SCOTTISH ARCHAEOLOGICAL INTERNET REPORTS
e-ISSN: 2056-7421
Doune Roman fort, Stirlingshire: excavations in 1999, 2008 and 2010
How to cite:
Hatherley, C 2020 ‘Doune Roman fort, Stirlingshire: excavations in 1999,
2008 and 2010’, Scottish Archaeological Internet Reports 92.
https://doi.org/10.9750/issn.2056-7421.2020.92
Visit http://journals.socantscot.org/index.php/sair to see the series
homepage
Please note:
This document is the publisher’s PDF of a report published in the Scottish
Archaeological Internet Reports series. This version has been peerreviewed and contains all final editorial corrections and pagination.
Copyright © 2020 rests with Candy Hatherley and individual
contributors. Except where otherwise noted, this work is licensed
under Creative Commons Attribution-NonCommercial-No
Derivatives licence.
http://creativecommons.org/licenses/by-nc-nd/4.0/
The permission to reproduce the Society's copyright-protected
material does not extend to any material which is identified as
being the copyright of a third party. Authorisation to reproduce such
material must be obtained from the copyright holders concerned.
Doune Roman fort, Stirlingshire: excavations in 1999,
2008 and 2010
Candy Hatherley
with contributions by
Jeremy Evans, Martin Goldberg, Kay F Hartley, Mhairi Hastie, Nicholas M McQ Holmes,
Fraser Hunter, Julie Lochrie, Gwladys Monteil, Effie Photos-Jones, Scott Timpany,
David F Williams and Steven H Willis
Address
Headland Archaeology (UK) Ltd, 13 Jane Street, Edinburgh, EH6 5HE
Author contact
candy.hatherley@headlandarchaeology.com
Funding
Stirling Council and Headland Archaeology (UK) Ltd
e-ISSN: 2056-7421
https://doi.org/10.9750/issn.2056-7421.2020.92
Published by the Society of Antiquaries of Scotland
Data Service archaeologydataservice.ac.uk
Society of Antiquaries of Scotland
National Museums Scotland
Chambers Street
Edinburgh EH1 1JF
United Kingdom
Managing editor: Adela Rauchova
Copy-editor: Susan Milligan
Production: Raspberry Creative Type, Edinburgh
Copyright © 2020 Candy Hatherley and individual
contributors. Except where otherwise noted, this work
is licensed under Creative Commons AttributionNonCommercial-NoDerivatives licence.
http://creativecommons.org/licenses/by-nc-nd/4.0/
Requests for permission to reproduce material from a SAIR
report should be made via the Society website: http://www.
socantscot.org/publications/copyrightpermissions/.
www.socantscot.org
Registered Scottish charity no. SC010440
TABLE OF CONTENTS
List of illustrations
List of tables
v
vi
1. Abstract .
1
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2
3. Background to the fort.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
7
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
9
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
2. Introduction.
.
.
.
.
.
.
.
4. Background to the publication.
5. Results . . . . . . . . . . . . . .
5.1 Pre-fort features
5.2 The fort
5.3 Abandonment of the fort
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
6. The finds. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
6.1 Pottery by Jeremy Evans with Kay F Hartley, David F Williams, Steven H Willis &
Gwladys Monteil
6.2 Fired clay sling shots by Jeremy Evans
6.3 The metalwork by Nicholas M McQ Holmes, Fraser Hunter & Julie Lochrie
6.4 Iron, glass and stone finds by Martin Goldberg, Fraser Hunter & Julie Lochrie
6.5 The industrial remains by Effie Photos-Jones
7. The charred plant remains by Mhairi Hastie & Scott Timpany.
7.1 Method
7.2 Results
8. Discussion . . . . . . . . . .
8.1 Defences
8.2 Intervallum structures
8.3 Buildings
8.4 Material culture
8.5 Conclusion
.
.
.
.
.
.
.
.
.
.
.
.
.
.
Appendix 2 Catalogue of illustrated coarse pottery vessels.
A.2.1 Amphorae
A.2.2 Flagon
A.2.3 Mortaria
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
18
18
20
21
23
29
.
.
.
.
.
.
.
.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix 1 Catalogue of samian ware by Steven H Willis (1999) &.
Gwladys Monteil (2008)
A.1.1 Dr 18
A.1.2 Dr 27
A.1.3 Dr 29
A.1.4 Dr 35
A.1.5 Dr 37
A.1.6 Unidentified forms
10
10
10
15
.
.
.
.
.
.
.
31
31
31
36
36
37
37
39
39
.40
.
40
40
40
40
40
40
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
42
42
42
42
SAIR 92 | iii
A.2.4 Bowls
A.2.5 Jars
42
44
Appendix 3 Fabric descriptions.
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
Appendix 4 Catalogue of sling shots.
A.4.1 1999 excavation
A.4.2 2008 excavation
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
47
47
47
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
.
48
. . . . . . . . . . . . . . . . .
57
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
58
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
59
Appendix 5 Composition of plant remains from Doune Primary School .
by Mhairi Hastie & Scott Timpany
Appendix 6 Composition of the charcoal from Doune Primary School .
9. Acknowledgements .
10. References .
45
SAIR 92 | iv
LIST OF ILLUSTRATIONS
1. Site location
2. Setting and contour map
3. Plan of archaeological features: 1999 and 2008
4. 2010 trench
5. Distribution of Flavian forts and fortlets in Scotland
6. Plan of Trenches 1 and 4
7. Plan of Trenches 3 and 5
8. Sections through demolition pits (Trench 1)
9. Sling shots from 1999 and 2008
10. Copper alloy and iron finds
11. The harness fitting
12. Iron finds
13. Pottery from 1999 and 2008 excavations
2
3
4
6
8
11
13
17
20
22
24
27
43
SAIR 92 | v
LIST OF TABLES
1.
2.
3.
4.
5.
6.
7.
8.
Doune fabric proportions (1999 and 2008 assemblages combined)
Functional analysis of vessels from Doune
Copper alloy finds from 2010 excavation (excluding coins)
Summary of ironwork in the Doune assemblage divided by season of excavation
Fabric descriptions
Composition of plant remains from Doune Primary School: Buildings 1–6
Composition of plant remains from Doune Primary School: other contexts
Composition of the charcoal from Doune Primary School, 1999 excavation samples
18
20
23
25
45
48
52
57
SAIR 92 | vi
Scottish Archaeological Internet Reports 92 2020
1. ABSTRACT
Three archaeological excavations were undertaken by Headland Archaeology (UK) Ltd within the grounds
of Doune Primary School in Stirlingshire, each located on the site of Doune Roman fort. These excavations
revealed sections through triple-ditched defences, elements of the turf rampart and the perimeter road (via
sagularis) on both the west and east sides of the fort. Within the interior of the fort the partial foundations
of seven buildings were recovered, including barracks blocks, a corridored building that may represent a
workshop (fabrica) and a stable-barracks to accommodate a cavalry squadron (turma). The everyday life
of the fort was also revealed, with a series of ovens and an iron-smelting shaft furnace, a first for Roman
Scotland. A range of pits were also identified, including some which are likely to be related to the demolition
of the fort as it was decommissioned. Artefacts confirm that the fort was built and occupied during the
Flavian occupation of Scotland between ad 80 and 86–7.
SAIR 92 | 1
Scottish Archaeological Internet Reports 92 2020
2. INTRODUCTION
Archaeological excavations were carried out in
1999, 2008 and 2010 in advance of development
within the grounds of Doune Primary School,
which partially overlies the Roman fort at Doune
(NGR: NN 7272 0130; NRHE No. NN70SW 36;
Canmore ID 24767) (Illus 1 & 2). All phases of this
work were commissioned by Stirling Council and
undertaken by Headland Archaeology (UK) Ltd.
Ardoch Burn
River T
eith
Doune
701400
Doune
Doune
Primary
School
coin found
701200
NRHE air photograph
transcriptions
701000
Doune Castle
272600
272800
273000
N
KEY
0
100m
excavation area
Illus 1 Site location. © Headland Archaeology (UK) Ltd
SAIR 92 | 2
Scottish Archaeological Internet Reports 92 2020
701600
30
50
701400
40
30
30
701200
20
40
Ardoch Burn
20
River T
eith
701000
20
20
30
40
40
20
272600
272800
30
273000
N
KEY
0
100m
excavation area
Illus 2 Setting and contour map. © Headland Archaeology (UK) Ltd
SAIR 92 | 3
701350
Illus 6
TR5
119
TR2 Trench 2
Building 5
117
117
116
Building 4
701325
115
127
Building 6
Illus 5
125
114
TR1
110
123
via
sag
ular
is
109
109
096
093
104
107
108
099 100
Building 3
Building 1
Building 2
TR4
272750
272775
272700
KEY
N
SAIR 92 | 4
0
10m
trench
feature
structure
gravel surface
Illus 3 Plan of archaeological features: 1999 and 2008. © Headland Archaeology (UK) Ltd
272725
272750
Scottish Archaeological Internet Reports 92 2020
701300
098
095
092
TR3
105
102
Scottish Archaeological Internet Reports 92 2020
The first phase of work was carried out in
1999 and was directed by Colm Moloney. This
excavation covered the footprint and services of
a new nursery located to the west of the main
school building (Illus 3, Trenches 1–4). The
excavation was located on the west side of the
fort and revealed evidence for internal buildings,
a row of ovens and a furnace. It also provided
a section across the fort’s defences and the
intervallum road.
A second phase of work was undertaken in 2008
and was directed by Paul Masser. This small-scale
excavation was undertaken prior to the construction
of a classroom extension adjoining the existing
school and revealed further evidence for buildings
located within the interior of the fort (see Illus 3,
Trench 5).
A third phase of work was undertaken in
2010, directed by Paul Masser. It was located
within a garden to the east of the primary school,
on the east side of the fort (Illus 4). The triple
ditch defences of the fort and the remains of
the turf rampart were exposed alongside cobbled
surfaces relating to the intervallum road. A
group of shallow pits between the road and the
rampart contained evidence for metalworking.
Within the interior of the fort, part of a timber
building was excavated, which is interpreted here
as a cavalry barracks block. Numerous large pits
were identified within the building, and between
it and the rampart. The pits located within
the building may represent gravel quarrying
or improvised latrines, probably excavated
after the building’s main phase of occupation
was over and possibly evidence of the partial
abandonment of the fort. Some features within
this trench clearly pre-dated the fort and are
thought to be prehistoric.
The 1999 excavation is archived with the NRHE
with Project Code DPS99, the 2008 as DPSE07 and
the 2010 as DPSX10.
SAIR 92 | 5
701360
Scottish Archaeological Internet Reports 92 2020
252
701340
270
293
325
285
324
244
323
322
260
turf
rampart
286
010
via
sag
ular
is
701320
257
009 049
146
4 poster
Structure 123
barracks block
272800
KEY
N
0
5m
272820
metalworking pit
trench
gravel surface
feature
prehistoric feature
cobbled surface
fort ditches
Illus 4 2010 trench. © Headland Archaeology (UK) Ltd
SAIR 92 | 6
Scottish Archaeological Internet Reports 92 2020
3. BACKGROUND TO THE FORT
The fort was initially identified during aerial survey
by the RCAHMS in 1983 (Maxwell 1984). The
aerial photographs identified one set of ditches on
the south-east turned inwards to unite in what is
known as a ‘parrot’s beak’, indicating the position
of the entrance. The probable location of the via
sagularis was also identified during this survey. A
small-scale investigation was carried out in 1984
by Gordon Maxwell, who identified a number of
construction trenches associated with single-phase
timber buildings, with 1st-century ad pottery
recovered (Frere 1985: 275). The pottery confirmed
the Flavian date of the fort, as previously suggested
by the parrot’s beak entrance (Maxwell 1998). An
archaeological evaluation in 1999 confirmed that
the foundations of buildings associated with the fort
were located within the footprint of a proposed new
nursery development (Moloney 1999a).
The Roman fort lies on a low promontory formed
by the River Teith and the Ardoch Burn (see Illus
2), close to the medieval Doune Castle. The ground
falls away sharply into the river valley on the south
side of the fort and it is bounded to the east side
by the burn. The interior of the fort occupies a
level plateau and it is conceivable that it guarded a
crossing of the Teith, which is fordable at this point.
The significance of the fort has been discussed by
Maxwell, who speculated that the Teith could be the
Tameia / Tamia in Ptolemy’s Geography (Maxwell
1984: 221–2).
Maxwell described the fort as a ‘route blocker’
(1984: 218), similar to the Flavian forts on the outer
line from Drumquhassle to Stracathro (Illus 5). Here
he also speculated that the forts of Drumquhassle,
Malling, Bochastle and Doune could have formed a
temporary frontier along the northern extremity of
the Forth–Clyde isthmus (ibid). The fort at Doune
also lies close to the presumed line of the Roman
road north of Camelon, which is thought to have
crossed the upper reaches of the River Forth close to
Stirling (Crawford 1949: 18–26; RCAHMS 1963:
112–15). This Roman road is presumed to run into
southern Stirling. Excavations across the probable
line of the road at Beechwood Park in Bannockburn
identified a cobbled surface (Cook 2014). Postmedieval pottery and coins were recovered which
could indicate that this cobbling was a later surface,
potentially built over the line of the Roman road.
Alternatively, the cobbles could be the original
Roman road surface with later material incorporated
into it through later use. The road is identified again
north-east of Dunblane, close to the Allan Water
en route to Ardoch fort (Woolliscroft & Hoffmann
2006: 85).
There is enough uncertainty about the route of
the Roman road between Stirling and Dunblane
that alternative routes through the extensive mosses
west of Stirling have been proposed, but none have
been recognised on the ground so far (eg Crawford
1949: 18–21; Woolliscroft & Hoffmann 2006:
80–1).
SAIR 92 | 7
Scottish Archaeological Internet Reports 92 2020
Stracathro
Inverquharity
Cardean
Inchtuthil
Cargill
Fendoch
Bertha
Dalginross
Strageath
Bochastle
Ardoch
Malling
Doune
Drumquhassle
Cramond
Camelon
Mollins
Elginhaugh
Barochan
Easter Happrew
Newstead
Castledykes
Loudoun Hill
Oakwood
Crawford
Cuppuck
Learchild
Milton
Drumlanrig
High Rochester
Dalswinton
Ladyward
Broomholm
Ward Law
Glenlochar
Birrens
Carlisle
Corbridge
N
0
25km
Illus 5 Distribution of Flavian forts and fortlets in Scotland. © Headland Archaeology (UK) Ltd
SAIR 92 | 8
Scottish Archaeological Internet Reports 92 2020
4. BACKGROUND TO THE PUBLICATION
The publication of these excavations at the
Roman fort of Doune has a convoluted history.
The publication of the initial 1999 excavation
(Moloney 1999a, 1999b) was in the process of
being refereed by Britannia when the second
phase of work was commissioned in 2008 (Masser
2008). It was therefore considered best practice
to delay publication until the two phases of work
could be combined into a single paper. During
the final preparation of this combined 1999/2008
paper, a third phase of work was commissioned
in 2010 and again publication was delayed, with
the hope of incorporating all three phases into a
single publication report. Unfortunately, due to no
explicit reference being made by Stirling Council
(as both the client and the curator) of a requirement
to fund the publication of the 2008 and 2010
excavations, Headland Archaeology (UK) Ltd was
only contracted to produce assessment reports for
these phases of work (Masser 2008, 2010). In a
bid to present here (at minimum) a summary of
the results for the entire archaeological fieldwork
undertaken within the fort between 1999 and 2010,
this paper is a detailed report of the 1999 and 2008
excavations (an edited version of the combined
1999/2008 paper) with a summary of the results of
the 2010 excavation.
SAIR 92 | 9
Scottish Archaeological Internet Reports 92 2020
5. RESULTS
5.1 Pre-fort features
A number of probable prehistoric features were
identified within the 2010 trench. Some of these
clearly pre-dated the fort, as they were cut by
the foundation of the barracks block located
within the excavation and were much lighter in
colour than the Roman features. Other probable
prehistoric features with no relationship to the
Roman archaeology were identified by this colour
difference (see Illus 4). Although none of these
features produced finds, they were interpreted
as prehistoric rather than representing an earlier
phase of Roman activity on the site. Most obvious
of these was a shallow linear ditch cut by the
foundations of the barracks building, and a possible
four-poster structure located within the footprint
of the building.
5.2 The fort
5.2.1 Ditches and rampart
A section across the defences on the north-west
side of the fort was excavated in Trench 2 (Illus 3).
Three parallel ditches were identified here, which
would have formed the outer defences of the fort
(Contexts 123, 125, 127). These ditches were all a
U-shaped profile with a maximum depth of 0.8m
and c 4m wide. A spread of soil was recovered from
the inner ditch, sealing the primary silting deposit
of probable turves which could have come from
the levelled remains of the rampart. The outer two
ditches showed no evidence of silting and appeared
to have been filled directly (and deliberately) with
the spread from the rampart.
In the 2010 excavation the three ditches covered
a total area of 13.5m from the inner edge of the
inner ditch (C293) to the outer edge of the outer
one (C252) (Illus 4). The inner and outer ditches
were similar in size, approximately 3m wide and
1.5–1.6m deep, while the middle ditch (C270)
was smaller, 2.3m wide and 1.1m deep. All three
ditches had steep, V-shaped sections with verticalsided ‘ankle-breaker’ slots, 0.35m wide, at the
base. The inner ditch was markedly asymmetrical,
sloping more gently on the inner edge towards the
rampart.
The fills of all three ditches contained a comparable
sequence of deposits, which suggested that they had
been deliberately partially backfilled at some stage to
leave shallow-sided ditches less than half the original
depth. Following this partial backfilling, which must
reflect the initial slighting of the fort’s defences, fine
silty deposits accumulated within all three ditches,
reflecting gradual silting and soil development over
an extended period.
To the south of these ditches were the remains
of the turf rampart. The rampart was 6m wide and
survived as an upstanding deposit up to 0.15m
thick, with a distinctive soft, silty and almost
stone-free texture. This material overlay a truncated
buried soil no more than 0.1m thick, with no turf
line visible, indicating that the area must have
been de-turfed prior to the rampart’s construction.
While no structure could be discerned within the
core of the rampart, which presumably consisted
of randomly dumped turf, it was edged on either
side by distinct lines of patchy, pale yellow silt
approximately 0.5m wide (C285 and C286). These
deposits must represent the facing or revetting of
the rampart with clay or stacked turf.
5.2.2 The intervallum road (via sagularis)
A gravel surface, interpreted as the via sagularis, was
identified running diagonally across Trenches 1 and
4 from north-east to south-west (Illus 6) and in two
areas in the 2010 trench (Illus 4). In Trenches 1
and 4 it measured a maximum of 6.2m wide and
appeared to turn towards the south-east. A narrow
drainage ditch was identified running along the west
edge of the road capped by stone slabs.
