[image ALT: Much of my site will be useless to you if you've got the images turned off!]
Bill Thayer

[image ALT: Cliccare qui per una pagina di aiuto in Italiano.]

[Link to a series of help pages]
[Link to the next level up]
[Link to my homepage]

[image ALT: link to previous section]
This webpage reproduces a chapter of
Roman Roads in Britain

by Thomas Codrington

published by the Society for Promoting Christian Knowledge
London, 1903

Text and maps are in the public domain.


[image ALT: link to next section]

 p285  Chapter IX: Roman Roads from London to Silchester and the West

(1) General Course
(2) London to Silchester
(3) Silchester to Winchester
(4) Silchester to Old Sarum
(5) Old Sarum to the West
(6) Old Sarum to Badbury
(7) Badbury to Dorchester
(8) Dorchester to Exeter
(9) Winchester to Cunetio and Wanborough

 p286  (1) General Course

The Roman road from London, which branching at Silchester communicated with the west of England and South Wales, is not referred to in the Law of Edward, though it was as important as any of the Four Ways. The road takes a direct course from London to cross the Thames at Staines, beyond which it bends slightly towards the south, and then turns due west, straight to Silchester, the Roman Calleva Atrebatum. From Silchester a road of which little trace remains led to the north. To the south a road went to Winchester, from which roads ran to Porchester and Chichester, to Bitterne near Southampton, and to Old Sarum. To the south-west the Portway led direct to Old Sarum, from which one road went to Dorchester and Exeter, and another westward, perhaps on to the Mendips. From Silchester to the west was a road to Speen, where it divided, one branch going to Bath and across the Severn to Caerleon and South Wales, and the other to Cirencester, Gloucester, Brecon, and to beyond Carmarthen. Crossing these roads was a road from Winchester by Marlborough to near Wanborough, which formed a part of Higden's Ermingº Street from St. David's to Southampton.

(2) London to Silchester

The course of the road from Roman London towards Silchester has been noticed as far as the original Watling Street at the south end of Edgware Road (p65). From the Edgware Road the course continues on to Notting Hill along the line of the Bayswater Road, which  p288 a parish boundary follows nearly all the way to the Westbourne stream. There is then a slight change of direction, and thence to Staines, 14½ miles, the course of the road is so direct that it is nowhere more than a quarter of a mile out of a straight line. It would seem that from Notting Hill, on the edge of a terrace 95 feet above the sea and over­looking the Thames valley, some landmark or beacon, on ground (175′) near Upper Bakeham to the south-west of Egham, was the point to which the course of the road was directed. From Upper Bakeham the towers of South Kensington and Westminster can be plainly seen, and the high ground at Upper Bakeham must have been equally visible from Notting Hill before houses obscured the outlook. From Notting Hill the Roman road followed this line, which is the general course of the present high road until the latter turns southwards towards Chiswick. Stukeley rode on by a narrow straight way to Turnham Green, where to a discerning eye the trace of the road was manifest.​1 Parish boundaries, which have followed the present road all along, continue straight on in the same line by Stamford Brook Road to Stamford Brook, through Bedford Park, across Acton (Turnham) Green, and south of Acton Green railway-station, to near Gunnersbury station, and then along the high road again; and indicate the course of the Roman road to near Kew Bridge railway-station. The straight line appears to have been departed from to keep clear of the Thames, and through Brentford, and on nearly to  p289 Hounslow, the course of the road is uncertain. It is probably followed by the present road, curving northwards near Spring Grove to avoid a stream. From the east end of Hounslow lengths of straight road, almost in the same line, followed almost continuously by parish boundaries, indicate that the present road follows the course of the Roman road to Baber's Bridge. The Roman road itself was uncovered by General Roy at the end of the eighteenth century on Hounslow Heath, at the side of the modern road. For two miles on to East Bedfont a parish boundary runs straight a little on the south of the modern high-road, and then the latter, with a slight turn, goes straight to Staines with parish boundaries along it for the last mile and a quarter. The straight line crosses the Thames to Hythe a little to the south of the present bridge, and near the site of the old bridge, to the west of which Stukeley​2 saw the old road very evidently go through the fields, the ridge being then visible; but no sign of it now appears. He traced it along a lane and a footpath towards Thorpe Lea.

In 1835 the officers studying at Sandhurst made a survey of the Roman road onwards to Silchester, and a memoir was furnished to the United Service Journal.3 At Bakeham House, now called Upper Bakeham House, the substratum of the road, and also the foundations of a building, and other Roman remains, had then lately been discovered, proving apparently that the  p290 straight course of the road had continued from Notting Hill to that point. In the valley, nearly half-a‑mile to the south of this line, a stone pillar erected near Great Fosters in 1850 records that it marks the site of a Roman road to Silchester, a portion of which remains in the adjoining meadow. If so the road did not continue straight on, but bent to the south after crossing the Thames; the pillar may however mark the site of a branch road. Beyond Bakeham the memoir referred to states that the direction was through the yard of the inn at Virginia Water, where according to tradition of foundation of gravel, supposed to be the Roman road, had formerly been discovered, and also that the line cuts Virginia Water, and that the ridge could be distinguished for 300 yards, where one of the drives in Windsor Forest ran along it. The yard of the inn seems to be out of a line across any part of Virginia Water, and no trace of the ridge is now distinguishable on to Belvedere, and the course of the road is uncertain.

It is likely that the hill on which the Belvedere Tower stands (260′) was the point made for from Bakeham Hill, though it was perhaps avoided by the road. Beyond, the course of the road lies in a straight line between it and Duke's Hill, Bagshot (300′), the direction changing slightly, but the road from Notting Hill to Duke's Hill is so nearly straight for 23½ miles that no part is as much as three-quarters of a mile away from a straight line between those places. At Sunningdale the road is found in digging in the allotments near the church, and it was until lately visible  p291 on by King's Beeches, and by Chater's Pond, and to the back of Windlesham Hall, where the county boundary marks the line of it for a mile and a half. Enclosing, planting, and laying out the grounds of new houses have now however almost effaced all trace of it. About a mile from Duke's Hill it is described in 1835 as being raised to a considerable height where it crossed a marsh.

At Duke's Hill (300′) there is a change of 27° in direction, and the road goes nearly due west for 16 miles to Silchester. Under the local name of the Devil's Highway, it passes over Easthampstead Plain in a straight line to a point (311′) called Crowthorne on the old Ordnance map, and then in nearly the same line to Ridge Farm, Finchampstead (331′), and with a slight turn southwards on by St. James' and West Court. On Easthampstead Plain it passes a mile south of a large intrenchment called Caesar's Camp, between which and the road, at Wickham Bushes, Roman coins and pottery are found. The Devil's Highway is said by Bishop Bennet​4 to have been raised with a trench on each side, and to have been 90 feet wide, which probably included the trenches. It was levelled at the beginning of the last century when the ridings were cut. In 1835 portions were still existing to the north of Finchampstead Church. It crossed the Blackwater at Thatcher's Ford, and the river Loddon at Stanford, near the north of Stratfieldsaye Park, and beyond that, Park  p292 Lane, also called the Devil's Highway, with the county boundary along it for two miles, runs in a straight line to the east gate of Silchester (Calleva).

