Short URL for this page:

[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 Era in Britain

by John Ward

published by Methuen & Co. Ltd.
36 Essex Street W. C., London

The text, and illustrations except as noted,
are in the public domain.


[image ALT: link to next section]

 p22  Chapter II


With comparatively few exceptions, the indications of a Roman road can only be appreciated by the experienced eye. A slight rise here, a hollow there, a difference in the colour of the herbage, or, in the absence of these, a length of road still used, a hedgerow, or a parish boundary which coincides with the conjectured line — these are the hints from which the archaeologist puzzles out the course of a Roman road. But if these roads are regarded as a whole — as a network of communications — no remains impress us more with the thoroughness of the masters of the ancient world and their high sense of organization. They are more akin to our railways than to our country roads, except such as happen to perpetuate a Roman course or to be the product of a modern engineer. They exhibit, too, in their distribution, their relation to the country at large, and their directness, the impress of a single authority, imperial in its comprehensiveness. We cannot conceive that the semi-isolated British tribes could ever have elaborated such a road-system; on the other hand, this system, once established, must have powerfully operated in breaking down their mutual antipathies and hastening the acceptance of Roman rule. These roads represent the most useful of the great works raised by the conquerors, and they are the most enduring in their effects; indeed, it can scarcely be doubted that our island never possessed finer roads, until the revival of road engineering under Telford and his contemporaries in the reign of George III.

Many of our Roman roads are still used, but these rarely show signs of their Roman origin beyond certain peculiarities of their courses, which will be referred to later. Many, on the  p23 other hand, fell out of use at an early period, probably owing to the changed conditions brought about by the English conquest. These disused roads have in a wholesale manner been levelled by the plough and plundered of their materials; but in most counties of England and Wales there are fragments that have escaped. If Nature may have been unkind to these fragments — if the roots of trees have loosened their structure and the moisture of the ground has softened their concretes — she also has gently buried them under a protective mantle of vegetable mould. Thus buried they continue, sometimes as perfect in form as when they were abandoned.

These intact roads usually show as low, rounded ridges, varying from a few inches to a foot or more in height and from 15 to 30 ft. in width; but these limits are sometimes not reached, or are exceeded. For instance, the Ermingº Street in Lincolnshire and the East Riding, the road from Silchester to Bath, and the 'Achling Ditch' between Old Satum and Badbury Rings, occasionally attain a height of 5 ft. or even more. On the other hand, the ridge may be so low as to be scarcely discernible. The actual roadway is rarely seen, as it is usually covered with turf, and as this is generally thinner on the summit than at the sides, it has the effect of softening the contour and reducing the relative height. Another feature may sometimes be noticed — side ditches. These, in rare instances, retain much of their original size; but usually they are so far obliterated as to show only as gentle and ill-defined hollows, less apparent in themselves, than in their effect of accentuating the relief of the ridge.

The structure varies greatly, even in the same road. It may be anything from a mere spread of gravel on the old natural surface to a causeway of Vitruvian complexity. The differences of construction are not susceptible to any satisfactory classification. We know that there were various kinds of roads — viae militares, viae vicinales, viae agrariae, etc. — and maintained in different manners; but we cannot say whether these were distinguished by any peculiarities of form and construction. No doubt the art of road-making underwent some change during the centuries of Roman rule. No doubt, also, the by-ways were  p24 less carefully and strongly constructed than the principal thoroughfares. But it is certain that the chief cause of structural differences lay in the nature of the materials used. The ancient road-makers depended, with rare exceptions, upon materials near at hand, and as these varied considerably, the mode of construction varied also. The differences are more noticeable in the treatment of the surface than in the under-structure. Where hard or massive rocks abound, paved or pitched roads preponderate; whereas in other districts, especially those of cretaceous or tertiary formation, gravelled roads are the rule.

