Professional Documents
Culture Documents
Housing in The Military Vici of Northern Britain: James Hunter
Housing in The Military Vici of Northern Britain: James Hunter
Housing in The Military Vici of Northern Britain: James Hunter
Northern Britain
James Hunter
12,459 words
September 2013
Contents
1. Introduction 2
2. Methodology 4
3. Literature review 6
3.1. The evidence available
3.2. Housing
4. The military vici: 8
4.1. Housesteads 8
4.1.1. Introduction to the site
4.1.2. The evidence
4.1.3. Discussion
4.2. Greta Bridge 19
4.2.1. Introduction to the site
4.2.2. The western vicus
4.2.3. The eastern vicus
4.2.4. Discussion
4.3. Maryport 25
4.3.1 Introduction to the site
4.3.2. The problem of size
4.3.3. The evidence
4.3.4. Discussion
4.4. Vindolanda 32
4.4.1. Introduction to the site
4.4.2. Vicus I
4.4.3. Vicus II
4.4.4. Discussion
4.5. Observations from other sites 41
5. Conclusion 45
Bibliography 49
Appendix 1: Tables 56
Appendix 2: Figures 64
1
1. Introduction
“In many cases important new insights are achieved simply by asking the right
questions of the existing material”. 1 Indeed, in this paper I intend to ask new
neglected until recent years. 3 By asking the question of how houses differed in form,
size and function, the paper will be an extension of the research framework outlined
in 2009, 4 which focuses more on civic and religious structures, and vicus layout. The
primary reason for such neglect has been a general scholarly focus on the army and
their forts, especially in the frontier zone of Britain. The information that vici
housing can provide is crucial for assessing social structures in the extra-mural
military setting. Studies along these lines have been undertaken on isolated sites, but
their results have not yet been collated for comparison. By using the archaeological
evidence, we can explore in greater detail than before what the difference between
houses and other domestic structures meant in terms of how we view two things in
brush over the intricacies of housing in the military setting (except for barracks). I
shall use the trends that appear, in terms of quantity and form, in order to elucidate the
The term ‘military vicus’ itself can be quite easily defined for the purposes of this
paper. Legally, for a settlement to be labelled a vicus it required its own independent
1
Breeze 2013, 7.
2
The term ‘vicus’ (and derivatives) is not italicized in this paper.
3
The study is focused in northern Britain. Where appropriate, comparisons will be made with sites
elsewhere in the Roman Empire.
4
Symonds & Mason 2009, II. 15.
2
council. In practical terms, the label “tends to be applied to the settlements outside
forts which provided accommodation for people who were attached to the military in
these forts by family ties or trading links”. 5 It should be added that military vici
outside legionary fortresses were labelled canabae (but could be subdivided into
smaller vici). 6 These distinctions, between Roman law and practical use, and between
legionary and auxiliary terminology, led Sommer to use the term ‘military vicus’ in
order to avoid confusion. 7 Therefore this is the term that will primarily be used in this
paper. 8
This paper is not intended as an exhaustive survey of every military vicus in Britain.
single province.
5
Allason-Jones 2013, 72.
6
Sommer 1984, 3: Salway 1981, 9.
7
Sommer 1984, 4.
8
Unless otherwise noted, the definition of ‘vicus’ in this paper will be that of ‘military vicus’.
3
2. Methodology
In order to keep this study manageable, the sites I have chosen needed to first of all
adhere to the definition of ‘military vicus’ set out in Chapter 1. They also had to
provide evidence for significant occupation during the 2nd and/or 3rd centuries AD.
geophysical survey, had to be very good. That is, a site needed to provide a sample of
at least six ‘houses’ for which a reasonable estimation of all dimensions could be
made. Sites that fulfil these criteria are quite rare in the prescribed region. However,
Houses also needed to provide some evidence of their function. First, and most
simply, dimensions and internal layout can be used as a preliminary indicator. 9 The
materials used for flooring can also suggest what a building or room was used for.
Next, the presence (or absence) of certain features can often signify a particular
function. 10 In particular, any oven equal to or greater than 1.52 metres in length will
be classed ‘industrial’ rather than ‘domestic’; 11 that is, used for commercial purposes
rather than simple domestic tasks. In the same vein, the area of a ‘domestic’ hearth
was probably between 0.8 and 1.5 m2. 12 Anything larger will be classed ‘industrial’.
Finally, the assemblages excavated in a specific building (of pottery, for example) can
provide evidence for what activity was taking place. All of these criteria will be used
to investigate the social structure of a vicus, and determine the type of people that
9
Unless specified, measurements refer to external dimensions. Where dimensions were not available, I
have produced estimations based upon scale drawings and plans.
10
These include sills, shrines and wells.
11
Based on the oven in Housesteads Building 1 (Birley et al. 1932, 227); and in Vindolanda Building
27, which was 1 m2 (Birley 1977, 38).
12
Based on specific evidence from Vindolanda (Birley 1977, 41): two hearths in Building 23, with
areas of 0.8 and 1.5 m2.
4
might have lived there. 13 They will also aid an assessment of the wealth of its
inhabitants.
For each case study, I have tabulated all the relevant archaeological evidence,
including statistical analysis of dimensions and the number of rooms in a house. This
data has been obtained primarily from excavation and geophysical survey. The
excavation to obtain an impression of the overall extent of the site. Furthermore, with
shall discuss what the significance of the findings might be. However, reliant on
publication history is not sufficient for the purposes of an in-depth study of vici
housing. As a result, I have sometimes been forced to make speculation based on the
13
Evidence for a collective name – vicani – at: Carriden (3503), Housesteads (1616), Leintwardine
(3117), Old Carlisle (899), and Vindolanda (1700).
5
3. Literature review
Throughout much of modern scholarship, vici have been subject to very little in-depth
study. 14 The limited resources available to early 20th century archaeologists in Britain
meant that any excavation that did take place was guided by the research priorities of
individuals, which often tended towards the aspects related most closely to wider
issues of military and political history. Consequently, the scholarly spotlight fell
largely upon the archaeology of forts and the more ‘Romanized’ extra-mural
issues and a lack of resources have meant “the only substantial vicus excavation in
modern times has been at Vindolanda”. 16 However, in recent years progress has been
made in the use of geophysical survey to explore further the nature and layout of vici
in Britain. 17 As a result, scholars have been able to build upon Sommer’s analysis of
Romano-British sites and begin to compare vicus with vicus in more detail. 18
3.2. Housing
when it comes to assessing the development and range of services available within
these settlements, as well as their wider economic role”. 19 This statement suggests
that the study of vici retains the primary aim of examining their relationship with the
fort itself and the frontier economy. Accordingly, vici housing has generally been
14
Burnham & Wacher 1990, 15.
15
The vicus at Castleshaw is a case in point (Redhead 2009, 80).
16
Symonds & Mason 2009, I. 65. Excavation is ongoing at Binchester vicus, however:
https://sites.google.com/site/binchesterromanfort.
17
England: Biggins & Taylor 1999; 2004a-c. Taylor et al. 2000. Wales: Hopewell 2005; 2006.
18
Sommer 1984; 2006.
19
Symonds & Mason 2009, II. 14.
6
used to demonstrate the wider commercial nature of such settlements. 20 The degree
of variability of the housing in each vicus has been brushed over, and its social
great detail. 22 In no case has a scholar attempted any significant assessment of what
There has also not yet been any in-depth study of variation within individual
found in a house are rife, but analysis of what their function might have been is rare;
comparative analysis is even rarer. By using the data assembled, I shall identify more
precise and quantifiable differences within vici housing. Instead of focusing on the
economic terms, I shall explore how the people of the vici might have actually lived,
putting the frontier population under a social microscope. How much space did vicus
residents have, both for domestic and commercial purposes? To what extent might
civilians have undertaken industrial activity? Would a vicus have been overcrowded?
population sizes. 24
20
Symonds & Mason 2009, I. 65.
21
E.g., Sommer 1984, 48: “There is no space here to discuss their various methods of construction,
appearance and internal layout”. Davies 1984: the nature of vici housing is mentioned only briefly.
22
Simplistic statements regarding the strip-house being the most common form of housing in vici:
Higham 1982, 111; Symonds & Mason 2009, I. 65.
Regarding the existence of large structures elsewhere: Symonds & Mason 2009, I. 65.
23
Sommer (1984) collates the dimensions of houses, but as part of an appendix that provides only basic
observations.
24
Birley 1973, 15f.; 1977, 72. This view has been contested by Sommer (1984, 30-1).
7
4. The military vici
Of the sites I have identified for further study in Britain, two are from the northern
‘frontier zone’ (that is, in the vicinity of Hadrian’s Wall and the Stanegate):
Housesteads and Vindolanda. The remaining two are from farther south: Maryport on
the Cumbrian Coast and Greta Bridge in County Durham. I shall also address in less
detail the pertinent features of Halton Chesters and Birdoswald. The sites outlined
above offer significant evidence about housing in vici, both through their geophysical
4.1. Housesteads 25
The vicus 26 at Housesteads was situated immediately to the south of the fort and
covered an area of approximately 1.92 hectares. 27 The first excavations at the site
were carried out in the 1930s, 28 with further work being undertaken in the 1960s.
Excavations completed between 1974 and 1981 have only recently been published. 29
Much of this work concerns the fort itself, however; scholarship that directly
addresses the vicus is less common and not very detailed. 30 For example, it was once
felt “enough to mention that the buildings had in general proved to be simple
survey has since provided a better idea of the extent of the vicus, and of the buildings
therein, 32 but the very fact that Housesteads has such a rich publication history has
25
All building numbers in this section are in accordance with Figs. 1 & 2.
26
RIB 1616: ‘… d(ecreto) vica[norum]’, for a suggestion of a communal vicus identity.
27
See Table 1 for a brief chronology of the site.
28
Birley et al. 1932-3; Birley & Charlton 1934; Birley & Keeney 1935; Birley 1937.
29
Birley 1961 & 1962; Rushworth 2009, I.
30
Birley & Keeney 1935; Birley 1961 & 1962; Crow 2004, 73-88; Rushworth 2009, I. 233-68.
31
Birley & Keeney 1935, 247.
32
Biggins & Taylor 2004a.