From Duke's Hill, Bagshot, to Silchester, 16 miles, no part of the road is a quarter of a mile out of a straight line joining those places.

From near Duke's Hill a Roman road has been supposed to have gone southward to Frimley, Farnborough and Farnham, and the change in the direction of the Silchester road has been thought to confirm the supposition, but without much reason. Stukeley tells us​5 that he traced a Roman road from Winchester to Farnham and Farnborough, and which he supposed went on to Staines. He says that between Farnham and Alton the bank was visible, and in several places between Alton and Alresford. There appears to be no evidence of this road.

Silchester, on ground 300 to 320 above the sea, is in shape an irregular polygon 820 yards from the east to the west gates, and 803 yards from the north to the south gates. Outside the town walls are earthworks of uncertain age. Five Roman roads converge to Calleva, approaching it in different directions, and it cannot be doubted that it owed much of its importance to its being the place where the road from London branched in many directions.

The excavations carried on in recent years show that the street in continuation of the road from  p293 Staines runs straight from the east gate to the Forum, and a parallel street about 93 yards to the north of it leads to the west gate. A street at right angles to these led from the north to the south gate. The east and west gates consisted of two covered passages, 13 feet wide, separated by a middle pier, while the north and south gates had only one passage of 13 feet.

There was no doubt a road from Calleva northwards, but there is nothing to show its course with any certainty, nor could Bishop Bennet trace it in the beginning of the last century. Faint traces appear to have been observed in 1837 in a line between the north gate of Silchester and Ufton Church, and 25 or 30 years before that the road is said to have been traced by excavations in that line for 800 yards.​6 The modern road by Englefield to the ancient ford across the Thames between Pangbourne and Whitchurch is a continuation of the same line. The indications of a Roman road in this direction on the west of the Thames through Streatley to Dorchester have already been noticed (p244).

(3) Silchester to Winchester

This road left Calleva by the south gate, and its course is marked by the road by Three Ashes and Hall-in‑the‑Hole to Latchmere Green. Sir R. Colt Hoare says​7 that the ridge was "seen very fine" behind a barn on the side of the road from Three Ashes to Scotchman's Green, but he appears to refer to an earthwork a quarter of a mile  p294 away from the line of the Roman road, and the pitching of the latter has since been opened on Latchmere Green.​8 For three miles from Latchmere Green the course is obscure, but on mounting the chalk at Sherbourne St. John the line is taken up by the present road, and by a track over Rook's Down, on the highest point of which (440′) a short bit of the ridge remains supporting two fir-trees. From this point Silchester is visible, and a lane, followed by a parish boundary, continues the line for a mile and a half to Worting Cross. The parish boundary continues on in the same straight line for four miles from Rook's Down, and no doubt marks the course of the Roman road, of which there is now no trace on the ground. Near Southwood Farm the parish boundary bends, and is soon joined by the Basingstoke and Winchester road, along which it runs to the Wheatsheaf Inn. It would seem that the Roman road was laid out in a straight line between Rook's Down (440′) and high ground (552′) between Southwood Farm and Kempshott House, and that the high ground was avoided on approaching it by the turn near Southwood Farm. The course of the Roman road appears to continue on in the same straight line from the Wheatsheaf to about a mile south of Popham, the present road lying a little to the west of it; there is then a slight turn, and the modern road, followed by a parish boundary for three miles, occupies the line of the old road along the west side of Stratton Park, and on in a straight line to King's Worthy. The road then turned away from the  p295 river and passed through Headbourne Worthy in a course not very plain, and then followed the line of the present road in a straight line as far as Abbot's Barton, through the grounds of which the same straight line is continued, and where some remains of the ridge can still be traced. The course onwards to the site of the north gate of Venta Belgarum (Winchester) is now covered with buildings.

The Roman city, on the west of the river Itchen, was a rectangle with rounded corners, measuring about 860 yards from the east gate near the river-bank to the west gate at the other end of High Street, and about 780 yards from the north gate to the south gate; the principal cross street of the Roman city being marked by Southgate Street and Jewry Street at right angles with High Street.

A Roman road left the south gate of the city, the course of which the present road follows through St. Cross to Compton between the latter place and Otterbourne the old Ordnance map shows the ridge of the road for half-a‑mile close alongside the present road on the east of it. It can be traced in Otterbourne Hill Wood, but not beyond, unless Stoneham Street indicates the line. It led to the Roman station on the east bank of the Itchen at Bitterne, where near the east end of Northam Bridge a vallum 460 yards long cuts off a promontory, generally supposed to be Clausentum.

Iter VII of Antonine from Regnum to Londinium passes by Clausentum, Venta Belgarum, Calleva, and Pontes. Between London and Clausentum the  p296 Itinerary distances agree fairly well with measurements on the map, bringing Pontes a mile or more to the west of Staines, a more likely situation than on the low ground by the Thames, but between Clausentum at Bitterne, and Regnum (Chichester), the distance is 28 miles, compared with 20 M.P. in the Itinerary. There appears to be now no trace of a Roman road from Bitterne in the direction of Chichester.

A Roman road, said to have come from Bitterne, but more probably from the Winchester road on the west of the river Itchen, has been described​9 as passing along Burgess Street and the north of Southampton Common, where the road is a parish boundary for a mile, over Shirley Heath to Nursling, where it crossed the river Test at a ford near the mill, and on to Tachbury Mount, near which a fragment remained in 1834. The Roman road, of which traces remain on Beaulieu Heath, presumably branched from this road to Lepe, whence it is supposed the Solent was crossed to the Isle of Wight. Rew Street, running southward from Gurnard Bay, has been thought to show the point of landing in the island, but there is no evidence for that or any other Roman road in the Isle of Wight.