The surface of the intervallum road survived
in two areas in the 2010 excavation, the central
section truncated away. In both locations, a cobbled
surface no more than 2m wide was flanked by much
more extensive, though patchy and discontinuous,
spreads of gravel. The cobbled surfaces were fairly
loose and unconsolidated, and it is possible that
the finer upper surface of the road had washed off,
resulting in the spreads of gravel to either side.
Several pits in the area between the turf rampart
and barracks building would seem to predate the
road surfaces, since layers of stone continuous with
the cobbled surfaces were slumped into the upper
fills of the pits.
SAIR 92 | 10
701320
125
TR1
054
123
003
008
005
via
sag
ul a
ris
006
011
024
002
055
023
092
018
038
061
090
701300
047
Building 1
015
036
Building 2
TR4
272740
KEY
SAIR 92 | 11
N
0
5m
trench
feature
structure
gravel surface
Illus 6 Plan of Trenches 1 and 4. © Headland Archaeology (UK) Ltd
272760
272780
Scottish Archaeological Internet Reports 92 2020
043/044
Scottish Archaeological Internet Reports 92 2020
5.2.3 Metalworking
A group of very shallow pits, cut into a layer of
gravel immediately north of the intervallum road
in the 2010 excavation, contained concentrations
of ironworking slag (Illus 4). No actual structures
relating to metalworking (such as smithing hearths
or smelting furnaces) were found and it is possible
that the main focus of this activity lay further to
the south-east, beyond the limit of excavation.
Fragments of slag were also found in a very shallow
but well-defined pit (C010), located approximately
7m further to the west; and in one of the fills of Pit
C146 to the south: these may derive from the same
source.
A narrow linear feature (C257) was excavated
between the metalworking features and the cobbled
road surface, running parallel to the line of the
defences. This appeared to be a foundation trench
similar to those defining buildings in the interior of
the fort, with a square profile 0.4m wide. A similar
feature (C244) was seen at the opposite end of the
site and it is possible that these were the remains of
timber buildings located between the via sagularis
and the rampart.
5.2.4 Ovens and furnace
Parallel to the north-west side of the via sagularis
in Trench 1 was a line of five ovens (C002, 006,
018, 023 and 024). These were in varied states of
preservation, but all were roughly circular in plan,
2–3m in diameter, and of identical construction.
The main structure in each case comprised a circular
wall constructed of roughly squared stone blocks
bonded with yellow clay. The ovens had paved floors
which, in all cases, were sealed by the collapsed
reddened clay domes that originally covered the
structures. It is likely that the ovens had been
built at the back (the south-east edge) of the turf
rampart and may have been slightly set into it, as
at Fendoch in Perthshire (Richmond & McIntyre
1939: 137–8). A large spread of burnt material was
identified to the east of the ovens as the accumulated
rake-out from the ovens (C008). This material built
up to such a degree that it eventually encroached on
the via sagularis. Analysis of the charcoal content
of the rake-out demonstrated that oak and hazel
were the dominant wood species present and
were probably the main source of fuel. The only
significant concentrations of charred grain (mostly
barley) associated from the ovens came from within
oven C024 and from the rake-out (C008). The grain
may reflect that ovens were used for roasting grain
as well as baking.
Adjacent to the ovens the badly disturbed remains
of a metalworking furnace were identified (C003).
This comprised a shallow irregular hollow 4.3 ×
3.3m wide, no more than 0.2m deep, filled with slag
and daub. Analysis of the industrial waste (6.5 ‘The
industrial remains’ below) suggests that the material
derives from an ironworking shaft furnace. It would
appear that the garrison had a smith among its
ranks, and that iron was being smelted on site and
not simply worked. The furnace also contained a
deposit with several iron objects, including a spiked
loop (SF202, probably from a timber structure) and
a bar fragment (SF003). The presence of pottery
sherds and concentrations of charred grain in the
same deposit implies that the industrial waste was
not in situ and this was a dump of material after the
furnace was no longer in use.
5.2.5 Buildings
The foundations of seven rectangular timber
buildings (Illus 3, 4, 6 & 7) were clearly identified
within the interior of the fort. These have each been
interpreted as barracks blocks to house the soldiers,
their equipment and horses. Fragments of further
buildings were also exposed within the trenches.
5.2.5.1 Building 1
Immediately to the south-east of the intervallum
in Trench 1, a series of construction trenches
were identified which formed two rectangular
structures aligned north-east to south-west (Illus
3 & 6). The west structure (C054) measured 23m
in length but continued beyond the edge of the
trench. It was 4.6m in width and subdivided
into rooms each measuring 3.6m in length. The
east structure (C036) was more complete and
measured 18.2m in length by 4.6m in width, and
was subdivided into rooms each measuring 3.6m
in length, with one double room in the centre
measuring 7.2m.
The similarity between the two structures is
striking. They lie parallel to each other, separated
SAIR 92 | 12
701340
TR3
025
TR5
037
029
Building 5
034
023
011
014
119
006
016
117
SW limit of layer 005
060
701330
116
052
modern drains
007
049
054
Building 4
115
056
114
272730
KEY
SAIR 92 | 13
N
0
trench
feature
2.5m
structure
Illus 7 Plan of Trenches 3 and 5. © Headland Archaeology (UK) Ltd
272740
272750
Scottish Archaeological Internet Reports 92 2020
Building 6
Scottish Archaeological Internet Reports 92 2020
by a central gap measuring 2.6m wide, and were
divided into uniformly sized rooms. It seems
likely that together they represent two wings of
a corridored building. The construction trenches
for both wings measured 0.4m in width and 0.2m
in depth on average. In places the base of the cut
had slight depressions, which were interpreted
during the excavation as the impressions of the
bases of upright timber posts. A single post pipe
for a squared post was identified, measuring
0.13m square. Bent nails and charcoal fragments
were recovered from the fill throughout the
construction trenches of the building. A sample
taken from the fill of C054 (C053) produced
charcoal which was identified as hazel and oak
with smaller quantities of alder, a probable
indication of the types of wood used in the
construction of this building.
5.2.5.2 Building 2
The north-west corner of a second building (defined
by construction trenches C043 and C044 – Illus
3 & 6) was identified at the south-east corner of
Trench 1 and in Trench 4. Two post holes associated
with Building 2 contained structural evidence. One
of these (C015) contained a post pipe, suggesting
that the post had rotted in situ. The post pipe was
square in plan and, as in Building 1, measured
0.13m square. The second post hole (C047) was
elongated and may have held a double post setting.
The posts here appear to have been removed and
the feature was backfilled with a deposit (C046)
containing burnt wattle and daub and a number
of bent iron nails. A sample from C046 produced
large quantities of carbonised cereal grain, which
may represent food stored in the vicinity, if not in
Building 2 itself.
5.2.5.3 Buildings 3 and 4
Features identified to the east in Trench 3
undoubtedly represent other buildings of similar
construction to those seen in Trench 1. Less can
be said about their layout due to their limited
exposure in a 2m-wide trench. Linear construction
trenches (C092, 095 and 098) appear to represent
a building aligned north-west to south-east, divided
along its length with larger rooms to the northeast and smaller rooms to the south-west (Illus
3 – Building 3). Another building with similar
layout and dimensions was seen to the north-east
(defined by construction trenches C056, 114, 115
and 116) (Illus 7 – Building 4). A number of other
construction trenches which probably represent
further buildings were recognised between Buildings
3 and 4.
5.2.5.4 Building 5
The foundations of a timber building covered
the north-east half of Trench 5, consisting of
vertical-sided trenches 0.35–0.50m deep (C023,
025, 034 and 037 – Illus 7). The foundation
trench along the south-west side (C023) probably
represents the front of the building. It was deeper
than the others, with the variation in depths
probably reflecting post-in-trench (rather than
sleeper-beam) construction, although no trace of
individual post holes or post pipes could be seen.
From the limited area exposed, the foundations
appear to define a row of rooms, at least one of
which was 3.7m wide and 2.4m deep, across the
front of the building, with another row of rooms
to the rear.
Parallel with C023, 2.4m to the south-west, was
another foundation trench (C006), only 0.25m
deep, punctuated by three post holes 0.6m deep
(C011, 016 and 060). Post holes C011 and C016
were exactly opposite the foundation trenches C025
and C034; this, along with the precise alignment
and spacing of C006 with C023 (identical to the
spacing between C023 and C037), strongly suggests
they all formed part of the same building, with
C006 perhaps representing an open veranda along
its front. The backfill of all the foundation trenches
was remarkably sterile, with no trace of charcoal or
artefacts.
Two steep-sided, sub-rectangular pits (C014
and 029) were located centrally within two rooms
of Building 5. Neither pit showed any sign of
weathering; their primary fills consisted of black
silt deposits with very high concentrations
of charcoal in the base, which merged into
paler upper fills containing lesser (though still
appreciable) quantities of charred material.
While numerous pits, thought to be related to
the demolition of the fort, were identified in
Trenches 1–4 (see 5.3.3 ‘Demolition pits’ below),
Pits C014 and C029 are distinctive in that they
seem to be precisely located within the rooms of
SAIR 92 | 14
Scottish Archaeological Internet Reports 92 2020
the building and thus potentially related to its
occupation, rather than destruction.
5.2.5.5 Building 6
Two linear features in Trench 5 (C007 and
049), located to the south-west of Building
5, are thought to be foundation trenches for
another timber building (Illus 7). Both features
were relatively slight, less than 0.3m deep, and
contained backfill deposits barely distinguishable
from the surrounding subsoil. Linear C007
terminated 4.8m from the south-east corner of
the excavation, linear C049 adjoining it at right
angles and extending to the south-west. A small
circular depression at the junction of the two
features (C052), and a pronounced widening
and deepening (C054) at the south-west end of
C049, may represent post holes related to the
building.
Building 6 is unlikely to be contemporary
with Building 5, since it has a markedly different
orientation. It is also close enough that it would
probably have interfered with access to this structure
if they had been contemporary builds. The slight
nature of the foundations of Building 6 and its pale
colour may indicate that this was an earlier building,
potentially a temporary structure occupied during
the construction of the fort, or an early building
within the fort that was abandoned before it was
completed.
5.2.5.6 Building 123
The south end of the 2010 excavation contained
the foundations of a timber building. Enough of
the ground plan was exposed to give its overall
dimensions as 18.2 × 9m. It was divided into five
two-room units, each 3.6m wide.
Fifteen large pits were cut into the floor of
Building 123, with only two of the eight rooms
more than half-exposed not containing pits.
These features had been dug within the rooms
of the building and mainly respecting the line
of the walls, indicating that they were excavated
while the walls were still visible, even if only as a
ruin. These pits were also characterised (as were
those in the intervallum zone) by the complexity
of their fills: all contained multiple deposits,
which suggested that they had lain open for a
time before being eventually backfilled, possibly
in several stages and with occasional cases of
recutting. One pit (C049) contained a large
quantity of nails and other items including sling
shots while another (C009) contained an early
1st-century ad strap junction decorated in the
Mirror-style of southern England (see 6.3.2.2
‘2010 finds’ below; Illus 11).
Three rooms, fully or partially exposed on the
south-west side of the building, each contained
linear trenches aligned north-east to south-west,
slightly offset from the centre of the room. These
features were detached from the foundation trenches
of the building at either end, and although they were
a similar width, they tended to have more rounded
profiles and had quite different fills, consisting of
dark brown organic silt. These potentially represent
drains in the floor of the building.
A section of foundation trench was identified to
the south-east of Building 123. This may represent
another building, or the continuation of Building
123.
5.3 Abandonment of the fort
There is strong evidence that the fort was
abandoned, with some of the buildings showing
signs of being removed and burnt, the ditches
deliberately backfilled and the ramparts slighted.
The identification of rubbish pits cutting through
the foundations of Building 1 and within the
interior of the fort, containing broken pottery,
ironwork and sling shots, shows that there
was effort put into concealing rubbish prior to
departure.
5.3.1 Ditches and ramparts
The ditches on both sides of the fort appear to have
been deliberately backfilled, which reduced their
depth considerably and presumably made them
ineffective as defensive works. Turf identified within
the inner ditch suggests that the rampart was also
slighted at this time. On the north-west side of the
fort the rampart survived as a spread of soil which
sealed the ditches, which may suggest that the bank
may also have been deliberately slighted here during
the evacuation of the fort.
SAIR 92 | 15
Scottish Archaeological Internet Reports 92 2020
5.3.2 Buildings
Some of the buildings appear to have been
deliberately dismantled prior to abandoning the
fort. The evidence of this is most striking in Building
5 where, overlying the majority of the cuts for the
foundation trenches, was a layer of dark, charcoalrich silt (C010, 012, 017, 032, 035, 038, 040, 042
and 059), nowhere more than 0.1m thick. This
layer appears to represent an episode of burning that
concluded the use of the building. Since no trace
of posts was seen in the foundations, it seems likely
that the superstructure of the building had been
dismantled and the posts removed when the fire
occurred. A dark brown layer (C005), up to 0.2m
thick, overlay much of the dark charcoal-silt layer
covering Building 5. This soil horizon was located
within a slight hollow and remained unaffected by
subsequent ploughing or modern disturbance and
contained finds of exclusively 1st-century Roman
date.
Almost all the finds from the 2008 excavation
derive from the charcoal-rich destruction layer
overlying the foundations of Building 5, the
backfill of Pits C014 and 029, and from the
overlying soil layer C005. The metalwork is of
particular interest, including a number of items
interpreted as tools – a blacksmith’s punch (Illus
10:4), a fine file (Illus 10:5) and a thin knife blade
tip (Illus 12:6), which may reflect activities carried
out by occupants of the building. The punch and
the file were both found in Pit C029, hinting at
a possible use of the pit as a tool store, although
their presence may be accidental. Concentrations
of hobnails and lorica hamata chain mail links
throughout the charcoal-rich destruction deposits
were probably items of broken and discarded
equipment not considered worth salvaging. A
copper alloy carrying handle (Illus 10:2), possibly
from a casket, came from C005. Quantities of
nails, an iron T-clamp (Illus 12:12) and fragments
of daub presumably derive from the structure
of the building itself. The pottery assemblage
included a high proportion of samian ware, which
is consistent with identification of the building
as living quarters where food preparation and
consumption were taking place. Environmental
samples from contexts relating to the charcoal-rich
destruction layer produced high concentrations of
cereal grain, mostly spelt wheat and barley, which
probably reflect food preparation and/or storage
within Building 5.
5.3.3 Demolition pits
Four large rubbish pits were identified cutting
the construction trenches of Building 1,
suggesting that they may be related to the
abandonment of the building and of the fort
(Illus 6 & 8). The majority of the artefacts from
this trench came from these pits. Two of the
pits were extremely large, very similar in shape,
and are likely to relate to the destruction of
Building 1. The backfill of oval Pit C055 (C056)
comprised dark brown sandy silt with numerous
sherds of pottery. The upper fill of oval Pit C061
(C060) comprised dark brown silty clay with
moderate inclusions of gravel and charcoal, and
contained quantities of pottery including samian
and mortarium sherds and an iron intaglio ring
(Illus 10:3).
Two less substantial rubbish pits also truncated
Building 1. Pit C005 (Fill C004) contained large
quantities of pottery and nails, and other finds
included two javelin heads (Illus 12:8, cat no.
9 not illus) and a chain link (Illus 12:15). See
6.4.4 ‘Catalogue of iron’ below. Charred barley
and (less common) wheat grains recovered from
the fill may relate to food production or storage
within Building 1. Pit C011 lay in the corridor
between the two wings of Building 1. As with
some of the other pits, the upper fill of the pit
(C010) contained the most finds, in this case
a fragmentary catapult bolt head (cat no. 10),
a C-shaped iron timber clamp (cat no. 13, not
illus) and amphora sherds. See 6.4.4 ‘Catalogue
of iron’. A sample of the fill contained significant
quantities of charred cereal grain, principally
barley, a similar assemblage to that from Pit C005.
Further pits were identified in a service trench
adjacent to the north entrance to the school
building (not illus), although the trench was
only 0.5m wide and did not go deep enough to
impact on the basal fills of the pits. However, the
upper fills of two pits were investigated and these
produced a number of clay sling shots (see Illus
9; 6.2 ‘Fired clay sling shots’ below) which were
presumably buried deliberately.
SAIR 92 | 16
Scottish Archaeological Internet Reports 92 2020
004
004
005
005
010
010
038
037
039
040
041
011
011
056
056
055
055
060
060
064
065
066
061
061
N
0
1m
Illus 8 Sections through demolition pits (Trench 1). © Headland Archaeology (UK) Ltd
SAIR 92 | 17
Scottish Archaeological Internet Reports 92 2020
6. THE FINDS
6.1 Pottery
Jeremy Evans with Kay F Hartley, David F Williams,
Steven H Willis & Gwladys Monteil
Some 284 sherds weighing 11,074g were recovered
from the 1999 excavation, 271 from stratified
contexts. The quantity of pottery is small but is just
about adequate to give some broad outlines to the use
of ceramics on the site. A further 29 sherds weighing
263g were recovered from the 2008 excavation, 24
from stratified contexts. The pottery from the 2010
excavation produced 513 sherds, which were assessed,
but a full report was never commissioned. See Illus
13 and the catalogue in Appendix 2 ‘Catalogue of
illustrated coarse pottery vessels’.
In summary, the pottery types identified in the
collection include amphorae, samian ware, mortaria
and Black Burnished Wares. The other sherds are
mostly coarse vessels of sandy, red, cream and grey
fabrics, although there are some finer redware
examples. The fabric proportions from the 1999 and
2008 assemblages are combined in Table 1.
Sherds of amphora are by far the most common,
at 155 sherds, with varying curvatures suggesting
a range of shapes. Black Burnished Ware is the
second most common type recovered, numbering
86 sherds. These vessels tend to have a light grey
fabric with characteristic black surfaces and feature
sherds including small flat bases and everted rims.
There is very little samian ware and few mortaria
in the collection. The samian ware numbers only 18
sherds. They are mostly small bodysherds but there
is a foot-ring base sherd and three decorated sherds.
There are four sherds of mortaria, one of which is a
large rim sherd stamped ‘MAVIVE’.
Only the 1999 and 2008 assemblages are reported
in detail below. The catalogue of samian ware is
in Appendix 1, a catalogue of illustrated coarse
pottery vessels is in Appendix 2 and a table of fabric
descriptions (Table 5) is in Appendix 3.
6.1.1 Date
All the material appears to be of Flavian date, with
the exception of the heirloom, Claudio-Neronian
mortarium (Illus 13:6), a number of which seem
to appear in Flavian contexts in northern England
and Scotland, and the Neronian Dr 29 (samian
Table 1 Doune fabric proportions (1999 and
2008 assemblages combined)
Fabric
A01
% no. of
sherds
51.4
% weight
F01
0.3
0.2
F02
0.6
0.0
M01
0.3
0.5
M02
3.2
8.1
M03
1.6
3.4
M04
0.3
0.5
O01
8.6
1.2
O02
0.3
0.0
O03
1.6
0.2
O04
1.0
0.1
O05
0.6
0.4
O06
2.2
0.1
Q01
2.6
0.4
R01
2.6
0.3
R02
0.6
0.1
R03
2.6
0.2
R04
1.3
0.6
R05
0.3
0.2
S10
6.1
0.5
W01
9.0
0.5
W02
1.9
0.3
W03
1.0
0.3
Total
313
11,337g
82.0
catalogue, no. 4). There is also a sherd, probably
of Cologne colour-coated ware, which perhaps
post-dates ad 80. The samian ware (see Appendix 1
‘Catalogue of samian ware’ below) gives the closest
dating evidence, although none of the material is
very closely datable, and the best range is ad 65–90,
although the historical context of the fort suggests
that like Inchtuthil it should be dated c ad 83–86/7.