The 'Fen Road,' which threads Fenland from east to west, is an example of simple construction. Dugdale described it as "a long causeway made of gravel about three feet in thickness and sixty feet broad, now covered with the moor, in some places three and in others five feet thick."​1 The road between Badbury and Pool Harbour in Dorset is similar. Near Corfe Hills, it consists of gravel, 1½ ft. thick and 18 ft. wide at the base, resting on the old heath. Sometimes the gravel was mingled with larger stones, as in a road cut through at Manchester in 1765, which was 4½ ft. thick and 42 ft. wide.​2 In chalky districts, the roads were often of chalk and flints mixed together, as the ridge of the Watling Street on Barham Down, described by Dr. Stukeley.​3 More often the finer materials rest on a platform or foundation of large stones. The Watling Street near Rugby, for instance, has 3 ft. of gravel on a layer of large cobbles laid on the clayey subsoil.​4 A section of the Akeman Street at Woodstock, in 1898, presented a ridge 17 ft. wide with a small ditch on either side, which was constructed of two well-defined layers — a lower, 10 in. thick of Stonesfield slates naturally split and laid sloping in the direction of the road with a few placed flat on the top, and an upper, of 6 in. of the local gravel.5

Two cuttings made by the late General Pitt-Rivers across the Roman road between Old Sarum and Badbury at Bokerly Dyke in 1889, indicated a more complex structure. The second section shown in Fig. 5 presents a central agger with a  p26 convex summit, 25 ft. 6 in. wide at the base and between 3 and 4 ft. high in the middle, between two ditches which were not visible before excavation, the width over all being 87 ft. The agger was constructed as follows: Upon a spread of nodular flints, about 3 in. thick and resting directly upon the old natural surface, were laid, in ascending order, 6 in. of rammed chalk, 10 in. of gravel, 6 in. of rammed chalk rubble, and finally 6 in. more of gravel, the last forming to actual road-surface. Over this had accumulated 5 in. of humus.6

[image ALT: missingALT]
Sections of Roman Roads at Bokerly Dyke, Dorset, and at White Cross, Cumberland. (6 ft. to 1 in.)

Some drainage operations at Strood, Rochester, in 1897, disclosed a fine section of the Watling Street, which presented, in addition to a similar stratified structure to the last, the unusual feature of a timber foundation. The land here was formerly a marsh, and in order to provide a suitable bottom, the Roman engineers first drove two rows (14 ft. apart, the width of the intended road) of stout oak piles into the mud, and spanned the intervening space with timber beams. Upon the platform thus formed, were deposited the following: 3 ft. 6 in. of large pieces of flint and Kentish rag with a few broken tiles; 5 in. of rammed chalk; 7 in. of finely broken flints; 9 in. of small pebble gravel and earth, rammed; and finally the pavement of Kentish rag in irregular pieces, and jointed together with fine gravel. This pavement had worn wheel-tracks, a single one on one side, and three close together on the other, the outermost being 6 ft. 3 in. distant from the first.7

Examples of road-structure might be multiplied indefinitely, but we must further consider the treatment of the surface. That of the Roman equivalent of Edgware Road, London, which was cut through a few years ago, was found to consist of large nodular flints laid with their smooth faces upwards on a bed of rammed gravel, and grouted into a rock-like mass, the whole being about 24 ft. wide. Such roads were not uncommon where flints were obtainable. We read of roads elsewhere being paved with boulders, cobbles, moor-stones, etc. Others are better described as pitched. The surface of the Foss Way near Ilchester was, according to Dr. Stukeley, constructed of flat quarried stones laid edgeways, and resembled "the side of  p27 a wall fallen down."​8 A mile west of Pontypool is a small Roman road, the pitching of which is of the local Pennant grit packed on edge, like the granite settings of our street-crossings (Fig. 6, C).9

[image ALT: missingALT]
Fig. 6. — Roman Roads. A and B, Blackpool Bridge; C, near Pontypool.
(Plan 100 ft., and details 4 ft. to 1 in.)

Others again, as indicated above, were metalled with gravel; but it is probable that in some cases the gravel was the agglomerate of concrete, the limy constituent of which has perished. In other cases, the fine superficial material may be the bedding from which paving or pitching has been stripped. From many of our Roman roads the surface stones have been removed in a wholesale fashion for building purposes.10