Winchester to Porchester

A Roman road left Venta Belgarum by the east gate, and after crossing the Itchen and turning south, curved round between St. Catherine's Hill and Deacon Hill, and then bent nearly at right angles towards the south. A parish boundary along the road for three miles and a quarter round the curve and on to Morestead, shows that the present  p297 Bishops Waltham road follows the same line. A straight line is then entered upon which ranges with high ground (471′) on Deacon Hill, and can be traced for about six miles. There are remains of the ridge in the wood a little south of the turn, and in the belt of trees on the west of the present road. South of Morestead a parish boundary runs alongside the road for half-a‑mile, and in Jackman's Copse there are remains of the ridge. The old Ordnance map shows it for a mile and a quarter near Owslebury, and it is still to be traced through Rowhay Wood for more than half-a‑mile. The same line is taken up by a lane a mile and a quarter further on, and the ridge is to be traced for half-a‑mile on to Wintershill Farm, a mile and a quarter question of Bishops Waltham. Beyond that, on the tertiary beds, there is no trace except perhaps the name Cold Harbour, a mile north-west of Wickham. Stukeley says that upon Portsdown Hill he found some of the Roman way, which he supposed to go through Fareham and Havant to Chichester, with a branch to Porchester. Porchester Castle is one of the most perfect Roman walled fortresses remaining. It stands on a low point of land jutting into Portsmouth Harbour, the east wall, with a water-gate, being washed by the tide. It is a square of about two hundred and ten yards, with corner towers and mural bastions. It appears to be of a late Roman period. A parish boundary runs along the top of Portsdown for three miles, which may indicate the line of the road to Chichester.

 p298  Winchester to Old Sarum

The Romsey road is nearly on the course of the Roman road from Winchester to Old Sarum for three-quarters of a mile from the west gate. It then leaves the line, which is followed straight on by a highway to high ground (484′) at the south-west corner of Teg Down, where there is a turn, and a straight course begins, which is followed for six miles to the west side of the Test valley. The ridge of the road is shown on the old Ordnance map for nearly the whole distance. It is still observable beneath the present road between Teg Down and Crab Wood; it can be followed through West Wood, and it is plain on the west of the wood. Beyond Garlick Farm the straight line is taken up by a hedgerow about half-a‑mile south of Kings Sombourne, and on the edge of the Test valley (200′) the hedgerow and a lane bend slightly towards the north, and descend to Horsebridge Farm. In the beginning of the nineteenth century the pebbles of which the upper crust of the road was formed were still apparent on the east side of the Test valley, and the piles of a Roman bridge are said to have been found in cutting the canal near Horsebridge lock.​10 A bank somewhat resembling the ridge of a Roman road in the meadow on the north side of the road from Kings Sombourne, was the boundary fence of John of Gaunt's Deer Park.

On the west of the river a piece of the ridge is shown on the old Ordnance map, and again towards  p299 Littlewood, on the north of which it is still visible.​11 The present road then joins the course of the Roman road, which lies in a straight line between Farley Monument (587′), erected on a barrow four miles east of the Test valley, and Middle Winterslow (500′ +), seven miles west of the valley. The old Ordnance map shows the ridge for the greater part of the way; and traces remain beyond Buckholt Farm and towards Winterslow, where the present road leaves the line. A parish boundary follows the road for a quarter of a mile near Noad's Coppice. From Middle Winterslow the ridge of the old road is shown on the old Ordnance map winding down the steep hill and then running straight for three miles across the lower ground, over Winterbourne Gunner Down, and through Stock Bottom on Winterbourne Down, half-a‑mile south of Figsbury Rings. The traces of the ridge are now effaced in the low ground, and a good deal of the down has been ploughed up, but on Winterbourne Down it is still to be seen for a mile and a half. There is then a bend, and the road makes straight for the south side of the inner mound of Old Sarum. It crosses the Bourne at Winterbourne Ford, and the lane which now marks the course may be seen from the railway, running straight up to Winterbourne Down. A lane continues on westward, straight to Old Sarum, followed for one mile by a parish boundary.

 p300  The 19 M.P. between Venta Belgarum and Sorbiodunum in Antonine's Itinerary measure 21½ miles. The intermediate station, Brige, was placed by Sir R. Colt Hoare at half-a‑mile east of Buckholt Farm, about 12 miles from Winchester.

(4) Silchester to Old Sarum

This road, called the Portway, was supposed by Sir R. Colt Hoare to branch off from the Winchester road outside the south gate of Silchester, in the wood below the wall, where he says it was visible. A comparison of his map with that of Maclauchlan (1850) shows that the Roman roads of the former are ancient intrenchments in the latter. The new Ordnance map, on what authority does not appear, shows the road by a dotted line from the West Gate, through Pamber Forest, and by Tadley Place Farm to Foscot. Sir R. Colt Hoare failed to trace it, and Maclauchlan, in 1851​12 found not the least vestige of the road eastward of Foscot. There the tertiary beds are quitted, and on the chalk escarpment, six miles from Silchester, the ridge of the road is shown on the old Ordnance map for one and a half miles pointing towards the south side of Silchester, and in the opposite direction to high ground (700′) at Freemantle Park Farm, but it is now hardly traceable. At Freemantle Park Farm (700′) there is a very slight change in direction, and the road appears to have been directed to the high ground (450′) on the south side of Quarley Hill, nearly 19 miles off. The ridge was formerly traceable by Walkridge Farm and over Ridge Heath for four and a half miles, but it has  p301 been effaced by planting a belt of trees along it. The line is fortunately preserved by parish boundaries for three and a half miles to Bradley Wood, beyond which traces of the road are lost for two miles and a half. About half-a‑mile on the west of St. Mary Bourne, the paving of the road was removed in 1879 for a length of a quarter of a mile on account of its interference with farming operations. It was found at from four to eight inches below the surface, and was about eight yards wide.​13 The outline of the road is traceable in the wood to the eastward, and the course of the Portway onwards is taken up by the present road to Andover by Finkley. At Finkley, eighteen miles from Calleva, Sir R. Colt Hoare placed Vindomis, a station named in Iter XV, but the distances from Calleva and from Venta Belgarum would place the station to the east of St. Mary Bourne.

The Roman road from Winchester to Cunetio and Cirencester crosses the Portway about one mile to the west of Finkley, but there are few traces of the latter for two and a half miles across arable land until the Andover and Weyhill road is reached. There, in a straight line with the course of the Portway, which can be seen by Finkley and towards St. Mary Bourne, is a road to the south-west followed by a parish boundary, and the same line is continued by the road onwards through Monkston and Amport to the south side of Quarley Hill, from which, looking eastward, the  p302 course of the road can be traced in a straight line to Freemantle Park, more than 18 miles distant. There is then a turn, and the line of the road lies straight between the south side of Quarley Hill and the south side of the central mound of Old Sarum, ten miles off. From the top of Quarley Hill Old Sarum is visible, as it would doubtless be from the line of the Roman road on the south slope of it if a plantation of trees did not intervene. The present road occupies the course for about half-a‑mile, and then there is a track over the downs, generally a slightly raised grass-covered ridge, but in places worn down to the flint surface of the old road. In about a mile the railway approaches it on the south and runs close alongside it for three and a half miles to near Idmiston. On down land the appearance continues the same, but where the land has been ploughed up the road is no longer traceable. On the down on the east of Idmiston the ridge of the road remained inside the railway fence until the railway was lately widened. At the descent into the Idmiston valley the track of the Roman road leaves the railway and diverges somewhat from the straight course in crossing the valley. It is shown on the old Ordnance map resuming the same straight line and passing round a barrow, but all traces of road and barrow are now effaced by ploughing. Through Porton and Gumbleton the course of the road is uncertain, but on the other side of the Bourne valley it is shown by a line of highways and tracks pointing  p303 straight to the south side of the central mound of Old Sarum, and continuing to within half-a‑mile of it.

