Roman Roads

by Colin Peal

Hill Farm, Gestingthorpe, is situated some 4 miles north-east of Halstead in north Essex, and close to the Suffolk border. The farm, of approximately 680 acres, has been owned and managed by Harold and Ashley Cooper, father and son for over 50 years. As soon as Harold Cooper took over the farm he began finding artefacts which alerted him to the likelihood that a Roman settlement of some kind may have existed on the farm. His fascination inspired him to study Roman history and archaeology and his subsequent investigations have established that there was indeed an important settlement there. Despite all the archaeological work that has been done on and around the site over the years and continuing efforts to locate the line of a major Roman road that was believed to have passed close to it, no trace of the road has ever been found. Although I have dowsed several sites and areas of landscape in North Essex, including what I believe were two military encampments, I had no experience of dowsing ancient roads or trackways, so some homework needed to be done first in order to brief myself on the general background. I try to avoid getting into a lot of detail too early on to avoid the danger of auto suggestion influencing my results.

The main reference work on the subject is still ‘Roman Roads in Britain‘ by Ivan D. Margary (Phoenix House Ltd), first published about 50 years ago. Copies are hard to come by and it was nearly 6 months before our county library was able to produce one for me, by which time I was well into my programme. Margary designates the road in question ‘Route 33a’. Starting from Chelmsford (Caesaromagus) it follows a virtually straight line north-east along the present A.131 from near Little Waltham through Braintree, where it crossed ‘Stane Street’ (now the A. 120) the east-west route from Colchester through to Braughing near Hertford. About 2 miles north-east of Braintree, at a point known as High Garrett, where the A.131. and A1017 divide, the line has disappeared and it doesn’t reappear on the official record until Long Melford in Suffolk, where it continues along the same alignment through Ixworth (on the present A.143) near Bury St. Edmonds, eventually becoming Peddars Way in Norfolk. Margary (Vol 1 p 224) considered this route to be of great importance, more so even than Stane Street, and thought that it might have linked up with a ferry service across the Wash. On all maps of Roman roads, the 8 mile (13 km) missing section of this road between High Garrett and Long Melford is shown as a dotted line: ‘course inferred’. This dotted line passes approximately 1 km to the east of Hill Farm.

At first glance it didn’t look as though it should be too difficult to pick up the line from where it has got lost at High Garrett, and dowse it through to link up with the point where it reappears at Long Melford. In order to get between the two the road would have had to cross the Colne River, the Stour River, and intersect with another major Roman road believed (though not universally) to have run north-west from Colchester through Cambridge and on to Chester. This road has been somewhat grandly called the ‘Via Devana’, or ‘Wool Street’. Margary designates it Route 24a. In addition, Route 33a would have had to cross the Bourne Brook, a tributary of the Colne, somewhere near Gosfield, and the Belchamp Brook, which runs into the Stour, at some point between Hill Farm and Long Melford. Research at the Sites and Monuments Record Office at County Hall in Chelmsford, and elsewhere, made it clear, rather to my surprise, that none of these points has ever been identified. A lot of fieldwork searching for the Colne River crossing between Halstead and Sible Hedingham was done in the 1940’s and 1950’s by a local author and amateur archaeologist, the late Jack Lindsay, with a substantial amount of professional help, and recorded in his ‘The Discovery of Britain’ (Merlin Press), but he was never able to find any trace suggestive of a major crossing anywhere near the line of the inferred route.

Much valuable fieldwork on Roman sites and roads has also been done by the Haverhill District Archaeological Group in their area in the upper reaches of the Stour valley nearer to Cambridge, and recorded in a series of excellent Journals. The following introductory comment in one (Vol IV No.2, 1986) provided some useful down-to-earth advice;-

“Unfortunately many local antiquarians have so desired the presence of a ‘Roman Road’ to use as ‘proof’ of the antiquity of a given area, as well as to form the foundations upon which they may build their historical research, that routes have sprung up like weeds! We read in so many accounts that a Roman road ‘must’ have once been routed through the area under consideration (preferably now submerged under modern thoroughfares, but with appellations of proof such as ‘-street’). There is nothing new about these desires, they have been manifest for centuries from the time when antiquarians espied Roman roads in every exposed strip of plough-soil showing colour differences, or gravel, and then ‘proved’ a route by assiduously labelling every stray find of pottery, metal work and glass as ‘Roman’….. In short such work is too often accepted with degrees of naivety ranging from the gullible to the blinkered, so let the reader beware!”