8
meant that assumptions made long ago regarding the identity and function of
enable us to at least acknowledge the alternatives, even if we ultimately agree with the
prevailing opinion.
structure (Table 2). Likewise, the internal arrangement of the buildings was varied.
a basement room was later created by digging down to the foundations at one end;
this was filled in and an oven built in the northeast corner of the building. Building 2
contained an internal partition 3.05 metres long running from west to east, and it too
contained a hearth, this time raised and in a central position. The small statue of a
genius discovered on the site was perhaps used to ensure the owner’s safety and
commercial prosperity (Fig. 4). 36 The northern end of Building 3 was open, and
indicate that, in its first phase, it was subdivided into at least three rooms, with a box
furnace for metal working in the most southerly. 37 In its second phase it was
converted into a single, long room. Building 5 was an unusually square structure
33
E.g. ‘The house of the beneficarius consularis’, labelled thus by Birley & Keeney (1935, 254-5).
34
= area contained by the red box in Fig. 3.
35
All buildings at Housesteads studied in this paper date to the 3rd century. The existence of more
buildings is strongly suggested by recent geophysics (Biggins & Taylor 2004a).
36
Crow 2004, 76. For the use of genii in the domestic setting, see Henig 1995, 168-9.
37
A similar furnace was found in the fort, in barrack 6 (Crow 2004, 77).
9
whose location near the fort may well indicate an official function. 38 Building 6 was
probably related in function, and perhaps similar in layout. Buildings 7 and 16 were
quite close to the fort and considerably larger than any others in the immediate
vicinity. The upper courses of the west wall of 7 (the other walls were subject to stone
robbing) were constructed using “finely dressed ashlar blocks and measured 6 feet by
importance. 41
Also perhaps of some significance was Building 8 which, despite being further from
the fort, displayed high quality masonry in its street frontage. This might indicate that
the owners wished to convey an image of wealth or high status. Buildings 12 and 14
both had an open western end, and their masonry seems to have provided foundations
Buildings 19A, B and C are a set of structures identified by geophysics. 42 19A, close
to the fort, might well have been an official building; on the other hand, like 1, 2, 3
and 4, it might have been a dual-function workshop/house. 19C seems to have been a
38
Birley & Keeney (1935, 254) suggest it was a storehouse, which seems likely considering its position
close to the fort and the road.
39
Birley & Keeney 1935, 253.
40
Crow 2004, 79.
41
Birley & Keeney 1935, 254; Crow 2004, 79.
42
Biggins & Taylor 2004a, 56.
10
What emphasises the diversity of the vici housing is the set of smaller buildings, 19
and 22-27. They have been described as “hovels” within a “shanty town”. 43 The
buildings are comparable in size to the ‘chalets’ in the fort, 44 but each has a highly
4.1.3. Discussion
It should be noted first that the buildings closer to the fort and/or on the main
they were the first to be constructed as the vicus developed, and therefore had a
longer and more elaborate construction history. The buildings possibly constituted
‘prime retail space’, closest to the source of wealth – the soldiers of the fort. 45 Any
direct challenge, not based on problems of dating, to the theory that a more prominent
location bestowed more importance on a building remains unknown to the author. The
latest geophysical evidence further supports the claim: 19A, B and C were of a
substantial size and next to a main road that probably led to Vindolanda. 46 The
exception to this ‘rule’ is 7. Despite being set back from the main road, it was a large,
well-built structure, a feature which, along with its relationship to 16, 47 hints at a
greater importance. 48
It has been suggested that the considerable size of 7 and 16 indicates an official
function. 49 However, it need not always be the case. Although Maryport Building 1
43
Birley & Keeney 1935, 252; Breeze 1982, 150.
44
Crow 2004, 79.
45
Rushworth 2009, I. 295.
46
Birley & Keeney 1935, 227.
47
Ibid., 255. That the two were connected in function is supported by their close structural relationship
(Birley & Keeney 1935, 253).
48
Birley & Keeney 1935, 254; Crow 2004, 79.
49
Crow 2004, 79.
11
was probably 50 much larger than Housesteads 7, it was situated around 80 metres
from the gate of the fort and was a constituent part of a series of similar-sized
buildings, none of which were likely to have an official function. 51 The function of 7
is therefore unclear. Two suggestions have been made: first, that it was the office of
the beneficarius consularis; 52 second, that it was the headquarters of the Frisian
auxiliaries stationed at the site. 53 The former theory is more convincing, if only
because there is no evidence to locate the Frisians in 7. It has also been argued that 16
was the house of the beneficarius consularis. 54 However, the lack of significant finds
means we cannot assign a singular function to each building with any confidence. 55
Indeed, we could quite easily assume that 7 was a house and 16 an office.
Nevertheless, it seems likely that within the vicus there was at least some interaction
between high-ranking military official(s) and the civilian population, and it was
probably not the case that the army was restricted to the fort and the occasional
military annexe. Not only did soldiers purchase goods from civilian merchants, but
the domestic structures. At Housesteads, more than one layout could fit a single
function, and likewise more than one function could fit a single type of layout.
arrangement was usually the most obvious indicator that it had two separate functions.
50
The northern limit of Building 7 is uncertain, but the existence of the road and Buildings 5 and 6
suggest it would never have reached 30 metres in length.
51
Biggins & Taylor 2004c, 114, fig. 5.9.
52
Birley & Keeney 1935, 255. One of these officials at Housesteads was Litorius Pacatianus, who
dedicated an altar to Mithras (RIB 1599).
53
Crow 2004, 80.
54
Birley & Keeney 1935, 255.
55
For high-ranking individuals, a dual role for domestic accommodation was very likely (cf. the
praetorium at Vindolanda (Birley 2009, 149)).
12
The main entrance into Building 2 comprises two stone sills with slots typical of
aspect to the house. Similarly, 3 had an open end facing the street.
However, the layout was not the only way in which a commercial aspect could
manifest itself. The raised hearth in the centre of 2 was not a typical feature of vici
comparable in size to industrial ovens found in the fort, and the great deal of pottery
found in the building’s basement, may well be evidence of the commercial production
of ceramics. 57 Since these two buildings were very near the fort, it is likely both were
shops that served soldiers, and perhaps other civilians. That commercial activity took
place in the vicus is confirmed by the box furnace discovered in 4, which was
probably used to produce small change for soldiers and traders to use. 58
It is not clear that these commercial buildings, which also include 8, 12 and 14,
with regard to production. 59 Indeed, there is little other evidence for industry within
the vicus, 60 and to describe the site as “more a centre of exchange than a place of
56
Crow 2004, 76; for a study of taberna structures in Britain, see MacMahon 2003.
57
Crow 2004, 76.
58
Ibid,, 77. Coin moulds for this purpose have been found between Buildings 3 and 4, and just north of
Chapel Hill.
59
Casey 1982, 130.
60
Birley & Keeney 1935, 240f.
61
Ibid., 241.
13
What the function of 19 and 22-27 was is uncertain. Even if it were true that
regularity in building programmes was rare at British vici, 62 this jumbled collection
of which it consists were laid out apparently with no thought for practicality in terms
Since the buildings were comparable in size to the fort chalets, we should consider the
possibility that they were ‘overflow’ houses for soldiers. The increase in the size of
the garrison during the 3rd century with the potential influx of two hundred Frisian
auxiliaries would certainly have meant extra space for accommodation was required.
63
The unearthing of ‘Housesteads Ware’ in the vicus excavations, which has “very
close similarities to the products of Frisia”, indicates that the inhabitants of this area
were most likely to have been soldiers. 64 The irregular dimensions of the buildings
and the inferior quality of their construction might at first suggest that they were not
designed by, or intended for, ‘Romanized’ residents. 65 However, the fact that the
further south at Chapel Hill, dedicated to the Roman/Frisian hybrid god Mars
existed within the cult of Mars across Britain. 67 Further geophysical survey has
shown that this type of poor quality building might well have existed further south
62
Sommer 1984, 24-5.
63
Crow 2004, 65. RIB 1593: … Ger(mani) cives Tuihanti / cunei Frisiorum / Ver(covicianorum)
Se(ve)r(iani) Alexand/riani votum / solverunt / libent[es] / m(erito).
64
Crow 2004, 80.
65
Birley & Keeney 1935, 252: “[the walls of buildings 24, 25, and 26] invariably consist of a single
line of stones … reminiscent of [those] which often form the foundation of a hut circle on a native
site”.
66
Crow 2004, 85.
67
Hunter 2013.
14
and southwest of this excavated area, 68 therefore this section of the vicus population
The varying internal arrangement of these buildings, which are externally relatively
similar, can provide information regarding who built and/or commissioned the
dwellings. At first glance, the planned street system and methodical arrangement of
opinion the differing internal layout of the houses so far discussed suggest that the
civilian inhabitants had a say in how their house would eventually be laid out. The
annexe; 4’s original tenants were perhaps a larger family requiring more separate
rooms or space for industrial work, 71 whereas the later residents might have been an
their requirements. These structural changes provide an insight into how the “tenural
[sic] manner in which the buildings were held by inhabitants” could manifest itself
structurally. 72
In order to better envisage the vicus population, it would be useful to estimate the
68
Biggins & Taylor 2004a, 56, with fig. 3.
69
Birley and Keeney 1935, 216. Buildings 19 and 22-27 are an exception.
70
Birley 2009, 161. Indeed, this is more likely if, as is probable, a military vicus was established in
parallel with the settlement of an increased garrison (probably the First Cohort of the Tungrians in the
early 3rd century (Crow, 2004, 64)) at a new fort (Birley 1977, 70; Sommer 1984, 9).
71
Note the box furnace for metalworking (Birley et al. 1932, 230; Crow 2004, 77).
72
Casey 1982, 126.
73
Here, ‘settlement’ = the area contained within the red box in Fig. 3. Those buildings to the east of the
fort were probably military.