SAIR 92 | 18
Scottish Archaeological Internet Reports 92 2020
6.1.2 Fabric supply
The 1999 excavation assemblage is dominated by
amphora sherds, unusually so even for a military site,
with 56% by sherd count, 83% by weight, and even
14% by minimum numbers of rims. Although there
is a concentration of amphorae in the ‘destruction
pit’ (C060), even without this amphorae would
comprise 47% by count and 71% by weight. Levels
of amphorae from other military assemblages by
count is normally in the 5–10% range, with weight
figures in the range 30–45%, as can be seen at
Binchester, Birdoswald, Catterick, Carlisle and
Walton-le-Dale (Hird 1997; Ferris 2011). Notably
higher figures by weight come from Flavian Brithdir
(Evans 1997) at 64%. At Brithdir, in north Wales,
the assemblage is associated with the rampart back,
as is the assemblage here and the group with the
higher amphora figure from Birdoswald (Hird 1997;
Wilmott 1997). The consistent presence of high
amphora levels in rampart back locations may partly
explain the high amphora level in this assemblage.
The assemblage is completely dominated by
Baetican Dressel 20 oil amphora sherds with no fish
sauce amphorae present and no wine amphorae. Two
fineware fabrics are present, making up only 1.1%
of the assemblage; one is probably Central Gaulish
Pompeian redware 3 and the other is probably from
Cologne.
Mortaria are well represented in the assemblage,
four sources being present: Noyon in north-east
Gaul; a source perhaps in Central Gaul;
Verulamium; and a local source, probably Doune
itself. The commonest fabric is that of Noyon,
followed by Verulamium region ware, with the local
and possibly Central Gaulish sources represented
by single pieces.
Some 13.3% of the assemblage is composed
of oxidised wares. The commonest is fabric O01,
which seems likely to be of local origin as may be
fabrics O02 and O05, which with O01 may form
a continuum. White-slipped oxidised flagon fabrics
occur in only a single fabric Q01, at around 3%.
Greywares make up 7.1% of the group by count,
a rather lower level than the oxidised wares, as might
be expected in a northern military assemblage of this
date. Forms represented are globular jars with short
everted rims which can be paralleled at other Flavian
forts in the region.
Whitewares apparently make up 9.9% of
the assemblage by count, but most of these are
accounted for by a large number of sherds from a
single vessel in W01. Much of the oxidised ware,
greyware and whiteware contain varying quantities
of fine gold mica, like the sling shots, and all these
fabrics are probably of very local origin.
Samian ware is surprisingly poorly represented
in the group. Military assemblages usually contain
10% or more of samian ware (Evans 1993; Willis
1998).
The 2008 excavation pottery changes the overall
site figures very little because the assemblage size is
much smaller. As just discussed, the 1999 excavation
group is heavily amphora-dominated, but this is
not the case with the 2008 excavation group. Here
amphorae levels are probably a little lower than
those that might be found on a typical fort (cf Evans
2001: fig 11). Oxidised wares are well represented
and outnumber greywares, as might be expected on
a military site of this date, and whitewares are well
represented. Mortaria sherds are absent. Samian
ware is very strongly represented, both by count
and weight, in contrast to the assemblage from the
1999 excavation (cf Willis 2006).
The basic differences between these two groups
are probably not the result of sample size or
happenstance. It has been becoming evident for
some time that rampart back locations in forts
seem to produce more amphorae. As Doune itself
shows, these are the locations for bread ovens and
seem likely to have been used for various elements
of food preparation. That mortaria are also well
represented here brings to mind the mortarium
from Usk (Wright et al 1976: 391, no. 66 Pe]lveis
Contubernio Messoris), showing they could be owned
communally by the contubernium, presumably for
their communal food preparation. That more samian
comes from Trench 5, well in the interior of the fort,
equally follows established patterns. Samian tends to
be scarcer on rampart back locations, where it is not
needed in food preparation, but commoner where
status display and food consumption are located.
6.1.3 Samian ware from the 2008 excavation
Gwladys Monteil
Eight pieces of samian ware were recovered from
three contexts at Doune Primary School in 2008.
SAIR 92 | 19
Scottish Archaeological Internet Reports 92 2020
Three decorated sherds, four plain sherds and a
scrap, all South Gaulish in origin, make up this small
assemblage. The samian group is quite fragmented
and consists of fairly small sherds. No evidence of
repair was recorded.
A single foot-ring fragment from a dish Dr
18, recovered from C010, is unfortunately too
fragmentary to display a stamp. C018 yielded
a slightly larger and more varied group. Two
decorated fragments from a dish Dr 29 in the
style of the T-1 decorative group are particularly
interesting. They are pre-Flavian in date (see
Appendix 1 ‘Catalogue of samian ware’). They do
not display particularly distinctive signs of wear or
use compared to the other samian sherds but could
possibly be part of an heirloom. The rest of the
samian group is more difficult to date as precisely.
The presence of a Dr 37, unfortunately too small
to provide a close date range, does suggest a Flavian
date for the deposition.
6.1.4 Discussion
As noted above, amphorae levels are high, even for a
military assemblage, and samian ware levels are low
in the 1999 assemblage primarily associated with the
rampart back. The small 2008 group from the fort
interior shows neither of these features and, indeed,
samian is very well represented.
The functional composition of the group (Table
2), although with a low number of vessels overall,
also appears rather unusual (cf Evans 1993), with
high levels of amphorae and mortaria, although the
low jar level relative to tablewares is usual for early
military assemblages. As noted above, there is some
suggestion of higher amphora levels in forts from
rampart back locations, and it might be that these
and the mortaria reflect food preparation here (the
rampart back ovens providing clear evidence for
this taking place here). Although surfaces are often
1
0
Table 2 Functional analysis of vessels from
Doune
Vessel type
Flagons / constrictednecked jars
Jars
Bowls
Lids
Beakers and cups
Mortaria
Amphorae
No. of rims
No.
2
Percentage
11.1
4
2
1
2
5
2
18
22.2
11.1
5.6
11.1
27.8
11.1
eroded, there is little evidence of sooting on vessels;
a comparatively low level of sooting seems to be the
case quite often on early military assemblages.
The limited evidence for supply here would
suggest that only specialist vessels, such as amphorae,
mortaria and finewares, were being brought in over
large distances with any frequency. This picture
appears to be typical for military installations of
the mid-Flavian campaign period in Scotland. The
coarsewares, including a minority of the mortaria,
would seem likely to have been of fairly local
manufacture (cf Darling 1985).
6.2 Fired clay sling shots
Jeremy Evans
Eight sling shots were recovered from the 1999
excavation, and an additional two from the 2008
excavation (Illus 9). A total of 135 were found in the
2010 excavation, mainly within the pits cutting the
floor of Building 123. The report below only details
the 1999 and 2008 finds. The catalogue for the sling
shots is in Appendix 4.
3
4
4cm
4m
Illus 9 Sling shots from 1999 and 2008. © Headland Archaeology (UK) Ltd
SAIR 92 | 20
Scottish Archaeological Internet Reports 92 2020
Nos 1–8 were all recovered from monitoring of a
service trench adjacent to the north entrance to the
school, rather than the main trenches. The weights
of the reasonably complete sling shots vary from 15
to 39g, although their sizes are generally rather more
consistent. A close set of parallels to these sling shots
comes from Strageath (Frere & Wilkes 1989: 177),
the second closest fort north from Doune, although
the 91 Strageath examples all come from Antonine
contexts. Many of the Strageath sling shots also share
the double-cone form generally employed here.
Greep (1987) has reviewed the distribution of
sling shots in Britain. Most clay sling shots from
military sites have been recovered from Wales and
Scotland, with shots of the double-cone form
coming from sites north of the Antonine Wall and
round shots coming from Antonine Wall sites. The
Doune examples add to this pattern and would
seem to suggest that the Strageath examples might
be residual Flavian pieces rather than Antonine ones.
6.3 The metalwork
Nicholas M McQ Holmes, Fraser Hunter & Julie
Lochrie
It is particularly unfortunate that an area of
corrosion on the obverse of this coin coincides with
the latter part of the obverse legend, commencing
at the very point where the consular numeral
potentially becomes extremely important for
dating purposes. If the inscription reads COS XII,
the coin falls into the long-recognised group of
virtually unworn bronze coins minted in ad 86
which represent the latest found on a number
of Flavian military sites in Scotland, including
Inchtuthil, Stracathro, Dalginross, Strageath,
Camelon and Crawford (Robertson 1983: 419)
and Elginhaugh (Bateson 1989: 167). These coins
have been quoted as evidence for the evacuation
of these sites in ad 87 or very shortly after. It is
impossible to be certain whether the consular
numeral on this coin is XII or XIII, although the
symbol following XII does look rather more like
a vertical upright than the C of CENS. Were the
coin to have been minted in ad 87, it would appear
to be the most northerly find of a coin of this date
on a Flavian site (Hobley 1989).
6.3.1.2 The 2010 coins
Fraser Hunter
6.3.1 Coins
Five coins have been identified in total, one prior to
the 2008 excavation by the school janitor (location
shown on Illus 1) and four during the 2010
excavation.
6.3.1.1 The 2008 coin
Nicholas M McQ Holmes
This was a stray find, retrieved by the school
janitor, Mr Robert Kinnaird, while laying lighting
cables some time previously and handed in to the
excavation team in 2008 for identification (Illus
10:1).
▶ Domitian copper as
28.5 × 27.5mm, 8.20g, die axis 180°; ad 86–7
Obv: [IMP C]AE[S] DOMIT AVG GERM COS
XII[I? ]; head laureate right
Rev: MONETA AVG[VST(I)]; Moneta standing
left, holding scales and cornucopiae; S to left and
C to right in field.
Surfaces oxidised, with some pitting and accretion;
apparently unworn.
The 2010 coins were two of silver denarii (SF137,
SF972) and two of copper alloy (SF008, SF747).
The silver coins were around 20mm in diameter and
the copper alloy examples were larger at 27mm and
28mm respectively. SF008 is an as or dupondius but
is unidentifiable. SF747 is a Flavian dupondius. The
2010 coins are in a terrible condition and no details
can be discerned on their surface. No conservation
has been untaken and none have been X-rayed.
6.3.2 Copper alloy finds
Fraser Hunter
6.3.2.1 2008 finds
The only copper alloy find from the 1999/2008
excavations (aside from the coin described above)
is a fine decorated handle, probably from a casket
or an item of furniture.
▶ Fine U-shaped decorated handle, the terminals
narrowed and out-turned, one broken. The facetted
section bears punched decoration on the central and
lateral spines, with chevron decoration on the latter
and sinuous diagonals on the former. The concave
SAIR 92 | 21
Scottish Archaeological Internet Reports 92 2020
1
3
0
1cm
2
4
0
5
2cm
Illus 10 Copper alloy and iron finds: 1 coin; 2 handle; 3 intaglio ring; 4 punch; 5 file. © Headland
Archaeology (UK) Ltd
SAIR 92 | 22
Scottish Archaeological Internet Reports 92 2020
Table 3 Copper alloy finds from 2010 excavation (excluding coins)
Small Find no.
804
189
823
829
876
438
868/905
Identification
Belt or strap mount
Enamelled circular harness mount with four-way distributor fixing
Unknown
Pin shank
Bent rivet
Headstud brooch
Two joining fragments of an enamelled headstud brooch
moulding where the loop narrows to the terminals
is defined by two or three incised transverse grooves.
A fine carrying loop, perhaps from a casket. L:
28.5mm; W: 17mm; Th: 4.3mm. DPSE07, SF001,
C005 (Illus 10:2).
6.3.2.2 2010 finds
Aside from the two copper alloy coins, there were
eight copper alloy finds in the 2010 excavation
(Table 3). As with the entire 2010 assemblage these
have been only briefly assessed and only the harness
fitting (SF189) is commented on below.
The harness fitting (SF189) from C005/009
is undoubtedly the highlight of the assemblage
(Illus 11). It is a strap junction decorated in
the Mirror-style of southern England, typically
early 1st century ad, so an heirloom by the time
it came north, perhaps from a solider recruited
from the south. There are two fastenings on the
rear for the strap. It was originally four-armed
but one arm has been broken off. It has a very
nicely done decoration of a three-arm whirligig,
engraved with parts highlighted with engraved
basketry hatching and red enamel, around a
central trefoil design. The arms have engraved
scrolling designs, again with red enamel. This
fits into a wider pattern of troops bringing
material from previous postings, and also (by
implication) of troops recruited in the south
serving in the northern conquest. In this specific
case of southern ‘Celtic metalwork’ on Scottish
Roman sites, you can find parallels in ‘lipped
terrets’ (a southern form) from Cargill and
Newstead (former unpublished; from David
Woolliscroft/Brigitta Hoffmann’s work there;
latter Macgregor 1976: no. 63).
6.3.3 Other non-ferrous finds
Julie Lochrie
The 2010 excavation identified 13 lead alloy finds
and one possible silver find (SF968). The lead alloy
finds all appear to be waste fragments, although
SF806 is plano-convex in shape and may be an
ingot.
6.4 Iron, glass and stone finds
Martin Goldberg, Fraser Hunter & Julie Lochrie
6.4.1 Iron
Martin Goldberg & Fraser Hunter
Below is a summary of the 1999 and 2008
assemblage. The 2010 excavation produced 641 iron
finds, of which 453 are nails and 87 are hobnails.
Much of the iron is currently unidentifiable due to
corrosion but there may be knives, mail fragments
and tools present. No further assessment has been
undertaken of this assemblage.
The 1999/2008 excavations produced a small
but diverse assemblage of ironwork. Notable finds
include a ring complete with intaglio, a range of
militaria and various tools. The assemblage (17
objects; 16 pieces of chain mail; 123 nails and 52
hobnails) is summarised in Table 4.
The most striking item of personal ornament is
an iron finger ring which still retains its chalcedony
intaglio (Illus 10:3). The design of two interlocked
cornucopiae symbolises prosperity and fertility.
Militaria are represented by loose chain mail
links, javelins and catapult bolts. The type of
light throwing spears found at Doune is typical of
auxiliary equipment; they fall within Manning’s
SAIR 92 | 23
Scottish Archaeological Internet Reports 92 2020
0
4cm
KEY
enamel
0
2cm
damaged enamel
Illus 11 The harness fitting. © Headland Archaeology (UK) Ltd
SAIR 92 | 24
Scottish Archaeological Internet Reports 92 2020
Table 4 Summary of ironwork in the Doune assemblage divided by season of excavation
Season
Jewellery
1999
Intaglio
2008
Military
Tools
equipment
2 javelin heads;
catapult bolt
Chain mail
Blacksmith’s
punch; file;
?file; knife
Group 1A (Manning 1985: 162–5), a common type
with parallels inter alia from Newstead (Curle 1911:
188–9). More unusual is the evidence for artillery, in
the form of the catapult bolt head; post-excavation
fragmentation renders identification awkward, but
it has all the characteristics of such a bolt head.
This need occasion no surprise, as the discovery
of catapult parts from Elginhaugh illustrates that
artillery was not the sole preserve of the legionary
(Allason-Jones 2007: 405–7; Hanson 2007: 658–9).
The range of tools indicates the variety of activities
taking place within the fort. Metalworking is
represented by the blacksmith’s punch. There are one
or two files; the identification of one is uncertain,
but the fineness of the other suggests a role in
metalworking. Carpentry is represented by nails
and other structural fittings such as various clamps.
The only knife fragment has a notably thin blade,
suggesting it is not from a robust multi-purpose tool
but a finer, more specialist implement.
Remarkably, only a single hobnail was recovered
from the 1999 excavation, compared to 51 from
the 2008 excavation; possible reasons are discussed
below. The hobnails (Manning 1985: 135, type 10),
have short shanks and small domed heads, suitable
for sandals and boots. Those from DPSE07 C010
are smaller than the majority, suggesting a smaller
or finer shoe than the normal military type. Notable
concentrations came from DPSE07 C018, 036 and
010, the latter clearly representing deposition of an
intact shoe, as several are fused together.
Structural fittings and fixtures are represented
by a common range of types, including a doublespiked loop, a C-clamp and a T-clamp with anchor
head (Manning 1985: 130–2). The fixtures and
fittings are dominated by nails, as is normal. Most
Fixtures and Hobnails
fittings
Nails; clamp;
1
chain link;
spiked loop
Nails;
51
T-clamp;
washer
Miscellaneous
Bar fragment;
fragment
Bar fragment;
decorative
mount
frequent are Manning’s type 1B (below 15cm in
length), square-sectioned with tapering shanks and
a sub-rectangular flat head. The 1999 excavation
had a fairly homogeneous nail assemblage of
Manning type 1B, with C004, 010, 046 and
056 producing the highest numbers. Many were
straight, indicating that they were deposited still
in the wood. Two of these contexts (C004 and
056) come from features interpreted as rubbish/
destruction pits; the nails presumably relate to the
discard of wooden structural elements with intact
nails. By contrast, DPSE07 C015 has several nails
with bent shanks, suggesting removal from the
wood and discard within the secondary deposit
of Pit C014. DPSE07 C036 had high numbers of
both nails and hobnails from the lowest levels of
foundation slot 034; the nails may be structural, but
the hobnails may represent the loss or deposition
of a shoe during construction.
Most intriguing among the miscellaneous items
is a fine broken diamond-shaped decorative mount
with an ornamental terminal. No parallels have yet
been noted, but it is likely to be a decorative mount
from an organic item.
6.4.2 The 1999 and 2008 iron finds in context
While fixtures and fittings (especially the ubiquitous
nail) are the most common finds from both
excavation phases, there are notable differences
between the two assemblages, as summarised in
Table 4. Most obvious is that the 1999 excavation
produced weapons but no tools, whereas the
2008 excavation produced tools but no weapons.
However, these differences are more apparent than
real; both assemblages are largely the product of
just two pits, DPS99 Pit C005 and DPSE07 Pit
SAIR 92 | 25
Scottish Archaeological Internet Reports 92 2020
C029, interpreted as rubbish deposits. The relatively
small scale of the assemblages should caution against
drawing any radical conclusions.