Where suitable stone abounded, the roads were often confined with kerbs. The military way of the Wall of Hadrian has edgings of large rough stones, as indicated in Fig. 5, a section made in 1894 at White Cross. This road is peculiar in having a double row of large stones bedded in it along the middle, possibly to support a fence to prevent the commingling of troops travelling in opposite directions. The kerbs of the Maiden Way are similar but roughly squared.​11 A neater kerb, and one that would better withstand the thrust of the agger, was a line of flagstones planted on edge in the ground. The road near Pontypool, referred to above, has kerbs of this description; so also has a Roman road at Blackpool Bridge in the Forest of Dean. This road, which is barely 9 ft. wide, is formed of a rough spread of stones, and the kerbs are supported externally by a row of blocks, as indicated in Fig. 6.​12 Piles occasionally take the place of stone kerbs. A good example occurs near Chats Moss, Lancashire, where the whole structure, consisting of a pavement of large stones, 18 ft. wide, a layer of sand, and a foundation of brushwood, is supported along the sides by stakes driven into the peat below.13

In marshy places, the roads sometimes consisted of a 'corduroy' of oak logs. Such roads have been found near Ambleside and Gilpin Bridge, Westmorland, the latter resting on three longitudinal lines of logs, and kept in position by stakes  p28 at intervals.​14 The Danes Pad, near Fleetwood, is probably a Roman footpath. It averages 20 in. in width, and is constructed of oak trees sawn asunder and laid end to end — a single piece, if wide enough; if not, two or more side by side — upon transverse oak sleepers, through which they are pegged into the peat below.15

The Roman engineers were careful to give their roads the necessary convexity to ensure the rapid removal of rain-water, and they often, perhaps always, provided side ditches. These were sometimes small and bordered the actual roadway. In other instances, they were large and set back from the road-sides, as at Bokerly Dyke. On the Northern Watling Street, near High Rochester, they are 8 ft. wide and 34 ft. apart. The Roman road between Old Sarum and Badbury, near Vernditch Wood, is 6 ft. high and 16½ ft. wide, and the ditches are 60 ft. apart. Another road on Durdham Down, Clifton, was found a few years ago to be 20 ft. wide between two ditches 50 ft. apart. Even on the same road the distance of the ditches frequently varies. On dry ground, a mere gutter on each side sufficed to carry away the rain-water. On swampy ground, it was necessary to drain not only the agger, but the soil below, and for this purpose the ditches were larger and deeper. Prudence demanded that they should be at some distance, especially if the road was highly raised.

The popular belief that undeviating straightness is the distinguishing mark of a Roman road is not borne out by the facts. The Foss Way nearest approaches this condition. Throughout its 200 miles between Lincoln and Axminster it never deviates more than 6 miles from a straight line joining these places. Its gentle sinuosities swing it from time to time across this line, but nowhere do road and line coincide. It provides a remarkably direct route, but not a straight line. On the other hand, in the hilly districts where such direct roads would involve impracticable gradients, they are notably winding, as in the case of the Northumbrian Watling Street and the Doctor Gate between Brough and Melandra Castle in Derbyshire. Another example of circuitousness is the Roman road between Lincoln and York. As the crow flies the distance between these places is 55 miles, but  p30 in order to avoid the swamps of Humberland, the road swings inland, and this adds 17 miles to the route.

A characteristic of Roman roads, but one which they share with many other ancient roads, is a decided preference for high rather than low ground, due to the swampy and wooded condition of the valleys at the time. A more characteristic feature is the mode in which the deviations are laid out. In a modern road or railway, this is effected by curves; in a typical Roman road, by straight lengths forming angles with one another. These angles generally occur on hills or other high landmarks; less frequently at rivers and stations. Reaching a hill, where there is one of these bends, the spectator may expect the road to make a bee-line for some conspicuous point, be it a hill or a gap in the hills on the horizon. The road there may make a fresh bend, or it may continue in the same line until a suitable point is reached, from which the engineers of old were enabled to determine the next stage of their work.