15
attempted to do so, 74 producing a figure of 2,000 for the entire vicus, a population as
dense as it was in the fort. 75 To find space for this many people, Birley had to accept
the existence of ‘married quarters’ and suggested also that several ‘families’ lived in
the same house. However, in light of the evidence for more housing that has been
For the purposes of this calculation, it shall be assumed that all houses in the vicus
were occupied at the same time and that, given the lack of evidence to the contrary,
houses only had one storey. 76 Sommer notes that probably four or five people,
reasonably assume that this was the maximum number who would have lived in a
accept that some soldiers resided in 19 and 22-27, then these buildings could have
accommodated a greater number. Therefore I shall also be working on the basis of the
following reasoned assumptions: first, the largest strip-buildings (8, 17 and 18) could
house a maximum of five people (inclusive of slaves); second, conditions were more
cramped in the ‘hovels’, which could therefore accommodate more people than
perhaps was healthy; finally, 16 was the home of the beneficarius consularis.
With these assumptions in mind, I have come to an approximate total for the
(Table 3). I have taken what I believe is the more sensible approach and, instead of
74
Birley 1973, 15.
75
Sommer 1984, 33.
76
cf. the unfounded assumption of Crow (2004, 77): “many of which probably had upper floors”.
77
Sommer 1984, 78, n. 79.
78
Strip-houses at Zugmantel often reached 30 metres in length (Sommer 1988, Abb. 10).
16
using the number of married soldiers as a starting point, 79 I have first made an
estimation of how many people each house could reasonably accommodate. 80 This
method has provided a more realistic final figure. The results suggest that the
population of the vicus in the 3rd century, when it was at its peak, was approximately
230. This is a considerably lower estimation than has previously been suggested, but
it is at least based upon the available evidence, from Housesteads and Zugmantel,
married. If we take into account both the military area to the east of the fort (with the
likelihood that any soldiers accommodated there would be in less than luxurious
conditions), and the possibility of timber homes that no longer survive, we might
Such figures would suggest that Housesteads at its peak was a bustling vicus, but
almost certainly not overcrowded. The evidence does not suggest the existence of any
heavy industry, but we might imagine traders and shop-owners living (and
and his fellow soldiers, who probably provided their most significant source of
income. With the influx of Tungrians and Frisian soldiers, we must imagine a
culturally diverse society, which also attracted civilians from far-flung provinces such
as Upper Germany. 82 Nevertheless, the widespread use of the Roman institution that
79
cf. Birley 1973, 16.
80
I have disregarded Birley’s assertion that a strip-house accommodated at least seven people (2000,
40).
81
See Crow (2004, 76), for the discovery of dice in Building 1.
82
RIB 1620: “… heres procuravit Delf/inus Rautionis ex G(ermania) s(uperiore)”.
17
was the strip-house suggests that the society was ‘Romanized’ to a certain extent, or
at least hybridised. 83
83
The genius from Building 2 and the relief of genii cucullati from the shrine in Building 9 are
examples of the practice of traditional Roman religion at Housesteads. The latter, in which the figures
bear the byrrus Britannicus, represents a specifically British form of the genius.
18
4.2. Greta Bridge 84
South of Hadrian’s Wall and the frontier zone was the small fort of Greta Bridge.
Until the 1970s, very little excavation work had been completed in either the fort or
the extra-mural areas. 85 The rescue excavations that took place in 1973-74 as a result
of impending road-works were briefly summarised, but not fully published until 1998.
86
The vicus itself lies north of the fort, and is a ribbon development approximately
600 metres in length, from west to east (Fig. 5). Prohibited by the requirements of the
road builders, the central section has been left unexcavated. On either side are the
That the settlement had the status of vicus, or was at least referred to in such a way, is
suggested, but not confirmed, by an inscription found at the site. 87 Buildings have
been identified from Phase II (late 2nd to the third quarter of the 3rd century) and from
Casey and Hoffmann are not clear in their dating of Buildings 1 and 2. The footings
were apparently cut in Phase IIB, which roughly corresponds to the early 3rd century,
however they are categorised as Phase IV buildings. The latter date is most likely
since it ties in with the date of the stone buildings in the eastern vicus, which suggest
a planned redevelopment of the settlement in the late 3rd century. 1 was subdivided
into three sections by internal cross-walls: the back room had dimensions of 6.5 by
84
Table 4.
85
Casey and Hoffmann 1998, 111-2.
86
Summaries: Wilson 1974, 413-4; 1975, 235. Report: Casey and Hoffmann 1998.
87
RIB 749.
88
Phase III is “only represented by the recovery of usable artefacts” (Casey & Hoffmann 1998, 120).
19
4.5m, the central room 13.5 by 4.5m. The front room was the same width but seems to
have been open-ended. Nothing remains of the flooring, and there was no hearth
excavated. No suggestion of function has previously been made, but it seems likely
that it was a dual-function structure. The rear room was almost certainly reserved for
accommodation. The central room probably also held a domestic function, since, with
no industrial hearth or oven having been excavated in the room, it is unlikely that it
was used for ceramic production. With no small finds having been uncovered,
optimism is required if we are to imagine another scenario, in which the central room
was used for the production of items that left no trace. The front room, however,
Building 2 contains at least one internal cross-wall and therefore might have been of
Alternatively, it might have been more like 1, with a tripartite division. Indeed,
judging by the overall width of 2 (Table 4), its front room might well have been of
accommodation were probably in the same building. The rear room, which was
probably the living quarters, 91 had a clay floor. The northern half, presumably the
commercial section, used for sales and/or production, had a flagged floor. Buildings
89
Section 4.1.2. (buildings 8, 12 and 14).
90
Ibid.
91
Casey & Hoffmann 1998, 123.
20
B1 and B2 exhibit the same phenomenon, and at one point were surrounded by a
cobbled courtyard.
Building D probably had a clay floor and, with no evidence for industry and no open
porch fronting the road, most likely acted solely as a house. Next door, Building E
possibly had an open porch facing the road and a loading-bay at the rear. 92 A
partition wall was constructed a third of the way back from the frontage. It contained
at least some flagged flooring, which lined a passage leading from the rear room to
the front of the structure. Building G had a clay floor, whilst Building F had a
4.2.4. Discussion
Notably, many of the buildings at Greta Bridge (Table 4) are considerably longer than
those at Housesteads (Table 8). The reason for this is unclear, but it is possible that
they had to house more people. More likely, however, is that they fulfilled a greater
of the buildings at the site provide any strong evidence of industrial activity, either in
There is evidence, however, that different types of flooring could denote the function
of a room or building. Clay was generally used for accommodation, while flagged
floors were used for industry. This suggests that, in the western vicus, the two
probably more common for these functions to be separated into two separate
92
Casey & Hoffmann 1998, 122.
93
Ibid., 131.
21
buildings. This observation raises a question of ownership, since the possibility arises
that E and F acted as the industrial premises of the same men who lived in D and G
respectively. The fact that the domestic halves of the arrangement (D and G) differed
so greatly in size must be a result of personal preference on the part of the owner.
We can also map this theory onto 1 and 2 in the western vicus and consider the
flooring that might have been used. 1 probably had an open-fronted room, and so
most likely the commercial building of the pair. 94 We might, therefore, expect it to
have had a flagged floor. 2, then, would have been accommodation, with a clay floor.
That the two buildings were set up at the same time is made more likely by the
existence of a paved alleyway between them. Such a feature has been suggested as
evidence for the existence of side doors – perhaps the two buildings shared the same
On the other hand, if the pattern in the eastern vicus cannot be mapped onto the
western, the general trend at Greta Bridge becomes one in which a building’s rear
room had a clay floor; therefore we might reasonably assume that 1 followed such a
trend. Likewise, we might expect the central room to have had a flagged floor and the
open porch a clay floor (similarly to AZ), or the entire building to have had clay
These more sophisticated living and working arrangements were possible at Greta
Bridge because it seems to have been a later planned development, and therefore
owners could more carefully draw up structural plans. At other more cramped vici, it
94
Evidence is scarce, but a chisel was discovered in the building (Casey & Hoffmann 1998, 137).
95
Casey & Hoffmann 1998, 122.
22
was often the case that owners had to ‘take what they could get’ in terms of the plots
available and their size (although within that plot they could design the internal layout
of the building). It was less likely that they would have been able to secure a space
performing two functions. In more spacious vici, like Greta Bridge, owners had more
choice. The layout of AZ, B1 and B2 must once again be down to the preference of
the owner, who presumably wished to adhere to the trend that existed across most vici
in northern Britain, as well as express his status and wealth. This element of choice is
made clearer by the fact that, even though these three buildings are ‘typical’ strip-
houses, they are certainly not identical in dimensions, and the classic design must
surely have been adapted according to the wishes of the owner. It is thus unwise to
The large plot sizes and, as far as we can tell, generally less cramped nature of the
vicus seem to have allowed owners to modify their buildings to better suit the
required function. It must be noted that there is no absolute proof that there was a
great deal of commercial or industrial activity at the site, since large ovens, large
hearths and finds of an industrial nature are few and far between. 97 But the layout of
E, in which a flagged passageway provides easy access to the back of the house from
the front, is rare in British vici and, I suggest, an example of shrewd design to
facilitate an industrial function. The rear section was probably a workshop, whose
supplies were transported up and down the passageway. In doing so, industry was
kept separate from the accommodation, which was situated in the partitioned area at
the front of the building. Such an intuitive use of space has not been found at
96
Casey & Hoffmann 1998, 131.
97
Ibid., 156: an oven and two querns in Building B1 do, however, suggest that a “bakery or food
preparation area” existed.
23
Housesteads or Vindolanda, 98 perhaps because these vici were so quickly and densely
Moreover, there might not have been the level of industrial activity that demanded an
internal layout of this kind, which provided easy access to workshops. Greta Bridge
demonstrates how the conditions in which a vicus was set up can affect the
98
Sections 4.1.2 and 4.4.
99
cf. Buildings B1 and B2, surrounded by a courtyard possibly used for industry, with those at
Vindolanda, which had almost no “space outside the buildings for parking wagons or animals, and
there were no gardens” (Birley 2009, 167).
24
4.3. Maryport
In spite of work towards the end of the 19th century, 100 it has only been in the last
twenty years that our knowledge of the vicus to the north east of the fort has
improved. Aerial photography 101 and geophysical survey have identified a much
larger settlement than first realised, 102 and a five-year excavation project began in
2011. 103 Modern buildings mean it is impossible to know how far the settlement
extended around the rest of the fort, but the vicus currently available for study is
approximately 16.40 hectares in area (Fig. 8). There is no certain date for the
construction of the vicus. However, the likelihood of significant rebuilding of the fort
in the 3rd century might indicate a similar date for the adjacent civilian settlement.
