The Roman road to Brighton: South of Croydon

Brighton Rd - Riddlesdown road

My previous post interrogated the route of the Roman road from London to Brighton as it heads north from Streatham. This time we are looking at the same road as it heads south from Croydon.

From Streatham Hill down to Broad Green the A23 follows the route of the Roman road. Beyond that point, it has two options: a high way along the line of North End and Croydon High Street leading to the modern Brighton Road; or a low way down Handcroft Road to Old Town and Croydon Minster.

In favour of the North End/High Street route is the fact that it stays on higher ground, away from the River Wandle. And there is a cluster of Roman burials in the High Street/George Street/Park Street area, vaguely reminiscent of the roadside ‘cemetery zones’ outside Londinium: if the Romans were happy to bury their dead alongside Ermine Street, maybe they did the same alongside Croydon High Street.

As for the other route, Handcroft Road’s descent towards the Wandle seems to count

Brighton Rd - Handcroft Rd

against it, because in general Roman road-builders sought firm dry ground. But the Wandle at this point is close to its spring, and hardly a formidable obstacle. It’s even possible that the spring provided a religious or ritual attraction. Certainly this area was settled: it was part of Roman Croydon (occupation sites have been found at Rectory Grove and Old Town) and later on it was the centre of Saxon Croydon and a major ecclesiastical estate.

Brighton Rd - Croydon Minster (2)

Manning & Bray, in their 1809 History of Surrey, reported a local tradition of the Roman road passing through Old Town; and Ivan D. Margary, twentieth century Roman-road-hunter extraordinaire, agreed.

Between Croydon and Caterham, Margary admits that “ … we can only trace the probable course … for it is represented almost throughout by existing suburban streets which have covered all traces of ancient work”. Nevertheless his suggested route is rather compelling, consisting of a series of terrace-ways along the hillsides, avoiding the damp valley bottoms. Much of this route is not only walkable, but enjoyably walkable, and certainly not confined to suburban streets.

So: from Croydon Minster and Old Town, if you go up Duppas Hill and through the underpass, you will come out near the northern end of Violet Lane.

Brighton Rd - Violet Lane 3

This is a residential road about half a mile long, and is the surviving fragment of the original Violet Lane which ran for two miles to Russell Hill above Purley. Margary suggested that it may represent a survival of the Roman road. It’s clearly visible on John Rocque’s 1768 map as the track heading south-south-west out of Croydon, brushing the western edge of Haling Park.

Rocque Croydon (3)

Rocque’s map also shows it meeting another road at a Y-junction: this other road used to be called Coldharbour Lane, and is now the Purley Way.

Walk on down Violet Lane as it is today, follow it round to the junction with Waddon Way, and turn left. You are now facing Purley Way playing fields stretching away into the distance.

The original Violet Lane, and (if Margary’s hunch was correct) the Roman road, run underneath the football pitches. Assuming nobody is in the middle of a game, if you head south-south-west across the playing fields, aiming to hit Purley Way in the far corner somewhere near the reservoir, you’ll be roughly on the right line.

A short way beyond the Violet Lane/Purley Way junction, the Roman road would have swung round south and east to head downhill into Purley. We don’t know its precise line: maybe it is represented by the sole surviving scrap of Coldharbour Lane, a bridle path which runs down Russell Hill.

Brighton Rd - Coldharbour Lane

But whichever way it came down the hill, the Roman road would then have crossed the line of today’s Brighton Road, to head south-east along the line of today’s Godstone Road.

You are now down on the valley floor, but not for long. After about a third of a mile you will reach Downs Court Road, climbing up the valley’s eastern flank towards Riddlesdown. Margary suggested that this may have been the way taken by the Roman road, and there is general agreement that Riddlesdown’s main north-south track represents the Roman route.

Brighton Rd - Riddlesdown road

After a mile or more you reach a bridge over a railway line. Here, the modern track heads down to the valley bottom, but this does not represent the Roman route which would have stayed higher up on the hillside. In fact the railway line may give a fair idea of the course of the Roman road for the next mile and a half – so long as we remember that the railway runs low down in its cutting, while the Roman road would have been higher up on a hillside which no longer exists, having been excavated away to accommodate that same cutting.

Brighton Rd - Riddlesdown railway

Further on, where the railway line swings away towards Woldingham, Margary thought that Court Bush Road may lie along the Roman route.

Just as it was obliged to come down to the valley floor at Purley before shifting to the south-east and climbing up again onto Riddlesdown, here again the Roman road comes down in the vicinity of Wapses Lodge Roundabout before turning south, and climbing up Tillingdown Hill, and on towards Godstone.

Brighton Rd - Tillingdown Hill

We are now well outside London – we crossed the Surrey county boundary back in Whyteleafe –so maybe it’s time to call it a day. Caterham town centre is close by, offering coffee and cakes, and trains back to South London, home and glory.

Brighton Rd - Croydon map (2)


The Roman Road to Brighton: North of Streatham


Brighton Rd - Rocque (4)

The Roman road from London to Brighton – or to be more precise, to Pyecombe or Portslade outside Brighton – was an industrial road like the Lewes Way, as opposed to a military highway like Stane Street or Watling Street. It linked London to the rich farmland of the South Downs. And like many Roman roads, it followed straight alignments where possible, but adjusted to local conditions where necessary. This is an important point to remember as we chart its route through South London.

