising the overall population at the settlement’s height to perhaps two or
three thousand. As well as a major population concentration the site possessed considerable
political and economic importance, being located adjacent to Eildon Hill North, one of
Scotland’s largest hillforts, and having acted for much of the period as a major production and
distribution centre. However the settlement was relatively short lived. After an occupation of
about a century the site was abandoned. Newstead therefore provides an opportunity to examine
material diversity of a sophisticated nucleated settlement without the complication of a massive
time depth.
Military and Civilian
While it is important to avoid the assumption that soldiers and civilians represented two rigidly
separate communities the most obvious contrast within the complex was between the fort and
its extramural settlements. Firstly there was the built environment. While the buildings of the
fort were large, generally, covering many hundreds of sq. metres, those outside the defences,
with the exception of the bathhouse and mansio, were generally small. Those identified so far
mostly fall into the 30 to 100 sq. metre range The building materials too were different.
Dressed stone, even for footings is confined to the fort, bathhouse and mansio. Tile was also
probably rare outside the fort, most extramural buildings probably had thatch or shingle roofs.
However these differences should not be over emphasised. Timber was extensively used in the
fort, particularly for earlier buildings, while thatch was probably not confined to the extramural
area. Certainly the contrast cannot be explained as a simple richJpoor dichotomy, as other
aspects of material culture suggest a good deal of wealth outside the fort. Recent studies have
shown a surprisingly high proportion of brooches and intaglios to have been recovered from
extramural areas (Mc Naught Elliot & Henig 1982) (see Figure 8.2). Coins too are well
represented outside the fort (see below). The distribution of intaglios is especially illuminating
as they have come largely from fieldwalking which has examined the complex as a whole with
less of a bias towards the Fort and South Annexe. A Quantitative Analysis a/the Findsofrom the Roman Fort of New stead. 73
Newstead Roman Fort And Surrounding Crop marks
N
1
I 100 200 JOO m
Fig 8. J. Newstead Roman Fort and Surrounding Cropmarh.
(Key: J = Oak.endean Recfilinear Enclosure: 2 = Oak.endean Curvilinear Enclosure; 3 = Soulerrain;
4 = BogIe Field Recfilinear Enclosure; 5 = Red Curvilinear Enclosure; A = 1987 Excavation;
B = 1989 Excavation, RGT 33; C = 199{) Excavaf7″on, RGT 34; D = 1991 Excavation, RGT 126;
E = 1991 Excavation, RGT 101; F = 1992 Excavation, RGT 141; G = 1993 Excavation, RGT 164;
H = 1993 Excavation, RGT 165.)
Of course without knowing the precise context from which these finds came there must be
some suspicion that these relatively high levels of wealth outside the fort represent dumping of
domestic and industrial waste generated by the military. Prior to Bradford’s excavations at
Newstead I would have assumed most finds to have come not from systemic contexts, but
secondary deposits created for example by organised refuse collection and landscaping prior to
building construction. There was no doubt an element of this operating, for example a large
midden is known in the East Annexe as a finds concentration and soil mark. However there has
not been the harmonisation of the artefact record which might have been expected. The
example of the 1993 excavations either side of the main road in the South Annexe illustrate this
point handsomely. The two trenches each of about 200 sq. metres, both provided evidence of
several phases of buildings, but the character of finds recovered on either side of the road could
not have been more different. To the west finds were about ten times more common than to the
east. Though not as dramatic, highly localised differences in the artefact assemblage have been
encountered in every area examined. While finds from these deposits 110 doubt do not exactly
reflect the range of activities practised in the immediate area, it is clear that at Newstead most
debris inside the occupied area has not travelled far. 74 Simon Clarke
Brooch Distribution at Newstead
North Field 1
Baths r art I:- ast Annexe
6
31 1
West Annexe
0
South Annexe
9 after McNaught 199 3
Intagl io Distribution at Newstead
North Field 0
Baths Fort East Annexe
2
5
West Annexe 4
0
South Annexe
4
after Flliot and He nig 1982