A straight line from Silchester to Old Sarum passes a mile to the south of the Roman road at Quarley Hill, where the road lies farthest away from a straight line in the 36 miles between those places.

It is to be noticed that this road, the Roman road from Winchester, and that on to the west, all three point to the inner of the two immense concentric ramparts of Old Sarum. The outer ring, which is supposed to have been strengthened by Alfred, has a mean diameter of about 520 yards.

(5) Old Sarum to the West

The course of this road for the first three and a half miles across the valleys of the Avon and the Wily is uncertain. A ford and causeway across the Wily about a quarter of a mile below South Newton Mill may possibly mark the crossing of that river. Sir R. Colt Hoare described this road, and gave a map of it,​14 which is based on the old Ordnance survey, with additions and slight alterations. Then, as now, the first traces of the Roman road appear as a ridge issuing from the north-east side of Grovely Wood, pointing to the ramparts of Old Sarum, and according to Sir R. Colt Hoare it continued an uninterrupted course through thick copse wood for several miles until it made its exit near Dinton Beeches, and he speaks of its well-known course. The old Ordnance map and Sir R. C. Hoare's map both show the ridge through Grovely Wood in a straight line between the south side of the inner rampart  p304 of Old Sarum, and high ground (648′) near Dinton Beeches, and then on in nearly the same straight line through Stockton Wood and Great Ridge Wood to a quarter of a mile north of Lower Pertwood Farm, sixteen and a half miles from Old Sarum. The new Ordnance map marks the course through Grovely Wood, where it is now very difficult to follow it, by a dotted line not straight, but changing its direction at Grovely Lodge, and bending in the wood to get to Dinton Beeches; and beyond, what is a very crooked ditch and bank on the south of Stockton Wood and Great Ridge Wood, is marked Roman road. Sir R. Colt Hoare carefully mapped and described both the ditch and the road, and he found the latter beyond Dinton Beeches distinguishable across arable fields by a line of large flints, and passing into Stockton Wood, where he shows it on his map of the Stockton earthworks.​15 It is not now traceable, but there is little doubt that the true line of the Roman road is that laid down on the old Ordnance map. Near Lower Pertwood Farm the ridge is shown on the old Ordnance map, and it is described by Sir R. Colt Hoare as passing round a tumulus in its course. Beyond that traces of the road are lost; Sir R. Colt Hoare conjectured that the course was along a road north of Kingston Deverill, and then confesses himself at fault for a very considerable distance; and the course he gives by Maiden Bradley and East Cranmore to the Mendip Hills appears to have little evidence on the ground, or from parish boundaries. The old Ordnance map shows a piece of ridge nearly in  p305 the line of the ridge at Lower Pertwood, on Long Knoll, one mile south of Maiden Bradley. A parish boundary follows the ridge, which points to a barrow on the west end of the knoll. It is six miles from the ridge near Lower Pertwood, and there are no traces of the road beyond. The continuation of Sir R. Colt Hoare's road over the Mendips has already been noticed in connexion with the Foss road (p265).

(6) Old Sarum to Badbury

This road quitted Old Sarum on a line straight from the south side to the crossing of Bokerly Dyke (410′), 10 miles distant. A lane called Portlane leads to a ford south of Stratford-sub‑Castle, where the Roman road crossed the Avon near Coldharbour Farm, and a track, and Folly Lane, mark the course on in the same line to Bemerton, where Stukeley tells us​16 a stony ford over the Nadder was still very perfect. A road up to Wilton racecourse, along which a parish boundary runs, follows the line, and beyond, a green track through ploughed land, a track across the down, and a lane, mark the course to near Toney Stratford, where the river Ebble was crossed. A track, now a good deal effaced, leads on to high ground (500′), one mile south-south‑west of Bishopstone, where a remarkable diversion from the straight line begins. The latter was no doubt laid out from points on the high ground (500′) intermediate between Old Sarum and Bokerly Dyke, but if it had been followed it would have crossed in the space of a mile and a half three steep-sided coombes 150 to 250 feet deep, separated by two  p306 spurs of similar height, before regaining the 500 feet level. The road in such a position would also have been exposed to hostile attack from the high ground commanding it on the south. The straight line was therefore departed from, and the road was kept on the high ground to the south of the coombes. The ridge of the road is shown on the old Ordnance map for nearly the whole length of the diversion. Sir R. Colt Hoare described it some few years later as in very perfect form on the down, and traces still remain. A narrow line, and a track beside a hedge where the land has been ploughed up, now mark the course. Stukeley, who describes this "diverticulum", say that the ridge was very perfect to the west of it, butting full upon the end of a vast valley very deep and of steep descent. The course of the road where it resumes the same straight line as before is now marked by a track along a hedgerow north-west of Knighton Woods, where traces of a paving are marked on the new six-inch Ordnance map, and by Vernditch Woods to the down, where the embankment is very conspicuous, and the ditches at the side remain, including which the total width is about 20 yards. A parish boundary follows the ridge here for a mile and a half. Where the embankment is away from the modern road it remains almost perfect, about five and a half yards wide at top, and as much as six or seven feet high, and where a drove-way has been cut through, it shows a coating of tertiary gravel, two feet six inches to three feet thick. This must have been brought some four or five miles from the south. A little further on the ridge  p307 has been entirely cleared away, the materials having been taken to make the embankment of the modern road where it crosses a hollow. There are no traces of side ditches on the surface, but sections cut by General Pitt Rivers​17 revealed them, and showed the ridge to consist of a layer of nodules of flint on the old surface, upon which were six inches of rammed chalk, ten inches of tertiary gravel, six inches of rammed chalk rubble, six inches of gravel with rounded (tertiary) pebbles, with five inches of surface mould over, making a total height of three feet from the original surface to the top. The gravel top must have been entirely removed where this section was made. Towards Woodyates, where the old and the modern roads approach, the gravel top has been dug out from the middle of the embankment to a depth of more than two feet. Bokerly Dyke, a rampart and ditch which for some miles constitutes the county boundary, here crosses the Roman road, and the excavations made by General Pitt Rivers throw light on both. He cut a section through Bokerly in the line of the Roman road, which proved that the dyke was raised upon and across the road. A flint pitching mentioned in connexion with the latter would seem to have been the lowest stratum, not the surface of the road. Coins found in the rampart proved it to be not earlier than A.D. 394‑423, and General Pitt Rivers considered it might possibly have been thrown up by the Romanized Britons, and he supposed that the Romano-British settlement excavated  p308 by him here was Vindogladia, a station in Antonine's Itinerary, 12 M.P. from Sorbiodunum, but also eight M.P. from Durnonovaria. Sir R. Colt Hoare placed Vindogladia at another settlement found by him on Gussage Down, some five miles further on; and he altered the Itinerary distances accordingly.