Firstly, let’s deal with the straightforward bit. Since the 1930s, it’s been widely accepted that the modern A23 between Streatham and Croydon follows the line of the Roman road. Ivan D. Margary, doyen of Roman road scholars, put the evidence together. He pointed to reliable eighteenth century antiquarian reports of Roman paving visible at Broad Green on the northern edge of Croydon; to nineteenth century builders who found Roman material at various sites close to the road in Streatham; and to twentieth century Post Office engineers (remember them?) who reported regular encounters with ancient hard-packed gravel along the line of the modern road. He also pointed to the straightness of the modern road, apart from a wiggle just north of Norbury Station to accommodate the crossing of Norbury Brook.

Margary fails to mention the more significant shift in direction, from NNW to NNE, at St. Leonard’s Church in Streatham. But here the presence of the church may itself signify that we are still on the Roman route. There is a pattern of medieval churches being located at junctions, turning points and river crossings on Roman roads. The present St. Leonard’s is fourteenth century, and may well be on the site of an early medieval chapel recorded in the Domesday Book. And further north on the A23 a 1967 excavation found signs of a Roman road at the top of Telford Avenue, opposite the bus station on Streatham Hill.

So Margary was right. For something over four miles, from Broad Green in the south to Streatham Hill in the north, the modern road follows the line of the Roman road.

The question is: where did it go next, north of Streatham?

Margary argues that it stuck with the modern road down Brixton Hill, along Brixton Road, to meet Stane Street opposite the Oval, at Kennington Park. In support of this he points out that Brixton Hill used to be called Brixton Causeway. He says this is “most suggestive”, since ‘causeway’ means a paved or pebbled road. But ‘causeway’ also implies a raised track over wet or marshy ground – which brings us to the River Effra.

People argue about the upper parts of the River Effra, but it’s an established fact that once it hit Brixton, it ran along the line of Brixton Road up to Kennington. It is clearly shown on Rocque’s map of 1746

Brighton Rd - Rocque Effra (2)

where it is called not the Effra but ‘The Shore’. At Kennington it turns west to enter the Thames near Vauxhall Bridge.


Because of the Effra, Brixton Road was notorious for being water-logged: it used to be known as the ‘Wash Way’, and it was only in the mid-nineteenth century that the river was contained within the new network of sewers. If the area around Brixton Road was water-logged as recently as the nineteenth century, it is likely to have been even wetter in Roman times, because the climate generally was wetter.

Roman road engineers hated wet, boggy, marshy places, and took drastic steps to avoid them. For instance: they built Stane Street along a lovely straight alignment down Kennington Road and Clapham Road – but they cheerfully abandoned that alignment in order to stay on dry ground at Newington Causeway, and then again at Clapham Rise. So: if we imagine Roman engineers standing at the top of Streatham Hill, looking north, arguing about the best way to link up with Stane Street and reach Londinium, we would expect them to choose the driest option.

The driest option was not Margary’s route, by Brixton Hill,  Brixton Road and the River Effra. The driest option was along the line now followed by Lyham Road and Bedford Road, heading down to connect with Stane Street at Clapham North tube station.

Brighton Rd - map

This would avoid the Effra entirely, while keeping to the east of another boggy area which Stane Street avoided by veering to the west at Clapham Rise. This Lyham-Bedford line has a long and continuous history in the landscape. A thirteenth century charter confirms it as the boundary between Clapham and Lambeth manors, and Michael Green suggests in Historic Clapham that it may have been a property boundary as far back as the seventh, eighth or ninth centuries. The connection between Roman roads and medieval property boundaries is well-known: Margary himself makes frequent use of it. The Lyham-Bedford line also served continuously as a road or track, and appears in Rocque’s map as ‘The Back Road’, which meets Stane Street at ‘Babilon’.

Brighton Rd - Rocque (3)

For all these reasons, it seems to me that the Lyham-Bedford route makes more sense than the Brixton Road route. Of course, I may be wrong. Someone, at some point, did run a road down Brixton Hill, River Effra notwithstanding, and that someone may have been the Romans. The old name ‘Brixton Causeway’ does suggest a paved or pebbled road, whose builders may have been the Romans. And there is no direct archaeological evidence for the Lyham-Bedford route, any more than for the Brixton Road route.

All I’m saying is that it would surely have been peculiarly perverse for pragmatic Roman engineers to run a road into a river, and condemn it to regular flooding, when they had a drier alternative.

A Modest Way: South-East London’s Roman road (part 2)

Lewes road Blythe Hill 2

In the previous post I walked the first part of the Roman Way from London to Lewes, from the New Cross/Peckham area to Blythe Hill – or rather, I tracked it at a distance, because what’s special about this Roman road is its stubborn refusal to conform to modern roads or paths. But that will change, as we shall see.

The engineers who built the Lewes Way shifted its alignment fractionally to the south at Blythe Hill, aiming across Stanstead Road and Catford Hill for a crossing point on the River Pool. Rather wonderfully, that crossing point is still in use 2,000 years later. The modern bridge is at exactly the same place, in the River Pool Linear Park.

Lewes Road Pool River bridge 2

In the 1930s our intrepid archaeologists, Davis and Margary, revealed the line of the Way here on the eastern bank of the river

Lewes Road Pool River bridge 1

where it cuts diagonally from right to left to cross the railway line and head into Bellingham.

