Roman Roads in Southeast Wales Year 3

Roman roads in Southeast Wales
Desk-based assessment with recommendations

for fieldwork
September 2004
A report for Cadw

by Andy Sherman BA and GGAT report no. 2004/073
Edith Evans BA PhD MIFA Project no.GGAT 75
GG
AT
The Glamorgan-Gwent Archaeological Trust Ltd
Heathfield House Heathfield Swansea SA1 6EL
Roman roads in South East Wales: Desk-based assessment
Summary
This report takes the form of a gazetteer of Roman roads within the former counties of
Glamorgan and Gwent. Where possible, information on the line has been presented in map
form, and the maps for each road are accompanied by a short description. There is a short
assessment of the potential of each proposed road for further work. Appendices present details
of roads for which there is insufficient information to permit mapping.
Acknowlegements
This report was written by Edith Evans BA PhD MIFA and Andy Sherman BA The maps were
compiled by Andy Sherman. The authors are grateful to the staff of the National Monument
Record for access to OS data and providing copies of relevant material, and to Bob Silvester for
advice on methodology. The report was read by Andrew Marvell, who made useful
suggestions; any remaining mistakes are our own. The maps are based upon Ordnance Survey
material with the permission of Ordnance Survey on behalf of the Controller of Her Majesty’s
Stationary Office © Crown copyright. Unauthorised reproduction infringes Crown copyright
and may lead to prosecution or civil proceedings. Glamorgan-Gwent Archaeological Trust Ltd
100017916 2004. Additional annotations © GGAT and Cadw.
1
CONTENTS
Summary ....................................................................................................................................... 1
Acknowlegements ..................................................................................................................... 1
Introduction ............................................................................................................................... 4
The Roman background ........................................................................................................ 4
Previous work in southeast Wales......................................................................................... 5
Methodology ......................................................................................................................... 5
Names and numbers ............................................................................................................. 6
Excluded roads ...................................................................................................................... 7
Terminology.......................................................................................................................... 7
Map format............................................................................................................................ 8
Conclusions and recommendations......................................................................................... 12
Recommendations ............................................................................................................... 12
Gazetteer ................................................................................................................................. 15
RR6d Monmouth - Chepstow........................................................................................ 15
RR60a Newnham-Caerleon............................................................................................. 17
RR60aa Crick - Sudbrook ............................................................................................. 19
RR60b Caerleon - Cardiff................................................................................................ 21
RR60c Cardiff - Neath .................................................................................................... 24
RR60d Neath - Loughor -Pontardulais -Carmarthen....................................................... 27
RR 612a Ariconium (Weston under Penyard) - Monmouth........................................... 29
RR612b Monmouth - Usk ............................................................................................. 31
RR62a Y Gaer (Brecon) - Caerleon ................................................................................ 34
RR62a (variant) Caerleon - Usk (River Usk east bank route via Llantrisant) ................. 38
RR621 Castell Collen - Cardiff ....................................................................................... 41
RR622 Neath - Y Gaer, Aberyscir................................................................................... 45
RRX65 Bassaleg - Lower Machen-Caerphilly ............................................................. 48
RRX80 Penhow - Magor Pill ....................................................................................... 50
RR GGAT 001 Wheel Lane............................................................................................. 52
RR GGAT 002 LG Semicon Road .................................................................................. 54
RR GGAT 003 Sor Brook Valley Road .......................................................................... 56
RR GGAT 004 Ffordd y Gyfraith.................................................................................... 58
RR GGAT 005 Heol Spencer .......................................................................................... 60
RR GGAT 006 West Aberthaw - Llanharry .................................................................... 62
RR GGAT 007 Road on east bank of River Neath .......................................................... 64
RR GGAT 008 Road on Stormy Down/Newton Down................................................... 66
RR GGAT 009 Road at Rumney ..................................................................................... 68
RR GGAT 010 Road at Llandaff ..................................................................................... 70
RR GGAT 011 Possible road at Pantllaca ....................................................................... 72
Appendix 1: Roads in OS files for which there is insufficient evidence to map..................... 74
Appendix 2: PRNs for isolated sightings of possible Roman road, and rejected sites............ 75
Bibliography............................................................................................................................ 77
Figures page
Figure 1: Roman roads and the distribution of known and possible forts and fortresses 10
Figure 2: Roman roads and the distribution of settlements and industrial sites 11
Figure 3: Roman roads 14
Figure 4: Roman Road RR6d (Monmouth - Chepstow) 16
Figure 5: RR60a Newnham-Caerleon 18
Figure 6: RR60aa Crick - Sudbrook 20
Figure 7: RR60b Caerleon - Cardiff 23
Figure 8: RR60c Cardiff - Neath 26
2
Figure 9: RR60d Neath - Loughor -Pontardulais -Carmarthen 1 28

Figure 10: RR 612a Ariconium (Weston under Penyard) - Monmouth 30
Figure 11: RR612b Monmouth - Usk 33
Figure 12.1: RR62a Y Gaer - Caerleon (Y Gaer -Clytha Park) 36
Figure 12.2: RR62a Y Gaer - Caerleon (Clytha Park - Caerleon 37
Figure 13: RR62a (variant) Caerleon - Usk (River Usk east bank route via Llantrisant) 40
Figure 14.1: RR621 Castell Collen - Cardiff (Taf Fechan reservoir - Gelligaer) 43
Figure 14.2: RR621 Castell Collen - Cardiff (Gelligaer - Cardiff) 44
Figure 15: RR622 Neath - Y Gaer, Aberyscir 47
Figure 16: RRX65 Bassaleg - Lower Machen-Caerphilly 49
Figure 17: RRX80 Penhow - Magor Pill 51
Figure 18: RR GGAT 001 Wheel Lane 53
Figure 19: RR GGAT 002 LG Semicon Road 55
Figure 20: RR GGAT 003 Sor Brook Valley Road 57
Figure 21: RR GGAT 004 Ffordd y Gyfraith 59
Figure 22: RR GGAT 005 Heol Spencer 61
Figure 23: RR GGAT 006 West Aberthaw - Llanharry 63
Figure 24: RR GGAT 007 Road on east bank of River Neath 65
Figure 25: RR GGAT 008 Road on Stormy Down/Newton Down 67
Figure 26: RR GGAT 009 Road at Rumney 69
Figure 27: RR GGAT 010 Road at Llandaff 71
Figure 28: RR GGAT 011 Possible road at Pantllaca 73
1
Incorporates RRX 41 and RRX41a (RCAHMW 1995,42 )
3
Introduction
The Roman background
Roman military action against the Silures, the tribe recorded as occupying southeastern Wales,
began in the late 40s and had been completed by the end of the 70s. The extent of the tribal
territory is uncertain, particularly to the west (see Manning 1981, 15-23), but the timescale of
establishment of Roman control over neighbouring areas appears from archaeological evidence
to have been completed by the latter date, when the whole of the area covered by this survey
was under Roman rule. The legacy of this conquest was a fortress for the Second Augustan
Legion at Caerleon and a network of forts consisting at least Abergavenny, Caerphilly, Cardiff,
Coelbren, Gelligaer, Loughor, Neath, Penydarren and Usk (the last replacing an earlier fortress,
probably for the Twentieth Legion Valeria Victrix), and possibly also Chepstow, Caerwanaf
(Miskin) and Monmouth. Earlier models of Roman administration in Wales envisaged that the
area had been largely demilitarised by the end of the 2nd century (see Nash-Williams 1969, 19-
27, figs 4-9), but more recent work has shown that some of the forts which had been thought to
have been decommissioned continued in use into the 3rd or even 4th century. A civilian
infrastructure, which was Romanised to at least some degree, developed alongside the military;
the familiar elements of villas and ‘small towns’ can be recognised, but the largest portion of the
population must have been engaged in subsistence agriculture and were probably living in a
way which had changed little from pre-Roman times (Evans 2001).
The Romans developed the road network initially as an instrument of conquest and
consolidation. It is generally agreed that the prime requisite was for a system of communication
between forts, and routes which have been generally accepted as Roman do precisely this: they
connect Caerleon, Cardiff, Neath and Loughor (RR60/RRX41); Neath, Coelbren and Brecon
(RR622); Cardiff, Caerphilly, Gelligaer, Penydarren and Brecon (RR621); and Caerleon, Usk,
Abergavenny and Brecon (RR62a). Other roads have been proposed to connect Caerleon with
Caerphilly (RRX65) and Monmouth with Usk (RR612b), but with less evidence on the ground.
The extent to which east-west communications north of RR60 were developed between forts
will have depended on the need to move troops quickly from one to another, mainly across the
grain the land. Roads have been proposed on the ridgeways of the Coalfield or climbing from
one valley to another across the ridges (RRX43 i and ii), but none of them have produced clear
evidence for Roman origins.
The completeness of the military network as currently proposed must depend on the extent to
which all the forts have been discovered. The possible fort at Caergwanaf, Miskin, proposed on
the basis of geophysical survey and under trial excavation at the time of writing, is well away
from any known road, unless the two fragments of metalling noted in RR GGAT 010 form part
of a road from Cardiff. This installation appears to be connected with the exploitation of iron
ore deposits, but would also be well placed to defend the gap in the coalfield escarpment where
it is breached by the River Ely, forming a natural communication corridor into the uplands: if
this is the case, another possible location for an undiscovered fort is at Coity, where a major
medieval castle 2 is located on the possible line of a road (RR GGAT 005) heading straight for
the point at which the next gap occurs in the escarpment a few miles away at the River Ogmore.
The conjectural road network presented below in this report will only have been a part of the
full picture in Roman times. It must have been supplemented by a network of minor roads,
2
There is significant overlap in the location of Roman forts and medieval castles. This would
appear to be because the strategic imperatives were the same for the Norman invaders as they
had been for their Roman predecessors.
4
lanes and trackways which were never deliberately surveyed, may never have been metalled,
and will in most cases not be recognisable on the ground as Roman, even with excavation.
Previous work in southeast Wales
The earliest scholarly work on the roads of the area was carried out by William Harris who in
1773 published an account of the Gloucester -Carmathen road (RR 60), known to antiquaries as
the Via Julia, in the second volume of Archaeologia (Harris 1773). He used a combination of
information from the Antonine Itinerary, place-names and archaeological evidence in
attempting to map its course. A certain amount of antiquarian investigation and conjecture took
place during the course of the 19th and early 20th century), and led to the identification of the
main generally accepted routes, as well as other which are much les well founded (see Bradney
1906; 1907; 1921; 1933; Coxe 1801; EH 1872; Wood 1922). However, the history of modern
work on the road network starts in the 1950s, when there were two developments. One was the
production of I D Margary’s Roman roads in Britain (Margary 1955) which codified all the
information known to date about the road network. His work in our area consisted of synthesis
only, and did not include any fieldwork; this was left to archaeological fieldworkers employed
by the Ordnance Survey, who systematically examined all proposed routes and recorded their
observations in the form of notes and annotated maps. For Glamorgan, this work was built on
by the Royal Commission whose investigators reviewed the evidence in the historic county of
Glamorgan for the first volume of the Glamorgan Inventory (RCAHMW 1976), and for the
parishes of Pendoylan and Vaynor for the Brecknock Inventory (RCAHMW 1986). Selection
by RCAHMW for both counties was rigorous: the criteria as laid down for the Brecknock
volume were (a) excavated structural remains; (b) unexcavated agger or paving; 3 (c) milestones,
not necessarily in situ but found close to a likely line; (d) terracing, usually in conjunction with
other evidence and with (e) fairly long alignments and changes of direction on sighting points.
There has been no comparable work for Gwent. 4 Manning (1981, 57-8) included a short
consideration of the roads leading to Usk as part of his report on the fortress excavations, but he
added little to Margary’s account, and there was no new field survey. Comparatively little work
has been done in either county by local enthusiasts, and much of that which has been carried out
has not been informed by sufficiently critical fieldwork. The main contributor of well-thought
out material has been Brian Trott, though some of his proposals unfortunately lack definitive
evidence.
Methodology
The methodology broadly follows that set out in Silvester and Owen (2003, 6). The data for this
survey has been drawn from two principal sources., the regional SMR and the files of data
originally compiled by the Ordnance Survey and now held in the National Monuments Record.
The latter consist of annotated strip maps covering the course or possible alternative courses for
each individual route, together with the standard OS cards, and sometimes supplemented
photographs, measured surveys and other notes. Most of these seem to have been compiled
initially in the 1950s, but some had been updated more recently. None, however, contained
much in the way of recent information. A catalogue of the files is available (RCAHMW 1994).
There is some overlap between the OS material and the SMR, some entries of which were also
derived from OS cards, but the SMR also contains other material, especially entries derived
from excavations and other projects carried out since the 1980s. A search was carried out using
the criteria: type - road, period - Roman. This produced 62 entries, some for lengths of roads of
variable length, others for isolated sightings of features believed to relate to Roman roads.
Separate searches were made for entries on Roman milestones and bridges, forts, settlements
3
This term seems to have been used for metalling.
4
Other than for the parish of Llanelli, now in Monmouthshire but previously in Brecknock,
which was surveyed for the Brecknock Inventory.
5
with urban characteristics, villas and other settlements with good evidence, as defined by the
Romano-British Lowland Settlement Survey (Evans 2001). Given the very large numbers of
entries for Roman finds, particularly in the lower-lying areas of the survey where they were
studied as part of the Romano-British Lowland Settlement Survey, it was not considered
worthwhile to consider their distribution in relation to the roads.
Data from both the SMR and the OS files were then plotted on Mapinfo, using the conventions
established by CPAT. A separate table was established for each road. The only roads to be
mapped were those where there was sufficient information to establish a reasonable line; other
postulated alignments. Plausible routes where there is insufficient evidence to allow for
mapping can be found in the Appendix 1, along with suggested road alignments which have
been examined in the field by the OS and RCAHMW and rejected with good reason. Isolated
sightings of road-related features with no context are listed in Appendix 2. Sections of road
within settlements have not been mapped unless they can be shown to be part of a through road
rather than a local street system.
We have included here some possible additional routes for which no fieldwork has been carried
out. The criteria adopted here were less rigorous than RCAHMW’s. Lengths of modern
alignment which may represent possible fossilised road have been included so that it may be
possible in the future to test them according to RCAHMW criteria. By their very nature, little
evidence can be expected to have survived centuries of use, repair and widening, although there
is scope for fieldwork at the ends of these sections, and on trying to establish possible sighting-
points. By and large, such alignments have only been accepted if there is supplementary
evidence. Straight alignments of modern roads following known Roman routes are the easiest
to accept, such as that over Crack Hill in RR 60c, or at other locations where the road makes for
a fort or settlement of known Roman date, such as Lower Machen (RRX 65). Other indicators
are the Roman style laying-out practice noted in Ffordd y Gyfraith by RCAHMW (1976, 105-
6), or where sections of road go up to and away from important features of medieval date which
stand on their line, with the present road forming a kink around them before resuming its
straight course. This has been noted at Llansannor (RR GGAT 006) where the road kinks
around a church of possible pre-Norman origin (Evans 2003, 71), and at Coity (RR GGAT 005),
where there is an important Norman castle. In other cases, a relatively straight alignment over
the medium distance has been forced on communications by the topography, but the non-
Roman nature of such roads is betrayed by the obvious lack of a coherent system of setting out.
Names and numbers
Roads are generally referred to by forts at the beginning and end of their routes. This can cause
some confusion, since different segments of the same road may be referred to by different fort
names. The solution adopted here has been to name all the forts along the route, enabling
segments to be identified more easily. In some cases, alternative lines have been suggested for
the same routes. These are referred to by topographical features, as ‘Caerleon - Usk west bank
route’.
No record has survived of the Roman name for any Roman road; the name Via Julia, sometimes
used for RR 60 is an antiquarian construct, presumably because it was thought to have been
made by Julius Frontinus. The Neath - Coelbren - Brecon (RR622) bears the traditional name
of Sarn Helen.
6
The numbering system follows that established by Margary and RCAHMW.

