Except where stated otherwise (e.g. when the copyright of photographs is retained), you are free to reproduce any of this work for non-commercial purposes only, provided proper and appropriate attribution, credit and citation is given to the author, any original contributor or source, and the RRRA, in accordance with the terms of this licence.

Creative Commons Licence
This work is licensed under a Creative Commons Attribution-NonCommercial 2.0 UK: England & Wales License

© Mike Haken & RRRA, 2018

Home Gazetteer of Roads Margary's Numbering Itineraries & Sources Glossary/Biography RRRA Website

Margary Number:

Other Numbering System:


45.5 Miles





Grimescar Tilery


HE Pastscape Mon.1326352

Lancashire, West Riding of Yorkshire


Whilst we know the course of at least five Roman roads leaving what is now the city of Manchester, the precise layout close to the Roman fort remains a bit of a guess, the evidence lying beneath the modern city, if it survives at all. David Ratledge’s illustration (fig.1) gives a good indication. RR712, in effect the main road to York, leaves Manchester along an alignment that starts at Gorse Hill in Stretford, near Old Trafford Cricket ground and some two miles south west of the fort, and runs to High Moor, ENE of Oldham. The fact that Gorse hill is close to the road from Chester, RR7a, might at first seem to suggest that the original road planning envisaged a single road linking the legionary fortresses of Chester and York, however it seems almost certain that this alignment was set out from east to west,ie from High Moor to Gorse Hill. The position of the fort, and the fact that the principal road through the vicus is clearly what we know as Deansgate, strongly suggest that the road to Manchester from Chester, and the road north to Carlisle, were both planned before the alignment underlying the road from York was laid out. Whether Ivan Margary was right to consider the Chester and Carlisle roads as part of the same route, RR7a and RR7b respectively, remains debateable.

There are no clear accounts of the road’s course near either the fort or the R. Medlock: even by the mid 18th century, the road was rapidly being buried beneath the developing city of Manchester, although at first the destruction was not quite as you might expect:

Arrowsmith, P., Burke, T., Redhead., N. (1996); Castleshaw and Piethorne - North West Water Landholding : Archaeological Survey; University of Manchester Archaeological Unit ; available at http://www.castleshawarchaeology.co.uk/documents/CASTLESHAW_AND_PIETHORNE-NORTH_WEST_WATER_LANDHOLDING_Nov_1996.pdf accessed, 1/1/18

Bidwell, P. & Hodgson, N. (2009); The Roman Army in Northern England ; The Arbeia Society

Brook, N. (2018); Roman Road Report, Upper Fell Greave; unpublished

Crump, W. B. (1925); Ancient Highways - Early Maps and Road Surveys, Trans. Halifax Antiquarian Society, Vol. 22 (1925) pp 189-256

Faull, M. L. (1981); The Roman Period in Faull, M. L.  & Moorhouse, S. A. (Eds.) (1981); West Yorkshire: An Archaeological Survey to A.D. 1500. Vol 1: West Yorkshire Metropolitan County Council; pp. 141 - 170

Haigh, D. (1982); Saddleworth Seven One Two; Donald Haigh and the 712 Group, Bradford. Now also available online at http://www.castleshawarchaeology.co.uk/documents/Saddleworth712.pdf

Haken, M. (2012); Cambodunum – a re-appraisal, Roman Yorkshire 2, pp. 8–15

Hobson, B., Clay, G., Brown, G. (2015); The Romans in Huddersfield - A New Assessment; BAR British Series 620; Archaeopress, Oxford

Jefferys, Thomas (1771); The County of York Survey’d in MDCLXVII, VIII, IX, and MDCCLXX, Sheet V

Lunn, N., Crossland, W., Spence, B., & Clay, G. (2008); The Romans Came This Way; Huddersfield & District Archaeological Society, Huddersfield

Lunn, K (2013); Field observations of proposed Roman Road 712, Pole Moor; Unpublished report, West Yorkshire Archaeology Advisory Service

Margary, Ivan D. (1973); Roman Roads in Britain, John Baker, London

Moorhouse, S. (Ed.)(1973); The Yorkshire Archaeological Register: 1972, Yorkshire Archaeological Journal, vol. 45, pp. 198-213