Here its course runs through houses and gardens as it heads south-south-east towards Beckenham, but we can still identify certain marker points. For instance, this is where it crosses Stumps Hill Lane

Lewes Road Stump Hill 1

before continuing over Southend Road, and hitting the old London Chatham and Dover railway line a couple of hundred yards’ east of Beckenham Junction Station.

Lewes Road Beckenham Junction

From here the Way tracks the course of the River Beck, before cutting across the south-east corner of Kelsey Park, conveniently close to the café where I reckon I had earned the right to a refreshing cuppa.

Lewes Road Kelsey Park

I was now deep in leafy suburbia, and the next marker point was St. Dunstan’s Lane which wanders quietly through a landscape of sports grounds and playing fields and golf. Just where the Lane turns from a south-westerly to a near-southerly direction, the Lewes Way comes through, still on its south-south-east trajectory.

Lewes Road St Dunstan's Lane 1

It ploughs straight through the golf course – no bad thing in my opinion – to cross the Charing Cross to Hayes railway line almost exactly at West Wickham Station.

Lewes Road West Wickham station

The next marker point was Sparrow’s Den, at the bottom of Corkscrew Hill, south of West Wickham, where the road cuts across the playing field. Here we are close to a Roman settlement just across the road, near St. John the Baptist church.

Lewes Road Sparrows Den

This is one of many sites in South London to be nominated as the ‘lost city’ of Noviomagus, which lay somewhere on an imperially-endorsed route from London to Rochester. Clearly there was a Roman settlement here – excavations prove that. But as I’ve argued at length in a previous post, the notion that West Wickham is Noviomagus makes no sense. Who would travel from London to Rochester via West Wickham?

The course of the Way meanwhile carries on, south-south-east, across the fields, gradually approaching the New Addington estate. And here at last, after all those miles of hiding in suburban gardens and scuttling across railway lines, the Lewes Way deigns to correspond to a modern way in the modern landscape.

From Addington Road I took a foot-path which climbs south-east through Birch Wood towards Castlehill Ruffs. It was not marked on the OS map as a public right of way, but neither was it marked on the ground as private or with warnings against trespass. It is  clearly regularly used as an informal way up to the New Addington estate. After a bit less than a mile the path crossed a rough vehicle track, and I turned left for a couple of hundred yards along this track towards an electricity sub-station. Just before the sub-station I turned right again onto a new foot-path heading south-south-east through Rowdown Wood.

I was now, finally, walking the line of the Lewes Way, the first time I had been able to do so since leaving Watling Street.

Lewes Road New Addington path 2

The path continued through the woods, not arrow-straight by any means but not deviating wildly either, edging closer to New Addington’s eastern border. It hit this border right at its most unattractive point, next to the industrial zone. Suddenly, what had been a pleasant enough woodland walk became a grim urban edgeland slog, pinned between an ageing concrete fence to the right, and shabby undergrowth to the left, littered and scattered with rubbish and debris of all kinds.

Lewes Road New Addington path 4 (2)

This grubby scramble didn’t last for ever. Once it got beyond the industrial area and backed onto housing, the path became more pleasant. And even when it was at its worst I tried to remind myself that I was walking the line not of a pristine Roman military highway, but of a working industrial road whose job was to link the manufacturing zone of the Weald with markets and barracks in Londinium. So maybe the waste scattered along the path was fitting, a grim nod of recognition from one industrial era to another, expressed in refuse.

This grubby scramble is also, by the way, the line of the modern borough boundary between Croydon and Bromley, which follows the much older county boundary between Surrey and Kent. And this in turn suggests that the course of the Lewes Way across this particular landscape was very clear in early medieval times, offering itself as a ready-made marker just when these two English counties were defining themselves and acquiring firm borders.

The path finally emerges on the southern edge of New Addington, a few yards from Fairchild School, and opposite a Bromley Borough post to confirm that we are indeed right on the old county boundary.

Lewes Road Bromley boundary

That’s as far as I intend to follow the Lewes Way. If you fancy carrying on, I suggest you try to get hold of a copy of Margary’s 1948 classic Roman Ways in the Weald.


It is of course out of print, but there were copies available through last time I looked, and if that doesn’t work, we still – just about – have libraries. Of course Margary was at work many decades ago, and his arguments have sometimes been superseded by more recent scholarship. But for me, none of this alters the fact that the book itself is a sheer delight, and with a bit of interpretation to make allowance for seventy years of suburban growth, its brilliant maps are still surprisingly useable.

A Modest Way: South-East London’s Roman road (part 1)

Lewes road Blythe Hill 1

Most of South London’s Roman roads coincide, roughly or in part, with modern roads: Watling Street with the A2; Stane Street with the A3; the Brighton road with the A23. But there is one Roman road which is positively modest, self-deprecating, leaving no visible mark on the modern townscape, remembered by no modern highway. This is the London to Lewes Way which cuts through South-East London, from Peckham through Nunhead and Brockley and Beckenham and West Wickham, and onwards to the south.

Lewes road map

One practical result of its modesty is that for a walker, following this route on foot is a far more pleasant experience than following the others. Rather than a charmless tramp along busy main roads infested with exhaust fumes and bad temper, the Lewes Way offers quiet streets, footpaths and parks.