Fig B.2. Brooch Distribution (after McNaught 1993) and intaglio Distribution (afler Elliot &
Henig 1982) in different zones at Newstead. A Quantitative Analysis of the Findsjrom the Roman Fort of New stead. 75
One of the most striking differences between inside and outside the fort was the proportion
of silver coins recovered. Bradford University’s excavations recovered nine silver and four
bronze coins inside the fort, but only six silver to twelve bronze outside the fort. This
assemblage is too small to be accepted as significant in its own right, but fortunately it is
possible to draw on the data provided by Curie. He himself noted that there were significant
variations in the proportion of bronze and silver from various parts of the complex (Curie 1911:
102). However there are some problems in using this data. Records of finds made casually have
been excluded as they are clearly highly biased towards the more exciting silver and gold
discoveries. Even coins from CurIe’s excavations present serious difficulties as the location of
only eighty-five percent of silver and fifty percent of bronze coins have been noted. This
necessitates the use of a multiplication factor to eliminate bias towards the silver issues. Once
adjusted Curie’s data suggests a similar trend to that noted in the recent excavations, with
silver slightly more common than bronze inside the fort, but outnumbered two to one outside
(see table 8.1)
Numbers Percentages
Fort Extramural Whole complex Fort Extramural Whole complex
Bronze 36 48 84 46.6% 67.6% 56.7%
Silver 41.2 23 64.2 53.4% 32.4% 43.3%
Total 77.2 71 148.2
Table 8. J. Coin Loss at Newstead: A Comparison of Fort and Extramural Areas.
(The actual number of silver coins recovered by Curie has been mulhplied by 0.6 to allow for the higher
incidence of a detailed account ofsilver’s location.)
corrected numbers Percentages
Silver Bronze Silver Bronze
FORT
defences 10.4 8 57% 43%
principia 5.3 8 40% 60%
west of reducing wall 6.5 9 42% 58%
rest of fort 19.1 11 63% 37%
EXTRAMURAL
north annexe 4.1 12 25% 75%
east annexe 2.9 4 42% 58%
south annexe 10.1 16 39% 61%
west annexe 0.6 2 23% 77%
bath house 5.3 14 27% 73%
Table 8.2. Coin Loss at Newstead by Area
With the much larger sample created by combining the data of Bradford’s excavations and the
corrected data of Curie, it is possible to break the complex down into much smaller units of
study (see table 8.2). This shows that bronze coins heavily outnumber silver except in the 76 Simon Clarke
eastern, larger part of the fort and in the fort’ s defences. This has probably resulted from a
number of factors. The troops may well have been paid in silver, so that a soldier’s purse will
have contained a higher proportion of high denomination issues than occurred in general
circulation within the economy as a whole. However it is probably also the case that the
function to which coinage was put differed between fort and vicus. Within the fort relatively
few financial transactions will have taken place, as payments for mess bills and equipment will
have been deducted at source. In such an environment coin loss might have been predominantly
from hoarded wealth (mainly silver) rather than spending money (mainly bronze). Outside the
fort a free market presumably existed causing a higher rate of small change loss. Two parts of
the fort have produced more bronze than silver: the principia and the area west of the reducing
wall. In the case of the latter the anomaly can be explained by the area’s status for part of the
period as an industrial annexe outside the fort proper. The high proportion of bronze in the
principia may have resulted from a relatively small sample size (seventeen coins), or may
genuinely reflect the use of the principia as a social venue in which small money transactions
were frequently performed.
Variations Within the Extramural Settlement
Turning now to the extramural population, if an assessment of lifestyle and cultural aspirations
were confined to a consideration of building plans and construction techniques it might well be
concluded that the community was relatively homogenolls with a low level of social
differentiation. Fieldwalki ng, excavation and geophysics have demonstrated that virtually all
the buildings were relatively small timber structures, of one or two rooms. While window glass
seems to have been widely available the structures are otherwise modest, without architectural
embellishments such as elaborate heating systems, pavements, wall paintings or even tile roofs.