In general, strip-houses in Britain are considered to have been smaller than those in
mainland Europe, 104 but the vicus at Maryport is a notable exception to this rule. A
great number of its strip-buildings are very large (Table 5); indeed, the number of
buildings along the line of the northeast road out of the fort with a width in excess of
fifty”. 105 There are several buildings that have a width of around 11 metres, a
dimension which at Housesteads has led to the near-automatic interpretation that such
buildings had an official function. 106 There are no known strip-buildings with such a
width at Greta Bridge (Table 4), and the most frequent width in the northern frontier
100
Robinson (1881) apparently uncovered several strip-houses, one of which was fully excavated.
101
Jarrett 1976, pl. III.
102
Biggins & Taylor 2004c.
103
http://www.senhousemuseum.co.uk.
104
Sommer 2006, 124.
105
Biggins & Taylor 2004c , 115.
106
Section 4.1.3.
25
zone seems to have been around 7 metres. 107 Why was Maryport different? One
suggestion is that, although many of the buildings were houses, “[they] were used as
workshops or for storage … with any partitioned area at the rear used for living
accommodation”. 108 The evidence for this is not particularly convincing, but the
possibility is made more likely by the fact that it fits with the general trend for dual-
Furthermore, the evidence for the industrial activity that would necessitate such
buildings is clear and suggests that the practice of metalworking was well established
at the site. 109 The problem encountered here is that, since most of the modern
evidence up to this point has been drawn from aerial photography and geophysical
survey, indications of industry come primarily from the work of Robinson in the 19th
century. 110 This of course raises the problem that, with no physical evidence to
support his claims, we must trust his judgement. By no means should we regard his
Maryport were so large because they were used as part of industrial activity, Biggins
and Taylor implicitly assume that there was less industry at other sites, where
buildings were generally smaller. This may well be the case – indeed, the sites so far
surveyed display little evidence for significant industry – but the lack of excavation at
107
Birdoswald: Burnham et al. 2001, 333 (with fig. 11). Carvoran: Burnham et al. 2001, 330-31 (with
fig. 10). Castlesteads: Burnham et al. 2001, 333 (with fig. 12). Halton Chesters: Taylor et al. 2000.
108
Biggins and Taylor 2004c, 125.
109
Robinson (1881, 248): coal and iron debris was found next to the main vicus road; 239: a significant
amount of charcoal was noted on the western fringe of the site. He also notes that both coal and
charcoal were found in the buildings excavated, as well as a hearth.
110
The geophysical evidence does not support Robinson’s claims.
111
If for no other reason than I intend to use his reports myself.
26
many sites means that the evidence is not always sufficient to come to the conclusion
that Biggins and Taylor have. This is certainly the case at Maryport.
In order to understand why many of the domestic vicus buildings are relatively large,
we need to first examine some of the structures in greater detail. Building 1, the
largest that fronts the main road running through the ribbon development, had at least
two internal cross-walls. 112 The most northerly seems to create a porch facing the
road, probably used as a shop frontage. We should not assume that the building was
purely for commercial purposes as has been suggested: 113 it is just as likely that it
stands around 85 metres from the curtain wall, which is surely too far for the building
long and narrow with a small section facing the road – is not at all similar to that of
other possible official buildings. 114 Likewise, the existence of a possible courtyard to
Building 2 was slightly smaller and probably had a raised floor, which suggests that it
was used, at least partly, as a store for perishable foodstuffs. 116 We might assume that
the smaller room at the back of the building was used for accommodation, but the
combination within one structure of raised storage and living quarters is rare, if not
112
Biggins and Taylor 2004c, 115.
113
Ibid.
114
Housesteads Buildings 7 and 16.
115
cf. Building 114 at Vindolanda: a dual-function commercial/residential structure with a large
courtyard.
116
Biggins & Taylor 2004c, 115.
27
unique, in Britain. It might have been a purely residential building, raised above the
ground to avoid potential flooding, but more likely it was a dedicated storehouse.
The evidence for the existence of plot boundaries, 117 along with the presence of the
ownership of a plot (the two buildings represent the homes of two individuals). But
7’s poor access from the road suggests instead that the two buildings belonged to one
owner and were used for different functions. I suggest that 2 was a solely industrial
building, while 7 was the owner’s living quarters. The issue of access was therefore
unimportant, since the same man would have owned both structures.
Building 3 had at least one internal cross-wall and was relatively similar in layout to
1. There was possibly a smaller adjoining building closer to the fort. This might have
been a similar arrangement to that at Greta Bridge, in which one owner owned two
buildings. 118 There is even a possibility that a courtyard enclosure existed between 3
and 4 (which itself seems to have had a small two-roomed building in its ‘back
garden’). If this were the case, there might have existed a distinct sense of cooperation
The building next-door to 2, further away from the fort and poorly defined in the
geophysics, was probably another long and narrow structure. More remarkable is
Building 8 directly behind, which is substantial, but not mentioned by Biggins and
Taylor. It might have been a storehouse for the owners of the house in front, which
faced the main road. However, it could also have itself had private access from a
117
Geophysical evidence suggests that ditches were used to demarcate individual plots of land.
118
Section 4.2.4.
28
minor road to the northeast. 119 In this case the two small southerly rooms would have
been partitioned accommodation, the negative anomaly in the centre of the building
would represent the base of an industrial hearth, and the northerly section would have
prioritised over accommodation when the orientation of the building was first
decided.
The irregularity of the structures suggests they are another instance of what Birley
described as ‘hovels’. 121 In the light of the evidence from Housesteads, 122 if these
buildings were homes, then they were probably reserved for auxiliary soldiers.
4.3.4. Discussion
The diversity in size and type of housing at Maryport is extraordinary, and hints at the
existence of a complex society comprising people of varying status and wealth. If so,
it is this complexity that might help us understand why some of the buildings are
larger than elsewhere. Although the geophysical evidence cannot distinguish between
phases, the vicus at Maryport seems to have been a highly planned development (at
least the structures that faced onto the road) and less organic than Housesteads, for
example. This might be part of the reason for the existence of larger buildings, since if
the vicus were planned out to a certain extent before any construction took place, plot-
119
Sommer 2006, 100.
120
Section 4.2.4.
121
I have not included these buildings in my data set for Maryport, since the evidence is not conclusive
that they were used for accommodation.
122
Section 4.1.3.
29
holders and homeowners – if these were different sets of people – would have had
more space for, and choice in, the construction of their buildings. Unfortunately there
is little evidence to suggest where the required wealth – if it existed – came from.
More likely, therefore, is that the buildings were an industrial necessity, rather than a
product of wealth. That is, they were required for storage and workshops, with
A final possible reason for such large buildings at Maryport is a desire on the part of
without secure dates, but it is certainly feasible that the large structures were
constructed as part of a chain of events. That is, having seen a wealthy neighbour
build himself a substantial home, a second resident felt it necessary for the sake of his
social standing to do the same. However, it seems odd that we do not see a
With all of this discussion, it is easy to get carried away by the extraordinary features
of Maryport. It is certainly unique amongst British military vici in terms of its range
of building sizes. 123 However, we must remember that these very large structures (1,
2 and 8) constitute only one-sixth of the total so far identified in the settlement.
Further work must be done at other British vici in order to find out how exceptional
Maryport truly is. In many other ways, Maryport is very similar to other vici. First,
there is evidence for the existence of the widespread dual-function strip-house. There
123
The range of areas is 319.79 m2 (Table 8). Cf. ‘giant’ strip-house at Binchester (38.10 by 13.72
metres) (Salway 1965, 167).
30
is also the possibility of owners occupying two separate properties on the same plot,
each with a different function. Finally, the existence of smaller, more ‘rough-and-
ready’ housing, probably for auxiliary soldiers, is reminiscent of an area of the vicus
at Housesteads. 124 Although the nature of geophysics means that material finds are
not available to provide insight into who the inhabitants of the vicus were, the variety
124
Section 4.1.2.
31
4.4. Vindolanda 125
The site’s chronology and garrison history has been laid out in Table 6, but it is useful
to note here that the primary evidence for vicus buildings comes from period VII; 126
that is, contemporaneous with Stone Fort II. 127 At this time there would have been
approximately 600 soldiers stationed at the site, 128 who required accommodation for
themselves and almost certainly their families. The ‘vicus’ status of the period VII
housing. The accepted chronology of the site is sufficiently complex that the exact
complicated nature of the site (many buildings were erected on top of previous
which building a certain piece of material evidence might have belonged. These
complexities have led to hasty statements, probably designed to simplify the issue. 130
However, all such claims achieve is to gloss over the deficiencies of the evidence,
which, as published, is not anywhere near sufficient to support such statements. 131
Nevertheless, the very fact that the site remains one of the best studied on the
125
Much of the excavation at the Vindolanda vicus took place in the 1960s and 1970s (Birley 1970;
1977); since then, work has taken place during 2003 and 2004 (Birley & Blake 2005). I shall utilise the
building numbers formalised by these previous publications.
126
Partial remains of larger properties to the west have been examined, but have suffered from the
effects of ploughing; at the western fringe of the settlement are traces of wooden buildings from the
late 1st to 3rd century (Birley 2009, 167). More work needs to be done to reach any firm conclusions
about these buildings.
127
Bidwell (1985, 88) argues convincingly for the vicus being constructed in the 220s and largely
abandoned in c. 270.
128
Approximately 3.8 hectares (Fig. 9).
129
RIB 1700: ‘… vicani Vindol/andesses …’.
130
E.g. Birley (2009, 165): “[it] was likely that the majority of civilian buildings were timber-framed
structures set on stone foundations”.
131
The publication of the excavations during the 1960s and 1970s was inadequate. Reports largely
focused on
the fort, and most discussion of vicus buildings was reduced to a few pages in summary.
32
Romano-British frontier means there is still a significant amount of evidence to be
examined. Indeed, it represents the most valuable source for remnants of industrial
4.4.2. Vicus I
On the south side of the road leading from the west gate of the fort, the points of
interest tend to come from the period of Vicus I (Fig. 10), in particular from the so-
called ‘married quarters’. 132 This is primarily because of their abnormal size. 133
Building 9 has been reasonably identified as the mansio, therefore I shall focus on
Buildings 32 and 33, the most comprehensively excavated of the ‘married quarters’.