Nash, V., Roberts, J., Redhead, N., (2014); Redefining Roman Castleshaw Project; The Centre for applied Archaeology Report No. 32/2014, University of Salford,  available at   http://www.castleshawarchaeology.co.uk/documents/Castleshaw_2014_Excavation_Report_for_web_compact.pdf accessed 1/1/18

Parsons, E. (1834); The civil, ecclesiastical, literary, commercial and miscellaneous history of Leeds, Halifax, Huddersfield, Bradford, Wakefield, Dewsbury, Otley.; F. Hobson, Leeds & Simpkin & Marshall, London   Available at https://archive.org/details/civilecclesiast05parsgoog accessed 16/1/18

Percival, Thomas (1752); Observations on the Roman Colonies and Stations in Cheshire and Lancashire,by Thomas Percival Esq; Communicated by Hugh Lord Willoughby of Parham, F.R.S.; Philosophical Transactions Vol. 47 pp.216-230

Redhead, N. (2017); Castleshaw Roman Fort: Archaeological Evaluation of land north of the fort defences ; report for Friends of Castleshaw Roman Forts; available at http://www.castleshawarchaeology.co.uk/documents/Land_North_of_Castleshaw_Roman_Fort_eval_rep_2016.pdf accessed 1/1/18

Richmond, I. A. (1946); A New Roman Mountain-road in Dunfriesshire and Roxburghshire ; Transactions of the Society of Antiquaries of Scotland, Vol. 80

Rivet, A.L.F. & Smith, Colin (1979); The Place-names of Roman Britain; B.T. Batsford Ltd., London

Roberts, I., Deegan, A. & Berg, D.; (2010); Understanding the Cropmark Landscape of the Magnesian Limestone; Archaeological Services WYAS, Leeds

Spence, B. (N.D.);The Dyall or Devil Stone; unpublished paper included on CD accompanying Lunn et. al. 2008

Thorp, F. (Ed.) (1974); The Yorkshire Archaeological Register: 1973, Yorkshire Archaeological Journal, vol. 46, pp. 141-157

Thorp, F. (Ed.) (1975); The Yorkshire Archaeological Register: 1974, Yorkshire Archaeological Journal, vol. 47, pp. 1-14

Toller, H. S., (2008); W Castleshaw 1.jpg; annotated Google Earth image in Toller digital archive; RRRA archive

Toller, H. S., (2013); Current Research into Roman Roads in Yorkshire Based on Lidar Evidence, Forum Vol 2 pp 141-145; CBA Yorkshire , York

Available at http://www.cba-yorkshire.org.uk/2014/03/the-forum-journal-new-series-volume-2-for-2013/

Watkin, W. T. (1881); The Roman Roads of Lancashire. Part I; Transactions of the Historical Society of Lancashire and Cheshire, vol. 33, pp. 195-222 Available at https://www.hslc.org.uk/wp-content/uploads/2017/07/33-8-Watkin.pdf, accessed 1/1/18

Watkin, W. T. (1883); Roman Lancashire;

Wilson, Pete (2016); The Roman Period Name for Adel; Britannia vol 47, pp. 280-285

Whitaker, J. (1771); The History of Manchester (Book one); London

RRRA Forum for RR712 (opens new tab)


The course of RR712 from Ancoats is marked today approximately by Store Street, Old Mill Street, Bradford Street, and most of Briscoe Lane as far as All Saints church, Newton Heath. After a short gap, Ashton Road west preserves the line, followed by Roman Road in Failsworth. The road in Failsworth was certainly impressive, and appears to have been built on a massive scale to traverse a large tract of marshland.

…..and crossing the inclosures on the south-east end of the town appears in an inclosure near Ancoats, then runs thro’ Bradford, and crosses the very middle of Newton-heath, Newton chapel standing on the very ridge of it. Standing at the west end of the chapel, you see the trace of it into Bradford-lane; standing at the east end, you see the trace of it go betwixt a house and a barn on the east end of the common. It then runs thro’ the inclosures to Mr. Wagstaffes house, where it enters a lane, and is visible enough. In about 400 yards more, being interrupted with a moss, it rises with a prodigious grandeur, and is the finest remain of a Roman road in England, that I ever saw. This is at the back of Mr. Jenkinson’s house in Failsworth, his land lying on both sides, and is now called Street.

Thomas Percival 1752, pp 218-9

The artificial soil referred to is almost certainly night soil, collected from the hundreds of earth closets and middens of the rapidly expanding town. From Ancoats north eastwards, the road was still largely visible and both Whitaker an an earlier writer, Thomas Percival, describe the route in some detail.