Prefix Description
RR Those routes considered to be Roman by Margary; the numbers correspond to
those given in Margary 1955
RRX This is described by Percival (1994) as ‘consisting of those routes considered
by the Ordnance Survey to be Roman or probably Roman but not listed by
Margary’. In fact they also include a number of roads proposed as Roman on
the basis of antiquarian research but rejected by the OS.
RRN Those routes considered to be Roman or probably Roman but which have not
been allocated RR or RRX prefixes by the OS
To these have been added a further class:

RR GGAT Roads not catalogued by RCAHMW
None of the roads prefixed RRZ (miscellanea) by RCAHMW occur within the survey area.
Excluded roads
The following categories of material have not been plotted (see Appendices):
1. Ridgeway routes, unless there is clear evidence that they were incorporated within the formal
Roman road network.
There is little doubt that all or most of these are of ancient origin, as evidenced for example by
the cross dykes at Bwlch yr Afan at the head of Mynydd Llangeinor and Clawdd Mawr at
Blaengarw. They continued in use until industrialisation of the valleys changed the emphasis
from communications trough the uplands to communications along the valleys themselves (see
Yates’s 1799 map of Glamorgan). In some cases, there is clear evidence that Roman roads ran
along the ridgeways. The Neath - Coelbren road runs along the crest of the ridge to the
southwest of the Vale of Neath; the Gelligaer - Penydarren road is clearly attested running along
the ridge of Gelligaer common; these two roads (RR621, RR622) are plotted here. Other
suggested roads have been excluded.
The PRNs located at Blackwood and Hafodyrynys are for areas of paving noted on ridgeways.
Paving is known on some Roman roads in the north of England (Davies, 2002) but there is no
clear evidence for its employment within our area, where all datable exposures have been of
metalling. Given this fact, and the long history of use for these routeways, extending into the
post-Medieval period, the existence of paving cannot be considered as a reliable indication of
Roman date.
2. Isolated exposures of metalling which could be Roman from their form, but which do not lie
on any known or conjectured routes.
3. Records of roads, or elements of roads, whose position is not recorded in sufficient detail to
allow them to be plotted accurately. These can be found in Appendix 1.
Terminology
The following terminology is used:
Status:
i) Known. Extant earthwork or well-recorded buried feature. Shown as solid lines on
OS strip maps.
7
ii) Proposed. Conjectural sections either linking known segments or as hypothetical

road alignments for which there is some physical evidence. Shown as dashed lines
on OS strip maps, and where evidence is reasonably convincing.
iii) Predicted. Virtually no substantive evidence for a road other than someone’s belief
and/or conjectural road alignment with no known traces. Also used in cases where
undated road metalling of Roman type, bearing no relation to the post-medieval
road network has been found during excavation or noted in watching brief, but there
is no further information concerning line or potential destination. Where the
authenticity of a road is in significant doubt, this is the highest level of status that
can be achieved.
iv) Rejected. Used for sites described as Roman roads in the SMR, but for which there
is no reliable evidence.
Each mapped route has been given an overall status, depending upon a general assessment of
the value of the status of individual sections.
Condition (based solely on information available from the existing record, or subjective
assessment of the record)
i) Near intact. Site has suffered only minimal damage
ii) Damaged. Site has suffered substantial damage.
iii) Near destroyed. Very little of the site remains
iv) Destroyed. No trace of the site (above or below ground)
v) Unknown. Condition not known, usually because of incompleteness of records
Each mapped route has been given an overall status, depending upon a general assessment of
the value of the status of individual sections.
Survival (used for mapping)

i) Earthwork. Surviving road showing as an earthwork
ii) Fossilised. Used where road or road line is believed to have been adopted by a
more recent road, path or hedge. (In some cases where there fossilisation is
considered likely but cannot be proved, this convention has been overlaid above the
‘unknown’ convention)
iii) Air photograph. Surviving road shows as a line on an air photograph unrelated to
any modern landscape features, and there is no evidence for an earthwork
iv) Buried. Identified only through probing, geophysical survey or excavation
v) Unknown. Survival not known, usually because of incompleteness of records
Each category is shown as a line of a different colour, except where spot points are involved;
these are marked by a symbol of the appropriate colour.
Map format
The maps accompanying this report are indicative only, and it is expected that, for any serious
study, the routes will be used in their GIS form and overlaid on full modern detail. The GIS
system linked to the regional SMR does not include a separate layer showing rivers, and the
constraints of the project did not permit the generation of such a layer as part of the project. The
only element of physical topography available as a separate layer is the 244m contour, though
the modern unitary authority boundaries have also been included for clarity. In the absence of
detailed physical topography, it was necessary to choose a scale for the maps which provided
sufficient information to make the route of each road intelligible. Scales therefore vary from
map to map, and most are at a relatively small scale.
Forts, fortresses and sites with urban characteristics, both known and suggested, have been
labelled, but other military sites and civilian settlements normally have not, unless this was
necessary for clarity. The industrial sites shown are limited to those where industry appears to
8
have been the most significant aspect of the site; these are the kiln sites, mines, and those sites
adjacent to mines where metal processing appears to have been carried out on a significant
scale. Although the presence of slags demonstrate that metalworking was being carried out at
Caerleon, Cowbridge and many of the rural farmsteads, it seems here to have been just one
component of a larger economy.
9
Figure 1: Roman roads and the distribution of known and possible Roman forts and fortresses
Legend
Fort ?
Fort/Fortress
Abergavenny
Monmouth Other military
Coelbren
Wonastow
Penydarren Land over 244m
Usk Unitary authority boundaries shown

Neath as thin black lines.
Chepstow
Loughor Caerleon
Caerphilly
Sudbrook
Caergwanaf
Kenfig
Glanwenny House
Cardiff
Cowbridge
This map is based upon Ordnance Survey material with the permission of Ordnance Survey on behalf of the Controller of Her Majesty's Stationary Office © Crown copyright.
Unauthorised reproduction infringes Crown copyright and may lead to prosecution or civil proceedings. Glamorgan-Gwent Archaeological Trust Ltd 100017916 (2004).
Annotations copyright of GGAT and Cadw.
Figure 2: Roman roads and the distibution of settlements and industial sites
Legend
Settlement with
Abergavenny ) urban characteristics ?
)
Settlement with
urban characteristics
Villa/Small settlement
Usk
Neath Industrial site
Barland's Farm boat
and waterfront structures
Great Bulmore
Lower Machen Caerwent Land over 244m
Unitary authority boundaries shown

Caerleon as thin black lines.
Cowbridge
Cardiff
Conclusions and recommendations

This survey is essentially an appraisal of existing data on Roman roads in southeast Wales, with
strong emphasis on mapping routes in a form which is readily accessible for SMR searches and
no fieldwork. Only limited conclusions can be drawn with respect to the road network itself.
The main basis of the road network has probably been established, with roads linking most of
the known and presumed forts, though in some areas there is very little physical evidence for
suggested routes, and there is as yet no road linking the Caergwanaf fort into the system. A
certain amount of excavation work has been done in the near vicinity of Caerleon, Usk and
Bulmore on roads as they come into the settled areas, but there has been very little work done
away from forts and towns. There is therefore little information on the form taken by roads in
the countryside.
Another probable lacuna is in communications with shipping places. The forts of Loughor,
Neath, Cardiff, Caerleon and Chepstow are on the lower reaches of navigable rivers, but these
are unlikely to account for all the havens for shipping. Although there are roads leading the
Sudbrook and Magor and possible roads to Aberthaw and Porthcawl, all documented as havens
for shipping in the medieval or post-medieval period, there were probably other havens in use,
for example at Llantwit Major and Swansea. No road is known connecting the Barlands Farm
boat, which presumably lay at a shipping place, within the wider road network.
Recommendations
Further fieldwork is required in respect of most of the routes before they can be properly
established. In respect of those already examined by RCAHMW and the Ordnance Survey
between the 1950s and 1970s, field visits would establish the present condition of the surviving
earthwork remains (where these exist), since during this period there has been considerable
degradation of many archaeological earthwork sites of all types as a result of afforestation and
changes to agricultural regimes. In respect of those sites where no field visits have been made
at all, or where there has only been rapid assessment, this will provide an opportunity to assess
the validity of the route, to look for surviving earthwork features and possibly to test for buried
surfaces by probing.
Aerial photography is another possible way forward. Although the geology and soil cover in
this area have not generally been suitable for revealing cropmarks, these do sometimes occur
when conditions are right, as for example on the River Neath east bank route (RR GGAT 007).
John Sorrell’s work on the Newnham - Caerwent road (RR 60a) has provided useful new detail
concerning earthwork and fossilised remains of this road.
Detailed recommendations are given under the entry for each road. In the case of routes where
extensive rapid field survey has already been carried out by OS fieldworkers and RCAHMW,
further work of this nature seems unlikely to produce more positive results. There are two
possible ways forward for such roads. One is to carry out a programme of monitoring from the
air in the hope of identifying possible stretches of alignment, which can then be tested by
fieldwork on the ground. The other is to test conjectured alignments by limited excavation at
suitable points along the length or at either end. There is a much greater potential for those
roads where very little fieldwork has been done previously. Although some of this group were
initially identified as minor roads, further consideration suggests that the majority run between
forts, or between forts and potential shipping places. Full detailed study is recommended,
involving regression, air photograph search and fieldwork.
12
Roads for which a full study is recommended

RR 60d Loughor- Pontarddulais section only
RRX 65 Bassaleg - Lower Machen - Caerphilly
RR GGAT 003 Sor Brook Valley Road
RR GGAT 005 Heol Spencer (including Coity Castle)
RR GGAT 006 West Aberthaw - Llanharry
RR GGAT 007 Road on east bank of River Neath
RR GGAT 008 Road on Stormy Down/Newton Down
RR GGAT 010 Road at Llandaff
01040.0w Neath - Llandeilo 5
For details of proposals, see individual entries
Roads for which more limited fieldwork is recommended

RR 60aa Crick-Sudbrook (geophysical survey and/or excavation)
RR 62 (variant) Caerleon - Usk (River Usk east bank route) (geophysical survey and/or
excavation)
RR GGAT 011 Possible road at Pantllaca (possibly geophysical survey)
Roads where work should be limited to monitoring from the air, or monitoring of groundworks
during any modification or land use change
RR 6d Monmouth - Chepstow
RR 60a Newnham-Caerleon
RR 60b Caerleon-Cardiff
RR 60c Cardiff -Neath
RR 60d Neath- Loughor section
RR612a Ariconium - Monmouth
RR612b Monmouth - Usk
RR 62a and RR 62a variant Y Gaer- Usk - Caerleon
RR 621 Castell Collen - Cardiff
RR 622 Neath - y Gaer
RRX 80 Penhow - Magor Pill
RR GGAT 004 Ffordd y Gyfraith
Roads where evaluation should be carried out prior to the determination of any planning
permission
RR GGAT 001 Wheel Lane
RR GGAT 002 LG Semicon Road
5
There was insufficient information to plot this road, and it was therefore not given an RR-
prefixed number.
13
Figure 3: Roman roads
Legend
RR612a Land over 244m

RR612b as thin black lines.
RR 621
RR62a
RR622 RR621
RR6d
RR GGAT 007
RR60d RR62a (variant)
RR GGAT 003
RR60a RR60aa
RR GGAT 004 RR60b
RR GGAT 011 RRX65

RR60c RR GGAT 005 RR GGAT 002 RRX80
RR GGAT 010
RR GGAT 006 RR GGAT 001

RR GGAT 008
RR GGAT 009
RR60b
RR60c
RR GGAT 006
Gazetteer
RR6d Monmouth - Chepstow
SO504124 to ST526931
PRN(s) 02460g, 06213g, 06214g

Mapped Figure 4
Status Proposed
Condition Unknown
Route recorded by both Margary and the Royal Commission, although the exact course is
unknown. This road probably follows the same conjectured route out of Monmouth as RR612b
(see below p28), following Cinder Hill Street and Beach Road before joining the B4293 to cross
the river. The route leaves the course of the RR612b at approximately SO5055711542 just
before the tollhouse, continuing to follow the course of the B4293. This road however was
turnpiked in the 19th century, the Roman road probably follows the course of the road which
was in use in 1765, and which still survives as a hollow-way running parallel to the turnpiked
section of the road. The course of the 19th century turnpike and its predecessor merge at
approximately SO5057311119 and the Roman road is generally accepted as following the route
of the B4293 to SO5004909381, where the route diverges to follow the line of the older road
through Lydiart and Redwern Wood to SO4970808401. At this point the route of the road
follows the line of a footpath until SO49695508016, where the footpath rejoins the line of the
B4293 to run through Trellech. Once through Trellech, Margary states that the route is probably
marked by a hollow-way and field boundaries to Trellech Cross; it is difficult to establish which
features Margary had in mind, although those between SO5000105114 and SO4963004240
would appear to run in the right direction. The route is then assumed to follow the line of the
modern road through Parkhouse and Old Park Farm to Tintern Cross. Between SO5081300171
and SO5134599548 the road is assumed to follow a hollow-way between Tintern Cross and Fair
Oak, where the road rejoins the line of the modern road running from Banton to St Arvans.
Between SO5193497188 and SO5205796810 Bradney believed the course of the Roman road
left the modern road and followed a hollow-way down to the line of the modern A466, and from
here the route is generally assumed to follow the line of the A466 into Chepstow. Bradney
believed that between ST5220795372 and ST5223994539 the course of the road briefly leaves
the line of the A466 to follow a winding hollow-way, before rejoining the course of the A466
into Chepstow.
Sources:
Bradney 1933, 36
Margary 1957, 54-5
RCAHMW 1994, 1
Ordnance Survey strip maps.