Immediately on the west of Bokerly the Roman road makes a turn through about 25° towards Badbury Rings, a conspicuous British camp 11 miles off. The first point made for was, however, on Harley Down (360′), five miles distant. For half-a‑mile from the turn the line of the old road is in enclosed land on the north of the present road, and nothing remains visible but an undulation of the surface. Woodyates, formerly Woodyates Inn, is on the course of the Roman road, and then for a mile the present road occupies the site of it, the ridge having been destroyed, but a parish boundary running along the road. The present road then turns towards the west, and the Roman road, here called Achling or Ackling Ditch, or Dyke, goes straight on, and is very perfect for four miles across the downs, a parish boundary following it for three miles. It is five yards wide across the top, and four, five, or six feet or more high. The gravel top has been dug away in places to a depth of two or three feet for the sake of the material. Traces of the side ditches remain, and in several places they cut the annular bank and ditch surrounding a barrow. This is perhaps the most striking example of the embankment of a Roman road remaining in the country. It runs for miles in a  p309 straight line in bold and sharp relief over the open down, and the magnitude of the work and its situation are alike imposing.

At the highest point on Harley Down (360′), there is a slight turn more to the south in the direction of the east side of Badbury Rings. When the down is left the ridge appears as a hedgerow at the side of a lane, worn away on the other side by the plough, and with a parish boundary along it. After crossing the Gussage brook the lane ends in an old chalk-pit, but the parish boundary continues on for a quarter of a mile to a lane, beyond which a footpath and a fence along the east side of the enclosed Holly Down carry on the line. In a mile the grounds of Crichel Park are entered, and the ridge is traceable through the rookery. The school on the small common to the south of the park appears to be on, or close to, the line of the road, which is no longer traceable across the fields beyond. Witchampton Common, where Sir R. Colt Hoare saw the ridge entire in parts, has been enclosed, and a road has been made along the course of it, which is then followed by a lane for two miles to near Badbury Rings, except at one place where it cuts across the corner of an arable field which the lane goes round. From near Broadford, where a small stream is crossed, a parish boundary follows the lane for half-a‑mile.

Badbury to Poole Harbour

On the north of Badbury Rings the Roman road divides, one branch continuing on with a slight turn to the south, and the other branching off at an angle of 40° to the  p310 south-west. The ridge of the road is plain a quarter of a mile north of King's Down Farm, and can be followed through a wood, on emerging from which it is again plain, first with the upper surface hollowed by digging out the coating of pebble gravel, and further on as a turfed ridge three yards wide and about two feet high. The direction is here a few degrees west of south, pointing between Badbury Rings and High Wood.

The great earthwork called Badbury Rings consists of three concentric ramparts and ditches which crown a detached hill of chalk (327′). It is rather larger than Old Sarum, the mean diameter of the outer defence being about 533 yards.

It is to be observed that for six miles north of Badbury the direction of the Roman road from Old Sarum is straight towards the east side of that earthwork, and with a slight turn, passes to the east of it, as if in the first place the road was laid out to communicate with the south rather than the west, and the road branching off towards the west was afterwards laid out.

Between Badbury Rings and High Wood the road bends to the south-east, and arable ground is entered upon, across which the course could be traced in 1847.​18 The road does not now appear to be traceable between Badbury and the river Stour, but on the south side of the river the ridge can be followed across Eye Mead along the track marked on the  p311 Ordnance map; and the gravel surface of the road can be found a few inches beneath the turf. After the Stour valley had been crossed the road seems to have turned toward the west of south, and on the high ground on the south of the river, a section of the road was formerly to be seen in Corfe Mullen gravel-pit. The position in the pit, which covers several acres, is now unknown, and there is nothing to indicate the course of the road from the river. On the south of the gravel-pit a parish boundary marks the line, along which the ridge can be traced on to a lane about a mile west of Merley House, and then it is plain, a cart-track running sometimes alongside, and sometimes on the top of it, and on the heath it is almost perfect, four yards wide across the top, and three feet high. On the descent into a valley the ridge disappears, but on the ascent of the opposite hill it is again seen, and is very plain on the heath at Corfe Hills, of the same dimensions as before. After again disappearing where the line crosses a hollow, the ridge re-appears on the brow of the next hill, and a section exposed in a gravel-pit in 1900 showed one foot six inches of gravel on the old heath soil, with a bottom width of six yards. The ridge is higher on the top of the hill, and is again visible on the high ground to the south of the road to Broadstone. From that, down the hill, and across the railway to Upton House on Poole Harbour, the course of the road is marked by a track alongside a straight fence. Parish boundaries follow the straight road continuously for three miles from Corfe Mullen gravel-pit to  p312 Upton. The same straight line is continued on the south of Upton, and a piece of the ridge remains between Hamworthy Junction and Holes Bay. Traces also exist on the higher ground of Hamworthy Heath. The remains of the ridge there show a turn to the south-east, and in cultivated ground a little further on the road was taken up not many years ago. It was also dug through about 20 yards west of Hamworthy Churchyard.​19 According to Warne​20 it crossed the Poole road, and after passing to the east of the old Manor House (now the Rectory), it terminated on the shore of the Pool estuary.

A Roman road leaving Badbury at or near the point where the road from Old Sarum divides has been traced northward to Donhead, by Hemsworth Down, to the east of Tarrant Monkton and Tarrant Hinton, through the grounds of Eastbury Park, and over Main Down, where the ridge is visible, to the high ground (850′) on the north of Ashmore over­looking the vale of Wardour, into which it descends by Donhead Hollow.​21 It preserves a straight course for 11 miles, which if continued would, in another 10 miles, join the Roman road from Old Sarum to the west at the last point to which it is traceable, near Pertwood.

There are indications of another road from Badbury to Hod Hill, where there is a Roman camp  p313 within earlier intrenchments. A parish boundary in a straight line along a belt of trees for a mile and a quarter, points to Hod Hill eight miles off, and there are tracks onwards in the same line.

(7) Badbury to Dorchester

The course of the Roman road in the direction of Dorchester, branching off on the north side of Badbury, where the ridge is plain, appears to be through Shapwick, nearly parallel to and on the north of the street, crossing the river Stour a little below the church, and on through Little Coll Wood to high ground (240′) on the south of Coll Wood near Mapperton, where a hedgebank is in the line. Warne​22 describes the ridge a little further on as highly raised for 300 yards. A slight turn brings the line in the direction of the road through Winterbourne Kingston, and also pointing to the north side of the hill crowned by Maiden Castle (400′) 15 miles distant, beyond Dorchester. A bit of the ridge remains in this line in Bagwood, half-a‑mile west of Winterbourne Kingston, and there are several other traces between that and Tolpuddle, and fences and footways and bits of parish boundaries follow the line. From half-a‑mile west of Tolpuddle a parish boundary marks the course for a quarter of a mile, and then there is no trace for several miles. The new Ordnance map show remains of the ridge on Puddletown Heath, not in the same line, and where it is entirely hidden by the thick heath and furze. A piece of the ridge remains along the north side of the road through Kingston Park. The present road from  p314 Grays Bridge to Dorchester appears to be on the line of the Roman road.