And in all this course, so contiguous to the growing town, the trace of it is wholly obliterated by the deep bed of artificial soil with which all the fields are enriched.

John Whitaker, 1771, p.82

Had this massive structure survived it would certainly have vied for the title of largest Roman agger in Britain, possibly in the empire. Sadly, but inevitably, no trace of the actual structure survives, even though long stretches of its line are preserved in modern roads, as described above. As Margary observed (Margary 1973 p.366), it is striking that all the many modern roads that lie upon it are just side streets, the modern main road being a little to the north west (Oldham Road, A62). There are no known Roman settlements or sites along this stretch of road which, given its clear importance, seems a little odd. The field name of Kesterfield appears close to the road in what is now Hollins Green on the Chadderton Tithe map, and there has been much speculation over the years about a possible Roman camp here (Haigh, 1982, p.18), without any known archaeological evidence. An enclosed roadside settlement would make a lot of sense, but it’s unlikely any evidence of such a site will have survived. The road is now largely obliterated by modern housing south of Oldham, although for a short distance Honeywell Lane preserves the line.

As the road started to climb up the Pennine foothills, it passed through Glodwick and Austerlands, where traces survived in 1857 (Watkin, 1881, pp.213 & 214), still maintaining the rigid alignment from Gorse Hill. It should be noted that Margary’s inference that Thorpe Lane marks its course (Margary, 1973 p.366) is incorrect: in fact it is not followed by any modern road at any point beyond Honeywell Lane in Oldham, as far as is known. This first straight length of road, all the way from Ancoats to High Moor, clung to its alignment almost perfectly: as far as is known it never deviated by more than 10m along it nine and a half mile length.   

The precise course of the road as it passed through Saddleworth was not known until the 1970s when work by a WEA archaeology class in collaboration with Donald Haigh and his Bradford Grammar School Archaeology Society firmly established its course through intensive fieldwork (Haigh 1982). Today, we have it easy in comparison as the road shows so clearly on LiDAR imagery (figs 2 & 4). On reaching the edge of High Moor, at about SD97010624, a new alignment was laid out of two and three quarter miles, directed towards the south west corner of the fort at Castleshaw, indicating that the fort was probably built before the road (Haigh, 1982, p.27). There is no direct line of sight between the two ends of this new alignment, but the fact that the road changes direction here after nearly 10 miles, coupled with its approach to the fort at Castleshaw being precisely upon the new alignment can leave no doubt that this laid out be the Roman surveyors.   The road leaves the alignment immediately, contouring around the hill towards Thurston Clough. At this point, the road again becomes extraordinary in its construction, as can be seen clearly in fig. 2 & 3. The entire construction is at least 28m across with what appear to be extremely large ditches, and a second ditch to the north formed from a string of quarry pits. Similar oversized road construction has been identified further along this road at Castleshaw and Manor Farm, Marsden, and on RR710 above Sheffield. All four locations are in prominent positions which can be seen for some distance, and the obvious interpretation is that they were a statement of Imperial intent made in the landscape.

Sir Henry Dryden has favoured me with drawings of sections of this portion of the road, taken in 1844. It averaged twelve and a quarter feet in height, measured about twenty feet in width at the crown, and at the base sixty. It was formed of earth, covered with gravel, which latter remained about two feet six inches in depth on the crown, and, one foot eight inches thick on the sides.

W. Thompson Watkin, 1833

Fig. 1 Known Roman roads approaching Manchester. ©David Ratledge 2016

Fig. 2 Lidar image showing RR712 skirting High Moor, Saddleworth. Image by Mike Haken

Fig. 4 Lidar image showing RR712 at Thurston Clough and descending to Delph, Saddleworth. Image by Mike Haken

Fig. 3 Aerial photo of RR712 on a grand scale approaching Thurston Clough, Saddleworth. Image by Mike Haken

Manchester to Castleshaw

Castleshaw to Slack

Slack to Thorner??

At the outset, I must declare a strong personal interest in RR712, which has become Yorkshire’s most investigated Roman road. Growing up in Saddleworth in the 1970s, I already had a strong interest in local history when investigations into the route of RR712 through Saddleworth began. Whilst I wasn’t directly involved, I was party to many conversations between Ken Booth, David Chadderton and Donald Haigh at Saddleworth Museum, where my mother and I volunteered. Their combined passion for the subject sparked an interest which has never left me and the rest, as they say, is history.  