OS Record Card 1958, SO51SW 27
Recommendations for further work

This road has already been the subject of fieldwork by OS fieldworkers, and further field
survey at this stage is unlikely to be very profitable. Air photography would appear to afford
the best means to provide new information, but could be followed by fieldwork if new
information emerges. Also monitor any ground disturbance along the route.
15
Figure 4: Roman road RR6d (Monmouth - Chepstow)
3
50 55 Legend
Predicted
Monmouth Fossilised
Wonastow Fort ?
Fort/Fortress
Millitary
10
Settlement with urban
characteristics
Land over 244m

as thin black lines
05
3
00
2
95
Chepstow
This map is based upon Ordnance Survey material with the permission of Ordnance Survey on behalf of the
Controller of Her Majesty's Stationary Office © Crown copyright.
Unauthorised reproduction infringes Crown copyright and may lead to prosecution or civil proceedings.
Glamorgan-Gwent Archaeological Trust Ltd 100017916 (2004).
RR60a Newnham-Caerleon
ST5299994799 to ST3417990380
PRN(s) 00561g, 01016.1g, 01160g, 05316.0g, 05319.0g, 07048g, 07049g, 01016.17w

Mapped Figure 5
Status Known/Proposed
Condition Damaged
This is the eastern section of a much longer route from Chepstow to Loughor, part of Iter XII of
the Antonine Itinerary. It was outlined by Margary, and the published RCAHMW analysis
mainly adheres to the Margary route. However the course through Chepstow is totally unknown
and it is only after the Beechwood roundabout that a route can be traced. The line here has been
established by Sorrell from the air, and confirmed by fieldwork to exist in earthwork form as a
short stretch of agger to the west of the roundabout; beyond Pwllmeyrick it is marked by field
boundaries. West of Crick, the road is assumed to follow the route of the modern A48 through
Caerwent, though the section between Crick and Cat’s Ash is almost all purely conjectural,
except where it passes through Caerwent where it is confirmed by the position of the gates.
Although modern A48 consists of a series of straight alignments, these may owe their origin at
least in part to turnpiking. To the west of Penhow Castle, Ordnance Survey fieldworkers
excavated a short section of agger but little further information was recorded about the results
of the excavation. To the west of Cat’s Ash the route becomes clearer, as metalled road surfaces
have been excavated by Maynard in 1996 and Yates in 1997. 6 The route has also been traced as
field boundaries by Richards in 1985 and Maynard in 1996. As it approaches Caerleon it
appears to be fossilised as Isca Road, which is on the direct line of the via praetoria. It is
assumed to enter the fortress at the porta praetoria over a bridge across the River Usk.
Sources
GGAT 1992,
Locock 1992a
Locock 1992b
Margary 1973, 56
Maynard 1996
Mein 1990
RCAHMW 1994, 2
Richards 1985
Sorrell 2004
Yates 1997
Ordnance Survey strip maps

Air photography is unlikely to provide further information for much of the route if the road
follows the postulated route along the A48 and other modern roads, although deviations might
be apparent as cropmarks under suitable conditions. The largest lacuna lies between Five Lanes
and Penhow Castle, where air photography might be particularly useful. The most likely
method of obtaining additional information is through the monitoring of any ground
disturbance along the route. Further excavation to the west of Catash would probably provide
additional information on the exact course of the road over Belmont Hill, but is considered to
be of relatively low priority in any programme of work on Roman roads as a whole, since this
end of the route is relatively well understood.
6
Three further sections of metalled road surface in the Cat’s Ash area have been excavated by
Mein in 1990 and Locock in 1992, but these probably belong to roads other than the RR60a.
17
Figure 5: Roman road RR60a (Newnham - Caerleon)
3
35 40 45 50
Legend
Short section buried
Short section rejected

Earthwork
Buried
Predicted
95 Unknown
Fossilised
Rejected
Fort ?
Great Bulmore Fort/Fortress
Caerleon Other millitary

Chepstow
Caerwent Settlement with
1
90 Land over 244m

as thin black lines
Sudbrook
RR60aa Crick - Sudbrook

ST4879087637 to ST4995787694
PRN(s) 07761g
Mapped Figure 6
Status Proposed
Condition Unknown
Road, proposed by both Margary and the Royal Commission, as leaving the main Newnham-
Caerleon road at Crick, heading towards a probable port situated at either Black Rock or
Sudbrook. The Iron Age fort at Sudbrook seems to have been the site of an early Roman fort.
The route is believed to be fossilised in the modern Crick Road as far south as St Mary’s
Church and then by the Sudbrook Road, though there is no other supporting evidence; however
once past Sudbrook the route of the road is not known.
Sources
Margary, 1957, 56
Manning 1981, 42
RCAHMW 1994, 2

be apparent as cropmarks under suitable conditions. The most likely method of obtaining
additional information is through the monitoring of any ground disturbance along the route. To
the south of Sudbrook church, however, it is possible that the Roman line continued the line
further north rather than following along the course of the modern road. If this is the case,
geophysical survey or excavation at this point would help to confirm that this road is Roman.
19
Figure 6: Roman road RR60aa (Crick - Sudbrook)
3
45 50 Legend
Fossilised
Predicted
RR60a
(Newnham - Caerleon)
Chepstow
Fort ?
Fort/Fortress
Other millitary
Settlement with
Land over 244m

Caerwent
1
90 Unitary authority boundaries shown
as thin black lines
Sudbrook
RR60b Caerleon - Cardiff

ST3374990779 to ST1800876456
PRN(s) 01016.2w, 01016.13w, 01016.16w, 03020s

Mapped Figure 7
Status Proposed
Condition Unknown
Route recorded by both the RCAHMW and Margary. It is assumed to form part of that stretch
of the Antonine Itinerary Iter XII between Caerleon and Neath, since there is a Roman fort at
Cardiff which must have been connected to the road network. However, there is little physical
evidence for its existence. The route that this road takes on leaving the north gate of Caerleon is
not known, although the presence of cemeteries over the lower slopes of Lodge Hill indicates
that it probably ran along here. It was originally thought to follow the line of a linear bank, but
Courtney has demonstrated that this is probably a medieval park boundary. Excavations in the
cemetery area by Evans and Maynard in the area of what was locally believed to be the road
found only areas of ill-defined and poorly founded metalling, which were considered to
represent handstandings rather than the road. In 1967 the Ordnance Survey excavated a short
section of agger between ST322907 and ST324908; unfortunately little further information
about this stretch of the road was recorded, and it is not clear whether this is the earthwork
which Courtney has rejected. In 1991 Maylan suggested that the route of the road might cross
the Pill-Mawr stream at approximately ST312904 by a wooden or stone bridge, for which no
evidence is known. After this point the route is unknown until approximately ST3029888465
where the modern road known as Allt-yr-yn Avenue begins. From this point Margary and the
Ordnance Survey record the route as following a series of modern roads into and through
Bassaleg: The Ridgeway, Glassllwch Lane, Park View and Caerphilly Road.
The route after Bassaleg is uncertain. The Ordnance Survey strip map shows the route leaving
the modern road known as The Griffin northeast of its junction with the lane leading to Croes-
heolydd, and then running north to an old quarry as a hollow-way and embanked road, before
turning northeast and running along a field boundary as a hollow-way to meet a footpath which
joins the Pentrepoeth Road just to the east of The Knoll House. However the Royal Commission
rejected this route in 1991, since no Roman material had been found along it. Locock followed
this route during fieldwork during 1994 and states that no definite evidence to support the route
was visible. Margary suggests that the route simply follows The Griffin for its entirety. No
doubt this opinion was influenced by its notable straightness, but the road was turnpiked in the
19th century and cannot therefore be used in evidence; however in default of any other plausible
route, the line of The Griffin road has been followed for the purposes of this gazetteer.
To the east of Michaelstone-y-Vedw the route of the road becomes more certain: the Ordnance
Survey based its probable route between ST2499084442 and ST2429384045 and again between
ST2424883756 and ST2422983647 on the evidence of aerial photographs. In 1978, Trott noted
the line of the road as it passes Michaelstone-y-Vedw, again from aerial photographs, at
ST244842 or slightly to the northwest of where the Royal Commission notes it. After the
second of these aerial photgraphs the route again becomes more uncertain. The general
assumption has been that the route is fossilised by the Druidstone Road until ST2376583089
where a course has again been plotted on aerial photographs as far as ST2370482918, where it
is again assumed to follow the route of the modern Druidstone Road. The route is then general
assumed to follow the modern line of Ty Winch Road, Newport Road, Chapel Row, Oldhill,
and the Newport Road again. In 1960 during the widening of Newport Road, a well-metalled
road surface was noted lying beneath turf and topsoil at ST21007875 by workmen. After tithe
maps for the area had been examined, it was decided that this surface marked the route of the
21
Roman road. After this point the exact course is lost again , but route is assumed to follow the
route of the Newport Road, Queen Street, Duke Street and Castle Street.
Sources
Courtney 1983
Evans and Maynard 1997, 172
GGAT 1998a
Locock 1994
Margary 1957, 57
Maylan 1991
RCAHMW 1976, 106
RCAHMW 1994, 2-3
Trott 1978

follows the postulated route along modern roads, since most of it lies within built-up areas. The
most likely method of obtaining additional information is through the monitoring of any ground
disturbance along the route. The two main exceptions to this are the eastern end where the
route leaves Caerleon, and the section between the built-up areas of Newport and Cardiff,
where a couple of deviations from the line of the modern road have been observed from the air
at Michaelston-y-Fedw and Druidstone Road. These cropmarks could be tested by excavation.
Little success has so far attended attempts over a considerable length of time to find the exact
course of the eastern end of the road on Lodge Hill, Caerleon, in spite of the fact that fairly
extensive sections of the eastern end of the hill have been developed. There are now few
development pressures on the remaining parts of the hill other than on the hospital site, where
predetermination archaeological assessment will be required.
22
Figure 7: Roman road RR60b (Caerleon - Cardiff)
3
15 20 25 30 35 40 Legend
Possible bridge at
Great Bulmore Pill-Mawr stream
Short section unknown
90
Short section
Caerleon air photograph
Lower Machen
Fossilised
Air photograph
Caerphilly Unknown
Predicted
Buried
85
Fort ?
Fort/Fortress
Other millitary
Settlement with
1
80 Land over 244m

as thin black lines
Cardiff
RR60c Cardiff - Neath

ST31819876506 to SS7481897770
PRN(s) 01016.14w, 02738w

Mapped Figure 8
Status Proposed
Condition Unknown
Route recorded by both the RCAHMW and Margary along the line of the modern A48. It is
assumed to form part of that stretch of the Antonine Itinerary Iter XII between Caerleon and
Neath, since there is a Roman fort at Cardiff which must have been connected to the road
network. Other than the line, and the fact that it was adopted as a boundary by several parishes,
there is little evidence for its existence. There are, however, four points on this route where
some physical evidence has been noted:
• In 1957 the Ordnance Survey record that sections of the road have reportedly been noted in
the banks of the River Ely, in Cardiff, close to the location of the modern Ely bridge at
approximately ST145768.
• What was interpreted as the ditch on the north side of the road, and buildings aligned on it,
have been excavated where it runs through Cowbridge (the most probable site of the
otherwise unrecorded Bomium the only stage between Caerleon and Neath given in the
Antonine Itinerary). 7
• Between the Rivers Ogmore and Ewenny, on the 200m length of agger running from
SS904782 to SS902783 (RCAHMW 1976, 107). Here, an unpublished excavation ‘close to
the correct line of sight’ is noted by RCAHMW (1994), but no details are available.
• In 1976 Trott noted that the crest of Stormy Down at SS845804 would appear to have been
a sighting point for the route of the RR60c as it ran east-west across Newton Down and
Stormy Down. (MM Record Card, OS 1957, Glam 41 SW 14) .
Passing out of Cardiff to the east of the Roman fort, the road is generally assumed to follow the
course of the modern A48 except in five places.
• Just east of St Nicholas, the route is assumed to follow a road known as Grants Field
between ST1150474669 and ST1040174529. This road follows the boundaries between the
parishes of St George-super-Ely and St Lythans.
• Between ST0552873939 and ST0432274200, the route again follows a road that marks the
boundary between two parishes, Llantrithyd and Welsh St Donats. However it should be
noted that in 1919, when Codrington investigated a quarry (now disused) on its assumed line,
no evidence of it had been noted by the quarry personnel.
• At Cowbridge the modern road bypasses the town, whose high street preserves the original
line.
• Between SS8718579447 and SS8613879179; in 1957 the Ordnance Survey suggested that
the route followed a series of aligned field boundaries.
7
RCAHMW (1976, 121-2) suggests a possible location further west at Glanwenny House, but
no evidence has been found to support the presence of a Roman site here.
24
• Between the NGRs SS84250180630 and SS8010085550, to the south of North Cornelly, the
route is assumed to take a more eastern course closer to the coast, crossing Kenfig Sands to
the east of Kenfig Castle and following the route of the modern road known as Water Street.
Kenfig has been suggested as the site of an early fort on the basis of finds, but the
besandment of this area has made it impossible to establish whether this is the case. We
would suggest an alternative route following the modern course of the A48 north through
Pyle and then running northeast towards Margam. This would bring the road closer to
SS82698222 and SS81608485, the locations at which two milestones, respectively RIB 2251
(PRN 00142m) and RIB 2255 (PRN 00790m) were discovered. 8
Sources
Margary 1957, 58
Marvell and Page 1994
RCAHMW 1976, 106-8
RCAHMW 1994, 3-4
Parkhouse and Evans 1996, 80-9

be apparent as cropmarks under suitable conditions. Also monitor any ground disturbance
along the route.The section of the road around North Cornelly, Pyle and Margam is least well
understood, and air photography may have the potential to provide most information. Also
monitor any ground disturbance along the route. Otherwise, this road has been extensively
studied over an extended period of time (most recently by RCAHMW for the Glamorgan
Inventory), and without any new information, further fieldwork is not likely to be very
profitable.
8
This alternative route incorporates two earthworks that have previously been identified as
possible elements of Roman roads, one at SS8095085170 (PRNs 01016.14w. 02738w) and one
between SS80608580 and SS79958605 (P-TM6). However, both of these earthworks were
excavated in 1994 by Marvell and Page and were determined to be of later date.
25
Figure 8: Roman road RR60c (Cardiff - Neath)
2 3
2 75 80 85 90 95 00 05 10 15
Legend
00
Neath Possible sighting point