Foundation trenches such as DPSE07 C034
(Building 5) occasionally contained nails and other
items (eg fragmentary knife blade, SF103), but
the majority of finds came from pits interpreted as
discard or destruction deposits. The key significant
difference between the finds from the two excavation
phases is that significantly more hobnails and chain
mail fragments were found in 2008. As sampling
strategies were consistent, the patterns are likely
to be meaningful, and probably indicate rather
more intense activity in the barracks area, with the
movement and everyday activities of the troops
leading to the incidental loss of individual hobnails
and loose chain mail links. The rear of the rampart
either saw less activity, or the users were rarely
armoured and booted.
6.4.3 Conclusion
While small, the Doune iron assemblage is an
interesting one, with a wide range of ironwork.
Little of it is surprising in itself, but it includes
some striking finds (notably the intaglio and the
militaria) which add colour to our picture of life
on the frontier, while different loss patterns among
the smallest of artefacts, the hobnails and chain mail
links, suggest varying activity patterns in different
areas of the fort. With the rest of the finds, it is only
as larger samples of the site are dug, or more work is
done on comparing assemblages from different sites,
that any wider patterns will emerge.
6.4.4 Catalogue of iron
6.4.4.1 Jewellery
to corrosion on one side. Max Diam: 27mm; inner
Diam max: 19mm; band Diam 4mm; bezel W:
8mm. DPS99, SF1, C060, upper fill of Pit 061
(Illus 10:3).
6.4.4.2 Tools
Fraser Hunter
▶ Blacksmith’s punch, rounded tip, tapering
cylindrical shank, expanded head cracked in half.
A common type (Manning 1985: 9–11), with
parallels from the Blackburn Mill and Carlingwark
Loch hoards (Piggott 1953: 38 & 48, cat C.64–7
& B.45–7). L: 96mm; Diam at tip: 21mm; Diam at
top of shank: 30mm; head Diam: 34mm. DPSE07,
SF101, C028, lower fill of Pit 029 (Illus 10:4).
▶ Fine file. No traces of teeth due to corrosion, as
is often the case, but the form is typical of a file;
parallel-sided rectangular-sectioned bar tapering
towards tip and narrowing at one end to a broken
tang. L: 132mm; 8 × 3mm, narrowing to 3 × 3mm
at tang. DPSE07, SF102, C028, lower fill of Pit
029 (Illus 10:5).
▶ Fine knife blade tip. Both edges of tip sharpened;
fine blade, form unclear due to corrosion. L: 60mm;
max H: 15mm tapering to 3.2mm at broken tip;
Th: c 1mm. DPSE07, SF103, C036, lower (backfill)
deposit in foundation trench 034 (Illus 12:6).
▶ Tool? Function unclear; form suggests tang and
broken blade, possibly from a file. Rectangularsectioned bar tapering and narrowing to a rounded
end; other end broken, with a large corrosion blister.
L: 64mm; 2.5 × 5mm at tip; 9 × 16mm at widest
point. DPSE07, SF104, C042, upper (destruction)
deposit in foundation trench 037 (Illus 12:7).
Fraser Hunter
6.4.4.3 Military equipment
▶ Iron intaglio-set finger ring, the circularsectioned hoop swelling to the bezel (Henig 1978:
fig 1, type III) in which is set an oval flat semitranslucent dark grey chalcedony intaglio with
bevelled upper edges (10 × 12mm). The intaglio
bears a pair of interlocked cornucopiae flanking an
ear of corn. Both are standard symbols of fertility
and prosperity, which are common on gemstones;
for parallels to their use in combination (although
with the corn set in a vessel), see Henig (1978: no.
401 and app 54). Unrelated traces of wood adhere
Fraser Hunter
▶ Javelin head with damaged socket.
Approximately symmetrical diamond-shaped
blade; socket fragmentary, with remains of
wooden shaft. Manning group 1A; close parallel
in a diamond-shaped blade from the Durden
collection (Manning 1985: 164, no. V53). Overall
L: 82mm; head L: 67mm; W: 22mm; Th: 3mm;
socket surviving L: 15mm, outer Diam: 10mm;
inner Diam: 5mm. DPS99, SF004, C004, fill of
Pit 005 (Illus 12:8).
SAIR 92 | 26
Scottish Archaeological Internet Reports 92 2020
6
15
12
0
8
7
17
2cm
Illus 12 Iron finds: 6 knife blade; 7 tool; 8 javelin head; 12 T-clamp; 15 chain link; 17 decorative
mount. © Headland Archaeology (UK) Ltd
SAIR 92 | 27
Scottish Archaeological Internet Reports 92 2020
▶ Damaged javelin head. Leaf-shaped blade with
broken tip; a second non-joining fragment from
the same context is probably from the same object.
The lentoid-sectioned tip is blunted and bent
from use. The blade has a low belly at c 10% of
blade length (70mm from tip); no mid-rib; heavily
laminated, making thickness difficult to determine.
The tapering closed socket has no apparent rivet
(Diam: 13mm at neck, flaring to 15mm); wooden
shaft fragments in the socket. The leaf-shaped blade
is more typical of Manning’s Group 1A (Manning
1985: 162–5). Fragment 1, L: 127mm; spear head
surviving L: 82mm; W: 26mm; socket L: 45mm;
outer Diam: 15mm; inner Diam: 9mm. Fragment
2, L: 46mm; W: 20mm. DPS99, SF006, C004, fill
of Pit 005.
▶ Fragmentary remains of catapult bolt head.
Heavily corroded tapering socket with wooden
shaft remains. Other fragments appear to form the
neck of spear head, its original form uncertain,
but the square section and short length suggest
a tapering pyramidal point typical of artillery
bolt heads (Manning 1985: pls 82–3). Postexcavation corrosion makes it too fragmentary
to illustrate. Min L: 92mm; head L: min 34mm;
D: min 14mm; socket outer Diam: 14mm; inner
Diam: 9mm. DPS99, SF060, C010, upper fill of
Pit 011.
▶ Lorica hamata – chain mail links. Two complete
small punched rings (SF106, SF109); three riveted
rings (SF108); fragments of 11 others (SF105).
Diameters vary between 4.5 and 7mm but thickness
of wire is consistently 1.5mm. These fragments were
all from sample residues, generally from the upper
levels of features (DPSE07 C010, 035, 046) or
later deposits (C005). Pit C029 contained chain
mail fragments from both its lower and upper fills
(Sample 21 from the lower deposit C028, Sample
22 in its later fill C046). DPSE07, SF105, 106, 108,
109, C005, 010, 028, 035, 046.
6.4.4.4 Fixtures and fittings
Fraser Hunter
▶ T-clamp with anchor-shaped head; broken
square-sectioned shank with perpendicular tapering
down-curved arms. Intact, no wood traces and
thus not in situ when buried. For parallels see
Curle (1911: 289); Manning (1985: 132). Shank
L: 42mm; W: 5mm. Head L: 40mm; W: 8mm.
DPSE07, SF110, C042, foundation trench (Illus
12:12).
▶ C-shaped timber clamp. Formed from a
rectangular-sectioned bar, the broad end turned
through 90° and the tip a further 90°; bent 90° at
other tapered end. Bar bowed from use. L: 134mm;
W: 52mm; bar section 13 × 10mm; flattened end
24mm wide × 9mm thick; clamped round timber
Th: 31mm. DPS99, SF201, C010, upper fill of Pit
011.
▶ Spiked loop. Heavily fragmented. A common
type of fitting for timber architecture; compare
Curle (1911: 289); Manning (1985: 130).
L: 57mm; spike L: 38mm; head W: 27mm; loop
Diam: 10mm. DPS99, SF202, C003, hollow
containing iron-smelting debris.
▶ Chain link. Complete figure-of-eight link with
a small fragment of a second link attached. For
parallels see Manning (1985: 139). L: 39mm;
W: 21.5mm; Diam: 6mm. DPS99, SF002, C004,
fill of Pit 005 (Illus 12:15).
▶ Nails. There were 77 iron nails recovered in
the 1999 excavation and 46 nails from the 2008
excavation, all of Manning type 1B (Manning 1985:
134): < 150mm in length with flat, sub-rectangular
or slightly rounded heads. The 52 hobnails were all
of Manning type 10. Only one came from DPS99;
those from DPSE07 were found in concentrations
(> 2) in C010, 018, 036 and 045. In a number of
these, corroded organics indicate the deposition of
complete or fragmentary shoes.
6.4.4.5 Miscellaneous
Fraser Hunter
▶ Fine decorative mount. Diamond-shaped fitting,
broken at one end, with a decorative terminal which
expands in angular bifurcated form; small triangular
mouldings flank it. Flat or slightly plano-convex
section, with no rivet holes surviving; probably a fine
decorative clamp or mount for an organic object.
L: 44mm; W: 4.5–12mm; Th: 1–2mm. DPSE07,
SF113, C018, lower fill of Pit 014 (Illus 12:17).
SAIR 92 | 28
Scottish Archaeological Internet Reports 92 2020
6.4.5 Glass
Julie Lochrie
There were 32 glass finds recovered in the 2010
excavation, including three turquoise faience melon
beads. Melon beads are a common type very popular
during the Roman period.
The other glass sherds in the collection are very
similar thin curving fragments of green or blue
colour. It is unclear if they belong to bottles or vessels
but a few of the sherds have raised ribs which may be
decorative. A collection of larger sherds are likely to
be from the same vessel and their retrieval from the
primary fill of a pit located between the intervallum
road and the edge of the rampart provides a Roman
date. Similarly, three sherds were retrieved from the
primary fill of a pit inside one of the buildings and
are also certainly Roman in date.
6.4.6 Stone
Julie Lochrie & Fraser Hunter
There are 16 pieces of one or more Niedermendig
lava stone querns, imported from the Rhineland,
Germany identified in the 2010 excavation. Similar
querns have been discovered before in Scotland.
One complete example was found at Newstead,
Roxburghshire (NMS Cat no. 000-100-037-477-C).
Other stone finds include two possible tool
fragments, two whetstones, a weight, a pivot stone,
a prehistoric flint tool and a worked shale fragment.
The fragment was originally thought to be part of a
large D-sectioned bangle but has been re-identified
as a rim of a vessel (Hunter 2014: 158).
6.5 The industrial remains
Effie Photos-Jones
Significant quantities of industrial waste were
recovered from only four contexts, all from the
1999 excavation. The majority (7.1kg) came from
a shallow scoop (C003) cut into the surface of the
intervallum road. Smaller quantities (less than 1.0kg
each) were recovered nearby from the spreads of
burnt material in front of the ovens (C008 and 029)
and the fill of a large demolition pit (Pit C055, fill
C056). Slag, likely to be derived from ironworking,
was found in various contexts across the 2010
excavation. The levels are not large, 1,799g, but
enough to suggest ironworking in the area. Three
examples of likely furnace lining and three crucibles
certainly seem to confirm this. All the crucibles
were recovered from C029, strongly suggesting
metalworking nearby, although surprisingly there
was no associated industrial waste from this context.
Only the 1999 industrial waste is reported on below.
The industrial waste was initially classified as slag
from hand specimens. On sectioning the slag it
became obvious that the majority was highly fired
metallurgical ceramic rather than metallurgical slag
proper. Metallurgical slag was peculiarly absent in any
form other than small prills adhering to the surface of
the metallurgical ceramic. SEM-EDX analysis showed
that the metallurgical ceramic had been involved in
an iron-making/smelting process due to the presence
of metallic iron adhering to the surface of the ceramic.
The molten metal had been deposited rather than
reacted with the clay matrix; no reaction is obvious
between matrix and metallic area other than iron-rich
‘penetration’ into the clay matrix.
A number of fragments of the ceramic showed a
pronounced gradient from a black and glassy inner
surface to a low-fired, red external one with a grey
and vitrified area in between. This suggests that
the original container or structure to which these
fragments belonged was free-standing with its outer
surface exposed to an oxygen-rich atmosphere and
the source of heat contained within. The thickness
of the fragments indicates that this was a furnace
rather than a crucible. They could not have been
part of a smithing hearth since these rarely tend to
be vitrified apart from the area around the tuyère,
the rest consisting simply of heated clay.
Thus the evidence points to the vitrified clay
fragments being part of a small iron bloomery
furnace. It could have been free-standing or
embanked like those shown by Tylecote (1986)
for the Roman period in England. However, no in
situ remains of the furnace were found within the
excavated area at Doune and it is not possible to base
a detailed reconstruction on the recovered fragments
of furnace wall. The only distinctive fragment was
one piece resembling a tuyère (an identification
based on its narrow inner diameter and the extent
of vitrification). Alternatively, it may have served as
an air inlet as part of the furnace construction.
The location of the majority of the furnace debris
in C003, close to the back of the rampart, suggests
that the furnace itself may have been built into the
SAIR 92 | 29
Scottish Archaeological Internet Reports 92 2020
rampart. This location was also used for the ovens,
and for both furnace and oven this may reflect the
need to keep processes involving fire away from the
highly flammable timber buildings of the fort.
This is the first example of a shaft furnace from the
Roman period in Scotland. The design is seen widely
in England at this time but the local tradition in
Scotland is simple bowl furnaces like that excavated
at Tarras Farm, Forres (Will 1998; Photos-Jones
1999) and indicated at the Roman fort at Rough
Castle on the Antonine Wall (MacIvor et al 1980).
The evidence from Doune therefore indicates the
presence of a smith in the fort using local materials
to manufacture iron in a Roman-style furnace.
SAIR 92 | 30
Scottish Archaeological Internet Reports 92 2020
7. THE CHARRED PLANT REMAINS
Mhairi Hastie & Scott Timpany
The charred plant assemblages from the 1999 and
2008 excavations at Doune are presented here.
Samples were taken in both excavations from the
main features across the site(s), including identifiable
buildings and associated pits, together with bread
ovens and the base of a possible furnace. Charred
cereal grains were present in most contexts from
across the site, with particularly rich assemblages
attained from contexts associated with Buildings 3
and 5.
7.1 Method
7.1.1 Plant macrofossil assessment and charcoal
identifications
Samples were processed in laboratory conditions
using a standard floatation method (cf Kenward
et al 1980). Identifications were confirmed using
modern reference material and seed atlases,
including Cappers et al (2006). Plant taxonomic
nomenclature used in Tables 6 and 7 follows the
order of Stace (1997). Charcoal identifications were
made using wood keys by Schweingruber (1990)
and IAWA (1989).
7.2 Results
The results for the two phases of excavation
are presented in Tables 6 & 7 (in Appendix 5:
Composition of plant remains from Doune Primary
School) and Table 8 (in Appendix 6: Composition
of the charcoal from Doune Primary School),
which show the materials recovered from both the
retent and flot samples and take into account the
suggestions of van der Veen et al (2007). All plant
material was preserved through charring.
7.2.1 Charred cereals
Charred cereal grains dominate the charred plant
assemblage from both phases of excavation (see
Tables 6 & 7). Grains of hulled barley (Hordeum
vulgare), including rare quantities of naked barley
(Hordeum vulgare var nudum) were recovered,
together with oat (Avena sp), club/bread wheat
(Triticum aestivo-compactum), emmer wheat
(Triticum dicoccum) and spelt wheat (Triticum spelta).
Where possible the hulled barley has been recorded
as either having a ‘straight’ or ‘twisted’ central groove
to potentially differentiate the presence of two-row
(straight) and six-row (twisted) barley. A number
of degraded cereal grains were also present, which
could not be identified to species or family level;
these are recorded as indeterminate cereal grains
(Cereal indet). Together with the cereal grain, rare
quantities of barley and possible spelt wheat rachis
fragments were recovered from two samples (see
Tables 6 & 7).
7.2.2 Wild taxa
A wide variety of wild taxa are present within the
assemblage, with generally increased numbers
in those contexts containing large numbers of
cereal grains. In general the wild taxa fall into two
categories; those relating to agricultural weeds and
those relating to damp/boggy ground. Agricultural
weeds are present in samples from both phases of
excavations and include taxa such as ribwort plantain
(Plantago lanceolata), corn spurry (Spergula arvensis),
sheep’s sorrel (Rumex acetosella agg), buttercups
(Ranunculus sp) and corn marigold (Chrysanthemum
segetum). A number of taxa relating to damp/boggy
ground are present in the assemblage, including
a suite of sedge species (Carex sp), wood-rushes
(Luzula sp), violets (Viola sp) and spike-rushes
(Eleocharis sp). Grasses (Poaceae sp) are also well
represented in the assemblage, with meadow grass
(Poa sp), brome/false brome (Bromus/Brachypodium
sp) and heath grass (Sieglingia decumbens sp) present,
together with a number of grains, which could
not be identified to species level and are listed as
indeterminate. The only arboreal taxon present is
hazel (Corylus avellana), through charred nutshell
recovered from two contexts (see Table 8).
7.2.3 Charcoal
A small number of arboreal species were identified
during the charcoal analysis. The assemblage is
largely dominated by oak (Quercus sp) and hazel
charcoal. Present in lesser quantities are: alder (Alnus
glutinosa), birch (Betula sp), willow (Salix sp) and
plum/rowan (Rosaceae indet). The results are shown
in Table 8.
SAIR 92 | 31
Scottish Archaeological Internet Reports 92 2020
7.2.4 Discussion
7.2.4.1 The intervallum: bread ovens and furnace
Samples were taken from contexts relating to each
of the five bread ovens present on the site (DPS99
C002, 006, 012, 018 and 023), together with
the associated rake-out deposits (C008 and 029)
surrounding the ovens. The assemblages recovered
from the ovens are generally poor, with all but
two contexts (C012 and 023) being sterile. C023
contained only a single charred indeterminate cereal
grain. The greatest concentration of cereal grain
was recovered from C012 (relating to oven C024),
which contained in excess of 70 grains of hulled
barley, together with a significant number of wheat
grains, some of which have been identified as spelt
wheat. Wild taxa are also present in this assemblage.
A similar grain assemblage is also shown in one of
the rake-out deposits (C008) although in smaller
quantities (see Table 7).
The charred cereal grains present in bread oven
C024 are likely to reflect the last use of the oven:
with bread ovens being cleaned out on a regular basis
there would be less build-up of grains and associated
wild taxa accumulating within them (Clapham
2007). As the primary use of the bread ovens would
have been for baking rather than cooking, the cereals
present within bread oven C024 are likely to have
derived from secondary sources, such as adhering
to the sides of bread (Clapham 2007) or accidental
transfer from people. However, roasting of grain
in the ovens to prevent spoilage cannot be ruled
out entirely (eg Robertson 1941–2) and this may
well explain the presence of arable weeds within the
assemblage.
It is likely that both barley and wheat were being
used in bread production at Doune, as is reflected
in the overall grain assemblage recovered from the
site. Barley bread is thought to have been inferior to
wheat bread but was used by the Roman military,
particularly to feed lower ranked soldiers, while
higher ranking soldiers would have consumed the
wheat bread (Davies 1971). The higher proportion
of barley at Doune could suggest bread was largely
being produced for the lower ranked soldiers or it
could simply reflect local availability of cereal types.