The assumption that these were buildings designed for the accommodation of
soldiers’ families is questionable at best. The existence of eight small rooms in the
eastern half, with average dimensions of 3.66 by 4.88 metres and internal hearths,
suggests that this was almost certainly accommodation for very small groups or
individuals, but there are no specifically gendered small finds to indicate that it was
At Housesteads, it was probably the soldiers themselves that were housed in very
cramped conditions, 134 but an ‘overflow’ area for soldiers has not yet been identified
at Vindolanda. Why should 32 and 33 not represent such a place? The size of the
garrison during this period is uncertain, so whether extra space was required is
equally so, but such a setup would have been a cost-effective and simple method of
132
Birley 1977, 46.
133
Ibid., 42: “the thirteen small buildings of [Vicus] II had been preceded by only four, earlier, [Vicus]
I structures”.
134
Section 4.1.3.
33
The discovery of a bronze military standard underneath the earliest floor of the
corridor house, later converted into Buildings 23-25, at first seems to suggest that it
Housesteads. 135 A more likely conclusion, however, is that the house had always
been a civilian building, probably with a commercial aspect, and that the standard was
an anomalous find. Indeed, it has been suggested that the narrow wall on the left of
the door “seems to have been part of a serving counter, and the open triple drains in
the floor … suggest that it was perhaps a butcher’s shop”. 136 Its residents seem to
have made a significant amount of money from this enterprise, since also discovered
in the house was “a small gold ring with burnt cornelian gemstones”, as well as
4.4.3. Vicus II
The two later strip-houses (23-25) were clearly of a domestic and probably civilian
nature. There are several hearths either remaining or having been robbed in each room
of 23, and in two of the rooms of 25, but there are no traces of industrial activity.
These were probably hearths used for heating the home and cooking. 138 The
conversion could represent several arrangements. The most simple is that a new
owner decided to completely split the living quarters from the commercial section, in
the process taking possession of two separate buildings for different purposes.
Another possibility is that the house was obtained by a ‘family group’, perhaps the
family of two different soldiers, that therefore required an element of personal space
and segregation. A final option is that the owner of the house wished to lease the
135
Section 4.1.3.
136
Birley 1977, 40.
137
Ibid., 39.
138
Ibid., 37: the existence of flues suggests an attempt to ventilate the house.
34
property to tenants, and the most lucrative way of doing so was to collect rent from
The most interesting aspect of 23 is its apsidal northern end, a feature “elsewhere
usually reserved for shrines”. 139 Unfortunately there is no evidence to indicate the
function of this particular apse, but it is certainly possible that it could have held a
household shrine, perhaps dedicated to Silvanus. 140 There is also one example at
Housesteads of the existence of private shrines, though there the shrine was an
Some of the first buildings discovered, but not fully excavated, were 4-7. It is
sufficient only to note that the structures in this area, outside the western gate of Stone
Fort II, “appear to be similar in design to those outside the south gate at Housesteads”.
142
Although individual designs could differ within the broader domain of the strip-
house, these few buildings suggest that there was a common approach utilised across
From Vicus II, there is evidence on the south side of the road for a small amount of
industrial activity (small bowl furnaces, remnants of iron slag, and fragments of
bronze moulds), but on the whole the area is quite a straightforward setup with eight
houses facing onto the road, and five more lining the west wall of the fort.
Unfortunately there is little in the way of a detailed report on these buildings. In 29B a
139
Birley 1977, 37. Simple apsidal temples were usually associated with the military: see Lewis 1966,
72-3, for examples from Benwell, Housesteads and Corbridge.
140
Section 4.4.2.
141
Housesteads Building 9 (Birley & Charlton 1934, 190-1).
142
Birley 1970, 110.
35
small furnace and a workbench were later additions to the building. 143 This suggests
that either its owner decided to take advantage of the opportunities for commerce, or a
new owner, responsible for extending the structure, purchased the building with
industry in mind. The amount of refuse left on the floors was greater than was typical
for a civilian structure, which also suggests that commercial activity took place. 144
Based on the small finds, Building 30 seems to have been the home of a particular
wealthy family. They include gemstones featuring Minerva, Mars and a Maenad, a
bronze purse, a pipeclay statuette of Venus, a pair of bronze tweezers, and two
brooches. 145 These suggest the existence of residents familiar with Roman customs
Birley comments briefly on Buildings 74, 75 and 78, 146 whose basic, one-room
layout cannot hint at any function for certain. Lacking any evidence of hearths or
has not previously been suggested is that the plot on which these buildings stood was
leased to a private individual. As part of this deal, Building 79 was used as the
individual’s home and a commercial (or perhaps in this case military) function was
maintained in a separate building. Buildings 2 and 3 might have also been used as
143
Birley 2009, 163.
144
Ibid.
145
Ibid., 164-5.
146
Birley 1977, 42.
36
The date of 2 is uncertain, but finds on the site suggest a construction date in the
middle of the 3rd century. 147 Two phases can be identified, but the later alterations
were minor and did not change the layout of the building to any great extent. The
building was not particularly large, but its location close to the southwest corner of
the fort might indicate an official function, or at least some level of interaction with
the significant evidence for industrial activity found within the building as well as in
its immediate environs. In the eastern half, a small furnace “bore witness of intense
fire”, and a platform close by might well have been used as a workbench. 149 In
conjunction with the well just outside the south wall, and the large quantities of
rubbish (including a great deal of bone) in the ditch below, the eastern half of 2
The western half is more intriguing, since it suggests that the label ‘workshop’ does
not describe the entire building. Birley has suggested, on the basis of evidence for
floor supports and the absence of broken flagstones, that there existed a raised
wooden floor “to offset the dampness of the site”. 150 We have already seen raised
floors at Maryport, where they might indicate the presence of storehouses, 151 but this
building at Vindolanda is the only structure in which the floor has been raised in just
one section. In light of this, the theory that the western half was used for storage
becomes less likely. There are two other options. First, the raised section might have
been the residential quarters of a dual-function domestic building. On the other hand,
the inconclusiveness of the evidence might suggest the floor was not ever raised at all.
147
Birley 1962a, 99.
148
Ibid.
149
Ibid.
150
Ibid.
151
Section 4.3.3.
37
In this case, the layout of the building would more closely conform to the trend for
During the excavations of 1970-76, several more vicus houses were examined.
Building 27 has not been studied in much detail: Birley has described it as a
However, it is notable that the cooking bench was in the rear room, with a clay/stone
built oven in the southern room. This suggests that if the house did have an industrial
aspect, it was at the rear of the building, contrary to the most common setup
elsewhere. It did, however, have a room facing the main road leading west from the
Birley has suggested that the vicus layout was like that at Housesteads, “where a
jumble of small rectangular buildings clustered round the roads or tracks leading from
the fort, displaying no signs of any attempt at orderly planning”. 153 It was Buildings
111 and 115-17 154 that were, in terms of size and internal layout, most like 19 and 22-
27 at Housesteads and those buildings southeast of 1 at Maryport. 155 They were very
small structures, but less likely to have been military accommodation due to their
being located approximately 70 metres from the western wall of the fort. Instead, they
might have accommodated poorer residents, or have been tombs or mausoleums. 156
152
Birley 1977, 37-8.
153
Ibid., 39.
154
Excavated in 2003 and 2004.
155
Sections 4.1.3 and 4.3.3.
156
Birley & Blake 2005, 17.
38
The increased number of houses in Vicus II when compared with Vicus I might not
individual to set up his own home and enterprise, and live in greater comfort. Indeed,
there are large buildings from Vicus II that are not divided for multiple occupancy in
the same way as 32/33 from Vicus I. Building 113 was a significant size on its own
but, in combination with 111 and 112, quite possibly made up a very large complex,
probably owned by a single individual. 157 Each building might have had a different
function, and so the complex becomes an extreme example of the concept we have
Building 114, on the other hand, seems to have been an extreme example of the more
one corner – this was presumably the domestic quarters – while a larger room housed
the foundations of a possible industrial-sized oven. Outside there was a large flagged
courtyard area. An opening onto the road suggests it might have been used for parking
or perhaps as a workshop area, since both would have required easy access. Outdoor
areas were rare at Vindolanda, therefore, in order to afford his courtyard, the owner of
114 must have been relatively prosperous and an ambitious man. But it is the rarity of
such external space that also suggests Vindolanda did not thrive economically after
the departure of Roman army in the same way as some other vici. The concept of high
levels of occupation continuing into the 3rd and 4th centuries is attested at Birdoswald
158
and Halton Chesters, where buildings were clearly altered and reconstructed for
157
Birley & Blake 2005, 7ff.
158
Wilmott 1997, 408-19.
159
Taylor et al. 2000, 44.
39
4.4.4. Discussion
Although auxiliary soldiers might have resided in a particular set of buildings, they
were in no way segregated from the civilian inhabitants. Vindolanda, therefore, seems
to have been a place of social interaction, where the wealthy might have been living
in the same area as auxiliary soldiers. The presence of Silvanus and other Roman gods
160
hints also that the wealthier residents had at least some knowledge of Roman
religion and that there was a certain level of Romanization. Whether the more
to be catered for is difficult to say, but a higher level of commerce certainly suggests
that the vicus was a centre for traders from across the northern frontier zone. Its status
as a well-established commercial hub was probably also a result of its being populated
160
Sections 4.4.2-3.
40
4.5. Observations from other sites
I shall now briefly examine two further sites that are not supported by a great wealth
The vicus at Halton Chesters (Fig. 11) extends southwards for approximately 275
“numerous stone buildings are evident between the fort ditches and the Vallum…; in
form they follow a typical configuration with narrow gables and long side walls,
many of the structures being c. 7-10m wide by c. 10-15m long”. 161 However, it has
not been noted that close to the fort, on the western side of the southward road, there
are a number of buildings that are not ‘typical’ in configuration. They seem to be very
small constructions, of varying sizes, and laid out in a haphazard manner. It is odd
that such a set of structures are in a prominent location close to the fort and the main
road southwards, since in general it was a position reserved for larger strip-buildings
designed for commerce. Perhaps these buildings were the homes of the less wealthy
members of the vicus, or the overflow accommodation for soldiers. However, based
on the evidence we have seen elsewhere, it is unlikely that either of those functions
would have occupied such an important site in the vicus. Therefore I suggest that
these smaller buildings were part of the larger complexes of individual homeowners,
Taylor et al. argue that the civilian settlement at Halton Chesters “lies outside the
limits of the survey”, to the west of the fort, because “what little can be traced appears
161
Taylor et al. 2000, 40.