But this apparent uniformity is an illusion. As already noted some trenches produced much
higher densities of finds than others. Figure 8.3 demonstrates this by plotting pottery sherd
numbers from six trenches, roughly comparable in size and intensity of excavation carried out.
2500 I
2000 -L
500
o
I
I l J
33 34 101 141 164
Trench Number
Fig 8.3. Pottery Recovered from Six Trenches at Newstead.
11 : I
I
I
I l I
165 A Quantitative Analysis of the FindsJj-om the Roman Fort o/Newstead. 77
This level of variation is not easily explained away by intensity or length of occupation as the
settlement had a short life and all areas considered had at least two phases of buildings. More
startling still is the variability of the make up of the not simply the proportion of
different types of ceramics, but also levels of bronze, lead and particularly iron loss, in relation
to other finds (see Figures 8.4-7). This is a preliminary study and it should be stressed that
I
I
25 +
!
‘” Cl
t: 15 ‘”
0; ”
Cl.
10
5
0
Object.
per 100
sherds
• Amphorae
o Samian
33 34 101 141 164 lffi
Trench Number
Fig. 8.-1. Samian and Amphorae as a Percentage of the Total Assemblage.
100
I
90 -+
80 ,
70 T
00 +
50 +
I 40 ‘;-
30 t
:tn
o ! I !
33 34
I
n I I I i iln i I I I
101 141 164 165
Trench Number
Fig. 8.5. Prevalence o.(iron Objects Relative to Coarse Ware. 78 Simon Clarke
their are a number of problems with this approach. For example, there is variability in the level
of object fragmentation across the site which must ultimately be addressed by the application of
similar studies using volume by weight. AJso in some cases there are statistical problems due to
the small size of samples. Nevertheless neither fragmentation nor chance variation seem to be
adequate to explain the high level of diversity in all instances. Clearly quite distinctive lifestyles
generated some of these differences.
70
60
Coin. 50
per 1000
sherds 40
30
20
10 I
____ ______ ________ -r ________ ____
33 34 101 141 164 165
Trench Number
Fig. 8. 6. Prevalence o/Coins Relative to Coarse Ware.
20 I
18 • bronze
,
16 o lead
14
12
Objects
per 1000 10
sherds
8
6
4
0
33 34 101 141 164 165
Trench Number
Fig. 8.7. Bronze and Lead Objects Relative to Coarse Ware. A Quantitative Analysis of the Findsfram the Roman Fort of New stead. 79
RGT 34
The most exceptional assemblage outside the fort has come from trench RGT 34, in the East
Annexe. In terms of quantity it was a relatively poor year, but this was more than made up for
in the quality of finds recovered. One of the most striking features of the assemblage was the
very high proportion of iron work, mostly nails. This is perhaps best explained by thinking of
the artefacts as having derived from two separate processes. Firstly the built environment,
which in the absence of tile consists mainly of bolts and nails, left after timber structures have
decayed. Secondly the occupation deposit; everyday items lost or discarded and not removed
from the area. As the volume of structural debris relative to occupation debris was very high, it
must be surmised that either the building was unusually elaborate or occupation debris was not
allowed to accumulate. What was left behind suggests a relatively high degree of lUXUry. Over a
quarter of the pottery was samian and there were almost as many sherds of vessel glass – more
than from all the other Bradford University excavations at Newstead put together. There were
also an unusually large number of decorative glass items: gaming pieces, melon beads and
bracelets. The numbers of coins and other bronze and lead objects were also very high
considering the small size of the total assemblage. This then was a wealthy area, generally kept
clear of debris. Interestingly the midden site previously mentioned is less than 70 m to the east.