It is thought cereals would have been requisitioned
from the local population by the Roman army
(Johnson 1983). Barley is known to have been the
more common cereal produced in the north of
Britain (Dickson 1989) and has been evidenced from
other Roman sites in Scotland dating to around the
time of the occupation of the fort at Doune such
as Bearsden (Knights et al 1983), Cramond (Hastie
2006) and Elginhaugh (Clapham 2007). Wheat
such as spelt wheat is inferred to have been grown
in Scotland during this period from its presence at
Iron Age sites such as Culduthel, near Inverness
(Timpany 2007). However, it is likely it was grown
in smaller quantities than barley and thus the spelt
wheat at Doune may reflect a combination of wheat
brought to the site by the Roman army together
with local production (Dickson 1989; Boardman &
Dickson 1995; Dickson & Dickson 2000). Together
with the baking of bread it is suggested that cereals
would also have been used to make foodstuffs such
as porridge and soups (Dickson 1989).
Charcoal fragments recovered from the bread
ovens and the rake-out show an oak and hazel
dominated assemblage (see Table 8), with smaller
quantities of willow, alder and plum/rowan,
indicating the use of local wood for fuel. The high
number of sedge nutlets within the assemblage from
bread oven C024 (C012) suggest that turf/peat may
also have been used on occasion to fuel the ovens.
One sample was taken from the base of the
possible furnace (DPS99 C003), which is thought
to have been used primarily for metalworking and
the manufacture of iron products on site (see below).
However, the presence of over 100 cereal grains
dominated by barley, together with occasional wild
taxa of arable weeds such as buttercups, knotgrass
and chickweed (Stellaria sp cf S media) from this
context (see Table 7) would seem out of place for
a furnace assemblage. Given the close proximity of
this context to that of the oven rake-out deposits
(C008), it seems more likely that this context
contains a mix of both oven and furnace waste.
Thus the shallow pit feature containing C003 may
represent an accumulation of rake-out material from
both the ovens and furnace. The charcoal assemblage
from this context is dominated by oak fragments
with rare quantities of hazel, plum/rowan and
possible alder (see Table 8), all of which were likely
to come from local sources. The dominance of oak
charcoal, which burns at high temperatures, reflects
the industrial nature of this area of the site for both
metalworking and baking.
SAIR 92 | 32
Scottish Archaeological Internet Reports 92 2020
7.2.4.2 Buildings and pits in the interior of the fort
Five samples have been analysed from Building
1. These were taken from one of the construction
trenches (DPS99 C053), Pit C005 (C004 and
005), the top fill of Pit C011 (C010), and Pit C055
(C056). Samples were also analysed from the fill of
the construction trenches of Building 2 (C042) and
Building 3 (C094), and Pit 105 (C106).
The samples associated with Building 1, the
most substantial structure excavated, show a
general spread of grain across the interior of the
structure, with grain becoming trapped in the fills
of demolition pits and beam slots, following the
destruction of the structure. The grain assemblage
for the structure is dominated by barley, with lesser
incidences of spelt wheat and oat (see Table 6). The
wild taxa present are all indicative of arable weeds.
Identification of charcoal fragments from this
structure show utilisation of oak and hazel together
with wood of alder, willow and plum/rowan (see
Table 8). The absence of any charcoal fragments
of substantial size suggests they are more likely to
represent the remnants of fuel wood rather than
building timbers, although it is likely that oak and
hazel would have also have been the main arboreal
taxa used to provide construction materials for the
buildings.
The arboreal taxa represented are suggestive of
utilisation of mixed dryland/wetland woodland,
with species such as oak, hazel and plum/rowan
preferring drier areas, and alder and willow
indicative of damp/wet areas. It is generally observed
that by the Roman period large-scale clearance of
woodland had already taken place to make way
for agricultural land (eg Tipping 1997; Clapham
2007). However, the location of the fort near to
the River Teith suggests that all of these species may
have been growing locally on the wetland around
the river and the immediate dryland. The presence
of oak woodland in this area during the time of the
Antonine Wall has been noted by Tipping & Tisdall
(2005), who suggest that where this woodland
persisted it would have been a managed resource
by local people to provide wood for fuel and timber.
The assemblage from Building 2 (C042) is
extremely limited, particularly in contrast to that
from the rubbish pit outside this structure (see
below). Only a very limited number of barley and
spelt wheat grains were recovered from this building
(see Table 6). Charcoal fragments analysed from this
structure show that oak, hazel and cf plum/rowan
timbers were utilised, probably for fuel.
The Building 3 assemblage (C094) is much
more diverse and far richer than that of either of
the previous two buildings. A large quantity of both
charred cereal grain and wild taxa were recovered.
Over 1,500 grains were identified from the fill, with
barley grains dominating the assemblage and smaller
amounts of spelt wheat and oat also present (see
Table 6). On the whole, the assemblage is similar
to that from Building 1, but on a much larger scale.
The large number of grains from this single context
suggests that Building 3 may have been used as, or
was near to, a granary on the site before it was burnt
down. Together with the grain, rachis fragments
of barley and possible spelt wheat were present in
rare amounts. This small quantity of chaff-related
material suggests the crop was relatively clean and
had thus been processed elsewhere. However, seeds
of arable weeds such as ribwort plantain, buttercups,
docks (Rumex sp) and pale persicaria (Persicaria
lapithifolia) within the assemblage indicate that the
remains of some ruderals were able to survive the
cleaning process. A large quantity of grass caryopses
also occur within the assemblage, the large majority
of which have been identified as small-grained grasses
(see Table 6). The presence of such large numbers
of grasses may relate to one or a combination of:
grasses growing as arable weeds and being collected
with the crop during harvesting; grasses being used
as bedding or flooring; grasses present in turf being
burnt as fuel and/or used as a construction material.
Of particular note is the presence of heath grass in
the assemblage, which does not commonly grow
in arable fields today, but is thought to have been
a common agricultural weed in the past, until it
was removed from the field ecosystem as a result
of the change from the ard to the mould-board
plough during the medieval period (Hillman 1981).
Charcoal analysis from this structure indicates that
oak and hazel were the primary fuel woods used
(see Table 8).
The assemblage from C106, the fill of pit C105,
within a potential structure was similar to that of
Building 3, containing a large quantity of cereal
grain. However, the assemblage here is dominated by
oat rather than barley, and wheat is absent altogether
(see Table 6). The presence of a large quantity of
SAIR 92 | 33
Scottish Archaeological Internet Reports 92 2020
oat grain, compared to the other plant assemblages
from the site, could suggest that the grain stored in
an associated building was being used for a different
purpose. One possibility is that the oat may have
been used for feeding livestock such as horses, thus
indicating the potential presence of a cavalry unit
housed at Doune. However, Huntley & Stallibrass
(1995) considered this regarding other Roman forts
and could not find an obvious correlation between
sites with oat grain and those which housed cavalry
units.
Only a small number of wild taxa were recovered
from C106, which include sorrel, wild radish
(Raphanus raphinistrum) and hemp-nettle (Galeopsis
sp). These again are largely indicative of arable
weeds, with wild radish in particular notorious as a
troublesome weed (Clapham et al 1962). Charcoal
fragments identified from this structure show that
alder is the dominant wood type used, with oak
and hazel present together with possible birch and
plum/rowan wood.
Two contexts from individual pits located between
Buildings 2 and 3 were sampled. The fill of Pit 090
(C091), that closest to Building 3, was found to
contain only a single grain of possible spelt wheat
(Triticum cf spelta) together with a rare quantity of
oak and possible hazel charcoal fragments, and as
such offers little insight into the function of the pit.
The fill of Pit 047 (C046) just outside of Building
2, however, contains a much larger assemblage of
cereals, dominated by barley (see Table 7). It is
thought this pit may have been a post hole relating
to Building 2 that had the post removed during the
destruction of this structure. The assemblage from
the pit fill contains a larger number of grains than
was recovered from the building itself. Together with
hulled barley, the assemblage also contains a small
number of oat grains and wild taxa associated with
arable land such as ribwort plantain, buttercups,
grasses and sheep’s sorrel. The pit is also noted to
contain fragments of burnt wattle and daub relating
to the destruction of the building and thus the grain
may relate to storage of cereals within Building
2, the majority of which was not exposed during
excavation. Charcoal fragments from Pit C047 were
found to consist only of oak, which may relate to
fuel; oak was also recovered from Building 2 C042.
Samples were analysed from 12 contexts (C005,
009, 012, 015, 017, 018, 021, 028, 035, 044, 045
and 046) relating to Building 5. Charred cereal
grain within the building appears to be concentrated
within the rectangular pits C014 (C015 and 018)
and C029 (C028, 044, 045 and 046), a buried soil
layer (C005) and the upper fill of construction trench
034 (C035), situated between the two pits. There is
a general scatter of grain across the other contexts
sampled, with grains present in low quantities,
together with occasional wild taxa of largely arable
weed species including fat-hen (Chenopodium sp cf
C album), mustards (Sinapis sp) and buttercups.
The grain assemblages from pits C014 and 029,
occupation layer C005 and construction trench
C034 are dominated by a mixture of spelt wheat and
hulled barley. The Pit C014 assemblage is seen to
be dominated by spelt wheat, with over 120 grains
present in one context (015), while that from Pit
C029 is dominated by hulled barley, with two-row
barley in particular appearing to be the favoured
variety used (see Table 7). The assemblages from
construction trench C035 and buried soil C005 are
a mixture of both spelt wheat and hulled barley,
with spelt wheat being the most abundant species
present.
The high numbers of spelt wheat found within
Building 5 contrasts with that of the assemblages
from the DPS99 excavation, where hulled barley was
generally the dominant taxon (with the exception of
Pit C105 (C106) where large numbers of oat grains
were recovered). The large quantity of spelt wheat
recovered from Building 5 may reflect a shift in
cereal production between the lives of the buildings
or differential storage of cereals within Building 5.
Along with spelt wheat and hulled barley, smaller
quantities of bread/club wheat, emmer wheat and
oat are also present in the assemblage from Building
5. A post hole from this building (C017) contained
only two cereal grains: one of naked barley (Hordeum
vulgare var nudum) and one of emmer wheat. The
presence of naked barley, which is more prevalent
on prehistoric sites, may represent either reworked
material or a remnant crop.
The two pits (C014 and 029) together with
high concentrations of grain also contain rich wild
taxa assemblages (see Table 7). The assemblages
contain a number of arable weed species, such as
buttercups, ribwort plantain, corn marigold, sheep’s
sorrel and fumitories (Fumaria sp), which are likely
to have been transported with the grain during
SAIR 92 | 34
Scottish Archaeological Internet Reports 92 2020
harvesting. Together with the arable component of
the assemblages, a significant quantity of taxa from
wet/damp habitats is present. These are particularly
well illustrated by the diverse sedge assemblage
recovered, with species present including possible
downy-fruited sedge (Carex sp cf C filiformis), large
yellow-sedge (Carex sp cf C flava) and slender sedge
(Carex sp cf C elongata), which grows in wet places
such as reed beds (Clapham et al 1962). Other damp
ground taxa present include cinquefoil (Potentilla
sp cf P erecta), great wood-rush (Luzula sp cf
L sylvatica) and common spike-rush (Eleocharis sp cf
E palustris). Also of note is the significant number
of grasses recovered within the pits. This mixture of
charred grain, arable weeds and damp/wet ground
taxa is likely to reflect the mixing of the occupation
and destruction layers following the destruction
of the building, which is thought to have been
burnt down. The sedges and other wet plant taxa
are likely to represent the use of turf/peat in the
construction of Building 5, such as for the wattle
(shown by the presence of daub) and/or for the
roofing of the building. The concentration of sedges
and grasses within the pits could also relate to some
form of lining, but the separation of materials from
the occupation and destruction layers is extremely
difficult (eg Gustafsson 2000) and thus it is difficult
to say with any certainty to which they may relate.
One context (C021) was sampled from a
construction trench fill within the partially exposed
Building 6 (see Table 6). The assemblage from
Building 6 contains only a very limited number of
charred plant remains consisting of a single hulled
barley grain and a single mustard fruit. As such, the
assemblage does not provide any information on the
function or construction of the building.
SAIR 92 | 35
Scottish Archaeological Internet Reports 92 2020
8. DISCUSSION
The 1999, 2008 and 2010 excavations have allowed
for important evidence to be recovered which will
assist our understanding of the activities of the
Roman army in Doune in the latter part of the 1st
century ad.
The true extent of the fort can now be
extrapolated from a combination of the cropmark
evidence, topography and excavated features. The
main entrance to the fort was already identified
as a cropmark, and now the ditches forming the
north-west and north-east side of the fort have been
revealed during the excavations. The southern extent
of the site is curtailed by a sharp drop in ground level
into the valley of the Teith and defines a maximum
extent to the fort in this direction (see Illus 2). It
would therefore appear that the maximum area of
the fort was 2.8–3ha, but it is possible that the area
contained within the ramparts was much smaller,
perhaps only 1.4ha.
8.1 Defences
The fort was provided with three ditches, identified
in excavation on the north-west and north-east
sides, and visible as cropmarks on air photographs at
the north-east corner and main entrance. Although
it may seem reasonable to suggest that the fort
had three ditches around its entirety, not all forts
demonstrate this level of consistency. For example,
the Flavian fort at Cargill (Perthshire) appears to
have three ditches around part of its perimeter and
two ditches elsewhere (RCAHMS 1994: 84–5).
In Britain in the 1st century it was common for
forts to be protected with double ditches (Jones
1975: 112), although triple ditch systems on at
least one side are known from a number of forts
in Scotland, including Stracathro (St Joseph 1961:
123). Furthermore, both Cardean (Robertson 1977:
67; Woolliscroft & Hoffmann 2006: 160) and
Elginhaugh (Hanson 2007: 124–33) had at least
four ditches on one side.
The ditch system appears to have extended less
than 18m beyond the rampart, thereby placing it
well within the norm for 1st-century forts (Jones
1975: 112). The width of the ditches, between 3m
and 4m, is also within the norm, and the depth
between 0.8m and 1.6m. The north-west ditches
are shallow on the north-west and, unlike the
ditches on the north-east, had no evidence for ankle
breakers. Although the natural defence offered by
the river valley on this side may have reduced the
need for deep ditches on this side, and ditches in
multiple systems tend to be slightly smaller (Jones
1975: 112; Johnson 1983: 48), it is equally likely
that their depth is due to a considerable degree of
later truncation, or that they were not completely
excavated in 1999.
The ditches on the north-east side appear to have
been deliberately backfilled at a point when the sides
had weathered to some extent, but when they would
still have posed a serious obstacle. The backfilling
would have reduced the depth of the ditches to
around 0.8m and was perhaps intended to make
them ineffective as defensive works. The presence of
turf fragments in the inner ditch at the same level
suggests that the rampart was also slighted at this
time. From this point onwards the ditches seem to
have been left to silt up naturally over an extended
period.
There was evidence of a rampart on both the
north-west and north-east sides. The only direct
evidence indicative of the rampart on the north-west
was a spread of soil which sealed the ditches,
thought to be the levelled remains of the bank,
which may indicate that it had been slighted during
the evacuation of the fort. The position of the five
ovens here could indicate the location of the inner
face of the rampart, as these were frequently built
into the lee of the rampart. If it is assumed that the
berm between rampart and ditch measured 1.5–2m
in width, this could give a rampart width of about
7m. On the north-east side the rampart survived as a
6m-wide upstanding deposit, likely the remnants of
a dumped turf-and-earth structure on a surface that
had been previously de-turfed. The material forming
the rampart base was remarkably stone-free, so if
any foundation was provided it must have been of
organic material (eg brushwood). Evidence of facing
or revetting, either of turf or clay, was seen on both
edges of the rampart.
No evidence for gates was recovered during the
excavations, but the cropmark evidence showing the
uniting of the ditches on the south-east side into a
‘parrot’s beak’ shows the location of the north side
of this entrance.
SAIR 92 | 36
Scottish Archaeological Internet Reports 92 2020
8.2 Intervallum structures
Evidence was identified for the via sagularis on both
the north-west and north-east. No other internal
roads were seen. The curve recorded at the southern
extent of the via sagularis in the north-west may
respect a corresponding curve of the defences and
therefore indicate the position of the corner of
the fort. The position of the ovens and associated
working area here appears to have encroached on the
width of the via sagularis in the area of excavation
and seems to have truncated its original width. On
the north-east the road is a heavily truncated spread
of cobbled and gravel surfaces. A cropmark visible
on the south-east side of the fort probably indicates
the position of the via sagularis on this side.
A row of ovens built into the inner face of
the rampart adjacent to the via sagularis on the
north-west side appears to have been intensively
used primarily for bread production, and the
identification of large quantities of amphorae
together with mortaria here would suggest that this
area of the fort was related to food production. The
ovens were heated by burning wood, turf or peat
inside. When the required temperature was reached
the fuel would have been raked out and the dough
placed inside. The door would then be sealed until
the bread was baked (Johnson 1983: 200).
Similar examples of ovens between the rampart
base and via sagularis are known from a number of
forts, and their location within the intervallum area
is common (Jones 2011: 81). At Elginhaugh fort
in Midlothian the excavator suggested that up to
two ovens may have served each barracks (Hanson
2007: 191–3), a more realistic ratio than the single
oven per century as suggested for Fendoch Fort
in Perthshire (Richmond & McIntyre 1939: 138)
and the fortress at Inchtuthil, Perthshire (Pitts & St
Joseph 1985: 200). Ovens of similar construction
have been found around the perimeter of forts at
Fendoch Fort in Perthshire (Richmond & McIntyre
1939: 138) and in pairs at the rear of the rampart at
Inveresk, Midlothian (Leslie 2002: 24). These were
associated with spreads of burnt material and could
be interpreted as the rake-out of the ovens (Leslie &
Will 1999). At Strageath, Perthshire, the rake-out
appears to have been piled against the back of the
rampart (Frere & Wilkes 1989: 62–3). At Birrens,
Dumfriesshire, a well-preserved oven of Antonine
date was excavated which was similar to the Doune
examples. The stone wall for the structure survived
over half a metre high and, as at Doune, part of the
clay dome had collapsed onto the floor (Robertson
1975: 19–20).
On the north-east side pits underlying the road
suggest that the area was heavily used prior to the
laying of the via sagularis. Between the road and the
turf rampart here, there was evidence that ferrous
metalworking was taking place, in an area which
must have afforded some shelter from the wind.
This may have taken place in a small building or
shelter represented by a foundation trench running
alongside the road. This use of the back rampart
area can also be seen in the north-west where a
metalworking furnace, probably an ironworking
shaft furnace, was built into the back edge of the
turf rampart and suggests that this area was used
for industry. Indeed, the presence of a furnace is
noteworthy, and indicates the level of industrial
activity that may have taken place within the fort.
A possible bowl furnace was proposed at Rough
Castle on the Antonine Wall (MacIvor et al 1980),
and a putative furnace indicated at nearby Inveravon
(Dunwell & Ralston 1995). A bowl furnace was also
located at Inchtuthil, and the large fabrica contained
a smithing hearth (Pitts & St Joseph 1985: 108,
199). If the remains at Doune are those of a shaft
furnace, this is the first example from the Roman
period in Scotland.