This was not a military highway but a working road, connecting Roman Londinium with the industrial zone in the Weald: charcoal-burners and iron-works. But this doesn’t mean that it was a cart-track, or that it took shape in a casual manner. It was constructed by professional engineers, along straight alignments where possible, using high ground as a vantage point to alter those alignments where necessary. At Blythe Hill for instance, the alignment shifts from a south-easterly to a more southerly direction.

Blythe Hill also serves as a break-point in my own walk along the Lewes Way as it progresses through South-East London. This post records the first stretch from Peckham. The next will carry on from Blythe Hill to the south.

The Lewes Way heads north to south, and its starting point is its junction with Watling Street, the Roman road to Kent, which goes west to east.

In broad terms it makes sense to think of the Old Kent Road or A2 as the successor to Roman Watling Street, but only in broad terms, because the Old Kent Road does not follow the line of the Roman road. In this part of London, the line of the Roman road is about 260 or 270 yards to the south of the Old Kent Road. Consequently its junction with the Lewes Way is not on the A2, but in a residential street running off it.

The junction is, in fact, somewhere around here:

Lewes road Asylum Road

This is Asylum Road, named after the very large and impressive almshouses built by the Licensed Victuallers Association in the 1820s/30s. (As reported in a previous post, at one point Penge’s Watermen’s Almshouses were going to be located here as well, and were only re-routed to Penge after a timely intervention by a wealthy local resident).

The view above is looking north-east, so we must imagine Watling Street cutting across in front of us to meet the Lewes Way behind the houses, with the Way itself setting off to the south-east, to our right.

It would be nice if the route of the Lewes Way ran neatly down the middle of Asylum Road, but it doesn’t. Instead, it runs through the back-gardens of the houses on its east side, and for much of its passage through South-East London it is equally self-effacing. It rarely coincides directly with a modern road or path. The best way to follow it, therefore, is to identify a few ‘marker points’ where we know that we are on its track.

For instance, after running down behind Asylum Road, the Lewes Way crosses Queens Road about here, at the junction with York Grove, a few yards east of Queens Road Peckham station:

Lewes road York Grove

It then cuts uphill to the east of Nunhead Station, to run through some more back gardens, this time behind the houses on the east side of Ivydale Road:

Lewes road Ivydale Road

The line of the Lewes Way runs eerily parallel to the line of the northern section of Ivydale Road, and this was noticed by the archaeologists – messrs. Davis and Margary – who first established its route in the 1930s. Some Roman routes are ‘remembered’ by parish or county boundaries, but this is not the case with Ivydale Road. Instead, the line of the Lewes Way may have been ‘remembered’ as a property boundary, carried down through the centuries as the plot of land was bequeathed and sold on, until eventually a commercial builder acquired it and unwittingly reproduced the old Roman line in the form of a nice new residential street.

Our next ‘marker point’ is the footbridge which crosses the railway line just east of Camberwell Cemetery, between Brockley Way and Eddystone Road:

Lewes road Brockley bridge

A bit of imagination is required here. The cutting through which the railway now runs was first made for the Croydon Canal in the early nineteenth century. When the Canal Company went bankrupt in the 1830s, the canal was drained and a railway was run along its bed. But this is all recent history compared to the age of the Roman road which we are trying to follow. We must try to forget the cutting, the canal, the railway, and the footbridge, and imagine this site as an unblemished woodland ridge. From Queens Road onwards the Lewes Way has been coming up a gentle slope, and this site marks its top. From here, the Way heads downhill towards Brockley Rise, running straight through the middle of St Hilda’s Church:

Lewes road St Hilda's Church

From St Hilda’s it ploughs on, up the hill on the other side of Brockley Rise, heading for the top of Blythe Hill. From here we can look back along the line of the Lewes Way, with St Hilda’s in the foreground, and the glass towers of the City beyond:

Lewes road Blythe Hill 1

And if we turn around to the south we can see where our modest way will take us next time, guided of course by the new alignment prescribed by the Roman engineers when they stood here two thousand years ago:

Lewes road Blythe Hill 2

Next time: Beckenham and beyond.

Main sources for this post:

Bernard F. Davis, ‘The Roman Road from West Wickham to London’, Surrey Archaeological Collections, vol. XLIII, 1935.

Ivan D. Margary, Roman Ways in the Weald, Phoenix House, London, 1948.

Noviomagus: South London’s Roman puzzle


 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.



The Lambeth Ford and Roman Watling Street


 The London region, before the Romans arrived, wasn’t a bad place to live. The clay soil was hard work, but close to the Thames and its tributaries the soil was better, and the river itself was rich in food and other resources. For travellers by boat the Thames was also a highway, but for those on foot it was more of a barrier. Long before the Romans arrived, therefore, people would have needed crossing-places on the Thames.

There is a longstanding tradition that there was a ford across the river, roughly between St. Thomas’ Hospital in Lambeth and the Houses of Parliament in Westminster.


And there are versions of the tradition which associate this ford directly with the Roman road, Watling Street, whose southern section comes in to Southwark from Kent, and whose northern section heads up from Marble Arch along the Edgware Road. For some, because the line of Watling Street appears to ‘point’ at the ford, it follows that the ford must be its intended destination. For others, Watling Street in its entirety was merely the Roman version of an ancient pre-Roman track-way which ran from Kent to Wales, and which crossed the Thames at Lambeth.

I have no problem with the notion of a ford at Lambeth. I have no problem with the notion that people would have used the ford in Roman times, just as they did in the centuries before and after. But I don’t believe that this ford was directly connected with the Roman road that we call Watling Street.

