In 1568 William Camden, antiquary and humanist, published his great work Britannia, a topographical survey of sixteenth century Britain. In discussing the area that we now call South London he made reference to:
“ … Woodcot, where by a tuft of trees upon an hil-top there are to be seene manifest signes of a prety towne and diverse wals built of flint stones … This in my conceit was that Citie which Ptolomee called Niomagus, and the Emperour Antonine Noviomagus”.
Woodcote today is a suburb merging into Purley to the east and Wallington to the north, and I doubt whether many residents are aware that it was once written of as a ‘Citie’ which caught the attention of a Roman emperor.
However, the key point of interest here is the reference to Noviomagus. Scholars were arguing about its location when Elizabeth I was on the throne, they were arguing about it a century ago when the Victoria County History described it as “one of the greatest puzzles of Romano-British topography”, and they are arguing about it still. Noviomagus may not be a lost Roman ‘Citie’, but it is a lost Roman settlement, and it’s lost somewhere in South London.
Camden’s mention of the ‘Emperour Antonine’ is an oblique reference to a Roman imperial document, the Antonine Itinerary, produced in the early 200s, which listed routes and distances across various provinces of the Western Roman Empire. The reference to Noviomagus occurs in Iter II or “Route 2” of the British section, which describes a journey from Londinium (London) to Durobrivae (Rochester). According to the Itinerary, travellers following this route from London would after 10 Roman miles reach Noviomagus; after another 18 miles, Vagniacis; and finally arrive after another 9 miles at Rochester. The overall distance from London to Rochester by this route would therefore be 37 Roman miles.
There is of course a well-known Roman road – Watling Street – which runs direct from London to Rochester, so our first thought is that Noviomagus and Vagniacis must lie somewhere on Watling Street. But the distances don’t work. Depending on where you measure from, the direct distance from London to Rochester along Watling Street is between about 27 and 31 Roman miles, not 37.
Our second thought, therefore, is that the Itinerary simply made a mistake and got the distances confused. But this doesn’t work either, because other parts of the Itinerary correctly state the direct distance between London and Rochester: Route 3 and Route 4 both describe this journey, without any mention of Noviomagus or Vagniacis, and both give the distance as 27 Roman miles.
Route 2 must therefore be describing a more indirect journey, 10 miles longer than those in Routes 3 and 4, which might include some stretches along Watling Street, but which also clearly involves one or more significant detours. Suddenly, the location of Noviomagus becomes much more intriguing because in principle, it might be anywhere in South London which (a) can claim to be a ‘Roman site’; and (b) lies about 10 miles from the city of London and no more than 27 miles from Rochester. A lot of places meet these conditions, and over the years scholars have performed heroic feats of advocacy in arguing for their personal favourites. Suggestions have included Bexley Heath, Charlton, Crayford, Croydon, Greenwich Park, Keston, Welling, West Wickham and (as we have seen) Woodcote. Some claims are rooted in solid argument, others in parochial loyalty. But for my money a particularly convincing case was made almost 90 years ago by F.C. Elliston-Erwood.
Elliston-Erwood was a respected archaeologist and historian who published a paper in 1928 in the Journal of the British Archaeological Association. In it, he argued that routes in the Itinerary must have run along recognised roads: that is to say, while they needn’t necessarily stick to high-quality military highways, they would at least follow established tracks or ‘minor roads’ connecting settlements and centres of population. They would not simply wander off into the countryside.
Secondly, he pointed out that a journey from London to Rochester involves four river-crossings, at the Ravensbourne, Cray, Darent and Medway. Each of these had an established crossing point, at Deptford, Crayford, Dartford, and Rochester respectively. So, he argued, any reasonable route would make use of these crossing points.
This approach throws a new light on Route 2, with its additional 10 miles and its references to Noviomagus and Vagniacis. It suggests that these references signify not just places, but roads; that Route 2 detours off Watling Street to follow “the Noviomagus road”; returns to Watling Street for a river crossing; and then leaves it again further on to take “the Vagniacis road”. And the distances suggest that we should expect to find “the Noviomagus road” between Deptford and Crayford, and “the Vagniacis road” between Dartford and Rochester.
If this is right, then Noviomagus cannot be several miles to the south at Croydon, Keston, West Wickham or Woodcote. The Roman site at West Wickham, for instance, has been championed in recent years as the site of Noviomagus, and it is indeed about 10 miles from London. But it is also far from any of the established river crossing-points. And the only way for a Roman traveller to get from West Wickham to Rochester, while covering no more than 27 miles, would have been to strike out across open country. It is hard to believe that a cross-country hike like this would be consecrated as an imperially-approved route when good firm roads were available a few miles to the north.
This leaves us with five places which have been proposed as Noviomagus, and which are sited on roads which connect with the established river crossing-points. They are Greenwich Park, Charlton, Crayford, Welling and Bexley Heath.
We can dismiss Greenwich because it’s far too close to London. And if we’re right in suspecting that Noviomagus and Vagniacis are associated with minor roads, then we can also dismiss Crayford, Welling and Bexley Heath, all of which lie on Watling Street itself. That leaves us with Charlton.
Elliston-Erwood pointed out that there was a Romano-British settlement at Charlton: he should know, because he helped excavate it. It was on Cox’s Mount, the highest point of Gilbert’s Pit.
Gilbert’s Pit is the sandy bluff which lies between the Woolwich Road and Maryon Wilson Park, whose sand-pits were quarried for the Woolwich Arsenal and for glass-making in the eighteenth and nineteenth centuries.
The busy A206 at its foot, now lined with warehouses and retail parks, is the modern version of an old road, with a good gravel surface, connecting a line of settlements – Greenwich, Charlton, Woolwich, Plumstead – all of which have produced Roman remains. And the distance from London is about right. This is why Elliston-Erwood had “no hesitation” in declaring that Charlton was Noviomagus, which means that the modern A206 was once “the Noviomagus road”.
Following the same logic, he suggested that Vagniacis was probably on another by-road at Greenhithe or Northfleet, rather than at Springhead on Watling Street, where it is usually placed.
So: unless and until I come across a better case for some other site, and a convincing rebuttal of Elliston-Erwood’s arguments on river-crossings and minor roads, I’m backing Charlton as the likeliest solution to the centuries-old puzzle of Noviomagus.