8.3 Buildings
All the identified structures were extremely regular
and appear to have been of post-trench construction
with upright timbers placed at intervals and wattle
and daub forming the wall in between. Hanson
(2007: 40) has suggested that the normal spacing
between posts was 0.6–0.9m, and recent excavations
at Carlisle recorded similar dimensions (Zant 2009).
The squared post pipes identified in Buildings 1 and
2 both measured 0.13m square, well within the
average range for such posts and close to five Roman
inches (Hanson 1978: 303). Although no trace of
posts could be located in Building 5, the dimensions
of the foundation trenches are consistent with posttrench construction found elsewhere. No evidence
was recovered for the roofing material, although
evidence for turf/peat was recovered from the pits
SAIR 92 | 37
Scottish Archaeological Internet Reports 92 2020
in Building 5 which may have come from the roof.
With regard to the function of the corridored
Building 1, similar structures have been interpreted
as hospitals in the auxiliary forts of Fendoch
(Richmond & McIntyre 1939: 132–4), Corbridge
(Richmond & Gillam 1952: 241–3), and at
Oberstimm and Künzing 1 in Germany (Johnson
1983: 163), and this was the initial interpretation of
the building at Doune (Moloney 1999b). However,
none of the artefacts recovered during the excavation
can support this interpretation. Furthermore, a
similar building at Red House, Corbridge, was
interpreted as a workshop or fabrica, due to its
association with industrial activity (Hanson et al
1979: 80–1). Building 1 at Doune is a little larger
than the fabrica at Red House, but the presence
of nails and charcoal fragments in the fill of the
trenches, combined with the evidence for a possible
furnace, row of ovens and layers of industrial waste
in the immediate vicinity, suggests that this building
was located within an industrial quarter of the fort
and its interpretation as a workshop is more likely.
Several of the other buildings may represent the
remains of stores and barracks, aligned north-west
to south-east. Building 5 appears to be part of a
barracks block fronted by a veranda. Each group
of eight soldiers (contubernium) would have been
housed in a pair of rooms, with parts of three sets
of rooms excavated in Building 5. It is interesting
to note that the size of the rooms of Building 5
(3.7m length and 2.4m in width) appears to be
much smaller than the average barracks block
rooms. At Elginhaugh the rooms were on average
4.1 × 3.3m, with similar sizes seen in Building 1 at
Doune (4.6 × 3.6m) and at other Flavian-period
barracks (Davison 1989: 89, 97; Hanson 2007:
fig 5.2). For Building 5, however, the front rooms
(arma) would have opened onto the veranda and
housed the possessions and equipment of the men
who slept in the room to the rear (papilio). With this
arrangement the smaller than average size for these
rooms is clearly due to separate sleeping and kit
rooms. Two large pits in two of the front rooms of
Building 5 could represent internal latrines, possibly
lined and covered (Johnson 1983: 171–2).
Building 123 shows the typical layout of a
barracks block, comprising a long rectangular
building divided into paired rooms (Johnson 1983:
166–76). It appears to be adjoined to another
building to the south-east which may have formed
part of the same barracks block. A barracks building
which consisted of only five paired rooms (as seen in
Building 123) would be unusually small. In theory,
an infantry century was typically divided into ten
contuberniae, each occupying a pair of rooms, while
a cavalry barracks housed two turmae in at least eight
pairs of rooms, although Johnson (1983: 172) notes
that the number of barracks rooms in known cavalry
barracks blocks in fact varies from six to ten. Taking
into account the cropmark evidence (see Illus 1), it
is clear that if the adjoining building was the same
length as Building 123 it would have reached almost
as far as the rampart on the south-east side, with
little room for the intervallum road. Alternatively,
this unseen building may have formed the officers’
quarters, which typically occupied a much shorter
block, either adjoining the main barracks, or slightly
detached from it.
Building 5 was a different form of barracks block.
It was fronted by an open veranda and, although the
rooms were identical in width to those of Building
123, the front rooms were only 2.4m deep. The
orientation of the two buildings is the same, but
they clearly form part of two separate ranges of
buildings. The presence of two different types of
barracks blocks suggests two different troop types
– perhaps Doune accommodated a part-mounted
cohort, with infantry occupying buildings such as
Building 5, and cavalry the alternative type such as
Building 123. Buildings in some forts have been
interpreted as cavalry barracks which would have
accommodated the soldiers in the room to the
rear of the building, and their mounts in the room
to the front (Johnson 1983: 176–82). While the
front rooms of Building 5 are surely too small to be
anything other than store rooms (arma), the rooms
along the south-west side of Building 123 might
have been large enough to function as stables, with
the horses facing along the long-axis of the building.
The size of the rooms in Building 123, at 3.8
× 3.2m, is similar to those elsewhere interpreted
as stable-barracks (eg Elginhaugh: Hanson 2007;
Wallsend: Hodgson 2003). It is assumed that in
order to accommodate a cavalry squadron (turma)
in a stable-barracks, three horses would be stabled
together in the front room. Hodgson (2003: 83)
has argued for a minimum of 1.2m to be allocated
to each horse. The gullies excavated in the floor of
SAIR 92 | 38
Scottish Archaeological Internet Reports 92 2020
the south-west row of rooms in Building 123, if
correctly interpreted as drains, would support this
interpretation, as drains, pits or soakaways were
commonly provided to collect the horses’ urine and
keep the floor dry. Pits similar to those seen within
Building 123 have been seen on Hadrian’s Wall,
and elsewhere have been interpreted as urine pits
connected with the stabling of horses (eg Hodgson
2003: 71–84). The finding of an ornamented horse
harness strap junction (SF007) in the building is
convincing evidence for a cavalry unit in the fort,
whether or not the horses were kept within the
building itself.
The foundation trenches which make up Building
6 are slighter than those recorded elsewhere in the
fort, and it was difficult to distinguish the features
from the surrounding subsoil. This pale sterile
backfill perhaps suggests that the building was only
used for a short period, with no cultural material
getting into the backfill. Sited next to Building 5, it
was located on a slightly different alignment. If both
buildings were upstanding at the same time, Building
6 would have blocked access to part of Building 5.
Building 6 may therefore represent an earlier phase
of use of the fort or a temporary structure possibly
erected during the initial construction of the fort.
8.4 Material culture
The Flavian date of the fort, previously attributed
by Maxwell, was confirmed through analysis of
the pottery. However, as none of this material is
very closely datable, the best date range gained
from the pottery for the fort is between ad 65
and 90. While the foundation date of the fort is
unknown, conventional analogies would suggest
that it was founded in the early 80s ad, although
debate currently rages on the dating of the first
Roman conquest of this area (Breeze et al 2009). It
is particularly frustrating that the as found by the
school janitor cannot be dated with any certainty
to either ad 86 or 87 – if it dates to ad 86 it fits
the general pattern of finds from Flavian forts in
Scotland (see 6.3.1.1 ‘The 2008 coin’ above).
However, if it dates to ad 87, it is the most northerly
find of this date in Scotland and would be a very
significant find indeed.
8.5 Conclusion
The excavations in advance of the development of the
primary school at Doune have provided supporting
evidence for the Flavian date initially given to the
fort on its discovery through aerial survey. It has
also furthered our understanding of the internal
organisation of the fort, including different building
forms, the location of an industrial quarter and the
main road. The finds recovered through this work
have added detail to our understanding of life in
this frontier region. The excavations have also shown
that within the school grounds, and potentially even
below the school buildings, the preservation of the
Roman fort at Doune is good.
SAIR 92 | 39
Scottish Archaeological Internet Reports 92 2020
APPENDIX 1: CATALOGUE OF SAMIAN WARE
Steven H Willis (1999) & Gwladys Monteil (2008)
Each sherd was examined, after breaking, under a ×20
binocular microscope in order to identify the fabric.
Each archive catalogue entry consists of a context
number alongside fabric, form and decoration
identification, sherd count, rim or base EVE when
appropriate, and weight. Rubbings of the interesting
decorated fragments were undertaken during analysis.
The Inventory Numbers (Inv No.) quoted for
the decorated South Gaulish vessels are taken from
European intake of Roman samian ceramics. https://
www1.rgzm.de/samian/home/frames.htm
A.1.1 Dr 18
▶ 1. DPSE07 Context 010, upper fill of foundation
trench 006
South Gaulish, Dr 18, base, one sherd, ad 50–100.
A.1.2 Dr 27
▶ 2. DPS99 Context 004, fill of Pit 005
Two joining sherds, Dr 27, South Gaulish, La
Graufesenque, ad 65–90.
▶ 3. DPS99 Context 004, fill of Pit 005
A cup bodysherd, South Gaulish, La Graufesenque,
either Ritt 9 or Dr 27, possibly the same vessel as
above, ad 65–90.
A.1.3 Dr 29
▶ 4. DPSE07 Context 018, lower fill of Pit 014
South Gaulish, Dr 29, two sherds.
T-1 style, ad 55–75
The sherds do not join but both have the same
decoration of a 13-petal rosette in a festoon. The
13-petal rosette belongs to a style found with stamps
of Bassus ii-Coelus (cf K52, Taf 58, T, V and Z). The
rosette is on a Dr 29 stamped by Bassus ii-Coelus from
la Graufesenque (serial no. 0000168) and on a Dr 29
from Colchester (Monteil 2008: vessel E6-E2048,
EL222). The rosette with the same festoon is on a
Dr 29 from Colchester whose style is attributed to
T-1 (Dannell 1999: no. 549). The festoon with the
poppyhead is found on stamped Dr 29s by Bassus
ii-Coelus (K19, Taf 13 and Serial number 0000163
from Vienne). The wreath of trifid motifs is on a
stamped Dr 29 from Rottweil (K52 Taf 58, Y) and
on a bowl from Colchester attributed to the style of
T-1 as stamped by Bassus ii-Coelus (Dannell 1999:
no. 553). There is a possibility that this vessel is later
and the work of Coelus ii when he worked alone.
The rosette is on a Dr 29 stamped by Coelus ii from
York? (serial no. 0000328) and the festoon with the
poppyhead on a Dr 29 stamped by Coelus ii (serial
no. 0000342). Coelus ii might have taken over
some of the moulds used by Bassus i, or obtained his
moulds from the same supplier (Polak 2000).
A.1.4 Dr 35
▶ 5. DPS99 Context 004, fill of Pit 005
A Dr 35 rim fragment, South Gaulish, La
Graufesenque, ad 65–90.
▶ 6. DPS99 Context evaluation 014, pit fill
A Dr 35 cup rim, South Gaulish, La Graufesenque,
c ad 65–90.
▶ 7. DPS99 Context 060, upper fill of Pit 061
Three South Gaulish sherds: one rim, one foot-ring
fragment and one bodysherd. The rim is a Dr 35
cup and the other two sherds are also from cup(s)
and probably the same vessel. c ad 65–90.
A.1.5 Dr 37
▶ 8. DPS99 Unstrat
A Dr 37 bodysherd, South Gaulish, La Graufesenque,
decoration unattributable, c ad 70–85/90.
▶ 9. DPSE07 Context 018, lower fill of Pit 014
South Gaulish, Dr 37, one sherd.
The sherd is small and only part of the ovolo
survived. The tongue is missing, which renders close
identification impossible. ad 70–100.
▶ 10. DPSE07 Context 018, lower fill of Pit 014
South Gaulish, dish, one sherd.
South Gaulish, bowl, one sherd, possibly from the
base of a Dr 37.
A.1.6 Unidentified forms
▶ 11. DPS99 Context 003, hollow containing
iron-smelting debris
SAIR 92 | 40
Scottish Archaeological Internet Reports 92 2020
A South Gaulish samian dish bodysherd, La
Graufesenque, ad 65–90.
▶ 12. DPS99 Context 008, rake-out from ovens
An eroded samian sherd, unidentifiable form,
South Gaulish, La Graufesenque, ad 40–100.
▶ 13. DPSE07 Context 035, upper fill of
foundation trench 034
South Gaulish, one very small chip.
SAIR 92 | 41
Scottish Archaeological Internet Reports 92 2020
APPENDIX 2: CATALOGUE OF ILLUSTRATED
COARSE POTTERY VESSELS
A.2.1 Amphorae
▶ 1. DPS99, Unstrat, ‘Service trench’ (Illus 13:1)
Fabric A01. A single sherd representing
approximately 20% of the rim. A similar form
at Augst was found in contexts dated ad 70–110
(Martin-Kilcher 1987: no. 534).
▶ 2. DPS99, Context 010, upper fill of Pit 011
(Illus 13:2)
Fabric A01. Two joining sherds representing
approximately 20% of the rim. Close parallels to
this form in Martin-Kilcher’s scheme occur in the
period ad 70–110 (Beilage D, nos 64–71). A similar
form of Dressel 20 rim has been found at the nearby
legionary fortress of Inchtuthil, which was probably
occupied from ad 83–86 (Darling 1985, no.
72).
A.2.2 Flagon
▶ 3. DPSE07, Context 004, topsoil (Illus 13:3)
Fabric O06. A rimsherd and three chips from a ringnecked flagon with a fairly straight, vertical neck.
Lid?
▶ 4. DPS99, Context 007, base of oven C002
(Illus 13:4)
Fabric F01. A brown colour-coated bead-rimmed
lid(?). M Darling and B Precious (pers comm)
suggest that this is a Pompeian Redware 3 lid from
Central Gaul. There are sherds from 10 vessels,
including a lid, from Inchtuthil (Darling 1985: nos
67–71).
fabric and context, fits well with Neronian-Flavian
production in a military workshop not far from
Doune. Variety of rim-profile, often unusual, and
unusual fabric is, in fact, characteristic of mortaria
made by the army in Britain, in the pre-Flavian and
early Flavian periods. The workshops involved were
small and served very local markets (eg Longthorpe,
Metchley, Trent Vale, Exeter, Wroxeter, Usk,
Inchtuthil and Elginhaugh).
▶ 6. DPS99, Context 008, rake-out from ovens
(Illus 13:6)
Fabric M02. A Claudio-Neronian mortarium rim,
ad 40–60, cf Strageath (Frere & Hartley 1989:
no. 1). Other examples of these vessels occurring
as heirlooms are known from Strageath, Camelon
(Frere & Hartley 1989: no. 1), Binchester (Evans &
Rátkai 2010: type M091.1), and York (Monaghan
1997: no. 3406).
▶ 7. DPS99, Context 008, rake-out from ovens
(Illus 13:7)
Fabric M02. A Gillam (1970) type 238 mortarium
rim, from Noyon, Oise, ad 70–100.
▶ 8. DPS99, Context 060, upper fill of Pit 061
(Illus 13:8)
Fabric M03. A Verulamium region ware mortarium
with evenly curving, down-pointing flange
(incomplete). ad 60–90.
▶ 9. DPS99, Context 001, Unstrat (Illus 13:9)
Fabric M03. A burnt Verulamium region ware
mortarium with evenly curving down-pointing
flange. ad 60–90.
A.2.4 Bowls
A.2.3 Mortaria
▶ 5. DPS99, Context 007, base of oven C002
(Illus 13:5)
Fabric M01. An oxidised mortarium with bead set
well below flange. The fabric, together with the
provenance, date and form, indicate manufacture
in Scotland. We know little about the production of
mortaria in 1st century Scotland except at Inchtuthil
and Elginhaugh, but there is no doubt that it
occurred. There is no parallel for this form in this
period, but fortunately the context leaves no doubt
of its Flavian date. Nevertheless, the form, given the
▶ 10. DPS99, Context 060, upper fill of Pit 061
(Illus 13:10)
Fabric M04. A segmental bowl flange/or a
mortarium flange. The form of the flange fragment
and the fact that it becomes thinner at the point
where it meets the normal bead would better fit
a segmental bowl than a mortarium. There are,
however, a very few mortaria which could be cited
as parallels, notably an unpublished mortarium,
stamped by one of the Sexti Valerii who worked
at Colchester in the 1st century. The slight distal
bead and the fabric would certainly fit with its
SAIR 92 | 42
Scottish Archaeological Internet Reports 92 2020
4
1
5
6
2
3
7
12
8
13
9
14
10
15
11
0
10cm
Illus 13 Pottery from 1999 and 2008 excavations. © Headland Archaeology (UK) Ltd
SAIR 92 | 43
Scottish Archaeological Internet Reports 92 2020
being a mortarium, but the form is exceptional for a
mortarium, wherever it was made. The fabric would
fit best with production at some such centre as the
Aoste (Isère), but there is, as yet, no parallel there
for this form. A date of ad 55–85 is indicated by
the form and the similarity to mortaria made by
one of the Sexti Valerii (Hartley 1999: 203, die as
S110–111).
▶ 11. DPS99, Context 060, upper fill of Pit 061
(Illus 13:11)
Fabric O01. A reeded-rimmed bowl rim fragment, cf
Inchtuthil (Darling 1985: no. 50), Flavian-Trajanic.
A.2.5 Jars
▶ 12. DPS99, Context evaluation 014, pit fill
(Illus 13:12)
Fabric O01. A constricted-necked jar with a beaded,
undercut rim, perhaps cf Strageath (Anderson 1989:
no. 53).
▶ 13. DPS99, Context 004, fill of Pit 005 (Illus
13:13)
Fabric R01. A small jar with everted, rising rim.
▶ 14. DPS99, Context 010, upper fill of Pit 011
(Illus 13:14)
Fabric R01. A small globular jar, with short, everted
rim, cf Inchtuthil (Darling 1985: nos 7–11), and
Strageath (Anderson 1989: nos 39–45), Flavian.
▶ 15. DPS99, Context evaluation 005, intervallum
road make-up (Illus 13:15)
Fabric R05. A jar with a short, fairly vertical, grooved
rim and globular form, cf Strageath (Anderson
1989: no. 50) and Inchtuthil (Darling 1985: nos
7–9), Flavian.
SAIR 92 | 44
Scottish Archaeological Internet Reports 92 2020
APPENDIX 3: FABRIC DESCRIPTIONS
Table 5 Fabric descriptions
Fabric Description
F01
A colour-coated fabric with buff-orange core and margins, with a thin reddish-brown slip.
The fabric has abundant fine gold and silver mica c 0.1–0.2mm. Possibly Pompeian Redware
3, Central Gaul.
F02
A colour-coated roughcast fabric, probably Cologne. A hard, white, ‘clean’ fabric with white
clay pellet roughcasting, with a brown, rather metallic, slip. Anderson (1981: 327–30) would
suggest a date range of ad 80–180/190 for this.
M01 An oxidised mortarium fabric with blue-grey core and orange margins and surfaces, with
some/common sand c 0.2–0.4mm and very occasional sub-rounded white quartz c 1–3mm
and very occasional sub-rounded igneous rock inclusions. Geological note by Dr G Gaunt
(pers comm): The best indicator of likely provenance on geological grounds is the coarsegrained micro-conglomeratic sandstone, supported by the fine to medium grained sandstone.