RGT 33 and 165
In contrast trenches RGT’s 33 and 165, in the South Annexe, seem to have been for much of
the Roman occupation remarkably squalid environments, which saw the rapid accumulation of
Newstead Project 1993 Phases 3 to 5
+ + +.
N
t • •
0
•
0
[ •• IV OUeh
+ +
P,o l.c: ted SIde ROAd
\ \ I
Fig. B.B. Newstead Project 1993, Phases 3 to 5.
(Key: 1 = Unphased Gullies; 2 = Unphased Metalled Surfaces; 3 = Phase 5 Ditches;
4 = Phase 4 Gully; 5 = Unphased Post Hole; 6 = Phase 3 Post Hole; 7 = Phase 4 Post Hole
8 = Phase 4 Cobble FOlUldation; 9 = Curie’ s Excavation Trenches)
1
.
,
.
, 80 Simon Clarke
debris in and around a series of timber strip buildings (see Figure 8.8). These are not
necessarily low status areas as the structures themselves appear to have been soundly built, in
some cases probably with wooden floors and window glass. Some aspects of the finds record
also imply a fajr degree of more portable an intaglio and a silver fitting off a piece of
high class furniture were recovered from RGT 165, while both areas saw a significant level of
coin use. On the other hand samian made up a relatively small part of the pottery recorded and
only a handful of vessel glass sherds were present.
RGT 141 and 164
The other two excavation areas in the South Annexe seem to have been kept free from debris.
RGT 141 and 164 both lay on the east side of the more easterly of the north-south roads (see
Figures 8.8 and 8 9). Unlike RGT 34 however the number of iron objects recovered from these
areas was relatively modest. Probably the use of nails in the buildings of these areas was more
restricted than in the East Annexe as other aspects of the material culture suggest a lower level
of wealth. The most striking difference was the almost total absence of non-ferrous metals,
including coins. While the presence of imported ceramics and other manufactured goods
implies participation in the money economy (cf. Griffiths 1989), coins were clearly not handled
with the same frequency as in other parts of the complex. The absence of fittings and jewellery
of bronze and glass (except for a melon bead recovered from RGT 164) suggests that there was
+
+
+
Newstead Project 1992 Excavations South of the Roman Fort
O !l 15
– – =
+ +
,oc SON
N
+ Key
= Mod.r n Foul”, …
o C: Cr””., .. tlo>n,
• Orip Uno
L Sulldlng O,,\Hnlf
+ . POll Hole
·;tl CObbl. Of” Cr ….. 1 Surr.ce
tOE “ON+
Fig. 8.9. Newstead Project 1992. Excavations south of the Roman Fort. A Quantitative Analysis a/the Finds/ram the Roman Fort o/Newstead. 81
much less in the way of personal adornment. On the other hand the level of samian recovered
from both areas is paradoxically higher than any area except RGT 34. Again it is necessary to
remember that elements of society cannot be described as simply rich or poor. Perhaps this part
of the Newstead population invested a disproportionately large percentage of its income in
status display at the dining table rather than on personal adornment. As coin use appears to
have been low and evidence for industrial processes is lacking it may well have been that these
areas were more rural than urban in outlook. The discovery of an animal pen in RGT 141 reenforces this impression. The absence of debris on the site may therefore relate more to the
practice of manuring the surrounding field system than any particular scruples about
cleanliness.