Mike Haken, 16 January 2018

This important trans-Pennine road has proved both fascinating and baffling at the same time. Its course from Manchester into Saddleworth (formerly West Riding of Yorkshire, now Greater Manchester), has been well known for centuries, although details of its course were only established in the 1970s. From Castleshaw over the hill to the fort at Slack via the A640 its course used to be regarded as certain, although nearly half a century of work by the Huddersfield & District Archaeological Society has established that the true course lies much further south in the Colne valley. From the fort at Slack, the question marks get bigger. The “accepted” course takes it through Brighouse and Cleckheaton to Leeds joining RR290 at Thorner, but recent evidence from LiDAR calls this into question. For ease of reading, the route is discussed in three sections as follows:

The road was excavated here at New Inn Farm in 1973 (Haigh, 1982, p. 35&36), although only the central portion was investigated, with no exploration of the ditches or the area either side of the road. The road itself appeared to have fairly typical, with a low agger creating the appearance of a terrace across the slope, some 6.5m wide, up to 0.6m high and revetted on its south eastern edge. After curving round the hillside, the road then makes a kink westwards in order to traverse the steep slopes and deep cut valley of Thurston Clough although its exact course is hard to determine amongst the remains of multiple terraced trackways which have crossed the Clough over two millenia. The road continued to follow the contours along a wide terrace approx. 15m wide, keeping north west of Thurston Clough Road, passing north of Stoneswood where the agger is c.10m wide and 1.5m high, gradually descending towards Delph. Prior to the work in the 1970s, many writers described a route up Knarr Barn Lane from Thurston Clough, and then down the north side of Knott Hill to Delph along a well engineered terraceway. Hugh Toller considered that this could possibly represent an alternative Roman route (Toller, 2008), perhaps an early one superceded by the much easier southern route .

The road probably crossed the R.Tame a few metres west of the modern bridge in Delph High Street, where the medieval saltergate and the Friarmere boundary both crossed, although the road’s route through the modern village is  not known with certainty. It seems to have rejoined the alignment between High Moor and Castleshaw just north west of Delph Lane at about SD98570812; the agger can be seen just south east of Hull Mill lodge (see lidar, fig. 5) as a low wide earthwork marching straight up the Castleshaw valley towards the fort.

The fort at Castleshaw is thought by many to be the Rigodunum described by Ptolemy although there is no convincing evidence for this assumption. The first fort was probably established in c.79 AD during the governorship of Agricola (Bidwell & Hodgson 2009, p. 73), although it seems to have been shortlived. It was replaced by a smaller fortlet inside the ramparts of the first, probably built between about 105 and c.120 AD (Arrowsmith et al. 1996, p.16). The road has always been assumed to have passed along the south eastern edge of the fort between the fort and the vicus, just to the south east of the holloway known as Dry Croft Lane, but no trace of it was found during excavations in 2014 (Nash et al. 2014 p.132).  A substantial length of road was uncovered leading out of the east gate in the 2014 excavations and it was confirmed that, when the smaller fortlet was built in the early 2nd century, this road was diverted slightly to skirt around it to the north. This is suggestive that RR712 may have originally passed through the fort, although it is worth noting that the via principalis is perpendicular to the line of the road. Whatever the original course, itresumed its planned alignment about 100m to the east (where earthwork remains survive for a short distance east of Cote Lane) before heading straight up the hill towards Stanedge (fig. 6). Another road known to leave the north gate, which some had speculated might head north to join RR720a near Blackstone Edge, was investigated in 2016 and shown to terminate a short distance from the fort (Redhead, 2017 pp. 13 - 20).

The Roman road from Manchester leaves the forts at Castleshaw and heads straight up the slope towards Stanedge. Whilst there has been much disturbance and damage by quarrying and drainage work, as it climbs the hill, it is made clearly visible from the Castleshaw valley by the use of oversized ditches (fig. 7), just as earlier in its course at High Moor. There also seem to be two carriageways, although this has not been established by excavation. As in the the other three instances where this phenomenon is known, the probable reason for this over-engineering was to make a visual statement to both the local populus, and to any travellers utilising this vitally important crossing point of the South Pennines.  About halfway to the summit, the road  changes direction and heads more easterly to the south of the ruins of Brown Rough Farm, and again there appear to at least two phases of agger, not necessarily contemporary.  From here it becomes difficult to see on the ground, which has resulted in much confusion as to the route. Percival merely described it going to the north of Marsden but, two centuries later, Crump was more specific and stated that