There were trackways across the country long before the Roman invasion. Some were reinforced and extended by the Romans. However, many new roads are known to have been built during the occupation. Margary points out (Vol. 2 p 232) that the Roman surveyors showed ‘consummate ability in choosing the most direct route’. He refers to the exceptionally steep approach to Lincoln as an example of the fact that they were never deterred by gradients and other hazards. They held their line and having large forces at their disposal they just brought up more troops or slaves to get vehicles over awkward places where necessary.

The main Roman (and some earlier) roads were of similar construction. Two parallel ditches were dug, the spoil heaped up in between and compacted. Finally a metalled surface was laid on top consisting of layers of stones, larger below, smaller on top bound together with a matrix of some suitable local material. Some cobbled surfaces which have survived are of remarkably high quality. The road itself, the bit in the middle between ditches, is known as the ‘agger’. My investigations around here have suggested that there were considerable variations in the width of Roman and earlier roads. It seems reasonable to assume that there would have been some relationship between the width of the road and the traffic that it was intended to carry. This, in turn, must have depended on political and economic factors, local, regional and national at the time the roads were laid down.

The number of gaps in our knowledge of the Roman road network in this area seems surprising, bearing in mind that this was the territory of a confederation of the Trinovantes/Catuvellauni who constituted the major military force in the island at the time of the Roman invasion in 43 A.D. and it was in this area that the great revolt of A.D. 60 took place. The story is, of course, well known at least in general terms.

Under Cunobelinus (Shakespeare’s Cymbeline) the confederation had become wealthy and powerful, bullying the neighbouring tribes, notably the Atrebates, in roughly, West Sussex / Hampshire, and the Iceni in north Suffolk-Norfolk. Fearful of their aggressive neighbour these tribes had been only too happy to ally themselves to Rome and had been granted client status. The main urban centre of the Catuvellauni/Trinovantes, Camulodunum , (the fortress of Camulos, the Celtic God of War) now Colchester, covered a very large area.

When the Romans landed they marched on, and quickly captured Camulodunum, after defeating the Catuvellauni and Trinovantes. The emperor, Claudius in person, was then proclaimed ‘Imperator’ at a great ceremony on Lexden Heath just to the west of Colchester. A fort was established at what is now the western part of the town inside the Balkerne Gate, but in 49 A.D. the Romans decided to convert the place into the ‘Colonia Victricensis’, a settlement for pensioned-off legionaries. Large areas of land were requisitioned for the purpose, causing widespread anger and resentment.

In 60 A.D. Prasutagus , King of the Iceni, died. Reneging on earlier Roman promises the Procurator, Catus Decianus, decided to annex the kingdom. The widowed Queen Boudicca and her daughters were raped and flogged. The enraged tribe descended on Colchester. Joined by dispossessed locals they torched the colonia putting every Roman and collaborator to the sword. They then marched on London and St. Alban’s, razing them to the ground and massacring all the inhabitants they could find. The Roman governor, Suetonius Paulinus, was campaigning with his legion in Wales at the time, but returned post haste. After defeating Boudicca, he subjected the area to a policy of brutal repression and famine. The number slaughtered in the Great Revolt will probably never be known but 70,000 is one estimate, and at least as many again by the Romans in retribution. The revolt seems to have had repercussions throughout the empire but despite the massive impact that the carnage must have had in the region, not a great deal seems to be known about the episode and its aftermath beyond the bare facts.