Let’s take things in order. The idea of a ford at Lambeth may seem unlikely to us because today, it would be suicidal to try to wade across the river between St. Thomas’ Hospital and the Houses of Parliament. Today’s Thames at this point is an urban, embanked and tidal Thames, deep, fast, and dangerous. But for most of its life the river here was wider, shallower and slower, and it wasn’t tidal because the tide exhausted itself further downstream. Its banks were made up of mudflats, marsh and beach, broken up by creeks and inlets into numerous small islands or eyots. One of these on the north bank was Thorney Island, formed by the two arms of the River Tyburn as it approached the Thames from the north.

Thorney Island no longer exists as an island, but we can still place it, because in the seventh century it was chosen as the site for the church which went on to become Westminster Abbey. We know that early medieval churches were often located on or near roads or tracks, so maybe one reason for choosing the site of Westminster Abbey was its proximity to the crossing-place on the Thames.

It is therefore perfectly possible that there was a ford between Lambeth and Thorney IsIand, and that it was already there long before the Romans arrived. But what connection, if any, might this ancient ford have with Watling Street?

Watling Street is a Roman road. There is no evidence that it was laid on top of an ancient British track-way, and Alec Detsicas, in his well-researched study The Cantiaci, firmly refutes this idea. It is also intuitively unlikely if we look at the line of Watling Street on the map, because it really is one of the straightest Roman roads in Britain, slicing across the landscape along uninterrupted alignments in both its southern and northern sections. It has all the signs of a route dictated by the preferences and prejudices of Roman engineers.

So let’s come at it from another angle. Let’s take two fixed points: the church of St George the Martyr, which marks the junction where Watling Street meets two other Roman roads, Stane Street and Borough High Street; and St Thomas’ Hospital, which we believe was the site of Lambeth ford. If we really want to believe in a link between Watling Street and the ford, then we must posit some sort of spur-road running between these points, from St George the Martyr to St Thomas’ Hospital.

On the map below,


the broken line represents the most direct route for such a spur: parallel with and a short way north of Borough Road, north of St George’s Circus, cutting across Westminster Bridge Road near Lambeth North tube station,


grazing the northern edge of Archbishop’s Park before passing through the Hospital and across the river. On Thorney Island it runs up the nave of Westminster Abbey,


then roughly along the line of today’s Tothill Street towards Buckingham Palace, where it veers somewhat north to track Park Lane up to Marble Arch and Edgware Road.


But this broken line is entirely speculative. The only tentative suggestion that such a route may have existed as a Roman road is a trace of an undated gravelled road in Lambeth Palace garden. There is nothing else. Maps offer no support: the earliest accurate street map of this part of South London, John Rocque’s map of 1746, shows no road corresponding even faintly to our posited route. Nor is there any trace of it in parish boundaries: the boundary between the old parishes of Southwark and Lambeth runs north-south, not east-west as it would need to do if it were following our broken line.

Instead, the evidence suggests that travellers along Watling Street would have passed through the Roman city of Londinium. Coming in from Kent, they would have turned up Borough High Street to go through Southwark, across the bridge and into the city. When they resumed their journey they would leave along the line of High Holborn and Oxford Street to Marble Arch, where they would turn up Edgware Road towards Verulamium (St. Albans) and the north.

On a modern map this may look like a convoluted detour. But modern maps do not reflect Roman priorities. I believe that this ‘detour’ makes perfectly good sense once we grasp that Roman roads were projections of urban-based Roman power.

Roman roads were about control exercised at many levels. Their solid foundations and surfaces allowed for rapid travel, especially rapid travel by soldiers bent on keeping the peace: this is well known. Their engineering, combining straight alignments and pragmatic deviations (discussed in another post), were both practical, and highly symbolic assertions of control over the landscape itself. And most important of all, the roads ran to and from towns and fortresses, the sites from which power was exercised. The roads served both to push the agents of Roman power outwards in the form of soldiers and administrators and tax-gatherers; and to pull its subjects inwards, into the towns, to gain access to markets and special services and the prestige that came from rubbing up against Romanitas. Roman power was an urban power, and its roads were the means by which that urban power projected itself out into the rest of the country.

So, when travellers coming along Watling Street from Kent in the second or third century arrived at the junction where St George the Martyr now stands,


they would not have regarded Londinium as a detour. They would have regarded it as a natural and welcome destination. For farmers or merchants or artisans it had markets. For artists it had rich clients. For soldiers It had barracks. For everyone it offered food, rest, creature comforts, and a renewal of the sense of belonging to a vast, cosmopolitan civilization.

I have no problem with the notion of a ford across the Thames at Lambeth. But I don’t believe that the Romans would have regarded the existence of such a ford as justifying the construction of a new road, when the alternative crossing-place at London Bridge gave access to the city.

Roman Stane Street: on firm ground through Ewell & Epsom


We saw in a previous post that in his book on Roman Stane Street published just before the First World War, Hilaire Belloc insisted that it followed a single straight alignment from Merton Priory to Pebble Lane, seven or eight miles to the south on the downs above Dorking. If true, this would mean that Stane Street was unconnected with the suspiciously straight stretch of the modern A24 from Morden to Nonsuch Park, and it would mean that it simply bypassed modern Ewell and Epsom.