41
to be restricted to a linear development following the road south”. 162 This is not a
legitimate justification for their conclusion, since, as we have seen at every other site,
such a development was often the heart of a vicus. Indeed, they even point out
themselves that “the development alongside the road to the south of the fort bears a
strong parallel to that leading west out from the fort at Birdoswald”. 163 If the vicus
were to the west of the fort, how do we explain the existence of the structures to the
south? Their size and internal layout are not such that we can attribute an official or
military function to every building. The existence of what may well have been
military barracks to the east of the fort 164 would make it less likely that soldiers were
accommodated to the south. On the other hand, if these buildings to the east were not
At Birdoswald, it had previously been argued that there was an “apparent absence of a
vicus”. 166 However, recent geophysical survey has revealed that there was almost
certainly a civilian settlement outside the walls of the fort (Fig. 12). 167 To the west of
the fort the vicus seems to have been an ordered ribbon development that stretched for
more lightly constructed, with the use of timber-framed construction”. 168 The east
side was less regular in its layout and the buildings were of more substantial
the date of the settlements on either side of the fort, therefore it is impossible to say
162
Taylor et al. 2000, 43.
163
Ibid.
164
Ibid., 40.
165
There was probably a significant cavalry presence at Halton Chesters (Taylor et al. 2000, 43).
166
Shotter 1996, 50.
167
Biggins & Taylor 1999; 2004b.
168
Biggins & Taylor 1999, 106.
169
Ibid.
42
whether they co-existed or succeeded one another. However, the very different
layouts seem to suggest that the west side had a different purpose from that of the east
Unlike at Housesteads and Vindolanda, 171 where military buildings were probably
incorporated into the main vicus, it has been suggested that at Birdoswald the western
vicus was a solely civilian settlement, with possible industrial connections. The
indications are that the eastern vicus was a military annexe. 172 But it was still rather
irregular in its layout; could it really have such a settlement? The probable presence
of a bathhouse and a signal tower suggests that this was the case. 173 What the
conform to the type of regular layout one might find within the fort itself.
If there was any interaction between the inhabitants of the two ‘sub-vici’, it might
well have come in the form of commerce, since at Birdoswald is the most enticing
piece of evidence for how residents used their own homes as part of a communal
economic arrangement. In the western vicus, roughly 100 metres from the fort gate, is
an elliptical space approximately 90 by 30 metres in size. 174 Most notable are the
large buildings on the northern side of this space. They are similar to those we have
seen at Maryport, but their function was surely different. 175 At Maryport, the
owners. Here at Birdoswald, however, it is possible that the space was used for a
170
Biggins & Taylor 2004b, 176.
171
Sections 4.1.2 and 4.4.2.
172
Biggins & Taylor 2004b, 176.
173
Biggins & Taylor 1999, 107.
174
Biggins & Taylor 2004b, 165. Sommer (2006, 117) describes it as triangular.
175
Section 4.3.3.
43
market, and these larger buildings were used either as production quarters, or sales
quarters. 176 The evidence is not sufficient to tell us a great deal about their internal
layout, but based on what we have seen, it is certainly possible, if not likely, that they
were divided in order to also provide domestic accommodation for their owners. This
elliptical space is a unique feature in military vici in Britain, 177 and perhaps hints at a
176
cf. Cassius Dio (56.18.2) on the Romanization of Germany: “The barbarians were becoming
accustomed to hold markets…”.
177
Sommer (2006, 118) suggests that a similar space existed at Maryport. In Germany, market spaces
of various shapes were common, therefore it is possible that the elliptical space at Birdoswald was an
offshoot of this development.
Triangular
Heddernheim: Huld-Zetsche 1994, 20;
Saalburg: Sommer 1997, 160-4;
Zugmantel: Sommer 1988, 565-7;
Rectangular
Ladenburg: Kaiser and Sommer 1994, 380-87;
Regensburg-Kumpfmühl: Faber 1994, 44-5.
44
5. Conclusion
In this paper, through analysis of the archaeological data currently available, I have
attempted to ask new questions about the housing in the military vici of northern
Britain. I have explored how houses differed in form, size and function from site to
site, and used the results to assess as far as possible social structures and hierarchies
in the extra-mural military setting. Though limited at times by a lack of evidence from
size of the vicus population at Housesteads. 178 Furthermore, I have been able to
establish probable trends in the function of domestic buildings and to suggest how
they related to each other, economically and socially. In particular, I suggest the types
of people who might have inhabited the settlement, and in which types of building.
The key feature of vici housing is diversity. 179 Even though the most common layout
was a ribbon development with an adjoining road following a wall of the fort, the
buildings that lined the roads were not always similar. The type of material used was
generally masonry or timber, but this varied in quality. Dimensions were also diverse:
although on average a domestic building was between 12 and 15 metres long and 6 or
7 metres wide, at both Maryport and Housesteads the length could vary by
approximately 25 metres (Table 8). 180 The most likely number of rooms in a
domestic vicus building was two; more or fewer than that was, although not out of the
realms of possibility, an infrequent occurrence. What all this seems to indicate is not
178
Section 4.1.3. I suggest a population of 600-700, approximately half of the most recent estimation.
179
The regularity in the construction of vici housing found at some sites in Germany is generally absent
in northern Britain (e.g. Zugmantel (Schoenberger 1951)).
180
The areas of the very large buildings at Maryport do not fit the trend across the other sites studied,
and must therefore be treated as exceptional. Without them, Maryport would have been more in line
with Housesteads and Vindolanda.
45
only a number of functions for the buildings, but also different priorities and
It is clear that strip-houses could have a dual function. They often fulfilled the need
Bridge and Maryport, 181 however, we have seen these two functions being carried out
in separate buildings, often adjacent to each other, and often of different sizes. This
suggests that a single owner in these vici could have owned more than one building in
the same plot. Each function might well have been denoted by the use of different
types of flooring – for example, at Greta Bridge clay was probably used for
The number of hearths and ovens, although not often industrial-scale, found at each
site studied suggests that there was at least some commercial activity taking place,
probably on a small scale. It is difficult to be certain what level such commerce ever
produce goods for export out of Britain. The possible existence of market-spaces at
Birdoswald and Maryport hints at economies that went beyond self-sufficiency for the
merchants. 183 The main customers of the tradesmen and merchants at vici were
almost certainly the soldiers, both those based inside the fort and those who lived in
the vicus itself. The lack of many large industrial buildings and associated finds
(debris, tools etc.) at these vici, 184 along with the presence of market-spaces, could
181
Sections 4.2.2-3 and 4.3.3.
182
More work needs to be done elsewhere before we can confirm or refute this. However, Building 3 at
Housesteads exhibits a similar phenomenon (Birley et al 1933, 86).
183
Section 4.5. There does not seem to have been space for such an area at Housesteads.
184
cf. vici in Gaul, 58% of which had pottery, metalworking, food-processing, or textile workshops
(Rorison 2001, 51).
46
indicate that the settlements were used for the trade of goods from elsewhere, but not
of those produced on site. Thus, as the British economy became more regional-centric
and local manufacturing replaced long-distance trade, 185 vici might have become
The diversity of building and plot sizes 186 in the vici suggest that there were varying
levels of wealth within society, but they can only tell us so much. For instance, the
very large buildings at Maryport 187 may well have been the homes of wealthy
veterans or traders who had made their money at the local market, but they may
equally have been storehouses with an official function for the supply of the army.
More useful for establishing levels of wealth and status are small finds and the
domestic embellishments that often came as a result (mosaics, baths, luxury foods
etc.). However, at all of the sites studied, such evidence is scant. The existence of a
magister vikanorum at Old Carlisle suggests that some form of structured hierarchy
was in place, though there is little evidence to indicate that it was implemented as an
official policy. 188 Nor is there evidence that such a system was used at any other vici.
The ‘hovels’ at Housesteads were more likely to have been homes for auxiliary
soldiers, 189 and similar areas at other sites were probably used in a comparable way.
This raises the problem of where the poorer residents were accommodated. I believe
that this issue can be easily solved by accepting that there was no exceptionally poor
element of society. Based on the evidence available, the society seemed to be made
47
the type of complex levels of wealth that existed in larger civilian settlements in the
province. A hierarchy might have existed thus: senior military officials and veterans
at the top, soldiers and merchants somewhere in the middle, soldiers’ families just
below, and slaves at the bottom. It is a simplified pyramid suited to a simple civilian
settlement.
As I stated at the beginning of this paper, this has not been an exhaustive survey of all
the military vici in Britain. Looking to the future, there is much work still to be done.
promises to provide a great deal more evidence for vici housing. As well as enlarging
the corpus of vici for which house plans are available, better documentation of finds
assemblages will enable a greater understanding of the gender, class and status of
revived (or perhaps created), since this paper has been shown what can be gained by
studying the sites in greater depth than has previously been attempted.
190
Further study along the lines of Allison 2006, in which she begins to interpret gendering artefacts
from forts in Germany, would be most useful.
48
Bibliography
49
Birley, R.E. (1962a). ‘Some excavations at Chesterholm-Vindolanda’, AA4, 40, 97-
103.
Birley, R.E. (1962b). ‘Housesteads vicus, 1961’, AA4, 40, 117–33.
Birley, R.E. (1970). ‘Excavations at Chesterholm-Vindolanda 1967-1969’, AA4, 48,
97-155.
Birley, R.E. (1973). Civilians on the Roman Frontier. Newcastle-upon-Tyne:
Graham.
Birley, R.E. (1977). Vindolanda: a Roman frontier post on Hadrian's Wall. London.
Birley, R.E. (2000). Civilians on Rome’s Northern Frontier. Greenhead: Roman
Army Museum Publications.
Birley, R.E. (2009). Vindolanda: a Roman frontier fort on Hadrian's Wall. Stroud:
Amberley.
Birley, A. & Blake, J. (2005). Vindolanda: the excavations of 2003-2004. Barton
Mill: Vindolanda Trust.