Dorchester, generally considered to be Durnonovaria, is bounded irregularly on the north side by the river Frome. Along the High Street the Roman city from the east to the west is 660 yards, and from the north to the south it measures from 360 yards at the east to 730 yards at the west. Though rectangular towards the south the general form approaches a quadrant of a circle.

The Itinerary of Antonine makes it 20 M.P. from Sorbiodunum to Durnonovaria, which is less than half the distance between Old Sarum and Dorchester. If the latter be Durnonovaria, 21 miles must be left out, with a station, probably Badbury, in the same way as a station and distance must have dropped out of Iter XIII between Cirencester and Speen.

Dorchester to the South

A Roman road from Dorchester to the south was very perfect between Maumbry and Winterbourne Monkton before the present turnpike road was made, "a high, broad ridge paved with stone."​23 Near the latter place it passes within half-a‑mile of Maiden Castle, a camp of oval form 1200 yards from east to west, and 550 yards from north to south, crowning a chalk hill (400′) with vast triple ramparts, enclosing 43 acres of ground. There are evidences of Roman occupation, and some traces of the ridge of a road on the Dorchester side of the earthworks. The present Weymouth road is straight for three miles to Ridgeway Hill (440′); and  p315 at the beginning of the last century the elevated ridge, pitched with stone and nearly covered with turf, was still visible for half-a‑mile beyond,​24 and the course of the road was traceable on the south side of the hill to the west of the modern road. A parish boundary follows the course for three-quarters of a mile. On the south of Upway the Roman road appears to be on the east of the present Weymouth road, and the ridge, 10 yards wide, was lately visible in the lane leading to Thornhill Farm, and could be traced in a meadow opposite Broadway Church.​25 On the south of Broadway a parish boundary follows the straight road for half-a‑mile to Redlands, and traces of the Roman road are said to have been found at the foot of the hill in Radipole parish. A parish boundary follows the present road for a quarter of a mile to the mouth of the Wey, near Radipole church.

Dorchester to Ilchester

The road from Dorchester to Ilchester, which has been already mentioned (p267), appears to have left by the west gate, passing to the south of Poundbury, following the course of a straight road for one and a quarter miles to Bradford Peverel, and crossing the river Frome to Stratton in a straight line to the high ground (611′) north of Frampton. It is described in 1796 as a "dorsum broad and high and  p316 paved with flints,"​26 and it is still plain in the fields beyond Stratton, and at Grimston Common Field as an elevated ridge paved with flints.​27 The present road takes the line beyond Grimston, and then there is a turn (611′), and Long Ash Lane, with a parish boundary along it, continues the course for four miles to high ground (800′), half-a‑mile east of Frome St. Quintin. Between this point and a point (400′) near Vagg, two miles north-west of Yeovil, and 11 miles distant, the course of the road seems to have been laid out. Roads and lanes here and there, sometimes with parish boundaries along them, lie in the line which passes one mile west of Yeovil. From Vagg to Ilchester the present road follows the line for three miles in a straight line.

(8) Dorchester to Exeter

To the west of Dorchester the modern road occupies the site of the Roman road for three and a half miles as far as Knowle Hill. From about half-a‑mile from the town it appears to be directed straight for two and a half miles to a tumulus on Bradford Down. A parish boundary follows it for a mile and a half along the straight road, and continues to follow the winding road for seven miles to Eggardun Hill (828′). When Stukeley saw the road between Dorchester and Bridport, then called the Ridgeway, the original ridge remained, made of flints laid in a fine bank and covered with turf.​28 He noticed that it frequently made great curves to avoid valleys, and kept on the highest ground; there are, however, pieces of straight road as well as curves. The ridge is still  p317 plain nearly all along, and in many places it is conspicuous.

To the north of the road, on Eggardun Hill, is an oval camp, triple trenched on the west, about half the size of Maiden Castle, which it somewhat resembles. It would seem that the Roman road took the course of an older road to this stronghold, and the Roman manner of setting out is not characteristic of the road onwards.

From near Eggardun the road turns towards the south-west and slants down the hill by Spyway Green, and is traceable through the arable land. When the old road was broken up it was seen to be "composed of a bed of large flints laid on the substratum of chalk, with a thick layer of smaller stones on the top, the whole almost as compact as a wall."​29 According to Davidson, the course of the road is joined by the present road to Bridport about a mile and a quarter west of Askerswell; but there is a good deal in favour of the course given by Warne on the authority of a local antiquary, through Up Loders and Yondover and along the road from Yondover to the main road one mile east of Bridport. From that point parish boundaries follow the present road to Bridport, and it appears to be on the line of the Roman road through Bridport, and on to half-a‑mile beyond Chideock, where the Roman road mounted Chardown Hill, and passed over Stonebarrow Hill, the course being by Cold Harbour and along a lane followed by a parish boundary for two miles on to the river  p318 Char. At Charmouth the Roman road seems to have divided, one branch continuing along the coast, and the other going by Axminster and Honiton.

The coast road passed north of Lyme Regis to Heathfield Cross in the course of the present road; parish boundaries follow for three-quarters of a mile, a quarter of a mile, and half-a‑mile, and on by Axbridge, Colyford, and Elverway, to Sidford, parish boundaries following the road for three miles on the east of Salcombe Regis, and for half-a‑mile on the east of the river Sid. At Stowford the modern road leaves the line of the old road, which is represented by a lane with a parish boundary along it over the shoulder of a hill, and is rejoined by the present road in three-quarters of a mile. The parish boundary follows the present road on to the river Otter, and then runs at the back of houses on the north of the street at Poppleford, and again joins the present road, along which it runs to Windmill Hill, having apparently followed the course of the Roman road continuously for six and a half miles from Stowford. Then, after an interval of one mile, there is a quarter of a mile of parish boundary, and after another interval of three-quarters of a mile, parish boundaries follow the present road cony for two and a half miles to the junction with the Roman road through Axminster and Honiton.

This latter road goes over Charmouth Hill by Penn Cross and Hunter's Lodge to Axminster, county and parish boundaries following the present road for two  p319 and a half miles from Penn Cross. Just before reaching Axminster a road called Stony Lane, turning off to the north to Millbrook, appears to mark the junction or crossing of the Foss Way. On the west of Axminster, after the valley of the Axe and the Yarty has been crossed, a straight road a mile and three-quarters long begins, pointing the high ground (700′) on the east of Axminster and continuing to Shute Hill (455′). The road winds, with a parish boundary following it for two miles, and descends Moorcox Hill, close to the railway four miles east of Honiton, where in 1828 the structure of the old road was seen, 18 inches thick with large stones at the bottom and smaller at the top, closely cemented as with mortar.​30 The present road occupies the line of the old road through Wilmington, parish boundaries following it for two miles, and then for one and a half miles to Honiton the course is not clear. Through Honiton the present road is straight for three miles, and it continues in nearly the same line to Patteson's Cross, and on to the river Otter at Fairmile, on the west of which a parish boundary follows the road for a mile and a half. From near Whimple the road, straight for two and a quarter miles, is followed by parish boundaries for nearly a mile and a half, and they continue along the road for another mile and a quarter through Rockbere; and again after an interval of a mile, parish boundaries follow the road from the river Clyst to the fork roads near East Wonford, where the Roman road along the coast joins. The  p320 road continues on through Heavitree to Exeter (Isca Dumnuniorum).