Fossilised
Gelligaer Predicted
95
Buried
Unknown
Rejected
Fort ?
90
Fort/Fortress
Caerphilly
Other military
85 ) Settlement with
urban charateristics ?
Settlement with
Caergwanaf urban characteristics
Kenfig
80 Land over 244m

as thin black lines
Glanwenny House
1
75
Cowbridge Cardiff
Annotations copyright of GGAT and C adw.
RR60d Neath - Loughor -Pontardulais -Carmarthen 9

SS7479097800 to SS5648098040
PRN(s) 01016.0w, 01040.0w, 05306w, 05311w

Mapped Figure 9
Status Proposed
Condition Unknown
Route recorded by both the Royal Commission and Margary; whilst little physical evidence is
known for this road and the exact route is uncertain its existence is attested by the Antonine
Itinerary (Iter XII). Very little of this route is known and the only generally accepted section of
the road is fossilised as the B4620 (Swansea Road and Loughor Road) between SS6195096580
and SS5648098040, where it passes to the south of the two practice works on Carn Goch
Common and to the north of the large practice work on Stafford Common. The route is then
generally assumed to follow the route of modern A484 into the east gate of Loughor fort. The
only feature probably relating to it that has been recorded is the well-constructed ford across the
Tawe, discovered whilst excavating for the Pottery Lock on the Swansea Canal; the exact
location is uncertain, but RCAHMW (1976) suggests that it was around SS66119468.
Beyond Loughor, Rivet and Smith suggest that the mileages in the Antonine Itinerary imply that
the most likely course of the road would have run northwards up the east bank of the River
Loughor to an easier crossing between Pontarddulais and Hendy at around SN5855002980.
Although there is no physical evidence for a road, its importance as a crossing in the Middle
Ages is indicated by the existence of two earthwork castles, one on either side of the river, and
the church of Llandeilo Talybont on the proposed line, which was a stage in the pilgrimage
route to St Davids.
It should also be noted that a section of this road is described as having been excavated in 1988
by the Neath Antiquarian Society somewhere in the region of Glyn Leiros, on a south-west
alignment (Neath Antiquarin Newsletter no:2). However, insufficient information was provided
to make it possible to map this section of road.
Sources
Margary 1957, 59
Marvell and Owen-John 1997, 222-3
RCAHMW 1976, 108
RCAHMW 1994, 4-5, 42
Rivet and Smith 1979, 174-5

Much of the any likely line of the road east of Garn Goch, where it is essentially unknown, is
built up and therefore unlikely to produce additional information from further air photography.
A possible exception is between Skewen and Bon-y-maen. Otherwise, this part of the route has
been extensively studied over an extended period of time (most recently by RCAHMW for the
Glamorgan Inventory), and without any new information, further fieldwork is not likely to be
very profitable. However, virtually no work has been carried out so far on the postulated
section between Loughor and Pontarddulais, and a full programme of map regression, air
photograph search and fieldwork is recommended. Elsewhere, monitor any ground disturbance
along the route.
9
Incorporates RRX 41 and RRX41a (RCAHMW 1995,42 )
27
Figure 9: Roman road RR60d (Neath - Loughor - Pontardulais - Carmarthen)
2
60 65 70 75 Legend
Possible ford
Fossilised
Unknown
Predicted
Fort ?
Fort/Fortress
2 Other military
00
Settlement with
Land over 244m

as thin black lines
Loughor
Neath
1
95
RR 612a Ariconium (Weston under Penyard) - Monmouth

SO5299616062 to SO5103313120
PRN(s) 02954.0g
Mapped Figure 10
Status Proposed
Condition Unknown
Route forming the first section of the Antonine Iter XIII, recorded by both Margary and the
Royal Commission; however only the last 4km of the route come within the boundaries of this
study. Very little of the route is physically confirmed, although it is generally assumed to follow
the course of the modern A40. RCAHMW notes that N P Bridgewater excavated a section of
the road in 1958, slightly to the west of the course of the A40, at SO526151. The route probably
leaves the course of the A40 briefly between SO5265114670 and SO5227313867 to follow the
course of the older Newton Court Lane. At SO5167113558 the route leaves the course of the
A40 and follows Dixton road into the Monmouth.
Sources
Margary 1957, 62
RCAHMW 1994, 5

survey at this stage is unlikely to be very profitable. Air photography is also unlikely to
provide further information if the road follows the postulated route along the A40 and other
modern roads. The most likely method of obtaining additional information is through the
monitoring of any ground disturbance along the route.
29
Figure 10: Roman roads RR612a (Ariconium - Monmouth)
3
50 55 Legend
Fossilised
Buried
Fort ?
Fort/Fortress
Other military
Settlement with
Land over 244m
2 Unitary authority boundaries shown

15 as thin black lines
Monmouth
RR612b Monmouth - Usk

SO5029012560- SO3378000600
PRN(s) 02460g, 4772.0g

Mapped Figure 11
Status Proposed
Condition Unknown
The existence of this road is attested by the Antonine Itinerary (Iter XIII) although the exact
route is unknown. Margary suggests a route which passes along Jingle Street, to Dingestow, and
on through Raglan to Usk. However he includes little specific route information, stating that
much more extensive fieldwork is required. For this reason, Margary’s route is not mapped for
this gazetteer beyond Jingle Street. Coxe preferred a route which crosses the river Trothy to the
south of Jingle Street, passes west of The Bourne and through Pen-y-clawdd on its way to Usk;
however his route is dismissed as illogical by the Ordnance Survey in 1958. In the same source
the Ordnance Survey also note and dismiss a possible route which runs almost due south after
leaving Pen-y-clawdd, passsing Trecastle to the east and running to London Cottage. Although
the above routes are all noted by the Royal Commission and mapped in the Ordnance Survey’s
linear files, the most likely route they record is that favoured by both Manning and Maynard.
This is the most easterly of the routes mapped and passes through Gwaylodyae, Great Llan
Thomas, London Cottage and The Crosshands on it way to Usk. It is this route which is
described below.
The exact course of the road as it leaves Monmouth is unknown; the Ordnance Survey records
two possible routes. The first of these routes leaves Monmouth to the east and follows the line
of the modern Wonastow road through Over Monnow, towards the recently discovered fort at
Wonastow and Jingle Street, turning south to cross the river Trothy and joining the Mitchel
Troy road at The Red House Farm. The second route leaves Monmouth to the south, following
Cinder Hill Street and Beach Road before joining the B4293 to cross the river. This route then
leaves the B4293 to follow the Mitchel Troy road, passing through Mitchel Troy to The Red
House Farm. The route then crosses the Cwm Carfan Brook before turning south at
SO4726310253 and passing through Gwaylodyae, at SO4721108614 it turns and heads in
westerly direction to run through Great Llan Thomas. At SO46240681 the route diverges from
the course of the modern road and follows the line of a 2m high, 10m wide bank until
SO46440681. At the point where the route leaves the modern road, the annotation on the
Ordnance Survey strip map records that light ploughing has uncovered patches of metalling.
This same ploughing has also disturbed pieces of slag supposedly of a Roman date from the
edges of the bank. The route then rejoins the line of the modern road and runs towards London
House. However, although significant stretches of the modern roads on this route are straight
and look as though they may have had a Roman origin, they have been turnpiked and cannot be
regarded as providing reliable evidence.
Between SO4545706365 and SO4521106171 the route follows the line of a spread hollow-way,
described as being 15m wide and between 2-4m deep; after the hollow-way the route is lost
crossing an open field. It is picked up again in a field boundary running between
SO4499006007 and SO4481705813, then follows a tree and shrub covered hollow-way, which
is on average 15m wide and between 2-4m deep, to SO4481705813. Here the route is
interrupted by an unnamed brook crossed by the London Bridge, where the route is again lost
for a short distance. It is re-established at SO4481105780 where it takes the form of a well-
constructed terrace which ascends a fairly steep hillside as far as SO4467005539. This terrace is
formed by cutting into the hillside to an average depth of 2m in the northwest, with a build up in
the southeast to an average depth of 2m to form a level surface. Probing of the topsoil along
sections of the terrace, not planted with crops, discovered no evidence for metalling. The route
then re-aligns with the course of the modern road to pass through Four Ash, Llanishen Court,
31
The Cross Hands, Pergoed Farm and Pont Pill to become the Monmouth Road at
SO4100001797. After this point the route of the road is totally unknown. 10
Sources
Margary 1957, 62-3
Manning 1981, 56-58
Maynard 1991
RCAHMW 1994, 5-6

survey at this stage is unlikely to be very profitable. Air photography would appear to afford
the best means to provide new information, but could be followed by fieldwork if new
information emerges. Also monitor any ground disturbance along the route.
10
It should be noted that a short section of road has been proposed leaving the Michel Troy road
at SO5056011541. This route has only be conjectured as far as SO50580211081 and is
supposed to follow the ‘pre-turnpike’ route of the B4293, now preserved as a hollow-way,
although no suggestion has been made as to its destination.
32
Figure 11: Roman road RR612b (Monmouth - Usk)
3
35 40 45 50 Legend
Monmouth Short section unknown
Short section fossilised
Wonastow
Short section earthwork
Predicted
Unknown
Earthwork
10 Fossilised
Fort ?
Fort/Fortress
Other military
Settlement with
Land over 244m

05

as thin black lines
Usk
2
00
RR62a Y Gaer (Brecon) - Caerleon

SO2458016350 to ST34130901810
PRN(s) 02902g, 04457g, 05892.0g, 05894.0g

Mapped Figure 12
Status Proposed
Condition Unknown
Route running from Y Gaer (Brecon) to Caerleon, recorded both by Margary and the Royal
Commission; the section of this route northeast of Gilwern is not within the boundaries of this
study area. Although the exact course of this route is unknown, it is generally assumed to follow
the course of the modern A40 trunk road, entering the study area at Glangrwyney and into the
town of Abergavenny along Brecon Road. It is then lost for a short distance, but leaves the town
along Cross Street and Monmouth Street. 11
At SO3075012480 the route of the Roman road leaves the A40 trunk road and follows the
B4598 to SO3217010488. At this point the Ordnance Survey believes the route diverges from
the modern road system to follow a series of field boundaries before again merging with the
route of the B4598 at SO3325010220. At SO3472009630 the route leaves the B4598 and
follows the minor road towards Clytha Park. Here, at SO3632008820 the road turns in a
southerly direction, running through Bettws Newydd and Troestry Court. Between
SO3650004100 and SO3670003600 the route follows the line of a bank which varies in width
between 17m and 22m; probing under the surface of this feature identified a metalled surface
approximately 8m wide at a depth of 0.2m below ground level. At SO3678003450 the route is
assumed to rejoin the line of the modern road through Llancayo, which rejoins the B44598 at
SO3687003040 to run into Usk.
The Royal Commission handlist notes that in 1877, W T Watkins excavated a section of the
road at SO364005 as it ran through Usk; unfortunately no further details are given. In the early
1970s, Manning excavated in a similar area and uncoverred what he believed to be a section of
the via decumana. Numerous other excavations inside Usk have located the via decumana as it
passes through the fort; these are omitted here since roads within forts have not been mapped
for this project. Although the crossing-point over the River Usk is not known, it is general
assumed that the road by NGR SO3755000160 it has crossed over the River Usk to follow the
course of the modern road through Llanbadoc and Llangybi. At ST3758399278 the Ordnance
Survey strip map notes two buildings, Helmaen Hotel and Helmaen Cottage; and suggests that
Helmaen = Heol Maen (Stone Road). At ST3682095310 the route leaves the course of the
modern road and follows the line of a footpath to the village of Croes Llywarch, from which
point it follows the line of the road through Llanhennock onto Caerleon, but is lost for a short
distance at approximately ST3323090190. At ST3429091020 a localised deep band of cobbles
can be seen eroding out of the east bank of the Afon Lwyd, and probably represents the road as
it crosses the river to enter Caerleon. In 1987 Evans excavated a section of road on this line,
between approximately ST3412790812 and ST3424790905, where it formed a continuation of
the via principalis sinistra running through the settlement outside the fortress. The first road laid
out in this section was constructed over a raft of brushwood, with a carefully laid and
11
It should be noted that whilst PRN 04457g (NGR SO30361474) might be associated with the
route of this road, it is not possible to be certain. Information on the SMR states that sections of
a Roman road (former linear PRN A33), associated with a coin hoard (PRN 1370g) and burials
(PRN 1375g), were found during the construction of housing on the outskirts of Abergavenny in
c1848. However, with only a single NGR it was not possible to determine whether it could have
been connected with a possible alternative course for this road.
34
compacted surface of graded cobbles and pebbles, which had a slight camber. An unlined road
ditch was recorded to the north. It had been remade at intervals. 12
Sources
Evans 2000, 35-8
GGAT 1991
Manning 1989, 56-58
Margary 1957, 65-6
RCAHMW 1986,
RCAHMW 1994, 6-7

Since most of the postulated line of this route is assumed to follow modern roads and footpaths,
there is little scope for further work, though the part immediately outside Caerleon as it crosses
the River Usk may produce results. However in this area the river has clearly a changing
pattern of meanders over time, and at least some of the road has probably been lost to erosion
on the lower ground, and to the 19th century abortive railway cutting on the higher ground.
Elsewhere along the route, monitoring from the air is a possible tool for determining whether it
did indeed run along modern roads, or whether any possible deviations can be seen which could
then be the subject of further fieldwork. Also monitor any ground disturbance along the route.
12
The map for this section includes part of the excavated street system of the civil settlement
associated with this road.
35
Figure 12.1: Roman road RR62a (Y Gaer - Caerleon) Y Gaer - Clytha Park
3
25 30 35 40 Legend
Possilble place
name evidence
15 Short section earthwork
Fossilised
Abergavenny ) Unknown
Earthwork
Buried
Predicted
Fort ?
Fort/Fortress
Other military
Settlement with
2 ) urban characteristics ?
10
Settlement with

as thin black lines
Figure 12.2: Roman road RR62a (Y Gaer - Caerleon) Clytha Park - Caerleon
10
Legend
3
35 40 Short section unknown
Possible place
name evidence
Short section earthwor
Fossilised
Unknown
Earthwork
Buried
Predicted
05
Fort ?
Fort/Fortress
Other military
Settlement with
) urban characteristics ?
Settlement with
Usk
2 Land over 244m
00

as thin black lines
95
Great Bulmore
1
Caerleon
90
RR62a (variant) Caerleon - Usk (River Usk east bank route via Llantrisant)
ST3475590254 to SO3791600377
PRN(s) 04921g, 04922.0g, 04426g, 05596.0g, 05597.0g, 05666.0g, 06053.0g