This predilection for high ground and straight sections may be well studied on the Foss Way between Lincoln and Leicester. As already stated, this road crosses the country diagonally in a singularly direct course, nowhere deviating more than 6 miles from a straight line; and the greatest deviation occurs in this portion. The road branches from the Ermingº Street about a mile and a half south of Lincoln, and instead of pointing to its destination approximately south-west by south, it takes a more westerly course, heading straight for Potter Hill, 8 miles away. This hill crests the watershed of the Trent and Witham basins. By this westerly trend, the road avoids the valley of the latter river, which it otherwise would follow for about 18 miles. To avoid the Trent, on the other hand, it now gently swerves somewhat to the south; but after 6½ miles of straightness, it again has a more westerly trend at Newark in order to reach that river at East Stoke, 2½ miles farther, where a bridge carried a branch-road to the north-west. The road here makes a more decided turn towards the south, in order to gain the high ground behind the east side of the valley. For 8 miles it goes straight ahead, until an eminence near Cropwell is capped, when its course becomes still more southerly. Then after another straight course  p31 of 2½ miles, the high ground of Cotgrave Gorse is reached, and here the road attains its maximum divergence from the ideal line. The configuration of the country is now favourable for a return to this line. Availing itself of the stretch of high land between the Devon catchment and that of several small streams which debouch into the Trent to the west, the road takes an 8½ miles' almost due east course, and perfectly straight except for an easterly detour near Willoughby, probably a deviation from its original line. Six Hills, the highest point between Lincoln and Leicester, is then reached. Here a slight westerly bend directs the road more towards the latter city, and at the same time enables it, in its generally descending course to the Soar valley, to take advantage of the spur between the tributaries of that river on the west, and those of the Wreak on the east. After 7 miles of straightness, the Soar is reached at Thurmaston, then another slight westerly deflection directs it to Leicester, 3 miles distant. Throughout its 42 miles of the Foss Way, the road is made up of straight lengths, and the changes in the course take place on the hills and brows, but in no instance at the intervening stations, of which the sites of three are known.

These peculiarities in the setting-out of Roman roads render valuable corroborative service where the actual remains are doubtful; and in the absence of such remains, will sometimes suggest the probable line. But it should be kept in mind, that the Romans did not always follow so exact a method in the setting-out of their roads as in the example just cited.

The distribution of our Roman roads now claims attention. If three maps of Britain, one showing the principal Roman roads, another the principal modern roads, and the third the railways, are compared, it will be observed that, in each, London is the grand centre from which the chief thoroughfares radiate. Further, it will be observed that many of these arteries follow similar courses on the three maps; also, that most of the places where the Roman roads intersected — 'junctions' in railway parlance — still fulfil the same function in the modern road and railway systems. The generally closer network of the modern communications and the multiplication of towns and villages, indicate a denser population, but the preponderance of the  p32 country's traffic still flows along the old lines; and this in spite of the radical change in the distribution of the population referred to on page 4, and indicated on the maps (Figs. 2 and 3). That our highways should reflect the Roman system is not surprising, for many of them perpetuate Roman lines; but that the railways should in any appreciable degree reflect that system, may seem extraordinary. The explanation lies in the paramount importance of London the physical features of the country; and as both the ancient and modern engineers have in the main followed the lines of least resistance, the results, not unnaturally, are also, in the main, similar. The Pennine Chain furnishes a simple illustration to the point. Both the ancient and the modern engineers have avoided it in their routes to the north. Our Great Northern and North-Eastern railway route is along the lowlands to the east, through east Yorkshire, Durham, and Northumberland; while the London and North-Western is through Lancashire, and the basins of the Lune and the Eden which divide the 'chain' from its more mountainous outlier, the Cumbrian Mountains. In both systems it was necessary to connect east and west by threading some of the transverse valleys of the 'chain.'

A comparison of some of the more important railway routes with the Roman will impress the reader's memory with not only the courses of the chief roads, but also with the places on them.

The Watling Street in its diagonal course across the country through St. Albans (Verolamium), Dunstable (Durocobrivae), and Towcester (Lactodorum), to Chester (Deva), may be regarded as the Roman London and North-Western main line, and its ramifications make the analogy all the more striking. Through Chester, the traveller could proceed to Carnarvon (Segontium) and the Menai Straits, by a route which the present Irish mail closely follows. Or, he could branch off en route for Viroconium, our Wroxeter, and there, as his modern representative does at Shrewsbury, 'book' for Caerleon (Isca Silurum), via Abergavenny (Gobannium), by an almost identical route with present, only Newport would be the destination instead of the 'City of the Legion.' Or, if he wished to go to Scotland, there was a line  p33 of road through Lancashire, Westmorland, and Cumberland, which eventually reached the vicinity of Glasgow, by a route singularly prophetic of that traversed by the London and North-Western expresses of to‑day. On this route Bremetonacum, our Ribchester, stood for Preston, and Carlisle (Luguvallium) was the junction, as now, for the Roman 'North-Eastern' for Newcastle (Pons Aelii). Or, again, the traveller could branch off for Manchester (Mancunium), either at Chester or a point farther south in the vicinity of Whitchurch, a Roman Crewe. This route continued, would conduct him to York (Eburacum), only without the modern detour by Leeds.