Birley, E., Charlton, J., & Hedley, W.P. (1932). ‘Excavations at Housesteads in
1931’, AA4, 9, 222–37.
Birley, E., Charlton, J., & Hedley, W.P. (1933). ‘Excavations at Housesteads in
1932’, AA4, 10, 82–96.
Birley, E. & Charlton, J. (1934). ‘Third report on excavations at Housesteads’, AA4,
11, 185–205.
Birley, E. & Keeney, G.S. (1935). ‘Fourth report on excavations at Housesteads’,
AA4, 12, 204–58.
Bishop, M.C. & Dore, J.N. (1988). Corbridge: excavations of the Roman fort and
town, 1947-80. London: English Heritage.
Blagg, T.F.C. & King, A.C. (eds.) (1984). Military and civilian in Roman Britain:
cultural relationships in a frontier province. BAR 136. Oxford.
Blake, B. (1960). ‘Excavations of Native (Iron Age) sites in Cumberland’, Trans.
Cumberland & Westmorland Antiq. & Archaeol. Soc.2 59, 1-14.
Branigan, K. (1980). Rome and the Brigantes: the impact of Rome on Northern
England. Sheffield.
Breeze, D.J. (1982). The Northern Frontiers of Roman Britain. London.
Breeze, D.J. (2013). ‘Preface’, in Collins & Symonds 2013, 7-8.
Breeze, D.J. & Dobson, B. (1987). Hadrian’s Wall. London.
50
Brown, A.E. (1995). Roman Small Towns in eastern England and beyond. Oxbow
Monograph. Oxford.
Burnham, B.C. & Davies, J.L. (2010). Roman Frontiers in Wales and the Marches.
Aberystwyth: RCAHMW.
Burnham, B.C. & Wacher, J.S. (1990). The Small Towns of Roman Britain. Berkeley.
Burnham, B.C., Keppie, L.J.F., Fitzpatrick, A.J., Hassall, M.W.C., & Tomlin, R.S.O.
(2001). ‘Roman Britain in 2000’, Britannia 32, 311-400.
Casey, P.J. (1982). ‘Civilians and soldiers – friends, Romans, countrymen?’, in Clack
& Haselgrove 1982, 123-32.
Casey, P.J. & Hoffmann, B. (1998). ‘Rescue excavations in the vicus of the fort at
Greta Bridge, Co. Durham, 1972-4’, Britannia 29, 111-83.
Clack, P. & Haselgrove, S. (eds.) (1982). Rural Settlement in the Roman North.
Durham.
Collins, R. & Symonds, M. (2013). Breaking Down Boundaries: Hadrian's Wall in
the 21st Century. JRA Supplementary Series 93.
Creighton, J.D. & Wilson, R.J.A. (eds.) (1999). Roman Germany: Studies in Cultural
Interaction. JRA Supplementary Series 32.
Crow, J. (2004). Housesteads: a fort and garrison on Hadrian's Wall. Stroud:
Tempus.
Daniels, C.M. (1980). ‘Excavations at Wallsend and the fourth-century barracks on
Hadrian’s Wall’, in Hanson & Keppie 1980, 173-93.
Darvill, T. & Thomas, J. (eds.) (2001). Neolithic Enclosures in Atlantic Northwest
Europe. Neolithic Studies Group Seminar Papers 6. Oxford: Oxbow.
Davies, J.L. (1984). ‘Soldiers, peasants and markets in Wales and the Marches’, in
Blagg & King 1984, 93-127.
Evans, J.G., Limbrey, S., & Cleere, H. (eds.) (1979). The Effect of Man on the
Landscape: the Highland Zone. London: Council for British Archaeology.
Faber, A. (1994). Das römische Auxiliarkastell und der Vicus von Regensburg-
Kumpfmühl. Münchner Beiträge zur Vor- und Frühgeschichte 49. Munich.
Frere, S.S. & St Joseph, J.K.S. (1983). Roman Britain From the Air. Cambridge.
Garlick, T. (1972). Romans in the Lake Countries. Clapham (Yorks.): Dalesmen.
Gibson, A.M. (1999). The Walton Basin Project: Excavation and Survey in a
Prehistoric Landscape 1993-7, CBA Research Report 118. York: Council for British
Archaeology.
51
Gibson, A.M. (2001). ‘Survey at Hindwell enclosure, Walton, Powys, Wales’, in
Darvill & Thomas 2001, 101-10.
Gillam, J.P. (1950). ‘Recent excavations at Birdoswald’, Trans. Cumberland &
Westmorland Antiq. & Archaeol. Soc.2 50, 63-8.
Hankinson, R. (2011). Hindwell Roman fort and vicus, Walton, Powys:
Archaeological Investigations 2010-11. CPAT Report No. 1092.
Hanson, W.S. & Keppie, L.J. (eds.) (1980). Roman Frontier Studies 1979: papers
presented to the 12th International Congress of Roman Frontier Studies. Oxford:
BAR International Series 71.
Higham, N. (1982). ‘The Roman Impact upon Rural Settlement in Cumbria’, in Clack
& Haselgrove 1982, 105-122.
Higham, N. & Jones, B. (1985). The Carvetii. Gloucester.
Hopewell, D. (2005). ‘Roman Fort Environs in North-West Wales’, Britannia 36,
225-269.
Hopewell, D. (2006). Roman Fort Environs: Geophysical Survey at
Trawscoed and Llanio Roman Forts, Report 623. Gwynedd Archaeological Trust.
Horsley, J. (1732). Britannia Romana. Newcastle-Upon-Tyne: Graham (reprint).
Huld-Zetsche, I. (1994). NIDA – eine römische Stadt in Frankfurt am Main. Schriften
des Limesmuseums Aalen 48. Stuttgart.
Hunter, J. (2013). ‘An assessment of the evidence for the cult of Mars in Roman
Britain’, coursework essay for module ‘Roman Britain’. King's College London
(unpublished).
James, S. (2005). ‘Limesfreunde in Philadelphia: a Snapshot of the State of Roman
Frontier Studies’, Britannia 36, 499-502.
Jarrett, M.G. (1976). Maryport, Cumbria: a Roman fort and its garrison. Kendal.
Jones, G.D.B. (1984). ‘“Becoming different without knowing it”. The role and
development of vici’, in Blagg & King 1984, 75-91.
Jones, N.W. (2012). Hindwell Double-palisaded Enclosure and Roman Vicus, Powys:
Archaeological Investigations 2012. CPAT Report No. 1138.
Kaiser, H. & Sommer, S. (1994). LOPODUNUM I. Die römischen Befunde der
Ausgrabung an der Kellerei in Ladenburg 1981-1985 und 1990. Forschungen und
Berichte zur Vor- und Frühgeschichte in Baden-Württemberg 50. Stuttgart.
52
Lax, A. & Blood, K. (1997). ‘The earthworks of the Maryport fort: an analytical field
survey by the Royal Commission on the Historical Monuments of England’, in
Wilson 1997, 52-66.
Lewis, M.J.T. (1966). Temples in Roman Britain. Cambridge.
Ling, R. (ed.) (1997). The Insula of the Menander at Pompeii. Vol. 1, The structures.
Oxford: Clarendon Press.
MacMahon, A. (2003). The Taberna Structures of the Roman Britain. BAR 356.
Oxford.
Maxfield, V.A. & Dobson, M.J. (eds.) (1991). Roman Frontier Studies 1989:
proceedings of the XVth International Congress of Roman Frontier Studies. Exeter.
McCarthy, M.R. (2002). Roman Carlisle & the Lands of the Solway. Stroud: Tempus.
Millett, M. (1984). ‘Forts and the origins of towns: cause or effect?’, in Blagg & King
1984, 65-74.
Millett, M. (1990). The Romanization of Britain. Cambridge.
Nevell, M. (ed.) (1999). Living on the edge of empire: models, methodology &
marginality; late-prehistoric and Romano-British rural settlement in North West
England. Archaeology North West, Vol. 3, Issue 13.
Olivier, A. (1982). ‘The Ribchester vicus and its context – the results of recent
excavations’, in Clack & Haselgrove 1982, 133-147.
Packer, J.E. (1971). The insulae of imperial Ostia. Memoirs of the American
Academy in Rome, vol. 31.
Potter, T.W. (1979). Romans in north-west England: excavations at the Roman forts
of Ravenglass, Watercrook and Bowness on Solway. Cumberland and Westmorland
Antiquarian and Archaeological Society.
Poulter, J. (2009). Surveying Roman military landscapes across northern Britain: the
planning of Roman Dere Street, Hadrian's Wall and the Vallum, and the Antonine
Wall in Scotland. BAR 492. Oxford.
Redhead, N. (1999). ‘Edge of Empire: Extra-Mural Settlement in a Marginal Context:
Roman Castleshaw’, in Nevell 1999, 74-81.
Richmond, I.A. (1929). ‘Excavations on Hadrian’s Wall in the Gilsland-Birdoswald-
Pike Hill Sector, 1928’, Trans. Cumberland & Westmorland Antiq. & Archaeol. Soc.2
29, 303-15.
53
Richmond, I.A. (1931). ‘Excavations on Hadrian’s Wall in the Gilsland-Birdoswald-
Pike Hill Sector, 1930’, Trans. Cumberland & Westmorland Antiq. & Archaeol. Soc.2
31, 122-34.
Richmond, I.A. & Birley, E. (1930). ‘Excavations on Hadrian’s Wall in the Gilsland-
Birdoswald-Pike Hill Sector, 1929’, Trans. Cumberland & Westmorland Antiq. &
Archaeol. Soc.2 30, 169-205.
Robinson, J. (1881). ‘Notes on the excavations near the Roman camp, Maryport’,
Trans. Cumberland & Westmorland Antiq. & Archaeol. Soc.1 5, 237-55.
Rodwell, W. & Rowley, T. (eds.) (1975). The Small Towns of Roman Britain: papers
presented to a conference, Oxford, 1975. BAR 15. Oxford.
Rorison, M. (2001). Vici in Roman Gaul. Oxford.
Rushworth, A. (2009). Housesteads Roman fort: the grandest station: excavation and
survey at Housesteads, 1954-95, Vols. I-II. Swindon: English Heritage.