After the Great War, Captain Grant published his damning rebuttal of Belloc. He argued, correctly, that the Morden-Nonsuch stretch of the A24 was an authentic part of Stane Street. He then projected this line to the south-west, and argued that the road ran through Ewell and Epsom and on to Ashtead, before turning south-east to cross the River Mole at Burford Bridge. In Grant’s view, Pebble Lane had nothing to do with Stane Street.


Grant was right in general, but wrong in this particular. Both the Morden-Nonsuch stretch of the A24, and Pebble Lane, are bona fide survivors of Stane Street. But they are clearly on quite different alignments. So we have to ask (a) how do they connect? and (b) why did the Roman engineers make things so complicated in this area?

First of all, let’s sort out the route of the Roman road through Ewell, which was clarified by excavations in 1970-75 and 2003.


The modern road runs down from Morden and alongside Nonsuch Park, then does a dog-leg before meeting the dual carriageway which by-passes Ewell. The Roman road, however, carried straight on along the Morden-Nonsuch alignment, crossed the dual carriageway, crossed St Mary’s churchyard, and continued to the Old Tower just beyond the church.


Here it shifted slightly to the west to cut through Ewell’s residential streets (Staneway, St James Avenue).

It ran on to cross the modern railway line close to Windmill Bridge, through St Martin’s churchyard in Epsom,

Stane St Epsom St Martin's church

and on to the top of Woodcote Park. Here it shifted east, ran on, shifted west, ran on, shifted east, and finally met Pebble Lane just beyond Thirty Acre Barn.


There is nothing unusual about a few changes of direction in a road or path over the course of four or five miles, the distance from Nonsuch Park to Pebble Lane. However, this is not just any road, but a Roman road which is notable for its long straight alignments. As we saw in the previous post, the Roman engineers liked their alignments – but they were also happy to make pragmatic compromises when circumstances justified it. So: what were the circumstances around Ewell and Epsom?

According to Ivan Margary, the doyen of Roman road studies, it was all about the road surface. We saw previously that the course of Stane Street at Newington, and possibly at Clapham, was diverted to skirt around areas of marshland. In the Ewell and Epsom area the issue wasn’t marsh but clay: the connecting route between Ewell and Pebble Lane was chosen to keep the road off the clay and on firm chalk.

Finally: note once more the recurring theme of churches. Medieval churches situated directly on Stane Street as it passes through South London include St George the Martyr, Merton Priory, St Mary’s Ewell, and St Martin’s Epsom.

Belloc made much of this connection and, however wrong he may have been about other things, he was right about this. There is a pattern here to which we will return.

Stane Street at Merton: the Wandle and the Priory

Despite arguments about the precise course of Roman Stane Street as it makes its way through South London, there has always been a consensus that the road crossed the River Wandle at Merton, close to the site which later became Merton Priory. However, for a long time no-one knew the exact location of the crossing.   

In trying to identify this location, we might expect the River Wandle itself to provide a reliable fixed point from which to start. But rivers are not fixed. It is in their nature to wander, to find new courses, and in Merton the course of the Roman Wandle was some way to the east of the modern Wandle. Two thousand years ago the river ran close to the line of today’s Christchurch Road, and passed in front of today’s Colliers Wood tube station. In fact Colliers Wood tube station is the site of the Roman river crossing. An extended and highly successful archaeological excavation, which ran from 1976 to 1990, answered this and other longstanding questions. 


The Roman river crossing was simply the first in a long line of interventions which have continued right up to the present, and which make this part of Merton into a perfect illustration of the rationale for this blog. My aim in PengePast is to look beyond and beneath the suburban familiarity of South London, and to reveal it as an ancient landscape of human endeavour, struggle and strangeness. Merton certainly does that. 

To illustrate the point: here is the southern edge of the retail park, next to Merantun Way, as it is today. 


Three thousand years ago, this site was a riverside meadow. A little less than two thousand years ago, it carried the Roman road. 800 years ago it was the west end of a large priory church, as shown in the map at the top. 100 years ago  


it was Merton Abbey railway station. And today it is a retail park of quite astonishing ugliness. But in its very ugliness it makes the point that even the most mundane and hideous expressions of suburban modernity are part of an historical process. Even a shoppers’ car-park represents another set of meanings, laid on the palimpsest of South London’s landscape. 

Roman Stane Street was first built around AD 50, with further work carried out between 150 and 200. It seems to have fallen into disuse – or at least, it wasn’t repaired any more – towards the end of the Roman period, from about 350. But the string of Anglo-Saxon settlements and medieval churches along its length show that even if its paving fell short of the Roman ideal, it continued to serve as an important highway long after Britain fell out of the Empire. 

The map at the top highlights two phases of the site’s history: the Roman phase, including the course of the Wandle, the line of Stane Street, and the river crossing; and the medieval phase, including Merton Priory church, and the boundary of the Priory precinct. Both phases were addressed by the 1976-1990 excavations.    

The Augustinian Priory at Merton was founded in 1117, and became an important ecclesiastical institution enjoying royal patronage. At this time abbeys, monasteries and priories acted as clerical, economic and educational centres in their surrounding areas. Merton certainly played this role, and in the Priory’s early years the young Thomas Becket was sent from his home in the city of London to be a scholar here. By the time the Priory church was built towards the end of the C12th, Becket was already a martyr, murdered at the altar in Canterbury Cathedral. 