In 77 A.D. Agricola became military governor, bringing in a more conciliatory policy. The temple of Claudius was rebuilt, the building of new temples, public buildings and villas was encouraged, and Agricola is said to have exhorted the locals to seek a more cultured and pleasurable life – a bit of ‘Pax Romana’. Although Colchester’s importance seems to have declined in relation to the rest of Britain during the occupation it nevertheless remained an important centre. The Romans established a naval base further down the Colne river at Fingringhoe. To counter increasing raids on the East Coast (the ‘Litus Saxonicus’) in the 3rd century the Romans established a chain of 10 shore forts between the Wash and Southampton Water. Colchester was protected by those at Walton and Othona (Bradwell-on-Sea). Walton Castle was claimed by the North Sea a long time ago, but on the Othona site still stands the Saxon Chapel of St. Peter’s-on-the-Wall. Said to be our oldest church still in use, it was built in 654 from materials robbed from the Roman fort. Believed to be substantially unaltered it is all that remains of a minster established there by St. Cedd. In ‘The Essex Landscape’ (E.R.O. ISBN 1 898529 15 9) John Hunter points to the decline that set in eastern Britain during the last years of Roman rule. Population was falling, the tied peasantry being governed by an oppressive bureaucracy and a small rich elite.

Fort sites within a 50 mile radius of Colchester were established at Stanway, Kelvedon and Chelmsford (all on the line of the present A. 12, the main road from London to the east coast), Great Dunmow, Baylham in Suffolk, Ixworth north-east of Bury St. Edmunds, two near Great Chesterford, south of Cambridge, Godmanchester near Huntingdon, and St. Alban’s. Villa sites in the vicinity of the particular roads under investigation have been identified at Ridgewell, Hempstead, Colne Engaine, Sible Hedingham and Little Yeldham. There are several in the Haverhill area, also around Witham on the road between Chelmsford and Colchester (now the A.12). On the north bank of the Stour about 5 miles east of Haverhill Wixoe was the site of a significant Roman township or semi-urban settlement with a possible military presence there and a number of roads radiating out from it.

Ashley’s request, and therefore my original objective, was to find the main road near Hill Farm, Margary’s Route 33a, but as soon as I started it quickly became clear that the whole countryside appears to be criss-crossed by old trackways.

Dowsing from the point where the route has officially got lost north-east of Braintree, the dowsed line immediately branches into two, casting a doubt, right from the start, on the official record. My doubts about it have continued to grow ever since.

I picked up the crossing of the Stour south-west of Long Melford exactly on the inferred line and coinciding with the present road bridge. I was able to follow it south-west through to Hepworth Hall close to the north bank of the Colne about a mile upstream from Halstead. Much Roman material has been found on the property over the years but the dowsed line stops at that point. I was equally unable to find a crossing of the Colne from the south anywhere near this line. One branch from where the dowsed lines divide at High Garrett follows quite closely, for most of the way, the present road (A. 1017) right through to Haverhill. A crossing of the Colne can be detected coinciding with the present-day road bridge over the river, equidistant between, and in precise alignment with, the churches of the two Hedinghams, Sible to the west and Castle to the east. Dowsing has given strong indications of Neolithic and Bronze Age activity along the river in the vicinity, suggesting that this crossing may well be Neolithic. Dowsing the other branch north-east from High Garrett, it follows the inferred route fairly closely, but before reaching the Colne it suddenly swings east into the centre of Halstead to where the present A.134 bridge crosses the river. Another river crossing about half a mile downstream from Halstead was detected later.

In addition to the section of Route 33a that is missing, a considerable stretch of the ‘Via Devana’ has got lost through this area as well as a sizeable chunk of the road between Haverhill and Long Melford (Margary Route 34a). It seemed sensible, therefore, to try and fill in the gaps on all the main routes from the points where they have got lost. In addition I decided to try and establish exactly how many roads entered and exited the Hill Farm site. This meant, in effect, covering the whole area between High Garrett, Haverhill, Long Melford, and Chalkney Wood near Earls Colne, roughly 5 miles north-west of Colchester. A section of Roman road still runs prominently through the centre of this important area of ancient woodland.

As I covered more and more miles I detected intersections with other roads and further branches. I also became aware of new and hitherto unrecorded sites along the way. Roads, after all, connect places where people lived and worked. Being able to locate new sites is a potentially valuable spin-off from dowsing ancient trackways. Perhaps we should do more of it.