North east from Slack, the line of this road has been well documented for centuries for the next two miles, maintaining a straight alignment across Lindley Moor and running just south of the A643. East of Crossland Road, the entire route of the road is far less certain. The route generally accepted since at least the 18th century is that described by Margary (Margary, 1973, pp. 366-7) which takes the road on the northern edge of Fixby ridge to Rastrick, then through Brighouse and on the line of Highmoor Lane and Moorside from Hartshead to Cleckheaton, then northeast to Birkenshaw where Whitehall Road (A58) supposedly represents it all the way to Leeds. It is then supposed to head north east to meet RR729 at Thorner.

Until relatively recently, the only evidence for this supposed route in the 28 miles between Fixby Ridge and Thorner came from Leeds. Edward Parsons, writing in 1834, recorded that “it traversed the centre of Leeds in a line a little to the east of Briggate  and its line is traceable in the neighbourhood of Morley and Gildersome” (Parsons, 1834, p.18). Parsons also noted that, in 1745, “an ancient pavement, strongly cemented” was discovered during the excavation of a cellar on Briggate in Leeds (Parsons, 1834, p.22) and, in the early 19th century, on the bank of the R. Aire, workmen uncovered what was taken to be a Roman ford “composed of a substance known only to that people, wonderfully hard and compact…”, presumably concrete (ibid.). The rectangular earthwork at Quarry Hill, Leeds, formerly known as Wall Flat, marked on the 1st edition OS map as “site of a supposed Roman camp” was entirely built over by 1851 (Faull, 1981, p.161). Whether or not these three accounts record Roman remains is unknown, although a find of a concrete like material is certainly highly suggestive.

In 1972, prior to the building of the M62, an agger-like feature was excavated at Hartshead Moor Top by Donald Haigh (Moorhouse, 1973, p.203). The remains were badly denuded with no road structure remaining and there was just one ditch, on the downslope side, as it traversed the slope. Whilst there must be a suspicion that this was actually a boundary bank of some sort, a Roman road cannot be ruled out.

Despite the lack of evidence for the route, we know for certain that a road existed between Slack and Tadcaster because of Iter II of the Antonine Itinerary. The iter is discussed fully here. It records a route from Birrens, north of Hadrian’s Wall, to Richborough in Kent and, crucially for us, part of that route runs from Tadcaster to Manchester. That can only be this road. Unfortunately, the section  contains major errors.

Back to Top Back to Top Back to Top

Fig. 5 3D Virtual Lidar image showing RR712 heading up the Castleshaw valley from Delph, the agger clearly visible. LiDAR coverage runs out before the fort is reached Image by David Ratledge

Fig. 6 Aerial view of RR712 heading past the forts at Castleshaw. The huge ditch as actually a later hollow-way, although the road can be seen proceeding further up the hill

Fig. 7 Aerial view of RR712 headingup the hill towards Stanedge, before turning east to Millstone Edge and then on towards Pule Hill

Fig. 8 OS 1:50000 map illustrating the”traditional” and the true routes

Fig. 9a & 9b Excavation of RR712 at Manor Farm, above Marsden, Huddersfield, in 2015. Note the monumental scale of the ditches, 1.2m deep, 2.5m wide at the bottom, and 5m wide at the top

Fig. 11 Excavation of RR712 at Upper Holme showing the cobble layer partially removed exposing the base layer of flagstones. Image reproduced by permission of  Huddersfield & District Archaeological Society

Fig. 13 Aerial photo of Moorside Edge showing the road curving round the hillside in a series of short straight lengths, before heading off towards Wholestone Moor

Crump didn’t describe the route in detail, nor for that matter has anyone else who has proposed this general route.  If the A640 (New Hey Road) east of Buckstones represents a Roman road, the terrain is so difficult that there is really only one possible route, turning north north west on Millstone Edge, in effect a zigzag, and then following the Pennine watershed to Dan Clough Moss, where its course becomes supposedly marked by New Hey road (A640). Evidence for this route is based on scant remains of a possible road on Millstone Edge (Thorp, 1975, p.4), and the fact that the A640 is notably straight for three miles between Buckstones and Worts Hill. Yet, despite a lack of evidence, the route became accepted as fact by authorities such as Ian Richmond (Richmond, 1946, p.115) and Ivan Margary (Margary, 1973, p.366), after which the notion became extremely difficult to question without attracting ridicule.