Perhaps, however, the traveller would prefer to reach Scotland by the more direct eastern route. Striking due north from London (Londinium), by the Ermingº Street, he would traverse for the first 90 miles a belt of country familiar to the passengers of the Great Northern Scotch expresses of to‑day. The Roman, however, bore eastwards to reach Lincoln (Lindum), and then by a counter swing crossed the present route at Doncaster (Danum), and continued his journey through Aldborough (Isurium), Catterick (Cataracto), Binchester (Vinovia), Corbridge (Corstopitum), and High Rochester (Bremenium), by the Northern Watling Street; whereas the present expresses take a more easterly course through York and Newcastle.

The Great Western route from London to South Wales was remarkably anticipated. Silchester (Calleva Atrebatum) played the part of our Reading, and Speen (Spinae), of Swindon. At the second place, the Roman, like the modern traveller, could proceed by a northerly route via Cirencester (Corinium), and Gloucester (Clevum), or by a more direct one via Bath (Aquae Sulis) and across the Severn, only he negotiated the water by boat instead of tunnel. If he chose the former, he struck a little more inland on the western side of the Channel than the modern railway, and he joined his friends, who preferred the sea-passage, at Caerwent (Venta Silurum); thence the road threaded Caerleon, Cardiff, Neath (Nidum), and Carmarthen (Maridunum), much as the South Wales section of the Great Western Railway does to‑day.

The Great Eastern perpetuates the Roman road from London  p34 to Norwich, and as far as Colchester (Camulodunum) no railway more closely hugs an ancient route; but the important Peddars Way, which connected Colchester with the north-west angle of Norfolk near Hunstanton, is quite unrepresented in the East Anglian railway system. The Foss Way, it may also be remarked, is another important Roman route which is similarly unrepresented.

The Watling Street, east of London, like its modern representative the London Chatham and Dover Railway, was the great highway for Continental traffic, and then, as now, Canterbury (Durnovernum) was the point whence branches radiated to east Kent ports, Dover (Portus Dubris) for one. The Stane Street from London to Chichester (Regnum), and Porchester, is tolerably well represented by the London and South-Western through Guildford and Petersfield.

Silchester was one of the most important 'junctions' in the Roman road system. Having reached it, as indicated above, the traveller could turn southwards and proceed to Winchester (Venta Belgarum) and Southampton (Clausentum), just as the present railway from Reading; or he could proceed in a south-west direction through Old Sarum (Sorbiodunum) and Dorchester (Durnovaria) to Exeter (Isca Dumnuniorum) by a route almost as direct as our London and South-Western Railway.

The Midland Railway is less reminiscent of Roman communications but one limb of its X roughly coincides with the Rykneld Street through Worcester, Birmingham, Little Chester near Derby (Derventio), and Chesterfield, York-wards; while the Derby and Manchester section of the other limb represents the Roman road between these placed through Buxton whose warm springs were frequented, as at present.


In a well-watered country as ours, fords and bridges must have been numerous, but remains of few have survived. The Roman fords were submerged portions of the roads, only more strongly constructed so as to resist the scour of the water. A good example  p35 — perhaps the best of any — was one across the Trent at Littleborough near Lincoln, which was removed as a hindrance to navigation in 1820. It consisted of a pavement of large squared stones, the whole being kept in place between two rows of piles, which carried horizontal beams to serve as kerbs.​16 Dr. Stukeley mentions a ford on the Foss Way across the Ivel at Ilchester, and another across the Ebble at Bemerton near Old Sarum, both strongly paved.​17 Another paved ford, 20 ft. wide, crossed the Calder on the Roman road between Manchester and Ilkley.18