Russell, L. (ed.) (2001). Colonial Frontiers: Indigenous-European Encounters in
Settler Societies. Manchester.
Salway, P. (1965). The Frontier People of Roman Britain. Cambridge.
Salway, P. (1980). ‘The vici: urbanisation in the north’, in Branigan 1980, 8-17.
Schallmayer, E. (ed). (1997). Hundert Jahre Saalburg. Vom römischen Grenzposten
zum europäischen Museum. Mainz.
Schönberger, H. (1951). ‘Plan zu den Ausgrabungen am Kastell Zugmantel bis zum
Jahre 1950’, Saalburg-Jahrbuch 10, 55-75.
Shotter, D.C.A. (1996). The Roman frontier in Britain. Preston: Carnegie.
Simpson, F.G. & Richmond, I.A. (1934). ‘Report of the Cumberland Excavation
Committee for 1933. Excavations on Hadrian’s Wall’, Trans. Cumberland &
Westmorland Antiq. & Archaeol. Soc.2 34, 120-30.
Simmons, B. (1995). ‘Sapperton’, in Brown 1995, 157-65.
Snape, M.E. (1991). ‘Roman and native: vici on the north British frontier’, in
Maxfield & Dobson 1991, 468-71.
Sommer, S. (1984). The Military Vici in Roman Britain: aspects of their origins, their
location and layout, administration, function, and end. BAR 129. Oxford.
Sommer, S. (1988). ‘Kastellvicus und Kastell’, Fundberichte Baden-Württemberg 13,
457-707.
Sommer, S. (1997). ‘Der Saalburg-vicus. Neue Ideen zu alten Plänen”, in
Schallmayer 1997, 155-65.
54
Sommer, S. (1999). ‘From conquered territory to Roman province: recent discoveries
and debate on the Roman occupation of SW Germany’, in Creighton & Wilson (eds.)
1999, 160-98.
Sommer, S. (2006). ‘Military vici in Roman Britain revisited’, in Wilson 2006, 95-
145.
Symonds, M.F.A. & Mason, D.J.P. (eds.) (2009). Frontiers of knowledge: a research
framework for Hadrian's Wall, Vols. I-II. Durham.
Taylor, D.J.A., Robinson, J., & Biggins, J.A. (2000). ‘A report on a geophysical
survey of the Roman fort and vicus at Halton Chesters’, AA5 28, 37-46.
Wacher, J.S. (ed.) (1966). The civitas capitals of Roman Britain: papers given at a
conference held at the University of Leicester, 13-15 December 1963. Leicester.
Wilmott, T. (1997). Birdoswald: excavations of a Roman fort on Hadrian's Wall and
its successor settlements 1987-92. London: English Heritage.
Wilson, D.R. (1974). ‘Roman Britain in 1973: I. Sites Explored’, Britannia 5, 396-
460.
Wilson, D.R. (1975). ‘Roman Britain in 1974: I. Sites Explored’, Britannia 6, 220-83.
Wilson, R.J.A. (ed.) (1997). Roman Maryport and its setting: essays in memory of
Michael G. Jarrett. Cumberland and Westmorland Antiquarian & Archaeological
Society.
Wilson, R.J.A. (ed.) (2006). Romanitas: essays on Roman archaeology in honour of
Sheppard Frere on the occasion of his ninetieth birthday. Oxford.
Wilson, R.J.A. & Caruana, I.D. (eds.) (2004). Romans on the Solway: essays in
honour of Richard Bellhouse. Cumberland and Westmorland Antiquarian &
Archaeological Society.
55
Appendix 1: Tables
Cohors I
Vicus at its fullest extent -
Tungrorum
Cohors I
Tungrorum;
Early 3rd century - additional 1000
units of
Frisians
Size of garrison reduced; vicus
Late 3rd century - -
gradually abandoned
56
Table 2: Domestic buildings at Housesteads
Width
Building Rooms Length (m) Area (m2)
(m)
57
Table 3: Approximation of the population at the Housesteads vicus
Building No. of residents
1 4
2 4
3 5
4 5
5 0
6 0
7 0
8 5
9 3
10 4
11 4
12 5
13 5
14 4
15 4
16 3
17 5
18 5
19 6
20 4
21 4
22 6
23 5
24 6
25 7
26 5
27 5
Non-excavated buildings 114
Total 227
58
Table 4: Domestic buildings at Greta Bridge
WESTERN VICUS
2* 2 - 6.50 -
EASTERN VICUS
59
Table 5: Domestic buildings at Maryport
Building Rooms Length (m) Width (m) Area (m2)
1 3.00 30.00 11.00 330.00
2 4.00 26.00 11.00 286.00
3 3.00 21.00 8.00 168.00
4 - - - -
5 2.00 10.71 5.36 57.41
6 2.00 17.85 8.93 159.40
7 2.00 12.50 7.14 89.25
8 - 28.56 10.71 305.88
9 2.00 14.28 5.36 76.54
10 3.00 17.85 7.14 127.45
11 2.00 5.36 3.57 19.14
12 2.00 8.93 3.57 31.88
13 3.00 17.86 3.57 63.76
14 2.00 3.57 2.86 10.21
15 2.00 7.14 5.36 38.27
16 2.00 10.71 5.36 57.41
17 - 16.15 6.10 98.52
18 - 14.48 4.42 64.00
60
Table 6: Chronology and garrison history of Vindolanda (after Birley 2009, 183)
No. of
Approximate
Period Fort Garrison soldiers
date (AD)
(approx.)
O 79-85 - Unknown Unknown
500 at any one
I 85-90 Timber Fort I Coh I Tungrorum
time
Coh I Tung, followed by
II 90-100 Timber Fort II 750-1000
Coh VIIII Batavorum
III 100-105 Timber Fort III Coh VIIII Batavorum 1000
IV 105-120 Timber Fort IV Coh I Tungrorum 800
V 120-130 Timber Fort V Coh I Tungrorum 500
VI 130-165 Turf-and-timber fort Coh II Nerviorum (?) Unknown
VIA 165-205 Stone Fort I Unknown -
VIB 205-212 The Severan Fort Unknown -
VII 213-300 Stone Fort II Coh IV Gallorum 600
Theodosian
IX 370-400 Unknown -
reconstruction of SF II
X 400- Decline of SF II Unknown -
61
Table 7: Domestic buildings at Vindolanda
Excavations Building Rooms Length (m) Width (m) Area (m2)
1959 2 2 12.09 6.25 75.56
1967 4 1 12.15 7.15 86.87
1967 5 2 - 6.43 -
1967 6 2 14.29 5.72 81.74
1967 7 - - - -
1970-1976 21 - 13.58 5.72 77.68
1970-1976 23 3 13.72 5.79 79.44
1970-1976 25 3 16.00 4.57 73.12
1970-1976 27 2 10.40 6.80 70.72
1970-1976 28 - 17.15 7.86 134.80
1970-1976 29(B) - 30.00 10.25 307.50
1970-1976 30(N) 1 19.29 8.57 165.32
1970-1976 30(S) 1 17.86 4.29 76.62
1970-1976 32A 1 15.72 6.43 101.08
1970-1976 32B 1 17.15 6.72 115.25
1970-1976 33 1 14.29 5.72 81.74
1970-1976 34A 1 12.86 4.29 55.17
1970-1976 34B 1 12.15 5.00 60.75
1970-1976 35 1 14.29 7.86 112.32
1970-1976 39 1 12.86 8.57 110.21
1970-1976 71 - 9.29 5.72 53.14
1970-1976 72 2 10.00 7.15 71.50
1970-1976 74 1 12.15 8.57 104.13
1970-1976 75 1 12.86 7.86 101.08
1970-1976 78 1 12.86 7.15 91.95
2003/4 111 4 7.83 5.10 39.93
2003/4 113 3 11.00 10.00 110.00
2003/4 114 2 12.01 9.28 111.45
2003/4 115 1 6.19 5.46 33.80
2003/4 116 1 5.46 4.73 25.83
2003/4 117 1 5.28 4.91 25.92
2003/4 120 - - - -
2003/4 121 5 16.15 4.30 69.45
2003/4 122 2 16.50 6.50 107.25
Mean 1.71 13.34 6.59 90.69
Mode 1.00 12.86 5.72 81.74
Median 1.00 12.86 6.43 81.74
Range 4.00 24.72 5.96 281.67
Interquartile range 1.00 4.36 2.49 38.54
Outlier if greater
3.50 22.39 11.60 166.44
than
OR less than -0.50 4.97 1.64 12.27
Standard Deviation 1.05 4.67 1.66 50.88
62
Table 8: A comparison of domestic buildings at British military vici
MEAN No. of rooms Length (m) Width (m) Area (m2)
STANDARD DEVIATION No. of rooms Length (m) Width (m) Area (m2)
63
Appendix 2: Figures
Figure 1: Plan of the vicus at Housesteads, denoting dates of excavation (Rushworth 2009, I. 294;
based on Birley & Keeney 1935, pl. 22)
64
Figure 2: Magnetic anomaly plan of the vicus at Housesteads (Biggins & Taylor 2004a, 57).
65
Figure 3: The vicus at Housesteads in its wider context (http://www.english-
heritage.org.uk/content/imported-docs/f-j/housesteads-pp.pdf)
66
Figure 4: Relief of genius (Crow 2004, 75).
67
Figure 5: Plan of the vici at Greta Bridge in their wider context (Casey & Hoffmann 1998, 113)
68
Figure 6: Plan of the western vicus at Greta
Bridge (Casey & Hoffmann 1998, 124)
Figure 7: Plan of the eastern vicus at Greta Bridge (Casey & Hoffmann 1998, 124)
69
Figure 8: Plan of the vicus at Maryport (Biggins & Taylor 2004c, 110); annotations in red are my
own.
70
Figure 9: The vici at Vindolanda, periods I and II (Bidwell 1985, 91)
71
Figure
10:
Plan
of
the
major
vicus
buildings
uncovered
during
the
2003-‐04
excavations
at
Vindolanda
(Birley
&
Blake
2005,
5).
72
Figure 11: Composite excavation and magnetic anomaly plan of the vicus at Halton Chesters
(Taylor et al. 2000, 42)
73
74
Figure 12: Magnetic anomaly plan of the vicus at Birdoswald (Biggins & Taylor 2004b, 163)