These occur in the following Devonian outcrop regions:
1) South-east Wales and Welsh borders, the region within the Brecon – Caerleon – Hereford
– Bridgnorth areas.
2) Northern and southern bordering areas of the Scottish ‘Midland Valley’, ie
a) north side – from just north of Cardross on the Clyde estuary, north-eastwards past
Callander, Doune, Crieff and Perth to the North Sea between Carnoustie and Stonehaven
b) south side – the Maybole area in Ayrshire, a narrow belt from just north-east of New
Cumnock north-eastwards via Crawfordjohn and Roberton to Wiston, the Dolphinton–
Carlops area, and the small area around Coldingham south of St Abbs Head.
M02 A buff mortarium fabric with some fine sand temper c 0.2mm and occasional coarse sand
c 0.3–0.5mm. Source: NE France, Noyon, Oise.
M03 Verulamium region mortaria; white core, margins and surfaces, with abundant moderatecoarse sand temper c 0.3–0.5mm. Trituration grits; angular grey and white flint.
M04 A whiteware with white core, margins and surfaces, with occasional sand c 0.2–0.5mm,
occasional red ironstone c 0.2–1mm, and occasional rounded white clay pellets c 4mm.
Either a whiteware or a mortarium, possibly Central Gaul.
O01
An oxidised fabric, soft, ‘soapy’ with orange core, margins and orange to orange-yellow
surfaces, ‘clean’, with some voids c 0.5–3mm, possibly for vegetable temper.
O02
An oxidised fabric, fairly hard, with orange core, margins and surfaces, ‘soapy’ and ‘clean’
with no visible temper.
O03
An oxidised fabric with orange core, margins and surfaces, with common moderate sand
temper c 0.3mm.
O04
An oxidised fabric with orange-brown core, margins and surfaces, with common fine sand
c 0.05–0.1mm.
O05
An oxidised fabric with orange-brown core, margins and surfaces with common very fine
silver mica, and occasional rounded brown ironstone c 0.5–2mm, and very occasional sand
c 0.2mm.
O06
An oxidised fabric with orange-brown core, margins and surfaces with some sand temper
c 0.3–0.5mm.
SAIR 92 | 45
Scottish Archaeological Internet Reports 92 2020
Table 5 cont
Fabric Description
Q01
An oxidised flagon fabric with a thick white slip, fabric has an orange core and margins with
some sand c 0.05–0.1mm.
R01
A reduced fabric with black core, grey or black margins, and brown or black surfaces, with
occasional sand c 0.1–0.2mm and very common very fine silver mica.
R02
A greyware with a grey core, sometimes orange margins, and grey surfaces, with common
coarse angular translucent sand c 0.3–0.5mm.
R03
A reduced, handmade fabric with grey-brown core and margins and black-grey surfaces,
‘soapy’, ‘clean’ with some very fine silver mica, some brown ironstone c 0.5–2mm and
occasional large carbonised inclusion c 5mm.
R04
A reduced fabric with dark, blue-grey core, sometimes orange margins and dark grey surfaces,
with occasional rounded translucent sand c 0.3–0.5mm and occasional organic voids and
carbonised inclusions c 0.5–2mm.
R05
A reduced fabric with blue-grey core and mid-grey margins and surfaces, fairly ‘clean’ with
occasional-some moderate sand temper c 0.3mm.
S10
South Gaulish, La Graufesenque, samian ware.
W01 A whiteware with a buff-white core, margins and surfaces with common fine sand
c 0.1–0.2mm.
W02 A buff whiteware with buff-white core, margins and surfaces with some fine sand
c 0.1–0.2mm.
W03 A buff whiteware with white core and pinkish-white margins and surfaces, ‘soapy’ with
occasional sand c 0.2mm and occasional rounded clay pellet(?) c 2mm.
SAIR 92 | 46
Scottish Archaeological Internet Reports 92 2020
APPENDIX 4: CATALOGUE OF SLING SHOTS
A.4.1 1999 excavation
▶ A complete reduced fired clay sling shot. L:
34mm; Diam: 31mm; W: 25g. Unstratified (Illus
9:1).
▶ A slightly damaged fired clay sling shot, core
reduced, surfaces oxidised. Diam: 30mm; W: 25g.
Unstratified.
▶ A slightly damaged fired clay sling shot, core
reduced, surfaces oxidised. L: 34mm; Diam: 27mm;
W: 15g. Unstratified (Illus 9:3).
▶ A slightly damaged fired clay sling shot, core
reduced, surfaces oxidised. L: 41mm; Diam: 33mm;
W: 36g. Unstratified (Illus 9:4).
▶ A slightly damaged fired clay sling shot, core
reduced, surfaces oxidised. L: 39mm; Diam: 29mm;
W: 23g. Unstratified.
▶ About two-thirds of a fired clay sling shot, core
reduced, surfaces oxidised. L: 32mm; Diam: 28mm.
Unstratified.
▶ About half of a fired clay sling shot, core reduced,
surfaces oxidised. L: 33mm; Diam: 30mm.
Unstratified.
▶ About half of a fired clay sling shot, core reduced,
surfaces oxidised. L: 36mm; Diam: 29mm.
Unstratified.
A.4.2 2008 excavation
▶ A complete biconical sling shot, surfaces oxidised.
L: 41mm; Diam: 26mm; W: 20g. C010, upper fill
of foundation trench 006, Building 5.
▶ A complete biconical sling shot of much more
squat circular form than the others, surfaces
oxidised. L: 29mm; Diam: 34mm; W: 39g. C005,
layer overlying Building 5.
SAIR 92 | 47
Context no.
Sample no.
Orig vol
(litres)
Latin name
Building 2
Building 3
Building 5
Building 6
004
005
010
053
056
042
094
005
009
010
012
017
035
021
–
–
–
–
–
–
–
1
13
12
8
14
6
24
20
10
10
10
10
10
10
30
10
20
10
10
10
10
Common
name
Corylus avellana L
nutshell
hazel
Ranunculus sp cf R
sardous
achene
hairy
buttercup
Ranunculus acris/
repens
achene
meadow/
creeping
buttercup
Chenopodiaceae
indet
seed
fat hen family
Chenopodium sp cf
C album
seed
fat hen
Spergula arvensis
L
seed
corn spurrey
Persicaria
lapithifolia
achene
pale persicaria
12
Rumex spp
achene
dock
19
Rumex acetosella
agg
achene
sheep’s sorrel
Sinapis sp
fruit
mustards
Raphanus
raphinistrum L
siliqua
wild radish
2
Potentilla sp cf P
erecta
achene
cinquefoil
1
Wild taxa
2
1
1
39
2
1
1
4
1
1
1
1
SAIR 92 | 48
Scottish Archaeological Internet Reports 92 2020
Plant part
APPENDIX 5: COMPOSITION OF PLANT REMAINS FROM DOUNE PRIMARY SCHOOL
Building 1
Mhairi Hastie & Scott Timpany
Table 6 Composition of plant remains from Doune Primary School: Buildings 1–6
Table 6 cont
Building 1
Context no.
Sample no.
Orig vol
(litres)
Latin name
Plant part
Common
name
Vicia hirsuta
fruit
hairy tare
Trifolium repens L
fruit
white clover
Plantago lanceolata
L
seed
ribwort
Carex sp cf C
lasiocarpa
nutlet
slender sedge
Carex sp cf C
acutiformis
nutlet
lesser
pond-sedge
Juncus/Poaceae sp
culm
rush/grass
Building 2
Building 3
005
010
053
056
042
094
005
009
010
012
017
035
021
–
–
–
–
–
–
–
1
13
12
8
14
6
24
20
10
10
10
10
10
10
30
10
20
10
10
10
10
1
2
71
2
10
Poaceae indet
(medium)
caryopsis
mediumgrained grass
9
Poaceae indet
(small)
caryopsis
small-grained
grass
Bromus/
Brachypodium sp
caryopsis
brome/
false-brome
Poa spp
caryopsis
meadow grass
SAIR 92 | 49
indeterminate
seed
1
1
180
1
5
3
1
2
2
2
Scottish Archaeological Internet Reports 92 2020
6
Indeterminate
1
1
large-grained
grass
heath grass
Building 6
004
Poaceae indet (large) caryopsis
Sieglingia decumbens caryopsis
(L) Bernh
Building 5
Table 6 cont
Building 1
Context no.
Sample no.
Orig vol
(litres)
Latin name
Building 2
Building 3
Building 5
Building 6
004
005
010
053
056
042
094
005
009
010
012
017
035
021
–
–
–
–
–
–
–
1
13
12
8
14
6
24
20
10
10
10
10
10
10
30
10
20
10
10
10
10
2
1
34
2
1
4
1
1
1
Plant part
Common
name
Avena sp
caryopsis
oat
cf Avena sp
caryopsis
oat
Hordeum vulgare
indet
caryopsis
barley indet
42
11
cf Hordeum vulgare
indet
caryopsis
barley indet
2
5
Cereals
2
7
naked barley
Hordeum vulgare
(hulled)
caryopsis
hulled barley
Hordeum vulgare L
(hulled – straight)
caryopsis
hulled barley
Hordeum vulgare L
(hulled – twisted)
caryopsis
hulled barley
Hordeum vulgare
indet
rachis
internodes
barley indet
Triticum sp
caryopsis
wheat
6
cf Triticum sp
caryopsis
wheat
2
Triticum
aestivo-compactum
caryopsis
bread/club
wheat
Triticum dicoccum
caryopsis
emmer wheat
71
24
2
559
2
22
21
6
1
32
1250
25
1
6
19
1
19
3
17
19
5
6
14
9
10
4
9
1
3
48
20
1
1
1
Scottish Archaeological Internet Reports 92 2020
SAIR 92 | 50
Hordeum vulgare var caryopsis
nudum
5
Table 6 cont
Building 1
Context no.
Building 3
Building 5
Building 6
004
005
010
053
056
042
094
005
009
010
012
017
035
021
–
–
–
–
–
–
–
1
13
12
8
14
6
24
20
10
10
10
10
10
10
30
10
20
10
10
10
10
6
4
1
27
85
33
33
66
27
28
Sample no.
Orig vol
(litres)
Building 2
Latin name
Plant part
Common
name
Triticum spelta L
caryopsis
spelt wheat
Triticum cf spelta
caryopsis
spelt wheat
Triticum cf spelta
chaff
fragments
spelt wheat
1
Triticum sp
chaff
fragments
wheat
2
Cereal indet
caryopsis
cereal indet
Straw
culm nodes
1
3
11
12
1
3
1
1
Percentages
Cereals
100
100
88
87
100
100
84
98
100
99
57
100
94
50
Wild taxa
0
0
12
13
0
0
16
2
0
1
43
0
6
50
Chaff fragments
0
0
0
0
0
0
0.2
0
0
0
0
0
0
0
Wheat
10
20
0
4
0
25
5
49
0
48
0
50
43
0
Barley
81
80
91
80
100
75
93
18
50
23
75
50
30
100
Oat
1
0
6
4
0
0
2
1
50
4
25
0
0
0
Indet
8
0
3
12
0
0
0
32
0
25
0
0
27
0
7.1
5.0
3.1
2.6
0.8
0.4
201.5
7.0
0.2
5.3
0.4
0.2
10.2
0.1
No. of cereals per litre
Scottish Archaeological Internet Reports 92 2020
SAIR 92 | 51
Type
Table 7 Composition of plant remains from Doune Primary School: other contexts
Context no.
Sample no.
Orig vol (litres)
Latin name
Wild taxa
Corylus avellana L
Ranunculus sp
Pit 14 fill
015 018
4
5
10
10
Pit 47 Pit 90
Pit 29 fill
fill
fill
028 044 044 045
046
091
21 19 22 20
–
–
10 10 10 10
10
10
Oven Furnace
Oven fills rake-out
base
012 023
008
003
–
–
–
–
10
10
80
40
Fill of
defensive
ditch
122
–
30
Plant part Common name
nutshell
achene
achene
achene
hazel
buttercup/
crowfoot
hairy buttercup
3
2
8
achene
fruit
seed
meadow/ creeping
buttercup
fumitories
common nettle
fat hen family
1
seed
fat hen
1
seed
chickweed
achene
pale persicaria
achene
achene
achene
achene
knotgrass
knotgrass
dock
sheep’s sorrel
fruit
fruit
violets
mustards
2
15
2
1
1
1
3
1
1
1
1
1
1
1
1
1
1
2
5
2
1
1
1
Scottish Archaeological Internet Reports 92 2020
SAIR 92 | 52
Ranunuculus sp cf
R sardous
Ranunculus acris/
repens
Fumaria sp
Urtica dioica
Chenopodiaceae
indet
Chenopodium sp cf
C album
Stellaria sp cf S
media
Persicaria
lapithifolia
Polygonum aviculare
Polygonum spp
Rumex spp
Rumex acetosella
agg
Viola sp
Sinapis sp
Possible
structure
106
–
10
Table 7 cont
Context no.
Sample no.
Orig vol (litres)
Plant part Common name
fruit
black mustard/
wild radish
siliqua
wild radish
achene
cinquefoil
fruit
fruit
fruit
vetches
hairy tare
grass-poly
fruit
seed
hemp-nettle
ribwort
achene
corn marigold
nutlet
nutlet
wood-rushes
great wood-rush
nutlet
common spikerush
club-rushes
sedge
star sedge
nutlet
nutlet
nutlet
Pit 14 fill
015 018
4
5
10
10
Pit 47 Pit 90
Pit 29 fill
fill
fill
028 044 044 045
046
091
21 19 22 20
–
–
10 10 10 10
10
10
Oven Furnace
Oven fills rake-out
base
012 023
008
003
–
–
–
–
10
10
80
40
Fill of
defensive
ditch
122
–
30
1
2
1
3
1
1
1
1
5
2
1
3
8
1
5
1
1
1
1
2
1
1
1
1
1
40
Scottish Archaeological Internet Reports 92 2020
SAIR 92 | 53
Latin name
Brassica/Raphunus
sp
Raphanus
raphinistrum L
Potentilla sp cf P
erecta
cf Vicia sp
Vicia hirsuta
cf Lythrum
hyssopifolia
Galeopsis sp
Plantago lanceolata
L
Chrysanthemum
segetum
cf Luzula sp
Luzula sp cf L
sylvatica
Eleocharis sp cf E
palustris
Schoenoplectus sp
Carex sp indet
Carex sp cf C
echinata
Possible
structure
106
–
10
Table 7 cont
Context no.
Sample no.
Orig vol (litres)
Pit 14 fill
015 018
4
5
10
10
Pit 47 Pit 90
Pit 29 fill
fill
fill
028 044 044 045
046
091
21 19 22 20
–
–
10 10 10 10
10
10
Plant part Common name
nutlet
elongated sedge
nutlet
nutlet
white sedge
slender sedge
nutlet
cyperus sedge
nutlet
nutlet
large yellow-sedge
downy-fruited
sedge
rush/ grass
medium-grained
grass
small-grained grass
culm
caryopsis
caryopsis
caryopsis
Oven Furnace
Oven fills rake-out
base
012 023
008
003
–
–
–
–
10
10
80
40
Fill of
defensive
ditch
122
–
30
1
1
1
1
1
1
1
3
8
6
9
4
6
2
3
10
2
1
1
1
caryopsis
brome/ falsebrome
heath grass
seed
indeterminate
1
caryopsis
caryopsis
oat
oat
5
2
1
1
1
2
1
297
46
5
4
3
2
3
6
1
Scottish Archaeological Internet Reports 92 2020
SAIR 92 | 54
Latin name
Carex sp cf C
elongata
Carex sp cf C curta
Carex sp cf C
lasiocarpa
Carex sp cf C
pseudocyperus
Carex sp cf C flava
Carex sp cf C
filiformis
Juncus/Poaceae sp
Poaceae indet
(medium)
Poaceae indet
(small)
Bromus/
Brachypodium sp
Sieglingia
decumbens (L)
Bernh
Indeterminate
Cereals
Avena sp
cf Avena sp
Possible
structure
106
–
10
Table 7 cont
Context no.
Sample no.
Orig vol (litres)
Plant part Common name
caryopsis barley indet
Pit 14 fill
015 018
4
5
10
10
Pit 47 Pit 90
Pit 29 fill
fill
fill
028 044 044 045
046
091
21 19 22 20
–
–
10 10 10 10
10
10
Oven Furnace
Oven fills rake-out
base
012 023
008
003
–
–
–
–
10
10
80
40
116
32
32
caryopsis
barley indet
7
2
6
caryopsis
hulled barley
69
97
70
caryopsis
hulled barley
1
1
2
caryopsis
hulled barley
rachis
barley indet
internodes
caryopsis wheat
caryopsis bread/ club wheat
caryopsis
caryopsis
caryopsis
caryopsis
culm nodes
emmer wheat
spelt wheat
spelt wheat
cereal indet
18
4
21
12
40
7
13
1
5
5
8
27
6
4
1
2
1
9
127
1
59
6
1
17
3
2
9
Fill of
defensive
ditch
122
–
30
71
14
8
36
3
1
18
16
118
41
1
16
18
2
12
7
1
3
1
1
16
2
22
1
1
3
39
1
SAIR 92 | 55
Scottish Archaeological Internet Reports 92 2020
Latin name
Hordeum vulgare
indet
cf Hordeum vulgare
indet
Hordeum vulgare
(hulled)
Hordeum vulgare L
(hulled – straight)
Hordeum vulgare L
(hulled – twisted)
Hordeum vulgare
indet
Triticum sp
Triticum aestivocompactum
Triticum dicoccum
Triticum spelta L
Triticum cf spelta
Cereal indet
Straw
Possible
structure
106
–
10
Table 7 cont
Context no.
Sample no.
Orig vol (litres)
Type
Cereals
Wild taxa
Chaff fragments
Wheat
Barley
Oat
Indet
No. of cereals per litre
Possible
structure
106
–
10
Pit 14 fill
015 018
4
5
10
10
Percentages
97
97
3
3
0
0
Pit 47 Pit 90
Pit 29 fill
fill
fill
028 044 044 045
046
091
21 19
22 20
–
–
10 10
10 10
10
10
Oven Furnace
Oven fills rake-out
base
012 023
008
003
–
–
–
–
10
10
80
40
Fill of
defensive
ditch
122
–
30
100
0
0
64
36
0
69
31
0
57
43
0
41
59
0
97
3
0
100
0
0
77
23
0
100
0
0
87
13
0
98
2
0
0
0
0
0
35
62
3
52
10
1
37
57
4
3
36
20
50
0
30
34
17
4
32
7
71
4
18
18
41
0
41
0
93
4
3
50
0
0
50
22
66
0
13
0
0
0
100
5
81
0
14
0
75
1
24
0
0
0
0
55.3
31.6
1.2
5.2
5.6
6.8
1.7
14.8
0.2
16.8
0.1
0.2
4
0
Scottish Archaeological Internet Reports 92 2020
SAIR 92 | 56
Scottish Archaeological Internet Reports 92 2020
APPENDIX 6: COMPOSITION OF THE CHARCOAL FROM DOUNE PRIMARY SCHOOL
Table 8 Composition of the charcoal from Doune Primary School, 1999 excavation samples (+ = rare;
++ = occasional; +++ = common; ++++ = abundant; ~ = only very small fragments of charcoal, not large
enough for charcoal identifications; cf = tentative identifications)
Feature
Building 1
Building 2
Building 3
Possible
structure
Pit fill
Post hole
Oven fill
Oven rakeout
Furnace
base
Clay fill
Ditch fill
Context
no.