Mapped Figure 13
Status Known
Condition Damaged
This road is an alternative route between Usk and Caerleon, running along the opposite side of
the river from RR62a, which was the only route considered by Margary. However, it appears in
the Royal Commission handlist, and the physical evidence supporting it is compelling. The
description given here runs south-north.
This route probably joins with RR60a at the junction between Isca Road and Bulmore Road,
and then follows Bulmore Road to Bulmore. Excavations carried out by Yates in 1999 at Little
Bulmore failed to find conclusive evidence for the road, though metalling in two trenches at
ST35679133 near the line of the modern road could be connected. At Great Bulmore, a long
section was examined by Zienkiewicz during excavations in 1984 between ST35839146 and
ST35949151 (approx). Further sections of metalling were recorded by Maynard during a
watching brief in 1988, at ST36259160 and at ST35989152; and during field evaluations in
1996, between ST36409166 and ST36859172. 13 Tuck recorded a further section of metalling on
the route at ST36999179 during observations on a service trench in 1999. Beyond Bulmore,
Copeland has documented a section of trackway which runs between ST3898697257 and
ST3916298149, and probably represents the fossilised course of the route between Llantrisant
and Llanllywell. The course of the route between White Hall and Llantrisant has also been noted
during recent aerial photograph flights by Sorrell (pers comm M Tuck), although no detailed
information has yet been made available to the SMR.
Further excavation information becomes available as the road approaches Usk. A section of
metalling on the northern end of this route was excavated by Maynard at SO37970025 (approx)
in 1993 (Maynard and Marvell 1994), and part of another road, possibly a legionary patrol track
running around the fortress, at SO37970025. A metalled surface noted in an east-west stream
bank at approx ST371916, and a short section of hollow way at ST36959165 may have been
part of a route linking the Bulmore road to the Caerwent-Caerleon road, but are more likely to
be sections of a trackway running to a Post medieval quarry at ST37089175. However, it is
possible that the Caerwent-Caerleon road and the Bulmore road might have been linked in some
way, and there may have been metalled trackways connecting the sites in this area of the Usk
valley (pers comm Tuck).
Sources
Copeland 1980
Maynard 1988
Maynard 1993b
Maynard 1996a
Maynard 1996b
Maynard and Marvell 1994
Maynard and Marvell 1995
RCAHMW 1994, 7
Tuck 1999
Yates 1999
Zienkiewicz 1984, 4-8
13
Maynard has also recorded a third section of metalling on this route in 1996, although the
exact location is not recorded due to an error in the NGR.
38

The southern end of the route as it passes through Bulmore, and the northern end as it
approaches Usk, make this one of the best understood and recorded of all Roman roads in
southeast Wales. The central section is, however, still largely conjectural. Air photography is
already providing additional information about this road. It may run, as shown, on the line of
the modern road, but Post-medieval improvements to this make it unclear whether it is actually
of Roman origin. There are however sections around Llantrisant and south of Llanllowell
where geophysical survey and excavation to confirm the line might be possible.
39
Figure 13: Roman road RR62a variant (Caerleon - Usk) River Usk east bank route via Llantrisant
3
35 40 45 50 Legend
Usk Short section buried
2
00 Short section
air photograph
Air photograph
Predicted
Fossilised
Burried
Rejected
RR60a
95 Fort ?
Fort/Fortress
Millitary
Settlement with
Great Bulmore urban characterirstics
Land over 244m

Caerleon
Caerwent Unitary authority boundaries shown
as thin black lines
1
90
RR621 Castell Collen - Cardiff

SO0505014970 to ST1908074870
PRN(s) 00898s
Mapped Figure 14
Status Proposed
Condition Unknown
The section of this route between Castell Collen and Pont Cwm-y-fedwen is not within the
boundaries of this study area, and the northern most section of the route within the study area is
totally conjectural. Both the Royal Commission and the Ordnance Survey state that there are no
visible remains north of SO099052 until past the Taf Fechan reservoir. Margary, who was
unaware of the fort at Penydarren, proposed a route running straight across Merthyr Common
and the northern portion of Gelligaer Common, between SO0609011480 and SO0955005730,
which is now known to be unlikely, particularly since recent survey work by GGAT on Merthyr
Common, although by no means exhaustive, failed to find any traces. 14 Most authorities now
postulate a route through Penydarren; however the rapid expansion of Meythr Tydfil in the 18th
and 19th centuries have disguised all evidence for the road and the route is purely conjectural.
The first physical evidence for this road occurs between SO099052 and SO099051 in the form
of a terrace. This line is continued by a cropmark which runs between SO099051 and
SO100050, and then by a hollow-way which runs between SO100050 and SO101048. These
three features are noted by the RCAHMW (1976), but no further information is given. A further
pair of features is also noted without further description: an agger/terrace running between
SO101037 and SO103032, and a terrace/hollow-way/agger which runs between SO103032 and
SO104030. The agger is described in more detail in the documentation supporting the
scheduling of the site (Cadw 2003): it takes the form of an agger some 4m wide and 0.5m in
height, flanked by ditches 2m wide and 0.5m deep, and provided with external counterscarp
banks 2m wide and 0.5m high. However the grid reference (SO10460308 to SO10150390)
differs from that recorded in 1994. The route is then supposed to follow the course of a
cropmark, which runs between SO104030 and SO105028, and then what is described as an
agger/terrace which runs between SO105029 and SO106027. There is however some
inconsistency with these grid references as the features run parallel to each other for a short
distance rather than leading one into the other.
From here the route follows a terrace 2.5m wide and then a bank which varies in width between
3-4m, with an average height of 0.3m. This bank has on both sides shallow side ditches, which
are on average 1m wide, and the bank runs between SO106027 and SO106026. The line of the
route is then lost for a short distance before continuing as a terrace and bank, which run between
SO108024 and SO108023. The terrace is 6m wide at its northern end, narrowing to 4m wide at
its southern end. The bank which runs along the terrace is almost 5m wide at its southern end,
narrowing as it runs south, flanked on both sides by shallow ditches 3m wide and 0.5m deep.
The route is again lost for a short distance before being visible as a section of agger running
between SO112010 and SO112009, which is described as a slight bank. No further physical
evidence is known for this road and the route is again lost before being assumed to merge with
the course of the modern road across Gelligaer Common at ST1173899981.
The route beyond this point is again totally conjectural being assumed to follow the course of
the modern Heol Adam into Gelligaer past the practice works and running to the west of the
forts. The route then turns south following the course of Church Road and Penybryn Terrace
before being lost at ST1344095590. The route is again lost between ST1325095231 and
ST1344395586, presumably has been destroyed by the spoil tip. Beyond this point, the route is
14
It is however possible that extensive peat fields may obscure any remains.
41
assumed to follow the course of a footpath, then the Heol Pont-y-Saison and a series of un-
named tracks and minor roads across Mynydd Eglwysilan before joining the course of the Heol
Las at ST1401089740. The more fragmented of thses have not be mapped as fossilised. Once
into Caerphilly the route is lost at approximately ST1481088060, being assumed to pass
somewhere to the southeast of the fort, being re-discovered at ST1553086870. From here it is
assumed to follow the route of the modern Castle Street, Cardiff road, Mountain Road and
Thornhill Road into Cardiff. The route is then assumed to follow Caerphilly Road and North
Road before turning onto the Kingsway and Working Street. From here it is assumed to run
down to the harbour following the course of The Hayes, Hayes Bridge and Bute Street.
Sources
Cadw 2003
Margary 1957, 68-9
RCAHMW 1976, 108-9
RCAHMW 1986, 163-4
RCAHMW 1994, 9

This road has been extensively studied over an extended period of time (most recently by
RCAHMW for the Glamorgan Inventory), and without any new information, further fieldwork
is not likely to be very profitable. Air photography is unlikely to provide further information
for much of the route if the road follows the postulated route along modern roads, although
deviations might be apparent as cropmarks under suitable conditions. Towards the northern
end of the route, more fieldwork on Merthyr Common in the area north of that examined
recently might identify any road bypassing the Penydarren fort. Also monitor any ground
disturbance along the route.
42
Figure 14.1: Roman road RR621 (Castell Collen - Cardiff) Taf Fechan reservoir - Gelligear
3
05 10 Legend
Short section
earthwork
Short section
air photograph
Short section
15 unknown
Short section
fossilised
Predicted
Fossilised
Earthwork
Air photograph
Unknown
Other military
10
Fort ?
Fort/Fortress
Setllement with
Penydarren Land over 244m

as thin black lines
05
2
00
1
95
Figure 14.2: Roman road RR621 (Castell Collen - Cardiff) Gelligear - Cardiff
3
15 20 Legend
Gelligaer
Short section
earthwork
95 Short section
air photograph
Short section
unknown
Short section
fossilised
Predicted
Fossilised
Earthwork
Air photograph
90 Unknown
Other military
Fort ?
Fort/Fortress
Caerphilly
Setllement with
Land over 244m

85

as thin black lines
80
Cardiff
1
75
RR622 Neath - Y Gaer, Aberyscir

SS7496098030 to SN8609510870
PRN(s) 01025.0w, 01025.1w, 03738m

Mapped Figures 15
Status Known
Condition Damaged
Route which runs between the forts of Neath and Y Gaer, conventionally known as Sarn Helen;
recorded by both Margary and the Royal Commission. Unlike the vast majority of roads in the
Glamorgan-Gwent area this route has a number of well defined physical features; however its
exact route out of Neath is not known and it is generally assumed to follow the course of the
modern A4230 as Cadoxton Road, Main Road, Church Road and Cadoxton Terrace, then at
SS7693499235 the route follows the A4109 through Aberdulais.
At SS7779999783 the Ordnance Survey suggests two different routes for the road. The more
obvious follows the course of the B4242 until SS7801799835 where it is assumed to run across
a series of fields to SN7843700060. Here this first route joins the course of an old lane, which
forms the boundary between two parishes and runs up a hill in what the Ordnance Survey
describes as being a ‘relatively easy ascent’. The second, less favoured, route leaves
SS7779999783 to run across a lightly wooded hillside to SS7779999783 along what the
Ordnance Survey describes as a slight and mutilated terrace which is considered to be ‘too steep
for vehicular traffic, general direction inconsistent with Roman line.’ Between SN7829200103
and SN7846200110 the course of this second route is confused by what the Ordnance Survey
describe as a series of ‘complicated terrace-ways and hollow-ways’, all running in slightly
different directions. 15 . These two possible routes merge at SN7846200111 to follow the course
of the old lane until SN7909800929.
Once the route reaches Lletty’r-afel-fach at SN7909800929, it becomes well defined. At

SN7913801004 the road is marked by a well-engineered terrace 7m wide. Between
SN7916101031 and SN8522108498 the route of the road lies on a modern trackway used by the
forestry commission and seasonally by recreational off-road vehicles, but despite this use
sections of agger can clearly be seen in places as follows:
• Between approximately SN7940501642 and SN7942501848; grass covered agger

averaging 6m wide and 0.4m high.
• Between approximately SN7945802169 and SN7948802231; grass-covered agger 6m wide

and 0.2m high; ditch 3m wide and 0.2m deep on the northwest side.
• Between SN7953902299 and SN7959802374; grass-covered agger 0.2m high and 6m wide;
ditch 3m wide and 0.2m deep on the west side.
At SN8024002916 the road enters the modern forest of Rheola and has been replaced by a
modern track for about 3km; at SN8123904009 the road passes to the northeast of Hirfynydd
signal station. At approximately SN8219005320 the track emerges from the forest back onto the
moorland to follow the forest’s western edge until approximately SN8497008520, and at
SN8284006630 the road passes to the southeast of a second signal station. As the road follows
the edge of the forest to the north its route is marked by the line of a modern wall, which
supposedly shows the extent of the road and its medieval successor. Along this stretch of road,
the agger can be seen in a number of places. This agger varies between 3m-11m in width and
15
None are noted by Margary or the Royal Commission, and the Ordnance Survey does not
know which of these features, if any, relate to the road.
45
between 0.2m-0.6m in height, but none of these streches have evidence for the accompanying
ditches. Centred at approximately SN8452107909 is a section of terracing which follows the
contours of a hillside with a 1 in 3 slope. This terrace is on average 4m wide and in places has
had large boulders placed along its outer edge for reinforcement.
Between SN8522608493 and SN8542328726 the course of the road has largely been obliterated
by open-cast coal mining; notes on the Ordnance Survey strip map state that although section of
the road can be seen ‘here and there’ it is difficult to trace its exact course. Between
SN8542328726 and SN8566209749 the route follows the line of the modern road through
Banwen. In 1997 Williams excavated a section of the road which ran between approximately
SN8566209749 and SN857009903. The road was constructed from a layer of medium sized
cobbles which lay directly on top of the natural clay. This layer of metalling extended c0.6m
beyond the eastern edge of the current road, and running parallel to the metalling was a clay cut
drainage ditch of unknown depth, c0.5m wide. From SN857009903 the route is generally
assumed to follow the course of the modern Camnant Road passing through the southern gate of
Coelbren fort before crossing the Nant y Bryn at SN8609510870 and passing out of the
boundaries of this study area.
Sources
CPAT 1998
Margary 1957, 70-1
Neath Antiquarian Society 1986
RCAHMW 1976, 109
RCAHMW 1994, 16-7
Williams 1997a

This road has been extensively studied over an extended period of time (most recently by
RCAHMW for the Glamorgan Inventory), and without any new information, further fieldwork
is not likely to be very profitable. Any ground disturbance along the route should be
monitored. Air photography is unlikely to provide further information for much of the route if
the road follows the postulated route along modern roads and tracks, although deviations might
be apparent as cropmarks under suitable conditions. However, the presence of extensive areas
of forestry along the much of the route makes air photography impractical except at the
southern end.
46
Figure 15: Roman road RR622 (Neath - Y Gaer, Aberyscir)
2
70 75 80 85 90 95 Legend
Coelbren Short section predicted
Predicted
10 Fossilised
Unknown
Buried
Fort ?
Fort/Fortress
Other military
05 Settlement with
Land over 244m

as thin black lines
2
00
Neath
RRX65 Bassaleg - Lower Machen-Caerphilly

ST274870 to ST153872
PRN(s) 00024g, 02892.0m, 07762g

Mapped Figure 16
Status Predicted
Condition Unknown
Route which runs from Bassaleg, through the metal working settlement at Lower Machen
(centred at approximately ST2289087780) connecting the road between Caerleon and Cardiff
(RR60b) to the fort at Caerphilly; proposed by Boon in 1972. There is no certain evidence for
this road. The Ordnance Survey road files accept as reasonably likely the section east of Lower
Machen, where there is a three-mile straight section of modern road (A468) in the valley bottom
which could conceivably eventually connect with RR60b (and Caerleon) at Basseleg, but
beyond Lower Machen they state that the terrain is difficult and no line can be established. The
line of the road has been taken from the 1st edn 6" OS map, which is rather less straight than the
present line. 16
In 1978 the Ordnance Survey noted the presence of linear cropmarks seen on aerial photographs
running to the south of the modern settlement of Lower Machen, at ST227877 (Record Card
1978, ST 28 NW). However there is little evidence to suggest that these represent the course of
a road, and they are slightly off the line of the A468. Capel and Marvell note an old trackway
running east-west between ST1797187513 and ST1797187513 through Gwern-y-Domen, which
they considered might follow in part the route of the Roman road.
Sources
Boon 1972, 16
Capel and Marvell 1991, 14
RCAHMW 1994, 46
Ordnance Survey road files.