It appears that no one wondered why, if it was heading to Buckstones, the road would turn east above Castleshaw, then turn back on itself to a north-north-westerly direction, when it would have been easier and much quicker to simply turn north to get to the watershed and then follow it, roughly approximating to what is now the Pennine Way (fig. 8). No-one seems to have done any map regression either. To have done so would have shown that on Jefferys’ map (Jefferys 1771) the only road heading west from Outlane and Slack is a track that becomes the modern Polegate and turns south west towards Marsden; there is no sign of any road in the direction of the A640 to Newhey. The straightness of the A640 is solely down to it being the Huddersfield and Newhey turnpike built in 1806, and nothing to do with Roman surveying. For the sake of completeness, it should also be added  that an archaeological assessment of a substantial part of this route alongside the A640 near Pole Moor was carried out in 2013 by the West Yorkshire Archaeological Advisory Service, which found no evidence for a Roman road (Lunn, 2013).

When we consider that the remains of a roadway uncovered by Donald Haigh were so scant they could be from almost any period, we are left with absolutely no evidence at all to support the idea of a Roman road heading via Buckstones. Where did it go, if not where it is marked on Ordnance Survey maps?

Work to locate the true course of RR712 began in 1973, after the discovery of a feature that seemed to contour round the hill on Moorside Edge, north of Slaithwaite (Lunn at al., 2008, pp. 1-5). When excavated, a 7m wide cambered road was uncovered, consisting of carefully laid flat stones forming a pavement-like structure, but with smaller graded stones above, The top layer had been largely ploughed away, but the road material was still 23cm thick in the centre. There were shallow 3m wide ditches either side of the road, with a three metre wide gap between ditch and road - usually termed a shoulder. Such a well engineered road, especially one laid out as it was, could only be interpreted as Roman.  This length of road can still be seen clearly on aerial photography (fig. 12).

So began nearly half a century of work by the Huddersfield and District Archaeological Society (Lunn, et al. 2008). They first of all proved beyond any doubt that the Roman road continued its alignment from Brown Rough without turning northwards and crossed Millstone Edge through a cutting, and continued along the straight alignment almost due east as far as Thieves Clough. The first Turnpike road over Standedge was built by Blind Jack Metcalf of Knaresborough for the Wakefield and Austerlands Turnpike Trust in 1759 and approaches the Roman road here from a more southerly line. The two then roughly correspond eastwards, crossing the modern A62 and then along Mount Road, to the south of Pule Hill.

The road heads past Gilberts to a point at about SE03141009 when it changes course to head north east again. Mount Road, the course of the first turnpike road, goes straight on. In typical Roman fashion, the road avoids any steep gradients and follows the contours in a series of short straight alignments, working around the east of Pule Hill. In 1983, as a result of their investigations, HDAS discovered the remains of a small Roman military site at Worlow on the east face of Pule Hill (Lunn et al. pp. 13-19) and beside the road. It is badly denuded by quarrying, and could be interpreted as either a fortlet or signal station, or perhaps both.  Leading down the hill towards modern Marsden, the road was confirmed again in 2015 (figs 9a & 9b) above Manor Farm, where the road was revealed to be of similar construction to that discovered further north east in the Colne valley. Unlike the road to the north east however, it is flanked by rock-cut ditches of a monumental scale, 5m wide, 1.2m deep and 2.5m across the base (fig. 9b). As at the stretches of road at High Moor and above Castleshaw, this is in a prominent and easily visible position - these huge ditches were intended to be seen as Roman features in the landscape, rather than merely serving as drainage. Based on the line of the road at Manor Farm, the road is presumed to descend northeastwards through the centre of Marsden, and then climb gently up the northern side of the Colne Valley, although its line has not yet been identified with any certainty.