The church was built in the Romanesque style, its west door standing directly on top of the Roman road, and for more than three centuries it stood as the most imposing element of a powerful institution. With the C16th Reformation, of course, like other abbeys and monasteries throughout the country, Merton Priory was closed down. But whereas the buildings on many other sites were simply left to moulder, Merton’s church was promptly dismantled, its stones robbed for re-use in Henry VIII’s new palace a few miles down the road at Nonsuch. Some buildings seem to have survived into the 1640s, when the site was used to garrison troops in the Civil War. But the whole area was transformed from the late C17th as the Wandle in general became increasingly industrialised. Among these industries was textile manufacture and printing, which eventually led William Morris to establish his Abbey Mills works here in 1881.  

By this time, any physical traces of Merton Priory were thought to have disappeared. But in 1914 a Romanesque gate-way was discovered, built into the fabric of a local house. It had clearly belonged to an important Priory building, such as a guest-house or chapel. It was removed and re-erected about a mile away in the church-yard of St. Mary the Virgin.  


Here it still stands, a robust piece of C12th masonry, its chevron pattern still sharp, a surprising and rather wonderful remnant of the great medieval Priory at Merton.  

Roman Stane Street: the Clapham Deviation

Ivan D. Margary was the doyen of Roman road studies from the 1940s to the 1960s. He is still recognised as a formidable authority, and his books such as Roman Roads in Britain and Roman Ways in the Weald


are a wonderful resource. Margary always stressed the disciplined pragmatism of the Roman engineers. They used straight alignments to identify ideal and economic lines across the landscape, and would happily run their roads along these alignments where practicable. But when local topography or soil conditions intervened, they adjusted.

So for instance, at the northern end of Stane Street close to the Thames, the Roman landscape was made up of small islands, creeks and marshes (for more on this see ‘South London Begins‘). Between Newington Causeway and Great Dover Street was a marsh later known as ‘Stewfens’; just to the north was the aptly-named Marshalsea, later the site of a famous debtors’ prison; and the name ‘Newington Causeway’ suggests a raised track around boggy ground. Pragmatic Roman engineers were hardly likely to run their road straight through a marsh, when a minor detour would keep it on firm ground. This is why Margary and others agree that the curve of Newington Causeway round to Elephant and Castle does represent the original route of Roman Stane Street.

But for the next six miles or so, from Kennington right down to the Wandle crossing at Merton, Margary argued that Stane Street was guided by a single straight alignment pointing to the road’s ultimate destination of Chichester.

The modern road more or less sticks to this alignment, with one major exception. It is pretty straight from Kennington to Stockwell; and despite some wandering it stays broadly faithful to the alignment between Balham and Merton. But between Stockwell and Balham we have the “Clapham Deviation”: at Clapham North tube station the modern road veers off to the west up Clapham Rise and Clapham High Street, and then along Clapham Common South Side and Balham Hill, only re-joining the alignment at Balham Station.

In Margary’s view, the route followed by Roman Stane Street was much more direct. It did not veer west at Clapham North, but ploughed straight on through the tube station, through the residential streets to the east of Clapham High Street, through Clapham Crescent, across Clapham Park Road, running just east of Klea Avenue to cross Cavendish Road, and on to meet the modern road close to Balham Station. The map below


shows the deviation in diagrammatic form: Margary’s alignment is in red, and the course of the modern road in black. It also shows the locations of the Northern Line tube stations. I like tube stations. They are friendly and familiar markers in the London townscape, and along this section of the Northern Line they represent a string of ancient South London settlements which sprang up along the length of the road. Kennington, Clapham, Balham, and Tooting are all recorded in the Domesday Book, and must go back many centuries before that.

Essentially Margary’s argument was an application of Occam’s Razor: since the Romans who built the road followed a clear alignment from Kennington to Stockwell, and resumed that same alignment from Balham, the most economical assumption is that the section between Stockwell and Balham was also on that alignment. However archaeological excavations, while not proving him wrong, have failed to prove him right. An excavation in the 1940s at Tableer Road/Worsopp Drive south-west of Clapham Park Road; and another in the 1960s at near-by Lambeth College; found no sign of a road, though both were close to his projected alignment. A 1970s dig on the north side of Gaskarth Road did find pebbles and gravel which seemed to indicate a road or path, but not necessarily a Roman one.

Consequently local historian Michael Green, in his 2008 book Historic Clapham, argues that Roman Stane Street passed through Clapham not on Margary’s alignment, but on the route of the modern road. And he suggests that the Roman road took this course in order to avoid boggy ground in the area which later became Clapham Park. If this is right, then the “Clapham Deviation” is no deviation at all, but a sensible adjustment undertaken for the same reason as the adjustment at Newington Causeway.

Next time: Stane Street, the Wandle crossing, and the tragic fate of Merton Priory.

Contested routes: the Stane Street controversy

Stane Street was the Roman road from London to Chichester – an important military highway from the very start of the occupation, when the Romans’ single most important British ally was Cogidubnus, whose power base lay in what is now Hampshire and West Sussex.

From London Bridge and Southwark, running down through South London, much of the route of Stane Street is on or near the modern A24. But at Merton things get messy. For many years there were very different views, and furious arguments, about its route beyond this point. Everyone agreed that Stane Street crossed the Wandle at the site of Merton Priory (usually called ‘Merton Abbey’). Everyone agreed that, 20 miles to the south, it crossed the River Mole at Burford Bridge just outside Dorking. But there was little agreement about its route between Merton and Dorking, or about its relation (if any) with Morden, Ewell, Epsom and Ashtead.