Relatively little work has been done on this road, and therefore a full programme of map
regression, air photograph search and fieldwork is recommended. Of particular interest is the
linear cropmark south of Lower Machen; excavation here would establish whether this is
indeed the line of the Roman road, or whether it is more likely to lie on the line of the modern
road.
16
No information was available as to whether or not this road had been turnpiked.
48
Figure 16: Roman road RRX65 (Bassaleg - Lower Machen - Caerphilly)
3
15 20 25 Legend
Short section
air photograph
Short section
90 fossilised
Fossilised
Air photograph
Unknown
Predicted
Fort ?
Fort/Fortress
Caerphilly Other military
Lower Machen Settlement with

Urban characteristics
Land over 244m

1 as thin black lines
85
RRX80 Penhow - Magor Pill

ST4091088185 to ST4356486847
PRN(s) 06257g
Mapped Figure 17
Status Predicted
Condition Unknown
A section of medieval-modern road running from the higher ground near Caerwent and across
the Gwent Levels; proposed as possibly Roman in origin on the basis of straight alignments and
its apparent destination point
This route, connecting the Caerwent-Caerleon road (RR60a) to the supposed port site at Magor
Pill, was suggested by Rippon in 1996. The RCAHMW record it in their handlist, (as ST409926
- 434845), but it is not included by Margary or the Ordnance Survey. It is formed by what
Rippon describes as ‘a remarkable alignment of roads and field boundaries’, which leaves the
Caerwent-Caerleon road just to the west of Penhow to run in straight line down to the edge of
the Levels at Magor. Depending upon the post-Roman history of the Level at Magor, the section
to the south of the village may potentially also be Roman.
Sources
Nayling and McGrail 2004, 102-109
RCAHMW 1994, 51
Rippon 1996a, 32-33

follows the postulated route along modern roads. The most likely method of obtaining
additional information is through the monitoring of any ground disturbance along the route.
50
Figure 17: Roman road RRX80 (Penhow - Magor Pill)
3
35 40 45 50 Legend
Predicted
Unknown
Great Bulmore Fossilised
RR60a
Caerwent
Fort ?
Fort/Fortress
90 Other military
Settlement with
Sudbrook
Land over 244m

as thin black lines
1
85
RR GGAT 001 Wheel Lane

ST2826082492 to ST2922383212
PRN(s) 05199.0g
Mapped Figure 18
Status Predicted
Condition Unknown
Minor road, Roman date uncertain; not recorded by the Orndance Survey, Margary or the Royal
Commission. A section of medieval-modern road running across the Wentloog levels. This
route is represented by a short section of trackway running northeast-southwest across the
Wentlooge levels, which formed an important communication route in the Medieval period.
While there is no direct evidence that this route is Roman, Rippon argues that the northeast-
southwest boundaries on the Gwent levels are continuations of elements of the Roman
landscape and that therefore Wheel Lane reflects this earlier landscape.
Sources
Rippon 1996b, 54
Yates 1996

Given the ground conditions on the Levels, there would not appear to be any scope for
fieldwork or air photography. However, the Levels are being affected by development, and any
proposed development on the line of this road or its possible continuation at either end should
be the subject of predetermination archaeological assessment to determine whether physical
remains can be identified.
52
Figure 18: Roman road RR GGAT 001 (Wheel Lane)
3
20 25 30 Legend
Fossilised
Lower Machen
RR60b
(Caerleon - Cardiff)
Fort ?
Fort/Fortress
Other military
85 Settlement with
Land over 244m

as thin black lines
1
80
RR GGAT 002 LG Semicon Road

ST28308484 to ST28328478
PRN(s) 06279.0g
Mapped Figure 19
Status Known
Condition Damaged
Minor road, Roman date certain; not recorded by the Orndance Survey, Margary or the Royal
Commission. Route running in a south-southwestern direction towards the Wentlooge Levels,
though entirely on the solid geology of the fen-edge; partially excavated by Williams in 1997.
Almost 90m of metalled road were excavated across two trenches. The first of these sections
consisted of a surface constructed from small, mainly rounded stones, which had been laid
directly onto the underlying natural. Evidence of two probable wheel ruts where noted within
this metalled surface. In the second trench a similar metalled surface of small, rounded stones
was overlain by a band of larger cobbles. Both sections of road had a narrow ditch running
parallel to them on the northern side. The only finds excavated from both of these trenches were
of Roman date. This section of road has been interpreted as part of a lesser route running from
the main Caerleon-Cardiff road towards the settlement site excavated at Great Pencarn Farm
(Yates 2000).
Sources
Sell 1997,
Williams 1997b
Yates 2000

Given the ground conditions on the Levels, there would not appear to be any scope for
fieldwork or air photography. However, the Levels are being affected by development, and any
proposed development on the line of this road or its possible continuation at either end should
be the subject of predetermination archaeological assessment to determine whether physical
remains can be identified.
54
Figure 19: Roman road RR GGAT 002 (LG Semicon)
3
25 30 Legend
Buried
RR60b
Fort ?
Fort/Fortress
Other military
Settlement with
1 urban characteristics
85
RR GGAT 002
Villa/Small
settlement
Land over 244m

as thin black lines
Great Pencarn Farm
RR GGAT 003 Sor Brook Valley Road

ST3296092639 to ST3268696854
PRN(s) 05229.2g
Mapped Figure 20
Status Predicted
Condition Unknown
Possibly a miinor road, Roman date uncertain; not recorded by the Orndance Survey, Margary
or the Royal Commission. A section of medieval-modern road proposed as possibly Roman in
origin on the basis of straight alignments, and because the line runs directly towards Caerleon.
It is not clear, however, where the destination would have been at the northern end, unless it is
an alternative route to Abergavenny. There is no physical evidence There are two straight
sections with a common alignment, one running between Ponthir and Common Cefn-llwyn and
the other on the other side of the valley at the northern end of the mapped section near
Whitehall. If this is a Roman road, it is unclear what route the intervening section would have
taken across the valley to cross the brook possibly by a ford; it is possible that it may have
followed the course of the modern road, which winds its way by easy gradients across the
topography, particularly since the section on the north side of the valley forms part of the
boundary of Llandegfedd parish.
Source
Lawler and Page 1992

regression, air photograph search and fieldwork is recommended, to see whether it is indeed an
alternative route to Abergavenny.
56
Figure 20: Roman road RR GGAT 003 (Sor Brook Valley)
3
30 35 40 Legend
Fossilised
2 Usk Unknown
00
Fort ?
Fort/Fortress
Other military
Settlement with
urban charateristics
Land over 244m

as thin black lines
1
95
RR GGAT 004 Ffordd y Gyfraith

SS8721980212 to SS8619185992
PRN(s) 03112m
Mapped Figure 21
Status Predicted
Condition Unknown
Minor road, Roman date uncertain; not recorded by the Ordnance Survey or Margary. A section
of medieval-modern road running to and up Mynydd Baiden; proposed by the Royal
Commission as possibly Roman in origin on the basis of straight alignments and the observance
of Roman principles of setting out, though it has no obvious destination points in either
direction.
In the introduction to the section on Roman roads in the Glamorgan Inventory, RCAHMW
describes it thus: ‘The relevant part runs from just north of Laleston to Mynydd Baiden, where it
becomes a normal ridgeway. This length of 5.5 km, is set out essentially in three straight
sections, with angles on Cefn Cribwr and on a southern spur of Mynydd Baiden; it also uses
sharp zig-zags to descend steep slopes. These features are characteristic of known Roman
Roads, but it is difficult to see how this track could fit into the general network.’ Because of the
doubt as to its origin, the road was not given a separate number in the Inventory, neither was it
included in the RCAHMW handlist; as a result this route has neither a RCAHMW number nor a
Margary number.
Source
RCAHMW 1976, 104-6

additional information is through the monitoring of any ground disturbance along the route,
including also the ridgeway along Mynydd Margam which is the potential continuation of the
engineered section identified by RCAHMW.
58
Figure 21: Roman road RR GGAT 004 (Ffordd y Gyfraith)
2 3
85 90 95 00 05 Legend
Fossilised
85 RR60c
(Cardiff - Neath)
Fort ?
Fort/Fortress
Other military
Caergwanaf
Settlement with
) urban characteristics ?
Settlement with
80 urban characteristics
Land over 244m

Glanwenny House
as thin black lines
1
75
Cowbridge
RR GGAT 005 Heol Spencer

SS9116479622 to SS9506277039
PRN(s) 04087m
Mapped Figure 22
Status Predicted
Condition Medium
Minor road, Roman date uncertain. Not recorded by the Ordnance Survey, Margary or the
Royal Commission. A section of medieval-modern road running from Llangan, through Coity
and across Cefn Hirgoed towards the point at which the River Ogmore and its tributaries emerge
from the Coalfield escarpment; proposed as possibly Roman in origin on the basis of straight
alignment and its potential strategic importance in controlling a series of major early routeways
out of the upland. There is no physical evidence. Coity Castle stands either astride or
immediately alongside the line of this road. This castle is one of those of strategic importance
founded by the Normans as part of their conquest of Glamorgan; other castles of this type (eg
Caerphilly, Cardiff, Loughor) occupy sites which had previously been used for Roman forts,
and this could also be true of Coity, through no excavation has been carried out there to confirm
or deny the hypothesis. Heol Spencer, and its southerly continuation Heol Simonston, are of
remarkable straightness and appear on the earliest maps of the area; part of the boundary of the
historic parish of Coety runs along Heol Simonston.
Sources
Lawler and Andrews 1992
Evans 2002

additional information is through the monitoring of any ground disturbance along the route.
The identification of Heol Spencer as being Roman in origin is partly dependent upon the
hypothesis that it served a fort controlling the mouth of the Llynfi, Garw and Ogmore valleys.
Investigation of the area of Coity Castle might establish whether this is the site of such a fort,
particularly the linear feature in the field to the north of the church which has previously been
interpreted as a siegework of early 15th century date.
60
Figure 22: Roman road RR GGAT 005 (Heol Spencer)
2 3
90 95 00 Legend
85
Coity Castle
Fossilised
Predicted
RR60c
(Cardiff - Neath)
Fort ?
Fort/Fortress
Other military
) Settlement with
urban characteristics ?
1
Setllement with
80
Land over 244m

as thin black lines
Glanwenny House
RR GGAT 006 West Aberthaw - Llanharry

ST0261966904 to SS9905078975
PRN(s) 01037.0s
Mapped Figure 23
Status Predicted
Condition Near destroyed
Minor road, Roman date uncertain; not recorded by the Ordnance Survey, Margary or the Royal
Commission. Route running inland from the mouth of the River Thaw to the settlement at
Cowbridge and beyond, roughly following the course of the river. It is not clear where the
destination would have been at the northern end. Proposed as possibly Roman in origin on the
basis of a number of straight alignments to the north and south of Cowbridge, and the fact that
significant proportion of excavated structures in Cowbridge (including the bath-house, which
was probably an official building) follow this alignment rather than that of the Cardiff - Neath
road (RR60c), along which the Roman settlement at Cowbridge was built. The mouth of the
Thaw was a haven for shipping in post-medieval times; there is evidence for settlement at East
Aberthaw including 4th century coinage not normally associated with simple rural sites.
However, the proposed route at this point runs up the western side of the river, following the
existing roads, though there is as yet no evidence for any other Roman activity here During
limited fieldwork carried out in 1985, a V-shaped ditch was noted in the section of the cutting of
the railway line from Barry to Llantwit Major at the point where it crosses the predicted route.
North of this point, the modern road between ST002704 and ST009695 forms part of the
boundary between the parishes of Eglwys Brewis and Flemingston. Other features noted were a
number of hollow-ways following hedge lines and a stony bank on the right alignment which
could form an old hedge bank. There is also the placename Pen-yr-heol (top of the road),
although heol names are not necessarily an indicator that a road is great age. Around
Llandough-juxta-Cowbridge (NGR SS995729), where the terrain is difficult, the line is lost. At
Llansannor, the church, which is a possible Early Medieval foundation, sits astride the line of
the road, which is diverted round it at SS99377751.
Source
Evans 1985

regression, air photograph search and fieldwork is recommended.
62
Figure 23: Roman road RR GGAT 006 (West Aberthaw - Llanharry)
80
2 3
95 00 05 10 Legend
Short section fossilised
Llansannor Church
Fossilised
Unknown
Cowbridge
RR60c
(Cardiff - Neath)
75
Fort ?
Fort/Fortress
Other military
Settlement with
Land over 244m