Although many small bridges are popularly regarded as Roman, very few of these appear to be so ancient. A small and narrow bridge of a single semicircular arch over the Cock near Tadcaster and on a Roman road to that place, was regarded as Roman by the late Mr. Roach Smith. Of Roman bridges of greater magnitude and importance, there are undoubted remains of several. Those of one over the North Tyne at Chollerford are noteworthy.​19 It was of four spans, and probably of timber, resting upon piers and abutments of fine and massive masonry, the length between the abutments being 184 ft. The existing masonry encloses the remains of a narrower and earlier bridge, possibly the work of Hadrian. When the ancient bridge over the Tyne at Newcastle was demolished in 1771, it was found that the medieval builders had availed themselves of portions of the piers of an older and Roman structure. Another Roman bridge, about 462 ft. long, with eleven waterways, crossed the Tyne at Corbridge, and its remains were examined and surveyed a few years ago. A Roman timber bridge on stone piers is known to have crossed the Nen near Caistor; and the old timber and stone bridge at Caerleon, which was destroyed about a century ago, is said to have been Roman. Some remains of a timber bridge buried in silt at Wallasey near Birkenhead were regarded as Roman by the late Mr. Thompson Watkin. The plan of the Roman road at Blackpool Bridge (Fig. 6) presents an interesting example of a small road which crossed a brook by a ford and a bridge, of neither of which, however, are any traces left.  p36 The ford was the older, and subsequently a loop-road was made which crossed the brook by a bridge, the present bridge being on the site of the ancient one.


The chief roads of Britain, as elsewhere, were equipped with milestones (miliaria). The Roman mile was 1000 paces, hence its name mille passuum (usually abbreviated to M.P.), and the pace was 5 ft.,​20 so that the mile was 5000 ft. It was thus considerably shorter than our mile, thirteen of the one being roughly equivalent to twelve of the other. The typical Roman milestone was a cylindrical shaft of stone about 6 ft. high, but square shafts were not uncommon in this country, and not seldom rough moor-stones of suitable sizes and shapes were used for the purpose. They were usually inscribed. The normal inscription set forth the name and titles of the reigning emperor, the number of miles, and the name of the place from which they were reckoned; but one or both of the latter details were often omitted. The inscription is no evidence of the age of the road to which the stone appertained. On the so‑called 'Via Julia' in South Wales, for instance, miliaria to Diocletian have been found, yet the road was in existence at least one hundred and fifty years before his time; and examples are known of old inscriptions replaced by those of later emperors.

In the British Museum, there is a good example of a cylindrical miliarium found at Rhiwiau Uchaf, near Conway, with the following inscription:—


'To the Emperor Caesar Trajan Hadrian, Father of his country, Consul for the third time. From Conovium, eight miles.' Another of similar form in the Leicester Museum was found on the Foss  p37 Way two miles from that town, in 1771. The inscription is partially effaced:—


These two examples indicate how milestones sometimes prove or corroborate the Roman names of places. Caerhûn, near Conway, and Leicester have long been identified as the Roman Conovium and Ratae respectively, and these milestones found in their vicinities confirm the identification.

A milestone found at Castleford about 1880 is an example of a reinscribed one. It was first inscribed to Decius Trajan, and then was inverted under his successors, the joint emperors, Gallus and Volusianus, and inscribed to them at the opposite end.

The Author's Notes:

1 History of Imbanking and Draining, p174.

2 History of Manchester, i, p120.

3 Itin. Curiosum, p127.

4 Codrington, Roman Roads, p73.

5 Proc. Soc. Antiquities, xvii, p333.

6 Excavations, iii, p74.

7 Proc. Soc. Antiquities, xviii, p36.

8 Codrington, Roman Roads, p268.º

9 Itin. Curiosum, p155.

10 Personal Observation.

11 Cumb. and Westmor. Archaeo. Soc. xiv, pp196, 461.

12 Personal Observation.

13 Watkin, Rom. Lancashire.

14 Proc. Soc. Ant. xviii, p268.

15 Roman Lancashire.

16 Arch. Journ. xliii, p12.

17 Itin. Curiosum, p154.

18 Codrington, p108.

19 For detailed particulars of bridges, see Romano-British Buildings and Earthworks, chap. ix.

20 The Roman foot was a trifle less than the English, being about 11·65 of our inches.

[image ALT: Valid HTML 4.01.]

Page updated: 14 Dec 20