Conclusion
In conclusion by considering the artefact record as a whole rather than breaking it down into
specialist fields it is possible to observe archaeologically what historians have known for
that Roman period society was riven by petty status distinctions and ethnic
divisions. This has tended to be obscured in the past by the much more obvious distinctions
between Roman and native. On Britain’s northern frontier the gulf between the material
cultures of these is so great as to make even the distinction between military and civilian, seem
petty. In reality it can now be seen that there is no such thing as Roman society, or Roman
culture, but rather many sub-cultures each with its own aspirations and values. At Newstead for
instance not only were there marked differences between the fort and extramural areas, but also
within the predominantly civilian population between high and low status, agriculturalist and
urbanite. Even though the buildings of the extramural settlement display a high degree of
uniformity still more subtle distinctions could no doubt be identified with further analysis. This
has implications far beyond this one settlement. While large for northern Britain Newstead is
modest in scale when compared to the cities of southern Britain and the Continent. Any
substantial Roman period settlement can be expected to have possessed considerable variations
in its material culture. Studies which fail to take this into account risk generating profoundly
misleading results. Milieu, for example, has used the proportion of samian recovered from sites
in southern Britain to suggest that a saturated market existed (Millett 1990: figure 54). But the
level of variation in samian use between these partially excavated sites is no more than between
the finds groups within Newstead. To consider the assemblage of a whole site without regard to
possible zoning of activities would be considered laughable in environmental archaeology,
where thoughtful modelling of the formation processes is now routine (cf. Jones 1986: figure
28) Most non-environmental archaeologists are familiar with the reasoning behind this
methodology. It is therefore extraordinary that almost without exception specialist reports,
especially coin reports, fail to comment on and attempt to explain artefacts’ associations with
particular contexts and other classes of finds. This is much more than a question of assessing
how representative a finds group is of the site as a whole, as intra site variability need not
always be seen as a pitfall. To take just the example of coins, differences in the distribution of
silver and bronze at Newstead may well give us an important new insight into the different
roles that coins perform, as well as emphasising social differences within the complex. If
repeated across Britain we might be in a much better position to discuss money supply and
circulation. 82 Simon Clarke
Acknowledgements
The Newstead Research Project has been jointly supported by the National Museums of
Scotland, Bradford University, the Borders Regional Council and the British Academy. My
thanks go to my fellow researchers within the project who have generously allowed me to use
data belonging to the group as a whole, much of which is not yet published. I am particularly
grateful to Rick Jones and Pete Rush for their comments and encouragement in the preparation
of this paper.
Bibliography
Curie, l. 1911. A Roman Frontier Fort and ils People. Edinburgh University Press. Edinburgh.
Elliot. W. and Henig, M. 1982. Engraved Gemstones from the Roman Frontier Post at Newstead,
Roxburghshire, Proceedings o/the Society 0/ Anliquaries o/Scolland 112: 295-99.
Griffiths, K 1989. Marketing of Roman Pottery in Second Century Northamptonshire and the Milton
Keynes Area, Journal 0/ Roman Pottery Studies 2: 6fr 77.
lones, M. 1986. England Before Domesday. Batsford. London.
lones, R.F.J., Clark, K, Clarke, S. and Rush, P. 1989. The Newstead Research Project. 19891nlerim
Report. Bradford University. Bradford.
lones, R.F.1., Cheetham, P., Clark, K. Dent, 1. and Rush, P. 1990. The Newslead Research Project. 1990
lnlerim Report. Bradford University. Bradford.
lones, R.F.J., Cheetham, P., Clark, K. Clarke, S. and Dent. 1. 1991. The Newslead Research Project.
1991 Field Season. Preliminary Report. Bradford University. Bradford.
lones, R.F.J., Cheetham. P .. Clark. K , Clarke, S. and Dent. 1. 1992. The Newstead Research Project.
1992 Field Season. Preliminary Report. Bradford University. Bradford.
lones, R.F.1., Clarke. S., Clark, K., Dent. 1. Lucy. D. and Cheetham, P. 1993. The Newslead Research
Projecl. 1993 Field Season. Preliminary Reporl. Bradford University. Bradford.
McNaught. R. 1993. A Catalogue and Sludy 0/ the Roman Copper Alloy Brooches from Ihe Fort
Newstead. Unpublished BSc Thesis. Dept. of Archaeological Sciences. Bradford University.
Millett, M. 1990. The Romanization o/Britain. Cambridge University Press. Cambridge.