In the years before World War One, the popular writer Hilaire Belloc walked the route of Stane Street – or at least the route as he understood it – and in 1913 he published a book about it.



Belloc was a bundle of contradictions: a devout Catholic of French descent who loved England and the English countryside; a former Liberal MP with a decidedly conservative streak. His book on Stane Street was readable, entertaining, and in large part wrong.

Like many of us, Belloc was impressed by Roman roads, especially those stretches where a long highway streaks, straight as an arrow, through the countryside. But from this he seemed to draw the conclusion that straightness in itself was the Roman engineers’ overriding ambition, and that their roads followed point-to-point “alignments” cutting through the landscape almost regardless of obstacles.

Finally, and quite separately, he argued that in the centuries after the Romans, their roads remained sufficiently important to act as magnets for the siting of medieval churches and abbeys.

Putting these convictions together, Belloc argued that Stane Street ran through South London on a single straight alignment from London Bridge to the downs north of Dorking, where it is represented today by a footpath called Pebble Lane. And he claimed that this alignment ran straight through Merton Abbey on the River Wandle:

“The whole of my argument is based upon the exact alignment of the Stane Street where it has survived with the direction of London Bridge, and upon the identity of the crossing of the Wandle with Merton Abbey, and with the royal land of Merton”.


 Merton Abbey Mills today, commemorating William Morris’s C19th printworks, which commemorated the C12th Priory

 By committing himself to this alignment, Belloc chose to ignore some inconvenient evidence. Firstly, his alignment by-passed Ewell and Epsom altogether, although he knew perfectly well that traces of Roman road had been found in both places. And more significantly, it required him to ignore long stretches of the modern road which, in his day as in ours, were generally accepted as following the route of Stane Street through South London. To turn a blind eye to this enormous fact on the ground, in favour of an abstract theory of ‘alignment’, was somewhat perverse. We might almost suspect that he was deliberately courting controversy.

Whether or not he was courting it, he got it. In 1922, a retired cavalry officer published a furious refutation of Belloc’s account. The very title of Captain W.A. Grant’s book made his purpose clear – The Topography of Stane Street: A critical review of ‘The Stane Street’ by Hilaire Belloc.



Blissfully unencumbered by any sense of literary courtesy, Grant laid into Belloc with a will:

“Mr Belloc … jumps to conclusions without taking the trouble to verify them, and gives bearings and distances as ‘exact’ when they are not even approximate … “;

“ … neither of these alignments points where Mr Belloc says it does … ”;

“Surely this must break all previous records for carelessness and inaccuracy”;

“If I seem to have handled Mr Belloc somewhat severely, it is, I think, no more than he deserves … “.

I suspect that, as a military man, Grant was piqued by what he saw as Belloc’s amateurism, and was keen to defend the professionalism of the Roman military engineers who had built the road:

“What Mr Belloc appears to have done was to start off from Chichester along the known course of the road, groping his way as it were from point to point … But what we have to do, I think, is to try and put ourselves in the position of the Roman engineers before there was any road at all … “.

I like this passage. Grant makes no claim to be an historian, but in these few gruff words he neatly summarises the importance of historical sensibility and historical empathy. He recognises the distance in time and circumstance which separates us from those Roman engineers, but also the shared humanity which unites us to them, and he invites us to try to view the world as they viewed it.

Two things in particular infuriated Captain Grant. Firstly, Belloc was cavalier with his data, and even altered it to fit his theory – for instance, by drawing a misleading map which surreptitiously moved Merton Priory a quarter of a mile to the east so that it would appear to lie directly on his pet alignment. And secondly, Grant was convinced that Belloc simply misunderstood alignments. Grant believed that the Roman engineers used them not as building instructions, but rather as practical points of reference:

“ … I am strongly of opinion that in every case the direct alignment would be plotted out, whether it were practicable for the road to follow that line or not” (my italics).

In the case of Stane Street he suggested that the engineers plotted three ‘great alignments’, ‘ideal lines’ across the landscape which they used to find a route which was as direct as possible while allowing pragmatically for local topography.

Grant’s critique of Belloc is unrelenting, entertaining, and mostly justified. Among other things, he confirmed that the modern road from Epsom to London Bridge does indeed follow the route of Stane Street through South London (give or take some wanderings and gaps which I will investigate in the next post). But he didn’t get everything right.



Grant argued that after passing through Ewell and Epsom, Stane Street ran on through Ashtead where there is a small Roman camp, and past Leatherhead to make its way down to the River Mole at Burford Bridge, and so on to the camp at Dorking. He therefore denied that Pebble Lane, which was so important to Belloc and lay further to the east, was part of Stane Street at all. In this, Grant was mistaken and Belloc was correct.

However when it came to the big picture, to the methods of the Roman engineers, an appreciation of the manner in which the road was made, and a practical sense of its route through South London, Captain W.A. Grant, formerly of the 13th Hussars, got it mostly right, while Hilaire Belloc, celebrated popular writer, got it seriously wrong.

Future posts: more on Stane Street’s route through South London including the Clapham Divergence, the Merton Tragedy, the Ewell and Epsom Mysteries, and several significant churches.