The predicted line must at some point cross Booth Bank Clough, and whilst the road has not yet been located, a cylindrical stone was found in the clough in 1587 and could perhaps be a Roman Milestone (Spence, N.D.). It shows no sign of any inscription, but it is known that inscriptions were sometimes painted on. Given the date it was found, it is difficult to interpret this stone as anything else. It is currently in the garden of Slaithwaite Manor House (fig. 10). A little further north east, at Upper Holme, the road was finally found in 1988, surviving remarkably well. Fig 11 shows the road with the upper rubble layer partially removed to reveal the well laid foundation: the camber is very clear to see. Beneath the foundation layer is a substantial layer of crushed shale forming the body of the agger, and beneath that was found a central spine (fig. 12) consisting a single row of stones. Such features have been observed elsewhere although they are far from common and are generally thought to have been laid to mark the centre line of the road prior to construction. From Upper Holme the road heads north towards Merry Dale Clough, crossing at Clough House Bridge, curving round the hillside to the north west, and can then be seen as a terraceway heading north from Heys towards Wilberlee, where it turns to a more easterly direction past West Top, where it was excavated in 2000. Once on Moorside Edge, it contours around the gentle slope to head in a northwesterly direction in order to approach Wholestone Moor in such a way as to easily skirt around its western face - the agger is plainly visible on aerial photographs (fig. 13).

The road probably remains quite straight as it descends gently down the north eastern flank of Wholestone Moor, bearing just east of north, before changing direction just south east of New Hey Road, opposite Springhead Farm, to head northwest on a well attested two and a half mile long straight alignment past the fort at Slack and across Lindley Moor to Ainley Top. Whilst the M62 has obliterated much of the road near Outlane, small segments do survive, notably at SE081 174 in a garden, which was was recorded during garden landscaping by Donald Haigh in 1973 (Thorp, 1974, p. 146). The road was shown to be 6.5m wide, with construction very similar to that at Upper Holme, with a well cambered shale/clay foundation supporting flat sandstone slabs with a layer of cobbles above.

The fort at Slack was only occupied for a relatively short time, probably c.85 - c,140 A.D., based on the ceramics found during excavation (Hobson et al, 2015, p.4). Work by the Huddersfield and District Archaeological Society however, has shown that the vicus remained in occupation for a considerable period, at least until the late 3rd/early 4th centuries (ibid. pp. 12 & 17), testament to the importance of the road and the need for travellers to rest and stay overnight. Given the proximity of the M62, it’s tempting to draw a parallel with modern motorway services.

In 2004, the Society deservedly won the Mick Aston Presentation Award at the British Archaeological Awards for their work on this road. Despite that, and the undeniable evidence, the Ordnance Survey still inexplicably insist on marking the course of RR712 as being along the A640 (fig. 8).

…. and climbing Stanedge the road came above Marsden to Buckstones and near the line of the existing New Hey road through Outlane by Fixby Ridge, and then by Rastrick to Cleckheaton.

W. B. Crump, 1925  p.220

Manchesterhistory.net  Pastscape

Friends of Castleshaw website




Margary's Roman Roads in Britain, road 712

Fig. 10 the Dyall Stone. A probable cylindrical Roman milestone found in Boothbanks Clough in 1587, now in the garden of Slaithwaite Manor House. Image reproduced by permission of  Huddersfield & District Archaeological Society

Fig. 12 Excavation of RR712 at Upper Holme showing the foundation layer of stones and the central spine, a feature probably used to assist with setting out the road. Image courtesy of Huddersfield & District Archaeological Society







xiv or xv

Adel or Leeds

14 or 15


m.p. xiv replaces  m.p. xx






stage added






m.p. xxiii replaces m.p. xviii

The biggest problem with this section of Iter II is that it is clearly missing a stage since the distance is about 20 miles short, depending on where Cambodunum might be. This missing stage is usually assumed to be Camelodunum, omitted by a medieval copyist because of the similarity between the two names. If we insert an extra stage, but keep the xx miles accorded to Cambodunum with the new entry, Camelodunum, things start to make sense - the scribe lost his place on the line he was copying and jumped a line, missing out the true mileage for Cambodunum along with the name of Camelodunum (Rivet & Smith, 1979, pp.158-9). It’s very plausible, and leaves us with a suggestion for the original text as below: but where is Cambodunum?