The Itinerary of Antonine makes it 51 M.P. from Durnonovaria to Isca Dumnuniorum, which agrees fairly well with the actual distance between Dorchester and Exeter by either route. The intermediate station, Muridunum, 36 M.P. from the former and 15 M.P. from the latter, is placed by Camden and others at Seaton, which is seven miles too far to the east. On the inland line the distances would place it near Honiton, a site which agrees better with the inland position of Ridunum, as it seems to be shown in the Tabula Peutingeriana.

There appear to be no roads set out in the characteristic Roman manner beyond Exeter. According to Bishop Bennet, an old road, converted by the Romans to their own use, went over Haldon to Teignbridge, and was quite plain at the beginning of the nineteenth century on the right of the turnpike road on the ascent to Haldon. It seems to have passed by Alphington, on the south of which a parish boundary follows the present road for a quarter of a mile, and on by Red Cross and Kennford. On Haldon a parish boundary follows the road for half-a‑mile, and then leaves it and joins the road branching to Teignmouth, which it follows for half-a‑mile, and then follows Haldon Lane for two miles to the east end of Ugbrooke Park, where it rejoins the Exeter road, and follows it to Sandy Gate, a mile and a quarter north of Teignbridge. This line of parish  p321 boundaries, more than five miles long, seems to mark the course of the old road. Teignbridge is claimed to be of Roman origin, and it is certain that remains of a very ancient bridge were found in rebuilding the bridge in the latter part of the eighteenth century. Bishop Bennet continued the road by Totnes to St. Michael's Mount, led so far probably by the itinerary attributed to Richard of Cirencester.​31 Another old road over Little Haldon, now called Portway, and the boundary of Dawlish parish, is mentioned in a deed of 1044 as "Stroete" and "Port Stroete," and has been thought to be the line of a road which crossed the Teign below Newton Abbot.​32 Bishop Bennet mentions the crest of a Roman road visible for a mile near Uffculme on the road from Exeter to Taunton, and he says there was a road from Exeter in a north-easterly direction, but there seems to be no trace of these, and of a supposed Roman road from Exeter to Stratton, on the Cornish coast, there seems to be no evidence at all.

Iter XII of Antonine has occasioned much discussion. Wesseling gives the heading Per Muridunum Viroconium, and it was made a Calleva per Muridunum Viroconium by Gale to suit the Iter as it stands. Parthey and Pinder give it a Muriduno Viroconium, and while printing the Iter continuously as it stands in all MSS., they suggest in a note that the first eight lines, enclosed within brackets, were by mistake transferred from Iter XV. On this supposition  p322 the first station after Muridunum would be Leucarum, and the length of the Iter by addition of the distances would be 166 compared with the 186 in the heading. The position of the three stations between Muridunum and Isca Leg. II Augusta (Caerleon) are unknown. Bishop Clifford​33 continued Iter XII from Exeter by Hembury, putting Nidum near Taunton and Bomium near the mouth of the river Parrett, 27 miles by water from Caerleon. No trace of either road or stations are known; the country about the river Parrett must have been a swamp in Roman times, and a 20 mile passage across the Bristol Channel does not seem likely. The suggested removal of these stations to South Wales will be reverted to (p346).

(9) Winchester to Cunetio and Wanborough

This is a road crossing the Portway, passing by the Roman station Cunetio near Marlborough, and joining the road from Speen to Cirencester about halfway between those places. The road left the north gate of Venta Belgarum, and the present Andover road, followed by a parish boundary for five miles, is on the line of the Roman road from the suburbs of Winchester to Barton Down Farm, six and a half miles from Winchester. The course is straight to Worthy Down (400′), where there is a slight turn, and a straight line begins which continues for 15 miles to the north side of Conholt Park, a highway following it for nearly the whole distance. A plantation on the north side of the present road interferes with the view  p323 along the line of the road, but from Worthy Down Conholt is plainly visible in clear weather, and it is quite possible that the long straight line was laid out between Deacon Hill (471′) two miles south-east of Winchester, which the line cuts, and Conholt (790′). From Chute Heath close to the latter the high downs beyond Winchester can be plainly seen in clear weather.

The ridge of the road has been destroyed in making the modern road, but it is visible between Barton Down Farm and the Test valley, and again on the north side of the valley through Harewood Forest, where a parish boundary follows it for a mile and a half. The line of the road is crossed by the railway a mile and a half east of Andover station, and the straight lanes which follow it can be seen from the train. Near East Anton the line of the Portway is crossed, but there is no sign of the latter. On approaching Charlton Woodlands Farm the course is across fields for half-a‑mile, and then it is again followed by a lane, along which the ridge is visible in places. At the south-east of Conholt Park the road bisects a circular earthwork, and through Conholt Park the embankment is very plain. It was described as long ago as 1760,​34 and it was cut into by Mr. G. Knowles, the owner of Conholt Park, in December 1898. In one place, where the old road is a terrace on sloping ground, some six feet high on the lower side, and four or more yards wide, it was found that beneath a foot or 15 inches of soil  p324 there are five and a half inches of white chalk compacted together, then three or four inches of clay, under which was found a layer, 18 inches thick, of large flints, apparently set by hand, the upper four or five inches of which appeared as if they had been burnt and were mixed with a black powder. We learn from the old account of 1760 that the upper stratum was a beautiful gravel, no parts of the country near producing any such material. That has since been removed, probably to make garden-paths. In other places the black powder appears immediately under the turf, the upper layers having probably been removed. Under the flint layer a bed of stiff clay with small flints embedded was found, but was not dug into for more than a few inches.

The appearance of black ash or cinder over the layer of flints apparently burnt is remarkable, and attracted the notice of R. Willis in 1760, who likened them to the cinder and ashes of a blacksmith's forge, but found that they washed white. They must date from the making of the road, as they are covered by the upper layers, and the supposition may be hazarded that we have here traces of fire lighted on the flint foundation of the road to make a smoke as a beacon for the laying out of the 16 miles of straight road from the south.