1
70
as thin black lines
RR GGAT 007 Road on east bank of River Neath

SS7723594923 to SS7819097569 17
PRN(s) None yet allocated (Sites 3 and 8 in Network Archaeologies report no: 303)
Mapped Figure 24
Status Proposed
Condition Near destroyed
Possible conquest-period road, but identification and Roman date uncertain; not recorded by the
Ordnance Survey, Margary or the Royal Commission. Proposed as Roman on the grounds that it
lies on a possible route between the fort at Neath and the Blaen-Cwmbach marching camp and
between Neath and the fort at Penydarren. The road as excavated is 11m wide across the
flanking ditches; deposits 0.88m thick were noted in both excavated sections, though neither
had any metalling, and there was no dating evidence. However, the line of this proposed road
can be made out as an earthwork on aerial photographs at approximately SS7819097569. It
should however be noted that during the Second World War a decoy site was constructed in this
area to lure enemy aircraft away from the petrol refinery at Jesery Marine, and this involved the
building of numerous unmapped roads in the vicinity, now grassed over (R Lewis pers comm).
Sources
Network 2004.

regression, air photograph search and fieldwork is recommended.
17
The second of the above NGRs is only approximate, as the information provided to the SMR
was inaccurate. The correct NGR was reconstructed by comparing, Network plans, air
photographs (taken as part of GGAT61 and not yet catalogued) and modern Ordnance Survey
maps.
64
Figure 24: Roman road RR GGAT 007 (Road on east bank of River Neath)
2
75 80 Legend
Blaen-Cwmbach Buried
Marching camp
Air photograph
RR60c
(Cardiff - Neath)
Neath
Fort ?
RR GGAT 007
Fort/Fortress
Other military
Settlement with
Land over 244m

as thin black lines
1
95
RR GGAT 007
RR GGAT 008 Road on Stormy Down/Newton Down

SS845804 to SS835790
PRN(s) 00333.0m, 02494m

Mapped Figure 25
Status Predicted
Condition Unknown
Minor road, identification and Roman date uncertain; not recorded by the Ordnance Survey,
Margary or the Royal Commission. Proposed on the basis of alignments of air photograph and
earthwork features. Possible route of road running northeast-southwest across Stormy Down
and Newton Down between the Cardiff - Neath road (RR60c) and the coast at Porthcawl, where
a haven for shipping is recorded from at least the early post-Medieval period, and there is a
known villa, at Danygraig (Higgins 1968, 20-1, 27, 32, 136).
Two parallel cropmarks, approximately 20m apart, were noted on aerial photographs by Trott at
SS840797, and confirmed on the ground. They were aligned at about 215 degrees to grid north,
pointing northeastwards at the proposed sighting point for the RR60c on the crest of Stormy
Down, and southwestwards towards a small valley on the same alignment at SS835790, which
descends a steep escarpment on Newton Down and is so straight as to suggest it is artificial.
Source
Trott 1976a

regression, air photograph search and fieldwork is recommended. Fieldwork should include a
thorough examination of the ‘small valley’ on Newton Down, suggested as an artificial cutting,
66
Figure 25: Roman road RR GGAT 008 (Road on Stormy Down/Newton Down)
2
80 85 90 Legend
Possible sighting
point
Porthcawl haven
Air photograph
Kenfig
Earthwork
RR60c
(Caridff - Neath)
Fort ?
1 Fort/Fortress
80
Other military
Settlement with
Danygraig Cae Summerhouse
Villa/Small settlement
Glanwenny House
Land over 244m

as thin black lines
RR GGAT 009 Road at Rumney

ST2127944 to ST21037893
PRN(s) 00816.0s
Mapped Figure 26
Status Predicted
Condition Unknown
Minor road, Roman date uncertain; not recorded in the Ordnance Survey road files, or by
Margary or the Royal Commission. This road was reported at the beginning of the 20th century
by Ward (1908) as being ‘a causeway buried some 4-5 ft (1.2-1.5m) below ground level’,
frequently encountered during ground disturbance; minutes of evidence presented to the Royal
Commission note it as not relating to any known road system and extending in a southerly
direction from Pen yr heol in Rumney village towards the River Rumney, where ‘it came to the
back of a public house known as the Rumney Castle’. A route from Pen-yr-heol to the Rumney
Castle public house leads away from the river, and it is more likely that the correct Rumney
Castle is the site of the early earthwork castle situated on the river to defend the crossing
(Lightfoot 1992, 96-7), presumably on an existing routeway. The only discrepancy is that Ward
describes the causeway as being ‘half a mile below Rumney Bridge’, whereas the 1st edition 6"
Ordnance Survey map shows it as being above Rumney Bridge. This map also marks the
modern road running in the same direction but slightly further east as ‘Roman road’, whereas it
is clear that this cannot be cannot be associated with the buried causeway. For the purposes of
mapping, the first part of the line immediately south of Pen-yr-heol has been projected down a
hedgeline aligned in roughly the right direction.
Sources
Ward 1908
OS record card ST 27 NW 9, quoting Minutes of Evidence to the Royal Commission on
Ancient Monuments in Wales and Monmouth (1912) vol 1, no.2001

No fieldwork is possible, since this area is now built up, but a watching brief condition should
be put on any developments along the proposed line of the ‘causeway’ and on any likely line in
the section immediately south of Pen-yr-heol, as well as any possible line immediately north of
Pen-yr-heol.
68
Figure 26: Roman road RR GGAT 009 (Road at Rummney)
3
20 25 Legend
Predicted
Fossilised
RR60b
Fort ?
Fort/Fortress
1
80
Other military
RR GGAT 009
Settlement with
Land over 244m

as thin black lines
Cardiff
RR GGAT 010 Road at Llandaff

ST13717881 to ST14147860
PRN(s) 00935.0s
Mapped Figure 27
Status Predicted
Condition Unknown
Road, Roman date uncertain; not recorded by the Ordnance Survey, Margary or the Royal
Commission. Two areas of possible road metalling noted in section, one in a north-south facing
section at ST14147860, during works to a footpath in Llandaff, and the other at ST13717881
during the widening of Waterhall Road. The metalling was 8m wide and 0.5m in thickness with
a convex surface. Trott believed them to be parts of the same Roman road, although there is no
evidence apart from general similarity to prove they are that parts of the same road, or that they
are of Roman date.
This has always been assumed to be a minor road, but may possibly represent the road between
the forts at Cardiff and Caergwanaf (Miskin).
Sources
Trott 1981
Trott 1982

The two sections of metalling noted lie on either side of two properties with large grounds
largely unencumbered by buildings. These should be assessed to see whether geophysical
survey is feasible.
Map regression should be carried out to determine whether this postulated road is likely to run
between the forts at Cardiff and Caergwanaf, followed by fieldwork and possibly additional air
photography in the area to the west where there has been no development.
70
Figure 27: Roman road RR GGAT 010 (Road at Llandaff)
3
15 20 Legend
Buried
RR60c
(Cardiff - Neath)
1 Fort ?
80
Fort/Fortress
RR GGAT 010
Other military
Settlement wih
Land over 244m

as thin black lines
Cardiff
RR GGAT 011 Possible road at Pantllaca

ST202817 to ST20228170
PRN(s) 00812.0s
Mapped Figure 28
Status Predicted
Condition Unknown
Minor road, Roman date uncertain; not recorded by the Ordnance Survey, Margary or the Royal
Commission. Two sections of metalling noted close to the site of a Roman kiln and two coin
hoards (all located around ST199813). The metalling at ST202817 was some 120ft (36m) in
length, and was 12ft (3.5m) wide, cambered and composed of small stones. The other section of
metalling, at ST20228170, was 200ft (300m) in length, straight, and descended gently to a
stream. Trott believed them to be parts of the same Roman road, although there is no evidence
apart from general similarity to prove they are that parts of the same road, or of Roman date.
Source
Trott 1976b

No fieldwork is possible between the two sightings and to the east, since this area is now built
up, but a watching brief condition should be put on any developments along the proposed line.
To the west is Llanishen golf course. This should be examined to see whether there is likely to
be any survival of road metalling or ditches below any landscaping, and if it seems possible, a
programme of geophysical investigation could be devised to try and locate it.
72
Figure 28: Roman road RR GGAT 011 (Possible road at Pantllaca)
3
15 20 Legend
Buried
RR621
(Castell Collen - Cardiff)
Fort ?
Fort/Fortress
Other military
RR GGAT 011 Settlement with

urban characacteristics
Land over 244m

as thin black lines
Treoda Enclosure
1
80
Appendix 1: Roads in OS files for which there is insufficient evidence

to map
RR630 Northwards from Abergavenny (SO299140)
Margary (1957, 75) suggests a line to the east of the Skirrid, then running over Campston Hill to
Llangua bridge. The OS file contains a map making the the present main road up the east bank
of the River Monnow Llanfihngel Crucorney – Pandy – Llangua Bridge. Neither has any
certain evidence, and the alignments are not particularly compelling.
RCAHMW 1994, 21
RRX11 Bryn – Tredegar (ST169946 - SO1441011)

A ridgeway route, a ‘causeway over Mynydd Bedwellty’ suggested by Coxe (1801, 261) and
Wakeman. Rejected by OS fieldworkers as exhibiting no signs of Roman origin, and more
likely to be medieval.
RCAHMW 1994, 41-2
RRX42 Monmouth - Abergavenny (SO441097 - SO362088)

Jingle Street – Dingestow – Wern-y-cwrt – Clythan Park. Suggested by Wakeman; entirely
conjectural. Rejected by OS as excessively winding, with no traces in fields at minor bends
where they could be continuing line, no reported Roman finds from the area, no parish
boundaries, and no other traces of antiquity.
RCAHMW 1994, 42
RR43ii Mamhilad – Gelligaer (SO306035 - SO219020)

Section of ridgeway Llanhilleth – Abersychan –Mamhilad is an antiquarian proposal (EH 1872),
followed by Bradney (1906, 463). Examined by OS fieldworkers who rejected it as being too
angular and meandering, and having no traces of Roman construction.
RCAHMW 1994, 42
RR43ii (PRN 05625.0g) Pontypool – Aberdare (SO 274017 - ST180953)

Pentwyn-mawr – Pontypool – Newbridge – Pontllanffraith – Nelson - Aberdare – Neath valley
suggested by Savory (1954) on the basis of a handful of finds, not all of which are necessarily of
Roman date. No evidence for Roman origins, and no possible course beyond Pontllanffraith.
RCAHMW 1994, 43
RRX44 Tintern – Usk (SO530001 - SO379007)

Brockweir – Angidy valley – Tintern Cross – Trellech Grange – Llanishen - Usk. Suggested by
Wood (1922). Rejected by OS fieldworkers who preferred an alternative route from Usk to
Lower Maerdy (SO40030126) and then on to Monmouth, but were not able to provide any firm
evidence. ‘Modern road on traces of paving’ was noted connected with a slight hollow-way at
SO480027 and SO404025, but there were no traces of any certainly Roman remains, and that
there did not appear to be any reason for this route.
RCAHMW 1994, 43
RRX72 Caerwent – Sea Mills (ST468905 - ST550750)

A series of conjectural passages across the Bristol Channel from Caerwent. These are based on
the premise that the coastal conditions in Roman times were identical to their present state. One
of the suggested road lines is Crick – Portskewett – Sudbrook; another follows the driveway
through St Pierre Park. The others appear to be entirely conjectural.
RCAHMW 1994, 46
74
Appendix 2: PRNs for isolated sightings of possible Roman road, and

rejected sites
02280g Rumble Street (SO33510281 - SO33530373)
Proposed by Bradney (1921, 71) on the basis of the name. OS fieldworkers examined the route
in 1957 and described it as ‘a narrow modern metalled road which is partly sunken. It winds
unnecessarily and follows no significant route’ (OS record card SO 30 SW 27). There appears
to be no reason to regard it as Roman. Rejected
02510g Site name not known (SO3220)

Proposed on the basis of a statement by Coxe (1801, 18): ‘In the course of my tour I observed
vestiges of several [roads] which appeared to be of Roman origin particularly in the
neighbourhood of Llanvihangel Crickhornell, where the names of Lower and Upper Stanton
seem to indicate the site of a Roman road, and where the remains of numerous entrenchments,
which have served as exploratory camps, are still manifest.’ OS fieldworkers examined the area
in 1957 and were unable to find any traces of Roman roads (OS record card SO 32 SW 19).
Rejected.
02986g Site name not known (SO21170148 - SO22530210)

This site is both paved and on a ridgeway and has therefore been discarded as a Roman road.
Rejected.
06258g Bishton Church West (ST38488729)

Proposed in an annotation on a map held by NMGW: Metalled road surface containing
limestone blocks found at depth of 3ft while making new approach road. Fieldwork carried out
as part of the Romano-British Lowland Settlement Survey indicated that the earthworks in this
field were too complex and extensive to suggest that this was a road; it is more probably part of
a settlement site. Rejected.
05895.0g Site name not known (ST3362090270 - ST3342090470)

While the electronic record lists an eight-figure NGR for both the start and finish of this
proposed road, the electronic record contains no description, route details or source. The further
information file for this PRN is also missing. It is therefore impossible to check the accuracy of
the report and it has not been mapped.
07643g Site name not known (ST33769064)

The electronic record for this site lists a single eight-figure NGR, the description given is of a
trial trench revealing, a well-made Roman road surface with some Medieval intrusion. The
further information file for this PRN is missing, therefore with no description, no reliable dating
evidence, no source material and only a single NGR it has been decided not to map this feature.
07779g Royal Road Hollow-way (SO5304014400 to SO5295013610)

A deep hollow-way recorded as a Roman road on the 1st edition OS map. Fieldwork by Clwyd-
Powys Archaeological Trust failed to find any supporting evidence, other than the fact that there
is a villa a short distance to the north (Thomas 1996). Rejected.
02680.0s Cliffwood Cottage (ST08656695)

A well-made and finely metalled road some 4-6m wide, identified in 1992. This appears to
continue the line of the medieval road leading past Cliffwood Cottage. Rejected.
75
01040.0w Mynydd y Gwair (Tor Clawdd) SN6607