Since completing my report to Ashley Cooper, I have been giving attention to one or two of the other roads that came to light and they are already posing some new questions. It became increasingly clear that Margary’s record of the network, in this area at least, is far from comprehensive, and I am finding it more and more difficult to understand why he attached such importance to Route 33a. An excavation was carried out in 1995 by the Essex County Council’s Field Archaeology Unit on a cross-section of this road where it crosses a brook approaching the village of Great Leighs between Chelmsford and Braintree. The width of the agger is recorded as 3 m. Dowsing the gap sections in Margary’s Route 33a and 34a indicated an agger of similar dimensions. As I covered more miles and noted the way in which these roads have been laid out I became more and more of the view that they are not Roman at all but Iron Age. Was 3 metres (10 ft) really wide enough to provide rapid transit for a legion of 10,000 men with all their baggage etc? Dowsing for a starting date for these roads has given me 2nd to 3rd century B.C. In contrast the road through Chalkney Wood is at least double the width of the earlier roads. The aggers of the various branches that spread out from it that I have been able to follow so far appear also to be of the general order of 5-7 metres wide, and the whole way in which they run through the landscape is clearly much more ‘Roman’. This is also true of a section running south-east from Cambridge until it peters out a couple of miles north-west of Haverhill. Dowsing this through from that point to Long Melford it runs mostly on higher ground well to the north of the Stour, much of it straight as a die, and on a remarkably precise east-west alignment . Did the Romans find the ‘old’ roads on lower ground closer to the river to be inadequate? At present, therefore, I have at least two generations of roads from the Roman period, and there may well be others – after all they were here for four centuries. The road out of Chalkney Wood divides into several branches to the west and north, one of them being a good candidate for the so-called Via Devana. It crosses one of the earlier tracks heading north east in Broak’s Wood, another delightful area of ancient woodland.

I was always doubtful as to the extent to which it would be possible to substantiate any of my findings. When I gave the final draft of my report to Ashley he got his mini-digger to run a series of trenches across eight of the dowsed lines I had marked out. In one of them we found the clear outline of a ditch along one side of a dowsed trackway, and in another, signs of a former palisade which I had marked in near the entrance to what is thought to be the main industrial area at the Hill Farm site. Of all the other trackways I had marked, including what I believe to have been the main roadway running through the settlement, however, no trace could be detected.

Ashley pointed out, however, that the field-walking programme over the years has produced concentrations of finds in two lines either side of the track lines that I had staked out. My dowsed lines also run nicely close to where the main Roman building stood. Any credibility which my report can attract will have to come largely from such indirect supporting evidence.

Dowsing ancient tracks, henges, cursuses etc. with angle rods, mine start deflecting together in parallel some distance before I actually reach the feature that I am looking for. When I reach the edge of a ditch the rods cross and then open out and go all listless and floppy until I reach the other side when they cross again. While I am over the agger of a road my rods remain firmly crossed until I hit the ditch the other side when they go all floppy again until I reach the outside of the ditch. Thus when I cross the line of an ancient road or track, I get four reactions, the two outer ones denoting the outside of the ditches, the two inner ones marking the agger.

Since it is not often these days that you can actually follow the line of any ancient road (long invisible) for any great distance across country in this area at least, it is important to establish its alignment at each dowsing point. I sometimes use a compass, though caution is needed as I find that my compass can be influenced by power lines, wristwatch batteries and even by dowsing rods.

I have been asked how much map dowsing I did. Not much. Desktop dowsing is an essential tool in our kit but whatever you find you have to get out there and check anyway. Ancient roads and trackways are so much more than lines on a map. As arteries linking places where people lived and died, worked, traded, and conducted ritual and other ceremonial activities, are they not every bit as important as the sites themselves? In this particular investigation I live fairly well in the middle of the area being covered, all of it comfortably within about 20 minutes drive. Also, I am an outdoor sort of person, instinctively happier out in the fresh air, rain or shine, than I am stuck indoors. I suppose this is one reason why I find landscape dowsing so appealing. There are often times out there when it is important, for me anyway, to forget the clock, put rod and pendulum away and make time ‘just to stand and stare’. How much did Alfred Watkins owe to rod and pendulum?

© 2002 Colin Peal & BSD EEG