In the north of the park the long straight line ends; if continued, it would lead into Hippenscombe, a steep-sided  p325 valley more than 200 feet deep, commanded by high ground on all sides, and by Fosbury camp on the east. To keep on the high ground the Roman road, still well defended, bends westward through an angle of 60°, and skirts round the heads of the two branches of the coombe, by Chute Heath and Scots Poor, bending round to the north and north-east until a prolongation of the original line is reached, more than two miles from where it was quitted. This remarkable bend is roughly a half-circle of one mile radius, and is followed for a mile by the county boundary between Wilts and Berks. The ridge of the road is still very plain over Chute Heath, raised three or four feet above the surface, and six or seven yards wide. The semicircle ends near Tidcombe, and another straight line is entered upon which lies between Haydown Hill (850′), near Fosbury camp on the north of Hippenscombe, and Barbury Castle (871′), 16 miles distant, and visible from Haydown. Highways having parish boundaries along them for a mile and a half, and traces of the ridge, mark the course of the road from near Tidcombe by Marten and Crofton to Savernake Forest, and the ridge is shown on the old Ordnance map nearly on to Savernake House, close to which Sir R. Colt Hoare found it visible. The course continues through the forest, on the north-east of the avenue, and nearly parallel with it. On reaching the high ground (620′) on the southern edge of the Kennet valley the road takes a more northerly direction, and it can be seen on the north side of the valley rising  p326 up from Werg, where the river was crossed, to Poulton Down. The Roman station Cunetio at Folly Farm (644′), over­looking the Kennet valley, and some 200 feet above it, lies to the west of the prolongation of the straight line from Haydown. The camp, so far as it can be traced, seems to have been about 300 yards by 250 yards. Roman coins, pottery, and other objects have been found within it from time to time, and a Roman well was opened to a considerable depth about 20 years ago. There are indications of roads slanting down from the camp towards Werg, and westwards towards Marlborough, communicating with the Roman road from Speen to Bath, which will be presently followed (p329).

A Roman road is shown on Sir R. C. Hoare's map of Cunetio (1819) coming from the direction of Old Sarum, and joining the road from Winchester on the south of Folly Farm. It cannot now be traced there, but across the enclosures at Braden Oak, half-a‑mile distant, the ridge is just observable on the same line. From Old Sarum, a road with parish boundaries along it runs in this direction for four miles to Porton Firs, and the road continues on, pointing to Sidbury (735′), a doubly-intrenched earthwork rivalling Old Sarum in magnitude. From the north entrance what Sir R. Colt Hoare calls a bold, broad, and straight raised causeway, resembling a Roman road, runs for more than a mile to Everley, intersecting a barrow in its course.​35 There is nothing definite beyond to indicate the course of a Roman road.

 p327  From the north side of the Kennet valley the course of the road is almost straight for six and a half miles to the edge of the chalk escarpment between Badbury and Chisledon. It is marked by a line of highways up to Poulton Down, across which is a green track with the ridge almost entirely effaced. At the highest point, on Poulton Down (700′), one can see Folly Farm, and northward through the notch in the escarpment near Badbury into the vale beyond, and there can be little doubt that the road was laid out straight between this notch and the high ground near Cunetio from intermediate points. A lane follows the line to Ogbourne, from which onwards the Roman road was evident when Sir R. C. Hoare saw the turnpike road being made in 1818. There is a very slight turn at a high point (560′) near Whitefield Farm, and beyond a parish boundary runs along the road for a mile. Near Badbury the road bends to the east and descends into the valley; and it then lies straight between a high point (556′) on the edge of the chalk escarpment and the point of junction with the Roman road from Speen. At the time of Sir R. Colt Hoare's survey the workmen making the new road described the old road at Badbury Wick, then called High Street, as having been composed of flint, 10 or 12 feet wide, with large sarsen stones for a foundation in wet and marshy places.​36 Across the fields the hard causeway was revealed in ploughing. A parish boundary runs along the present road for three-quarters of a mile towards the junction of the Roman roads, three  p328 miles due east of Swindon station. Here, where many remains of Roman pottery and coins have been found, Sir R. Colt Hoare fixed a station to which he gave the name of Nidum. In Iter XIII of Antonine Spinae figures as 15 M.P. from Durocornovium, whereas Speen is really 34 miles from Cirencester. Fifteen miles brings us to the junction of the roads, and adding 19 miles, the distance on to Speen, makes up the total at the head of the Iter. It is a reasonable supposition that a station has dropped out, but if so it is more likely that the missing station was 15 miles from Speen, where on Wanborough Plain there are also Roman remains. A Latin name given to a supposed station is misleading, and Nidum is the more unfortunate as it is the name of a station in the Itinerary of Antonine, the site of which is uncertain. Unfortunately too the name has been perpetuated on the new Ordnance map.

The Author's Notes:

1 Itinerarium Curiosum, p205.

2 Itinerarium Curiosum, p205.

3 Gent.'s Mag., 1836, part I p535.

4 Lysons, vol. I p201.

5 Itinerarium Curiosum, p203.

6 Gent.'s Mag., 1838, part I p192.

7 Ancient Wilts, Roman Æra.

8 Maclauchlan, Archaeol. Journ., vol. VIII.

9 Maclauchlan, Journ. Archaeol. Assoc., 1891, p182.

10 Sir R. C. Hoare, Ancient Wilts, Roman Æra.

11 A pig of lead, found at Bossington on the west of the river in 1783, which is in the British Museum, bears a stamp showing the date of it to be A.D. 60. (Gent.'s Mag., 1783, part II p935) It apparently is evidence of traffic along the road from the Mendip mines at that early period.

12 Journ. Archaeol. Inst., vol. VIII.

13 Journ. Brit. Archaeol. Ass., vol. XXXV p92.

14 Ancient Wilts, Roman Æra.

15 Ancient Wilts.

16 Itinerarium Curiosum, p187.

17 Excavations in Bokerly Dyke, pl. CLXIII.

18 Dr. Wake Smart, Proc. Dorset Nat. Hist. and Antiq. Field Club, vol. XV p22.

19 Wake Smart, Proc. Dorset Nat. Hist. and Antiq. Field Club, vol. XI p21 (1889).

20 Ancient Dorset, p183.

21 Proc. Dorset Nat. Hist. and Antiq. Field Club, vol. IX p147.

22 Ancient Dorset, p174.

23 Hutchins' Dorsetshire (1796), p. viii of 3rd edition, 1861.

24 Hatcher, quoted by Warne, Ancient Dorset, p186.

25 Warne, 1872, p186.

26 Hutchins' Dorsetshire.

27 Warne.

28 Itinerarium Curiosum, p161.

29 Davidson, British and Roman Remains in the vicinity of Axminster, 1833.

30 Davidson, loc. cit.

31 Lysons, vol. VI p. cccxiv.

32 Davidson, Trans. Devon. Ass., XIII p106.

33 Somerset Archaeol. Proc., vol. XXII, pp. 2, 22.

34 Archaeologia, vol. I and vol. VIII.

35 Ancient Wilts, p180.

36 Ancient Wilts, Roman Æra, p94.

Page updated: 23 May 01