A short section of possible roadway visible as an earthwork flanked by ditches running
northwest-southeast across moorland. Usually regarded as part of the Tor Clawdd dyke system,
but potentially part of a road between the forts at Neath and Llandeilo. Not mapped as none of
the rest of a potential route has been established.
76
Bibliography
Published and printed
Armour-Chelu, RJ and Williams DN, 1997, LG Electronics (Phase 1), Stage 4: Watching brief,
GGAT report no.97/023
Boon, GC, 1972, Isca, The Roman legionary fortress at Caerleon, Monmouthshire, National
Museum of Wales
Bradney, J, 1906, A history of Monmouthshire vol 1 part 2b The Hundred of Abergavenny (Part
2)
Bradney, J, 1907, A history of Monmouthshire vol 2 part 1 The Hundred of Raglan
Bradney, J, 1921, A history of Monmouthshire, vol 3 part 1, The Hundred of Usk (Part 1)
Bradney, J, 1933, A history of Monmouthshire, vol 4 part 1, The Hundred of Caldicot (Part 1)
Capel, R and Marvell, AG, 1991, Gwern-y-Domen Farm, Rudry: Archaeological study. GGAT
unpublished report
Courtney, P, 1983, The rural landscape of eastern and lower Gwent AD 1070-1750.
Unpublished thesis, University of Wales
Cox, W, 1801, An historical tour of Monmouthshire
CPAT, 1998, Forest enterprise project phase II
Davies, H, 2002, Roads from Rome, British Archaeology, 67, 8-13
EH, 1872, Notes of an antiquarian ramble among the Monmouthshire hills, Archaeol
Cambrensis 4 ser 3, 153-8
Evans, EM and Maynard, DJ, 1997, Caerleon Lodge Hill cemetery: the Abbeyfield Site 1992,
Britannia 28, 169-243
Evans, EM, 2000, The Caerleon Canabae, Excavations in the Civil Settlement 1984-90,
Britannia Monograph Series no.16
Evans, EM, 2001, Romano-British south east Wales settlement survey: final report, GGAT
unpublished report no.2001/023
Evans, EM, 2002, Coity Link Road, Bridgend, archaeological desk-based assessment, GGAT
report no.2002/063
Evans, EM, 2003, Early medieval ecclesiastical sites in southeast Wales: Desk-based
assessment. GGAT unpublished report no.2003/030
Filmer-Sankey, WP, and Collcutt, SN, 1991, Fairwater Farm, Michaelston-y-Fedw, Gwent
preliminary desk-top study, Oxford Archaeological Associates unpublished report
GGAT, 1990, Watching brief on the Margam-Pyle mains duplicate sewer, GGAT unpublished
report no.174
GGAT, 1992, Summary of archaeological resource Celtic Manor Resource, Coldra Wood,
Gwent, GGAT unpublished report project no.A066
Harris, W, 1773, Observations on the Julia Strata, and on the Roman stations, forts, and camps
in the counties of Monmouth, Brecknock, Carmarthen and Glamorgan, Archaeologia 2, 1-24
Higgins, LS, 1968, Newton, Nottage and Porthcawl: From prehistoric times to 1950
Jones, J and Mattingly, D, 1990, An atlas of Roman Brittain, 103, 176
Lawler, M and Andrews, D, 1992, Waunbant-Cefn Hirgoed trunk main, Mid Glamorgan,
archaeological desk-based assessment, GGAT unpublished report, project no.A090
Lawler, M and Page N, 1992, Sluvad Farm to Court Farm pipeline, Archaeological desk-based
assessment, GGAT unpublished report, project no.A085
Lewis R and Dunning R, 2003, Sunnybank Farm, Aberdulais, Neath, GGAT unpublished report
no.2004/032
Lightfoot, KWB, 1992, Rumney Castle, a ringwork and manorial centre in South Glamorgan,
Medieval Archaeol 36, 96-163
Locock, M, 1992a, Archaeological desk top study: Celtic Manor Hotel, Coldra Woods, Gwent,
GGAT unpublished report, project no. A066
Locock, M, 1992b, Archaeological field evaluation: Coldra Woods, Christchurch, Gwent,
GGAT unpublished report, project no. A074
77
Locock, M, 1994, Archaeological desk-based assessment, Pentrepoeth (A468 Bassaleg By-

Pass), GGAT unpublished report no.94/007
Locock, M, 1999, Baglan Bay gas pipeline, Neath and Port Talbot archaeological assessment,
GGAT unpublished report no.99/048
Locock, M, 1999, Monmouthshire historic settlements: an archaeological survey, Part 3: The
Hundreds of Caldicot and Raglan, GGAT unpublished report no.99/010
Manning, WH, 1981, Report on the excavations at Usk 1965-1976 (The fortress excavations
1968-1971), University of Wales Press, Cardiff
Manning, WH, 1989, Report on the excavations at Usk 1965-1976 (The fortress excavations
1972-1974 and Minor excavations on the Fortress and Flavian Fort), University of Wales
Press, Cardiff
Margary, ID, 1955, Roman roads in Britain, vol 1, London
Margary, ID, 1957, Roman roads in Britain, vol 2, London
Marvell, AG, 1993, Archaeological desk-based assessment Port Talbot-Margam duplicate
watermain, GGAT unpublished report no.93/008
Marvell, AG, 2001, Investigation along Monnow Street, Monmouth, BAR 320, 7-9, 40-45
Marvell, AG and Owen-John HS, 1997, Leucarum, Excavations at the Roman auxiliary fort at
Loughor, West Glamorgan 1982-84 and 1987-88, Britannia Monograph No.12
Marvell, AG and Page, NA, 1994, Port Talbot – Margam Duplicate Main: phase 1
archaeological watching brief, GGAT unpublished report no.94/061
Marvell, AG and Maynard, DJ, 1998, Excavations south of the legionary fortress at Usk,
Gwent, 1994, Britannia 20, 247-67
Maylan, CN, 1991, Proposed Usk Barrage, initial archaeological assessment, GGAT
unpublished report
Maynard, D, 1988, Bulmore, Archaeol Wales 28, 62
Maynard, D, 1993a, Celtic Manor Golf-Course, Christchurch (ST350 901), Archaeol Wales, 33,
61
Maynard, D, 1993b, Archaeological field evaluation, Usk sewerage pumping main, Usk, Gwent,
Maynard, D, 1996a, Archaeological field evaluation, Celtic Manor Golf Course No. 3 near
Caerleon, Gwent, GGAT unpublished report no.96/016
Maynard, D, 1996b, First interim report, archaeological watching brief, Celtic Manor Golf
Course No. 3, near Caerleon, GGAT unpublished report no.96/058
Maynard, D, and Marvell, AG, 1995, Usk Rising Main archaeological excavations, GGAT
Mein, G, 1977, Hadnock, Archaeol Wales, 17, 36
Mein, G. 1990, Chepstow Hill, Cat’s Ash, Christchurch, Archaeol Wales, 30, 57
Nash-Williams, V E, 1969, The Roman frontier in Wales (2nd edn revised M G Jarrett)
Nayling, N and McGrail S, 2004, The Barlands Farm Romano-Celtic Boat, CBA Research
Report 138, York
Neath Antiquarian Society, 1986, Newsletter no.1
Neath Antiquarian Society, 1988, Newsletter no.2,
Network Archaeology, 2004, Baglan Bay Gas Pipeline, archaeological watching brief,
Network Archaeology unpublished report no.303
O’Dwyer, S, 1937, The Roman roads of Wales, A historical survey, vol VI Brecknock and
Glamorgan, Newton
Ordnance Survey, 2001, Map of Roman Britain, 5th edition, Southampton: Ordnance Survey
Owen-Jown, HS, 1981, A possible stone cross base and Roman road at Margam Park, GGAT
Ann Rep 1980-81, 26-27
Pearson, AF, 2002, Roman roads and vici in Southeast Wales, GGAT unpublished report
no.2002/061
Parkhouse, J and Evans, EM, 1996, Excavations at Cowbridge, South Glamorgan, 1977-88
BAR Brit Ser 245
Pearson, AF, 2003, Roman roads and vici on Southeast Wales: Year 2 report, GGAT
78
RCAHMW, 1976, An inventory of the ancient monuments in Glamorgan, vol 1, part II, The
Iron Age and the Roman occupation, Aberystwyth
RCAHMW, 1986, An inventory of the ancient monuments in Brecknock, The Prehistoric and
Roman monuments, part ii: Hill-forts and Roman remains, Aberystwyth
RCAHMW, 1994, Roman Roads in Wales, A handlist based upon the archive holdings of the
RCAHMW, RCAHMW unpublished report
Rippon, S, 1996a, Gwent Levels: The evolution of a wetland landscape, CBA Research Report
105, York
Rippon, S, 1996b, The Gwent Levels historic landscape survey, Cadw, Cardiff
Rivet, ALF and Smith, C, 1981, The place-names of Roman Britain, London
Savory, HN, 1954, Bronze figurine from Abercarn, Bull Board Celtic Stud 15, 72-3
Sell, SH, 1997, LG Europark Development: Archaeological scheme of investigation, Stage 4
watching brief works, LG Semicon Site: Roman Road, GGAT unpublished report no.97/088
Sell, SH, 2001, Croesheolydd Farm, Bassaleg, Monmouthshire: archaeological watching brief,
Silvester, R and Owen, W, 2003, Roman roads in Mid and North-East Wales: The first report.
CPAT unpublished report no.527
Thomas, D, 1996, Priory Grove, Monmouthshire, Archaeological Assessment, CPAT
unpublished report no.211
Trott, B, 1976a, Stormy Down, Archaeol Wales, 16, 34-35
Trott, B, 1976b, Pantllaca (Cardiff), Archaeol Wales, 16, 47
Trott, B, 1978, Michaelstone-y-Vedw, in Britannia 9, 409
Trott, B, 1981, The footpath between ST14187861 and 14037856 at Llandaff, Archaeol Wales,
21, 83
Trott, B, 1982, Waterhill Road, Llandaff, Cardiff, Archaeol Wales, 22, 44
Tuck, M, 1999, Abernant, Kemey’s Inferior (ST36999179), Archaeol Wales, 39, 100
Tuck, M, forthcoming, Abernant, Kemey’s Inferior, Roman cemetery, Archaeol Wales 44
Ward, J, 1908. Pen yr Heol, Archaeol Cambrensis, vol IIX, 43
Williams, DN, 1995, Archaeological field evaluation, Land at Heatherington Lodge,
Michaelston-y-Fedw, GGAT unpublished report, project no.A250
Williams, DN, 1997a, Archaeological watching brief Roman Road, Banwen, West Glamorgan,
Williams DN, 1997b, Archaeological field evaluation (Stage 3), LG Semicon – Roman Road,
Wood, JG, 1922, The island chapel of St Twrog in Severn and the manors of Tintern Parva and
Trellech
Yates, AM, 1996, Archaeological desk-based assessment Western Valley Sewer, St Brides,
Wentlooge, GGAT unpublished report no.96/067
Yates, AM, 1997, Celtic Manor Golf Course No. 3, Catsash Roman Road excavation, Cat’s
Ash, Monmouthshire, GGAT unpublished report no.97/038
Yates, AM, 1999, WRU Centre of Excellence, Bulmore, Caerleon, Newport, Archaeological
assessment, Stage 3: field evaluation, GGAT unpublished report no. 99/077
Yates, AM, 2000, Excavation of a Roman site south of Great Pencarn Farm 1997: Coedkernew,
Newport, Studia Celtica 34, 49-80
Zienkiewicz, JD, 1984, Excavations at Caerleon and Great Bulmore, GGAT Ann Rep, part 2,
1983-84, 2-30
Miscellaneous notes held in regional SMR

Cadw, 2003, Scheduling document for monument GMM556(MER), GGAT PRN 00898s
Copeland, T, 1980, Manuscript in the GGAT SMR
Evans, EM, 1985, Document held in GGAT SMR describing route of West Aberthaw -
Llanharry Roman Road, GGAT PRN: 01037.0s
GGAT, 1991, electronic record for PRN 04457g
GGAT, 1998a, electronic record for PRN 03020s
79
GGAT, 1998b, electronic record for PRN 02892.0m

Maynard, D, 1995, Document held in GGAT SMR describing route of the Monmouth to Usk
Roman Road, GGAT PRN: 4772.0g
Maynard, D, 1996c, information in GGAT SMR, PRN 01016.1g
Richards, JI, 1985, Unpublished notes for desktop assessment carried out around Coldra Wood
Sorrell J, 2004, Aerial archaeological photographs of Gwent (folder of annotated colour prints)
80

Roman Roads in Southeast Wales Year 3

Uploaded by

Copyright:

Available Formats

Roman Roads in Southeast Wales Year 3

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Roman Roads in Southeast Wales Year 3

Uploaded by

Copyright:

Available Formats

What is the purpose of the report?

What is the purpose of the report?

What methodology was used?

What methodology was used?

Roman roads in Southeast Wales

Desk-based assessment with recommendations

A report for Cadw

Figure 9: RR60d Neath - Loughor -Pontardulais -Carmarthen 1 28

The Roman background

The numbering system follows that established by Margary and RCAHMW.

To these have been added a further class:

The following terminology is used:

ii) Proposed. Conjectural sections either linking known segments or as hypothetical

Survival (used for mapping)

Usk Unitary authority boundaries shown

Lower Machen Caerwent Land over 244m

Unitary authority boundaries shown

Conclusions and recommendations

Roads for which a full study is recommended

For details of proposals, see individual entries

Roads for which more limited fieldwork is recommended

For details of proposals, see individual entries

For details of proposals, see individual entries

Unitary authority boundaries shown

RR GGAT 011 RRX65

RR GGAT 006 RR GGAT 001

PRN(s) 02460g, 06213g, 06214g

Ordnance Survey strip maps.

Recommendations for further work

Land over 244m

Unitary authority boundaries shown

PRN(s) 00561g, 01016.1g, 01160g, 05316.0g, 05319.0g, 07048g, 07049g, 01016.17w

Ordnance Survey strip maps

Recommendations for further work

Short section rejected

Caerleon Other millitary

Unitary authority boundaries shown

RR60aa Crick - Sudbrook

Ordnance Survey strip maps.

Recommendations for further work

Land over 244m

RR60b Caerleon - Cardiff

PRN(s) 01016.2w, 01016.13w, 01016.16w, 03020s

Ordnance Survey strip maps.

Recommendations for further work

Unitary authority boundaries shown

RR60c Cardiff - Neath

PRN(s) 01016.14w, 02738w

Ordnance Survey strip maps.

Recommendations for further work

Neath Possible sighting point

80 Land over 244m

Unitary authority boundaries shown

RR60d Neath - Loughor -Pontardulais -Carmarthen 9

PRN(s) 01016.0w, 01040.0w, 05306w, 05311w

Ordnance Survey strip maps.

Recommendations for further work

Land over 244m

Unitary authority boundaries shown

RR 612a Ariconium (Weston under Penyard) - Monmouth