005
010
053
056
042
094
106
004
046
065
089
091
101
031
063
002
006
012
018
023
008
Alnus
(alder)
Betula Corylus
(birch) (hazel)
+
+
++
cf +
++
+
+
+
++
+
+
+
+
+
Quercus
(oak)
+
+
+
+
+
+
+
Rosaceae indet
(plum/rowan)
Salix
Comments
(willow)
+
cf +
cf +
cf +
+
+
+
+
+
+
cf +
+
+
cf +
cf +
+
~
+
~
+
+
++
+
~
+
~
+
+
++
029
003
cf +
++
+
+
++++
+
+
+
058
122
124
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
~
126
~
~
~
~
~
~
~
~
~
~
~
~
+
~
~
~
~
+
Bark +
Cinders +
Bark +
SAIR 92 | 57
Scottish Archaeological Internet Reports 92 2020
9. ACKNOWLEDGEMENTS
The content of this paper has relied heavily on
two previous publication drafts by the directors
at Doune, Paul Masser and Colm Maloney. The
excavations were funded by Stirling Council and
this publication has been funded in part by the
Council and in part by Headland Archaeology
(UK) Ltd. Thanks are due to Joyce Wighton, Lorna
Main, Gordon Maxwell, Jane Beattie and the staff
of Doune Primary School for their assistance
during the 1999 excavation, and additionally
to Ron Spalding and Robert Kinnaird for their
help in 2008. Rectification of aerial photographs
was undertaken by Kevin Macleod of the Royal
Commission on the Ancient and Historical
Monuments of Scotland. The illustrations are the
work of Rafael Maya-Torcelly, Mike Middleton,
Jonathan Millar, Tom Small, Anna Faras-Pagowska
and Mano Kapazoglou.
SAIR 92 | 58
Scottish Archaeological Internet Reports 92 2020
10. REFERENCES
Allason-Jones, L 2007 ‘Small objects’, in Hanson,
W S Elginhaugh: A Flavian Fort and its Annexe,
396–442. Britannia Monograph Series No. 23.
London: Society for the Promotion of Roman
Studies.
Anderson, A C 1981 ‘Some continental beakers of
the first and second centuries AD’, in Anderson,
A & Anderson A S (eds.) Roman Pottery Research
in Britain and North-West Europe: Papers
presented to Graham Webster, 321–43. Oxford:
British Archaeological Reports, International
Series, 123.
Anderson, A C 1989 ‘The other Roman pottery’ in
Frere, S & Wilkes, J J 1989: 243–65.
Bateson, J D 1989 ‘Roman and medieval coins
found in Scotland, to 1987’, Proc Soc Antiq Scot
119: 165–88.
Boardman, S & Dickson, C 1995 ‘Botanical remains
from the inner ditch’, in Keppie, L J F, Bailey,
G B, Dunwell, A J, McBrien, J H & Speller,
K 1995 ‘Some excavations on the line of the
Antonine Wall, 1985–93’, Proc Soc Antiq Scot
125: 601–71 (653–6).
Breeze, D J, Thoms, L M & Hall, D W (eds) 2009
First Contact: Rome and Northern Britain.
Perth: Tayside & Fife Archaeology Committee
Monograph 7.
Cappers, R T J, Bekker, R M & Jans, J E A 2006
Digital Seed Atlas of the Netherlands. Groningen:
Barkhuis Publishing and Groningen University
Library.
Clapham, A J 2007 ‘Plant remains’, in Hanson, W S
Elginhaugh: A Flavian Fort and its Annexe, 571–
613. Britannia Monograph Series No. 23. London:
Society for the Promotion of Roman Studies.
Clapham, A R, Tutin, T G & Warburg, E F 1962
Flora of the British Isles, 2nd edition. Cambridge:
Cambridge University Press.
Cook, M 2014 ‘Stirling, Beechwood Park Roman
Road, Excavation’, Discovery and Excavation in
Scotland 14: 182.
Crawford, O G S 1949 Topography of Roman
Scotland North of the Antonine Wall. Cambridge:
Cambridge University Press.
Curle, J 1911 A Roman Frontier Post and its People:
The Fort of Newstead in the Parish of Melrose.
Glasgow: Maclehose.
Dannell, G B 1999 ‘South Gaulish decorated
samian’, in Symonds, R P & Wade, S Roman
Pottery from Excavations in Colchester, 1971–86,
13–74. Colchester Archaeological Report no.
10. Colchester Archaeological Trust.
Darling, M 1985 ‘The other Roman pottery’, in
Pitts, L F & St Joseph, J K Inchtuthil: The Roman
Legionary Fortress Excavations 1952–65, 323–38.
Britannia Monograph Series No. 6. London:
Society for the Promotion of Roman Studies.
Davies, R W 1971 ‘The Roman military
diet’, Britannia 2: 122–42. https://doi.
org/10.2307/525803
Davison, D P 1989 The Barracks of the Roman Army
from the First to Third Centuries AD. Oxford:
British Archaeological Reports, International
Series, 472.
Dickson, C 1989 ‘The Roman army diet in Britain
and Germany’, Archeobotanik, Dissertationes
Botanicae 13: 135–54.
Dickson, C & Dickson, J 2000 Plants and People in
Ancient Scotland. Stroud: Tempus.
Dunwell, A & Ralston, I 1995 ‘Excavations at
Inveravon on the Antonine Wall 1991’, Proc
Soc Antiq Scot 125: 521–76.
Evans, J, 1993 ‘Function and finewares in the
Roman north’, Journal of Roman Pottery Studies
6: 95–118.
Evans, J 1997 ‘The pottery from Brithdir Roman
fort’, in Hopewell, D ‘Archaeological survey and
excavations at Brithdir’, Journal of Merioneth
Historical and Record Society 12(4): 310–33.
Evans, J 2001 ‘Material approaches to the
identification of different Romano-British site
types’, in James, S & Millett, M (eds) Britons
and Romans: Advancing an Archaeological
Agenda, 26–35. CBA Research Report 125.
York: Council for British Archaeology.
Evans, J & Rátkai, S 2010 ‘Roman coarse pottery’,
in Ferris, I (ed.) The Beautiful Rooms are Empty:
Excavations at Binchester Roman Fort, County
Durham 1976–1981 and 1986–1991, 113–217.
Durham: Durham County Council.
Ferris, I M 2011 Vinovia: The Buried Roman City of
Binchester. London: Amberley Publishing.
Frere, S S 1985 ‘Roman Britain in 1984. I Sites
explored’, Britannia 16: 252–316.
Frere, S S & Hartley, B R 1989 List of samian
sherds, in Frere & Wilkes, 1989: 204–11.
SAIR 92 | 59
Scottish Archaeological Internet Reports 92 2020
Frere, S S & Wilkes, J J 1989 Strageath: Excavations
within the Roman Fort 1973–86. Britannia
Monograph Series No. 9. London: Society for
the Promotion of Roman Studies.
Gillam, J P 1970 Types of Roman Coarse Pottery
Vessels in Northern Britain. Newcastle-uponTyne: Oriel Press.
Greep, S 1987 ‘Lead sling-shot from Windridge
Farm, St Albans, and the use of the sling by the
Roman army in Britain’, Britannia 18: 183–200.
Gustafsson, S 2000 ‘Carbonized cereal grains
and weed seeds in prehistoric houses –
an experimental perspective’, Journal of
Archaeological Science 27(1): 65–70. https://doi.
org/10.1006/jasc.1999.0441
Hanson, W S 1978 ‘The organisation of Roman
military timber-supply’, Britannia 9: 293–305.
Hanson, W S 2007 Elginhaugh: A Flavian Fort and
its Annexe. Britannia Monograph Series No. 23.
London: Society for the Promotion of Roman
Studies.
Hanson, W S, Daniels, C M, Dore, J N & Jobey, G
1979 ‘The Agricolan supply base at Red House,
Corbridge’, Archaeologia Aeliana 5(7): 1–97.
Hartley, K F 1999 ‘Mortaria’, in Symonds, R P
& Wade, S Roman Pottery from Excavations
in Colchester, 1971–86, 165–211. Colchester
Archaeological Report no. 10. Colchester
Archaeological Trust.
Hastie, M 2006 ‘Plant remains’, in Masser,
P ‘Cramond Roman Fort: evidence from
excavations at Cramond Kirk Hall, 1998 and
2001’, Scottish Archaeological Internet Reports
20: 15–16. https://archaeologydataservice.
ac.uk/archives/view/sair/contents.cfm?vol=20.
Accessed 30 March 2020.
Henig, M 1978 A Corpus of Roman Engraved
Gemstones from British Sites. Oxford: British
Archaeological Reports, British Series, 8.
Hillman, G 1981 ‘Reconstructing crop husbandry
practices from charred remains of crops’, in
Mercer, R (ed.) Farming Practice in British
Prehistory, 123–62. Edinburgh: Edinburgh
University Press.
Hird, L 1997 ‘The Roman ceramics’, in Wilmott,
T Birdoswald: Excavations of a Roman Fort on
Hadrian’s Wall and its Successor Settlements: 1987–
92, 233–67. English Heritage Archaeology
Report, 14.
Hobley, A S 1989 ‘The numismatic evidence for
the post-Agricolan abandonment of the Roman
frontier in northern Scotland’, Britannia 20:
69–74.
Hodgson, N 2003 The Roman Fort at Wallsend
(Segedunum). Excavations in 1997–8. Tyne and
Wear Museums Archaeological Monograph No.
2. Newcastle-upon-Tyne.
Hunter, F 2014 ‘Stories from black bangles:
Jewellery and other finds of jet-like material in
Roman Scotland’, in Collins, R & McIntosh, F
(eds) Life in the Limes: Studies of the People and
Objects of the Roman Frontiers presented to Lindsay
Allason-Jones on the occasion of her birthday and
retirement, 152–65. Oxford: Oxbow Books.
Huntley, J P & Stallibrass, S 1995 Plant and
Vertebrate Remains from Archaeological Sites in
Northern England: Data Reviews and Future
Directions. Architecture and Archaeological
Society of Durham and Northumberland
Research Report, 4.
IAWA Committee: Wheeler, E A, Bass, P & Gasson,
P E (eds) 1989 IAWA List of Microscopic Features
for Hardwood Identification. Published for the
International Association of Wood Anatomists.
Johnson, A 1983 Roman Forts of the 1st and 2nd
centuries AD in Britain and the German Provinces.
London: A & C Black.
Jones, M J 1975 Roman Fort-Defences to AD 117,
with special reference to Britain. Oxford: British
Archaeological Reports, British Series, 21.
Jones, R H 2011 Roman Camps in Scotland.
Edinburgh: Society of Antiquaries of Scotland.
Kenward, H K, Hall, A R & Jones, A K G 1980
‘A tested set of techniques for the extraction of
plant and animal macrofossils from waterlogged
archaeological deposits’, Science and Archaeology
22: 3–15.
Knights, B A, Dickson, C A, Dickson, J H &
Breeze, D J 1983 ‘Evidence concerning
the Roman military diet at Bearsden,
Scotland, in the 2nd century AD’, Journal of
Archaeological Science 10: 139–52. https://doi.
org/10.1016/0305-4403(83)90048-1
Leslie, A 2002 ‘The Roman fort at Inveresk’, in Bishop,
M C (ed.) Roman Inveresk: Past, Present and Future,
17–28. Berwickshire: The Amatura Press.
Leslie, A & Will, R 1999 ‘St Michael’s Church,
Inveresk (Inveresk parish), Roman fort’,
SAIR 92 | 60
Scottish Archaeological Internet Reports 92 2020
Discovery and Excavation in Scotland 1999:
29–30.
Macgregor, M 1976 Early Celtic Art in North Britain:
A Study of Decorative Metalwork from the Third
Century BC to Third Century AD. Leicester:
Leicester University.
MacIvor, I, Thomas, M C & Breeze, D J 1980
‘Excavations on the Antonine Wall fort of
Rough Castle, Stirlingshire 1957–61’, Proc Soc
Antiq Scot 110: 230–85.
Manning, W H 1985 Catalogue of the RomanoBritish Iron Tools, Fittings and Weapons in the
British Museum. London: British Museum.
Martin-Kilcher S 1987 Die römischen Amphoren
aus Augst und Kaiseraugst. Ein Beitrag zur
römischen Handels- und Kulturgeschichte,
1: Die südspanischen Ölamphoren (Gruppe
1). Forschungen in Augst 7/1. Augst:
Römermuseum.
Masser, P 2008 ‘An archaeological excavation at
Doune Primary School (classroom extension),
Doune, Stirling’, Headland Archaeology Data
Structure Report, unpublished report.
Masser, P 2010 ‘An archaeological excavation at
Doune Primary School (Janitor’s garden),
Doune, Stirling’, Headland Archaeology Data
Structure Report, unpublished report.
Maxwell, G S 1984 ‘New frontiers: The Roman fort
at Doune and its possible significance’, Britannia
15: 217–23.
Maxwell, G S 1998 ‘Agricola and Roman Scotland:
some structural evidence’, in Bird, J (ed.)
Form and Fabric: Studies in Rome’s Material
Past in honour of BR Hartley, 13–20. Oxbow
Monograph 80. Oxford: Oxbow Books.
Moloney, C 1999a ‘An archaeological evaluation
of the playing fields of Doune Primary School,
Doune, Stirling’, Headland Archaeology Data
Structure Report, unpublished report.
Moloney, C 1999b ‘An archaeological excavation
of the playing fields of Doune Primary School,
Doune, Stirling’, Headland Archaeology Data
Structure Report, unpublished report.
Monaghan, J 1997 Roman Pottery from York
(Archaeology of York). York: Council for British
Archaeology.
Monteil, G 2008 ‘Samian pottery (St Mary’s
Excavations, Colchester)’, unpublished report
for Colchester Archaeological Trust.
Photos-Jones, E 1999 ‘The metallurgical
characterisation of the Tarras Farm, by Forres,
bloomery iron furnace and industrial waste:
Scotland’s earliest dated bloomery furnace?’,
Internal Report SASAA 40.
Piggott, S 1953 ‘Three metal-work hoards of the
Roman Period from southern Scotland’, Proc Soc
Antiq Scot 87: 1–50.
Pitts, L F & St Joseph, J K 1985 Inchtuthil: The
Roman Legionary Fortress Excavations 1952–65.
Britannia Monograph No. 6. London: Society
for the Promotion of Roman Studies.
Polak, M 2000 South Gaulish Terra Sigillata with
Potters’ Stamps from Vechten. Rei Cretariae
Romanae Fautorum Acta, Suppl 9. Nijmegen:
Katholieke Universiteit Nijmegen.
RCAHMS 1963 Stirlingshire: An Inventory of the
Ancient Monuments. Edinburgh: HMSO.
RCAHMS 1994 South-east Perth: An Archaeological
Landscape. Edinburgh: HMSO.
Richmond, I A & Gillam, J P 1952 ‘Further
exploration of the Antonine fort at Corbridge’,
Archaeologia Aeliana 4(30): 239–66.
Richmond, I A & McIntyre, J 1939 ‘The Agricolan
Fort at Fendoch’, Proc Soc Antiq Scot 73: 110–54.
Robertson, A S 1941–2 ‘A Roman oven at Mumrills,
Falkirk’, Proc Soc Antiq Scot 76: 119–27.
Robertson, A S 1975 Birrens (Blatobulgium).
Edinburgh: T & A Constable.
Robertson, A S 1977 ‘Excavations at Cardean and
Stracathro, Angus’, in Haupt, D & Horn, H
G (eds) Studien zu den Militärgrenzen Roms II,
65–74. Köln: Rheinland-Verlag.
Robertson, A S 1983 ‘Roman coins found in
Scotland, 1971–82’, Proc Soc Antiq Scot 113:
405–48.
St Joseph, J K 1961 ‘Air reconnaissance in Britain
1958–1960’, Journal of Roman Studies 51: 119–
35. https://doi.org/10.2307/298845
Schweingruber, F H 1990 Microscopic Wood
Anatomy, 3rd edition. Stuttgart: Paul Haupt
Publishers.
Stace, C 1997 New Flora of the British Isles, 2nd
edition. Cambridge: Cambridge University
Press.
Timpany, S 2007 ‘Environmental samples report’,
in Murray, R ‘Culduthel Mains Farm, Inverness
Phase 5 Excavation of a Later Prehistoric
Settlement: Assessment Report’, unpublished
SAIR 92 | 61
Scottish Archaeological Internet Reports 92 2020
Client Report, Headland Archaeology Ltd.
Tipping, R 1997 ‘Environment and environmental
change in eastern Dumfriesshire’, in RCAHMS
Eastern Dumfriesshire: An Archaeological
Landscape, 10–25. Edinburgh: HMSO.
Tipping, R & Tisdall, E 2005 ‘The landscape of the
Antonine Wall: a review of the literature’, Proc
Soc Antiq Scot 135: 443–69.
Tylecote, R F 1986 The Prehistory of Metallurgy in
the British Isles. London: Routledge.
van der Veen, M, Livarda, A & Hill, A 2007 ‘The
archaeobotany of Roman Britain: current
state and identification of research priorities’,
Britannia 38: 181–210.
Will, R 1998 ‘Tarras by Forres’, GUARD 567
Internal Report, unpublished report.
Willis, S H 1998 ‘Samian pottery in Britain:
exploring its distribution and archaeological
potential’, Archaeological Journal 155: 82–133.
https://doi.org/10.1080/00665983.1998.1107
8847
Willis, S H 2006 ‘Samian Pottery, a Resource for the
Study of Roman Britain and Beyond: the results
of the English Heritage funded Samian Project.
An e-monograph’, Internet Archaeology 17/1.
http://intarch.ac.uk/journal/issue17/willis_toc.
html. Accessed 28 June 2019.
Wilmott, T 1997 Birdoswald: Excavations of a
Roman Fort on Hadrian’s Wall and its Successor
Settlements, 1987–1992. London: English
Heritage.
Woolliscroft, D J & Hoffmann, B 2006 Rome’s First
Frontier: The Flavian Occupation of Northern
Scotland. Stroud: The History Press.
Wright, R P, Hassall, M W C & Tomlin, R S O
1976 ‘Roman Britain in 1975’, Britannia 7:
378–92.
Zant, J M 2009 The Carlisle Millennium Project:
Archaeological Excavations in the Roman Fort
and Medieval Castle, 1998–2001. Vol 1: The
Stratigraphic Sequence. Lancaster Imprints 14.
Lancaster.
SAIR 92 | 62