True Distance








(distance from York)



Adel or Leeds

15 or 14

+5 or 6

m.p. xiv replaces  m.p. xx






m.p. xxiii replaces m.p. xviii

These observations, along with the few scattered Roman finds in the city and the possible Roman site at Wall Flat, have led to the assumption that Cambodunum is within the modern city of Leeds (Faull, 1981, pp. 157-163). This would mean that the road went through the centre of Leeds presumably to join RR729 near Thorner, or possibly further east. More recently, however, the substantial Roman settlement at Adel has been suggested as a possible alternative for the site of Cambodunum (Haken, 2012 & Wilson, 2016) which would mean that the road must have passed west of Leeds, probably near Kirkstall Abbey, to join RR72b at the Roman settlement at Adel. The distances work reasonably well for either alternative. It must be pointed out that even if Adel is Cambodunum, that does not mean that there is no Roman road through the centre of Leeds joining RR729 somewhere along its course since there could easily be a branch road to Adel, followed by the itinerary.

Despite all the uncertainty, there has been some major progress in understanding this road. Recent evidence from lidar observed by Hugh Toller (Toller, 2012) tells us that the supposed route through Rastrick and Brighouse is almost certainly incorrect (fig. 14). The Roman road can be seen clearly running along the southern edge of Fixby ridge and across Huddersfield Golf Course (fig. 15). In February 2018, just to the east of the area covered by fig.15, the agger was located in the woodland at Upper Fell Greave (Brook, 2018), surviving very well as an earthwork 156m in length and seven metres wide. Projecting its general line, it would seem most likely that it follows a route close to the modern Bradley Road, along the ridge of high ground descending to a crossing of the R. Calder just west of Cooper Bridge. North east of Cooper Bridge at Moor Top is one short, but clear, linear feature which could be a short length of surviving agger, which has not yet been examined on the ground (fig. 16). From its alignment, it seems to be heading towards Leeds through Liversedge, Birstall and Gildersome. This would fit with Parsons’ 1834 account of it being discernable through Morley and Gildersome, assuming he meant that it passed between the two parishes, somewhere between Gelderd Road and the M621.

As discussed, a branch road to Adel cannot be ruled out. A feature can be  seen clearly on aerial photographs leading SSW from the Roman settlement at Adel looking very much like a Roman road (fig. 17). It is probably pure coincidence but, if the line is projected south, it aligns almost perfectly with Roberttown Lane, the straight road that can be seen heading NNE from the possible agger illustrated on fig.16.

Back to Top

Fig. 14 Lidar image showing the probable course of RR712 from Lindley Moor to Roberttown

Fig. 15 Lidar image showing the  course of RR712 from Lindley Moor across Huddersfield Golf Course

Fig. 16 Lidar image showing a possible length of agger at Moor Top, Roberttown, nr Huddersfield.

Fig. 17 Aerial photo showing cropmarks suggesting probable Roman road leading SSW from the Roman settlement at Adle.

The route of RR712 can now be said to be known with absolute certainty from Manchester all the way to the east of the Pennine watershed near Pule Hill. Whilst there remain a few gaps in our knowledge, the general course of the road is known for certain from Pule Hill, across the Colne valley, past Slack and as far as Huddersfield Golf Course. It seems probable that it heads along the ridge to Cooper Bridge and the short feature at Moor Top may be a survival. From there it is fair to say there is no real evidence to determine where it is going.

A road must head to Leeds or Adel but so many of the routes of the Antonine Itinerary are far from direct and can take diversions from main roads in order to take in key destinations, and so even if Cambodunum is Adle, the road to it could be a branch from RR712 rather than the main road. Even if Cambodunum is beneath Leeds, which seems less likely, it is simply impossible to say that RR712 must go to Thorner. Indeed, as it was the main route from Manchester to York it would seem more sensible if it headed to meet RR28, the main road from Lincoln, just north of Aberford but there is no supporting evidence of any kind. We can’t even rule out the seemingly unlikely possibility that RR712 is heading to Castleford, and is actually the same as the recently discovered road heading west from Castleford, RR283 (Roberts et al, 2010, p.71) .

Paraphrasing the maxim frequently put forward by Hugh Toller, “start with what you know with certainty, then follow the evidence”, the two key foci for fieldwork need to be the feature SSW of Adel (fig. 17) and the possible agger at Moor Top, Roberttown (fig. 16). Geophysical survey and excavation could at least tell us whether or not the two features are lengths of Roman road, and further fieldwork may be able to trace them further across the landscape. However, given the built up nature of so much of this part of West Yorkshire, we have to accept the possibility that we may never be able to locate the north eastern end of RR712.

Historic Counties:

Roman Sites on Route:

Historic Environment Records, HE Pastscape and other records

Entry compiled by Mike Haken - Last updated, 21 February 2018